Oxidation treatments affecting Sauvignon blanc wine sensory and chemical composition

Transcription

1 Oxidtion tretments ffecting Suvignon blnc wine sensory nd chemicl composition by Crien Coetzee Disserttion presented for the degree of Doctor of Philosophy (Agriculturl Science) t Stellenbosch University Deprtment of Viticulture nd Oenology, Fculty of AgriSciences Supervisor: Prof. Wessel Johnnes du Toit Co supervisor: Prof. Antonio Césr Silv Ferreir 214 i

2 Declrtion By submitting this disserttion electroniclly, I declre tht the entirety of the work contined therein is my own, originl work, tht I m the sole uthor thereof (sve to the extent explicitly otherwise stted) tht reproduction nd publiction thereof by Stellenbosch University will not infringe ny third prty rights nd tht I hve not previously in its entirety or in prt submitted it for obtining ny qulifiction. Dte: 9/12/213 Copyright 214 Stellenbosch University All rights reserved ii

3 Summry This study focussed on the effect of oxygen on the chemicl nd sensory evolution of Suvignon blnc wine under controlled oxidtion conditions. The sensoril interctive effects between Suvignon blnc vrietl rom compounds nd compounds tht typiclly rise during oxidtion of white wines were lso investigted. In the first reserch chpter the sensoril interctive effects of Suvignon blnc impct compounds with ldehydes typiclly originting from oxidtion of white wines were investigted. Four compounds, 3 mercptohexn 1 ol (3MH), 3 isobutyl 2 methoxypyrzine (IBMP), 3 (methylthio) propionldehyde (methionl) nd phenylcetldehyde were dded together in model wine medium t vrying concentrtions. The concentrtions chosen were ccording to those reported in literture to occur in Suvignon blnc wines. The sensory effects of these compounds were profiled using trined sensory pnel. Compounds were first profiled individully nd results showed tht the chnge in comcpound concentrtion not only led to chnge in intensity rtings but lso in some cses chnge in the descriptor. All four compounds in the sme smple showed complicted interctive effects. Dt were sttisticlly nlysed using reltively novel techniques such s sttisticl networks tht llowed deeper insights into the interctions involved. Vrious observtions were mde such s the contribution of 3MH to the green chrcter of the wine, the potent suppressing bilities of methionl on 3MH nd IBMP nd the dditive effect of methionl nd IBMP contributing to the cooked chrcter of the smple. In the second reserch chpter the effect of repetitive oxidtion on fresh nd fruity style Suvignon blnc wine ws investigted. Results showed the progress nd evolution of romtic nd non romtic compounds during n oxidtive ging period. A lrge rnge of chemicl nlyses were conducted together with extensive sensory profiling. Results showed decrese in voltile thiols responsible for the fruity nunces nd n increse in oxidtion relted compounds, such s cetldehyde, during the course of the oxidtion. Sensory profiling showed the evolution of the wine rom during oxidtion. The wine evolved from fresh nd fruity wine to slight oxidtion nd then developed extreme oxidtive chrcteristics. The Control smples (no oxygen dded) developed cooked chrcter which could indicte the formtion of reductive compounds in these wines. Conversely, the wines tht received single dose of oxygen did not develop this flvour nd were perceived to be more fresh nd fruity thn the Control smples. The evolution of the wine colour ws lso monitored using spectrophotometer s well s sensory pnel nd results suggest tht the colour of the wine evolved before the disppernce of the plesnt rom. The dvntges nd disdvntges of oxygen exposure to this type of wine style re discussed s well s the complexity of the wine mtrix nd sensory interctions occurring in the specific wine. iii

4 The im of the third reserch chpter ws gin to investigte further sensory interctions between Suvignon blnc vrietl rom compounds nd n oxidtion relted compound. After the observtions reported in the second reserch chpter, the inclusion of cetldehyde in n interction study seemed to be of criticl importnce. Acetldehyde cn rech significnt concentrtions during oxidtion nd cn hve detrimentl effect on wine rom. This interction study included three compounds, 3MH, IBMP nd cetldehyde. Acetldehyde ws ble to effectively msk the green chrcter of the smple, while it lso enhnced certin fruity nunces when present t specific concentrtions. 3MH ws ble to suppress the oxidising chrcter of cetldehyde when present t sufficient concentrtions. The results from this study clerly showed the complexity of the wine medium nd the interctions involved. It lso highlighted the importnce of performing these types of sensory studies in simple medium opposed to complex mtrix such s wine. The effect of oxygen on vrious spects of the wine ws investigted nd the combintion of chemicl nd sensory dt delivered some interesting conclusions lso involving interctions tht occurred. This study pved the wy for future investigtions on the sensory reltionships of Suvignon blnc rom compounds nd the role of proper oxygen mngement in the production of qulity wines. iv

5 Opsomming Die doelwit vn die studie ws om die effek vn gekontroleerde oksidsie op die ontwikkeling en verloop vn verskeie chemiese komponente sook die ontwikkeling vn die sensoriese profiel vn n Suvignon blnc wyn te ondersoek. Bykomende studies ondersoek die interksie/wisselwerking tussen romkomponente in n eenvoudige mtriks. Interksies tussen spesifieke Suvignon blnc kultivrgessosieerde romkomponente en komponente wt normlweg tydens die oksidsie vn witwyne ontwikkel, word ondersoek. Die eerste nvorsingshoofstuk het ten doel om die interksie tussen tipiese Suvignon blnc romkomponente (of impkkomponente) met ldehiede wt normlweg tydens oksidsie vn witwyne ontwikkel, n te vors. Vier komponente is in n modelwyn gevoeg teen verskeie konsentrsies, wt oor die lgemeen in die litertuur gerpporteer is om voor the kom in Suvignon blnc wyne. Die komponente wt ondersoek is, is: 3 merkptoheksnol (3MH), 3 isobutiel 2 metoksiepirsien (IBMP), 3 (metieltio) propionldehied (methionl) en fenielsetldehied. Die sensoriese effekte vn die komponente is deur n opgeleide sensoriese pneel geëvlueer. Komponente is eers individueel genliseer en die resultte het getoon dt dt die intensiteit vn die spesifieke rom vernder nmte die konsentrsie vernder. In sommige gevlle het die beskrywende woord vir die rom ook vernder. Deur l vier komponente in dieselfde monster te voeg word die ondersoek gekompliseer. Die dt is sttisties genliseer deur gebruik te mk vn reltiewe nuwe tegnieke soos sttistiese netwerke wt dieper insig in die betrokke interksies bewerkstellig. Verskeie wrnemings word gerpporteer onder ndere die bydre vn 3MH tot die groen krkter vn die wyn, die krgtige onderdrukkingsvermoë vn methionl op 3MH en IBMP sook die opbouende effek vn methionl en IBMP wt bydr tot die gekookte krkter vn die monster. Die tweede nvorsingshoofstuk is drop gemik om die effek vn herhlende oksidsie op 'n vrs en vrugtige styl Suvignon blnc wyn te ondersoek en om die vordering en ontwikkeling vn romtiese en nie romtiese komponente gedurende hierdie tydperk te nliseer. n Wye reeks chemiese komponente is genliseer tesme met omvngryke sensoriese nlise. n Afnme in die vlugtige tiole, wt verntwoordelik is vir die vrugtige geure, is gevind tesme met n toenme in oksidsie verwnte komponente (soos setldehied). Sensoriese ondersoeke toon ook die evolusie vn die wynrom tydens oksidsie. Die wyn het ontwikkel vn n vrs en vrugtige styl n effense oksidsiegeure wrn ekstreme oksidsiekrkters wrgeneem is. Die Kontrole monsters het n gekookte krkter ontwikkel wt n nduiding vn die ontwikkeling vn reduktiewe komponente in hierdie wyn kn wees. An die nder knt het wyne wt een suurstofdosering ontvng het, geen vn hierdie geure ontwikkel nie en die wyn is s vrser en vrugtiger beskryf in vergelyking met die Kontrole monsters. Die ontwikkeling vn die wynkleur is ook gemonitor deur gebruik te mk vn n spektrofotometer sook 'n sensoriese pneel. Resultte stel voor dt die kleur vn die wyn ontwikkel voor die ngenme geure begin verdwyn. Die v

6 voor en ndele vn suurstofblootstelling n hierdie tipe wynstyl word bespreek sook die kompleksiteit vn die wynmtriks en sensoriese interksies wt in hierdie spesifieke wyn voorkom. Die derde nvorsingshoofstuk is weereens drop gemik om die sensoriese interksie tussen tipiese Suvignon blnc kultivr gessosieerde romkomponente en nog n oksidsie gessosierde romkomponent te ondersoek. Die resultte vnuit die tweede nvorsingshoofstuk het die insluiting vn setldehied in die interksiestudie genoodsk. Asetldehied kn betekenisvolle konsentrsies tydens oksidsie behl en kn ook ndelige effekte op wynrom hê. Hierdie interksiestudie het die volgende drie komponente ingesluit: 3MH, IBMP en setldehied. Asetldehied het die groen krkter vn IBMP effektief gemskeer terwyl dit die wrneming vn die vrugtige rom ondersteun en selfs verhoog het wnneer dit teen sekere konsentrsies teenwoordig ws. 3 Merkptoheksnol het die oksidsiekrkter vn setldehied onderdruk wnneer dit teen genoegsme konsentrsies teenwoordig ws. Die kompleksiteit vn wyn s n nvorsingsmedium is duidelik vnuit die studie verl in die ondersoeking vn interksie effekte tussen komponente. Die belngrikheid vn die gebruik vn 'n eenvoudige medium teenoor n komplekse medium vir soortgelyke studies is dus duidelik. Die effek vn suurstof op verskeie spekte vn witwyn is ondersoek en die kombinsie vn chemiese en sensoriese dt het interessnte gevolgtrekkings gelewer. Die studie het die pd vir toekomstige studies gebn in terme vn sensoriese interksies met betrekking tot Suvignon blnc rom. Die belngrikheid vn oordeelkundige suurstofbestuur tydens die produksie vn kwliteit wyne is ook uitgelig. vi

7 Hierdie proefskrif is n my ouers opgedr. This disserttion is dedicted to my prents. vii

8 Biogrphicl sketch Crien Coetzee ws born on 31 Jnury 1986 in Bellville, South Afric. She went to DF Mln High School nd fter mtriculting in 24 enrolled for BScAgric degree, mjoring in Viticulture nd Oenology t Stellenbosch University. In 29, she enrolled for MScAgric degree in Oenology t the sme University. She completed both degrees cum lude in 28 nd 211 respectively. She enrolled for PhD in Oenology in 211. As prt of her PhD studies, she visited other Universities such s the University of Aucklnd in New Zelnd s well s Universidde Ctólic Portugues in Porto, Portugl to enhnce her reserch experience nd obtin further results regrding her studies. In 213 she ws wrded the TATA Afric Scholrship for Women in Science s prt of the Deprtment of Science nd Technology s Ntionl Women in Science Awrds for outstnding bility nd potentil in reserch. viii

9 Acknowledgements I wish to express my sincere grtitude nd pprecition to the following persons nd institutions: Prof. Wessel du Toit, for his input, support nd continuous encourgement throughout the project. For vluble discussions nd debtes nd llowing me to voice my own opinion. Mostly for believing in my reserch bilities nd providing mny opportunities to expnd my experience. Prof. Antonio Césr Silv Ferreir, for his input nd vluble discussions. Mrs. Jenne Brnd, for her unlimited ptience nd vst mount of help during sensory nlyses. Prof. Mrtin Kidd, for ssistnce with sttisticl nlyses nd interprettion nd for lwys being ptient nd friendly. Dr. Dn Jcobson nd Mr. Guy Emerton, for their sttisticl input in the sensory studies. The Deprtment of Viticulture nd Oenology t Stellenbosch University, to my friends for listening to me vent nd for constntly reminding me of the bigger picture, for vluble discussions, support, ssistnce nd understnding in the lbortory. To the stff for ll the help during the projects. The stff nd students t the University of Aucklnd in New Zelnd, for teching me vluble nlyticl skills nd providing support even fter returning home. Also for gret friendships formed. The stff nd students t the Universidde Ctólic Portugues in Porto, for help during nlyses nd being friendly hosts. Elizm vn Wyngrd, for her vst mount of help nd willingness to lern. De Grendel Wine Estte, for sponsorship of wine. The South Africn Wine Industry (Winetech), THRIP, nd the Ntionl Reserch Founttion for finncil support. My sisters nd friends, for support nd encourgement throughout the study. My mother, Annette vn der Merwe, for being my inspirtion every dy. For lughter nd ters nd for being the loving nd cring mother in everything I do. My fther, André Coetzee, most importntly, for wkening my interest in reserch nd for lwys believing in me. For lwys being enthusistic bout my work, being n exceptionl role model nd encourging me to lwys pursue excellence. Secondly, for his understnding nd finncil support for the pst yers without which the completion of this study would not hve been possible. Also to Suznne Coetzee, for her interest, support nd dvice during the yers. ix

10 Prefce This disserttion is presented s compiltion of six chpters. Ech chpter is introduced seprtely nd is written ccording to the style of the South Africn Journl of Enology nd Viticulture. Ech chpter should be regrded s n individul entity nd therefore some repetition between chpters my occur. Chpter 1 Chpter 2 Chpter 3 Chpter 4 Chpter 5 Chpter 6 Generl introduction nd project ims Literture review Suvignon blnc wine: Contribution of ging nd oxygen on romtic nd nonromtic compounds nd sensory composition Reserch results Sensory interction studies combining vrietl nd oxidtion relted compounds tht occur in Suvignon blnc wines Reserch results A chemicl nd sensory study on the evolution of romtic nd non romtic compounds during the progressive oxidtive storge of Suvignon blnc wine Reserch results Acetldehyde nd its msking/enhncing bilities A sensory study in model wine Generl discussion nd conclusions x

11 Contents CHAPTER 1: Generl introduction nd project ims 1.1 INTRODUCTION PROJECT AIMS REFERENCES 5 CHAPTER 2: Literture Review Suvignon blnc wine: Contribution of ging nd oxygen on romtic nd non romtic compounds nd sensory composition 2.1 INTRODUCTION WHITE WINE OXIDATION Oxygen during wine processing Oxygen nd phenolics The role of ntioxidnts Effect of oxygen on white wine colour SAUVIGNON BLANC AROMA AND EVOLUTION DURING AGING AND OXIDATION Voltile thiols Methoxypyrzines Esters, higher lcohols, ftty cids Monoterpenes REDUCTIVE AROMA OXIDATION AROMA Aldehydes Other crbonyl compounds nd cetls SENSORY SCIENCE Descriptive nlysis Sensory interction studies CONCLUSION ABBREVIATIONS USED REFERENCES 39 ii

12 CHAPTER 3: Reserch results Sensory interction studies combining vrietl nd oxidtion relted compounds tht occur in Suvignon blnc wines 3.1 INTRODUCTION MATERIALS AND METHODS Medium Chemicls nd spiking Experimentl design Sensory nlysis Dt Anlysis RESULTS AND DISCUSSION Pnel evlution Individul compound evlution Multiple compound evlution CONCLUSION ABBREVIATIONS USED REFERENCES 92 CHAPTER 4: Reserch results A chemicl nd sensory study on the evolution of romtic nd non romtic compounds during the progressive oxidtive storge of Suvignon blnc wine 4.1 INTRODUCTION MATERIALS AND METHODS Oxidtion of the wine Smpling Voltile thiol nlysis Methoxypyrzine nlysis Ester, cid nd lcohol nlysis Monoterpene nlysis Sulphur dioxide nd scorbic cid nlysis Glutthione nlysis Phenolic nlysis Crbonyl compounds nd cetl nlysis Spectrophotometric mesurement 11 iii

13 Sensory romtic descriptive nlysis Sensory colour nlysis Dt nlysis RESULTS AND DISCUSSION Oxygen concentrtion Chemicl nlyses Voltile Thiols Methoxypyrzines Esters, cids nd lcohols Monoterpenes Sulphur dioxide nd Glutthione Polyphenols Aldehydes Sotolon Acetls Colour Combined chemicl content Sensory evlution Aromtic descriptive nlysis Sensory nlysis ccording to wine colour Chemicl nd sensory dt combined CONCLUSION ABBREVIATIONS USED REFERENCES 17 CHAPTER 5: Reserch results Acetldehyde nd its msking/enhncing bilities A sensory study in model wine 5.1 INTRODUCTION MATERIALS AND METHODS Medium Chemicls nd spiking Experimentl design Sensory nlysis Dt Anlysis 191 iv

14 5.3 RESULTS AND DISCUSSION Pnel evlution Individul compound evlution Multiple compound evlution CONCLUSION ABBREVIATIONS USED REFERENCES 28 CHAPTER 6: Generl discussion nd conclusions 6.1 GENERAL DISCUSSION AND CONCLUSIONS REFERENCES 218 v

15 Chpter 1 Generl introduction nd project ims 1

16 Chpter 1: Generl introduction nd project ims 1.1 INTRODUCTION PROJECT AIMS REFERENCES 5 2

17 1.1 INTRODUCTION South Africn Suvignon blnc wine hs distinctive sensory chrcteristics. Both fruity ( pssion fruit, guv, grpefruit, tropicl ) nd green ( green pepper, grssy, sprgus ) descriptors hve been used to describe the unique rom of this cultivr. Virtully nothing ws known bout the voltile compounds responsible for these ttributes until the mid 199s, except for the role of 3 isobutyl 2 methoxypyrzine in the green pepper chrcter thnks to the work of Augustyn nd Allen in South Afric nd Austrli (Augustyn et l., 1982; Allen et l., 1991). Tody, the compounds responsible for Suvignon blnc flvours hve been, for the most prt, identified. Vrious grpe nd fermenttion derived compounds contribute to the rom of Suvignon blnc wine. However, mny of the mechnisms responsible for the formtion of these compounds re still not fully understood. Compounds such s the voltile thiols nd methoxypyrzines re key rom compounds responsible for the typicl Suvignon blnc rom contributing to fruity nd herbceous nunces respectively, while other rom compounds such s esters, lcohols nd cids cn lso contribute to the wine rom (Swiegers et l., 26; Lund et l., 29). Oxidtion of white wines is constnt problem for winemkers worldwide. Oxidtion my drsticlly ffect wine qulity nd is identified by loss in plesnt rom together with n increse in unwnted rom nunces. In generl, it is ccepted tht certin grpe vrieties re especilly sensitive to oxygen, suggesting tht some of the chemicl components key to their sensory ttributes re strongly modulted by oxygen. Suvignon blnc is well documented exmple of n oxygen sensitive wine (Coetzee & Du Toit, 212). Understnding the stbility of vrious compounds is crucil in order to preserve the fresh nd fruity chrcters of Suvignon blnc wines over long period of time. The term oxygen mngement refers to one or multiple opertions in which well controlled mount of oxygen is delivered to the wine in order to chieve optiml expression of desirble sensory ttributes, however it is lso importnt to mintin the optiml chemicl composition, especilly ntioxidnt content. Obtining better understnding on this topic will give wine producers the upper hnd to compete on commercil scle nd plce on the mrket product with qulity tht is better nd more consistent thn their competitors. Any defect in wine qulity should thus be identified rpidly nd rectified if possible. In the cse of oxidtion, the prevention of defects by proper oxygen mngement is importnt. This study on oxygen mngement is not only linked to wht hppens to the wine during ging, but tkes holistic pproch to oxygen in wine, s we know tody tht the winemking process nd oxygen history of the wine is extremely importnt (Uglino, 213). It is essentil to understnd the chemistry behind oxygen exposure to control nd mnge the development of vrious chrcteristics in wine. Given the intrinsic diversity existing cross individul wines, the gret chllenge in this regrd remins 3

18 the bility to predict the tendency of wine to develop certin chrcters nd, in generl, to define how much oxygen wine might need to express improved sensory chrcteristics. There seems to be lck of informtion especilly on South Africn Suvignon blnc wine nd the rection of vrious chemicl compounds during oxidtive processes. To dte, no scientific literture hs focussed on reltively lrge rnge of rom compounds during oxidtion of South Africn wines nd the combintion of extensive chemicl dt with sensory profiling cn deliver vluble informtion regrding the evolution of these importnt compounds. A number of studies hve investigted the effect of enhnced oxidtion on the chemicl composition of white wines (Ferreir et l., 1997; Escudero et l., 22; Silv Ferreir et l., 23; Loscos et l., 21; Cejudo Bstnte et l., 213), but little is known bout the effect of controlled repetitive oxygen dditions on Suvignon blnc wine. A need lso exists to elucidte the interctive effect between Suvignon blnc rom impct compounds nd oxidtive derived voltiles t concentrtion rnges found in white wines. 1.2 PROJECT AIMS The objectives of this study were to improve the understnding of: The evolution of lrge rnge of compounds (romtic nd non romtic) during controlled Suvignon blnc wine oxidtion The sensory impct of controlled Suvignon blnc wine oxidtion nd how the sensory reltes to chemicl dt Sensory interctions between Suvignon blnc impct rom compounds nd rom compounds typiclly linked to oxidtion Other objectives to which the results from this study could contribute re to formulte possible mechnisms to improve Suvignon blnc wine qulity nd longevity to provide credible informtion nd integrl solutions on how to mnge the impct of oxygen on the wine, thereby improving the winemker s control over the finl wine profile through the use of oxygen mngement to reduce wine fults 4

19 1.3 REFERENCES Allen, M.S., Lcey, M.J., Hrris, R.L.N. & Brown, W.V., Contribution of methoxypyrzines to Suvignon blnc wine rom. Americn Journl of Enology nd Viticulture. 42, Augustyn, O.P.H., Rpp, A. & Vn Wyk, C.J., Some voltile rom components of Vitis Vinifer L. cv. Suvignon blnc. South Africn Journl of Enology nd Viticulture. 3, Cejudo Bstnte, M.J., Hermosín Gutiérrez, I. & Pérez Coello, M.S., 213. Accelerted ging ginst conventionl storge: Effects on the voltile composition of Chrdonny white wines. Food Chemistry. 78, Coetzee, C. & Du Toit, W.J., 212. A comprehensive review on Suvignon blnc rom with focus on certin positive voltile thiols. Food Reserch Interntionl. 45, Escudero, A., Asencio, E., Ccho, J. & Ferreir, V., 22. Sensory nd chemicl chnges of young white wines stored under oxygen. An ssessment of the role plyed by ldehydes nd some other importnt odornts. Food Chemistry. 77, Ferreir, V., Escudero, A., Fernnádez, P.E. & Ccho, J., Chnges in the profile of voltile compounds in wines stored under oxygen nd their reltionship with the browning process. Z. Lebensm. Unters. Forsch. 25, Loscos, N., Hernndez Orte, P., Ccho, J. & Ferreir, V., 21. Evolution of the rom composition of wines supplemented with grpe flvour precursors from different vrietls during ccelerted wine ging. Food Chemistry. 12, Lund, C.M., Thompson, M.K., Benkwitz, F., Wohler, M.W., Triggs, C.M., Grdner, R., Heymnn, H. & Nicolu, L., 29. New Zelnd Suvignon blnc distinct flvor chrcteristics: Sensory, chemicl, nd consumer spects. Americn Journl of Enology nd Viticulture. 6, Silv Ferreir, A.C., Hogg, T. & Guedes de Pinho, P., 23. Identifiction of key odornts relted to the typicl rom of oxidtion spoiled white wines. Journl of Agriculturl nd Food Chemistry. 51, Swiegers, J.H., Frncis, I.L., Herderich, M.J. & Pretorius, I.S., 26. Meeting consumer expecttions through mngement in vineyrd nd winery: the choice of yest for fermenttion offers gret potentil to djust the rom of Suvignon blnc wine. Austrlin nd New Zelnd Wine Industry Journl. 21, Uglino, M., 213. Oxygen contribution to wine rom evolution during bottle ging. Journl of Agriculturl nd Food Chemistry. 61,

20 Chpter 2 Literture review Suvignon blnc wine: Contribution of ging nd oxygen on romtic nd non romtic compounds nd sensory composition 6

21 Chpter 2: Literture Review Suvignon blnc wine: Contribution of ging nd oxygen on romtic nd non romtic compounds nd sensory composition 2.1 INTRODUCTION WHITE WINE OXIDATION Oxygen during wine processing Oxygen nd phenolics The role of ntioxidnts Effect of oxygen on white wine colour SAUVIGNON BLANC AROMA AND EVOLUTION DURING AGING AND OXIDATION Voltile thiols Methoxypyrzines Esters, higher lcohols, ftty cids Monoterpenes REDUCTIVE AROMA OXIDATION AROMA Aldehydes Other crbonyl compounds nd cetls SENSORY SCIENCE Descriptive nlysis Sensory interction studies CONCLUSION ABBREVIATIONS USED REFERENCES 39 7

22 2.1 INTRODUCTION Arom is n importnt fctor in the qulity of ll foods, but in wines the rom is probbly one of the most importnt spects contributing to the overll qulity. A lrge number of chemicl compounds with different voltilities nd polrities re responsible for the rom of wine (Arrehnius et l., 1996) nd during winemking nd ging vrious rections nd interctions occur tht cn influence the perception of the wine bouquet. One of the most importnt oenologicl problems in winemking is premture wine oxidtion, prticulrly the oxidtive spoilge of young white wines. With the exception of sherry nd sherry like wines, white wine qulity, in generl, will decrese with oxidtion (Singleton et l., 1979). This oxidtion cn tke plce in short mount of time during which significnt loss of fresh nd fruity rom tkes plce, followed by colour chnge s well s the development of unwnted oxidtion odours. Arom ttributes tht develop due to the formtion of new chemicl compounds hve been described s honey like, frmfeed, hy, woody like, tosted, dry fruits, crmel, overripe fruit, pple, oxidised pple, cetldehyde, cooked, ldehyde nd liquor (Toukis, 1974; Noble et l., 1987; Renouil, 1988; Hllidy & Johnson, 1992; Chrisholm et l., 1995; Escudero et l., 22; Silv Ferreir et l., 22b). These types of unplesnt descriptors re unwnted in wines nd winemkers go to gret lengths to void the formtion of the chemicl compounds responsible for these ttributes. Young wines contin high concentrtions of oxygen rective species, triggering chin of chemicl rections consequently ltering the wine content. Vrious closure permebility studies hve been performed to ssess the effect of oxygen nd ging on the qulity of young white wines (Godden et l., 21; Brjkovich et l., 25; Lopes et l., 25; Skouroumounis et l., 25). Generlly speking, the conclusion of these studies re tht less permeble closures llow better preservtion of the fresh nd fruity chrcter of the wine, presumbly by preventing the oxidtive loss of plesnt rom compounds (Lopes et l., 29). Too little oxygen exposure hs been ssocited with reductive off odours. Closures not llowing sufficient oxygen ingress could led to the formtion of the compounds contributing to the reductive rom (Godden et l., 21; Skouroumounis et l., 25; Kwitkowski et l., 27; Lopes et l., 29). Conversely, oxidtive loss of romtic compounds nd the evolution of undesirble rom compounds will occur with excessive oxygen exposure. Figure 2.1 shows the evolution of rom compounds during ging of wines exposed to low nd high oxygen concentrtions. 8

23 Figure 2.1 The effect of oxygen exposure on the rom of bottled wine (Uglino et l., 21). Reproduced with permission from The Austrlin Wine Industry Technicl Conference Inc. It cn be concluded tht moderte degree of oxygen exposure llows expression of optiml rom ttributes (Brjkovich et l., 25; Lopes et l., 29; Uglino et l., 29; Nygrd et l., 21), however the ctul degree of oxygen exposure tht is necessry to chieve this hs not been determined. Very few such studies hve ctully included precise oxygen mesurements (Brjkovich et l., 25; Lopes et l., 29; Uglino et l., 29; Nygrd et l., 21), thus still mking it difficult to define how much oxygen would be beneficil for wine during bottle ging. The type (vriety nd composition) nd style (winemking prctices such s wood contct) of the wine would lso ply crucil role in the mount of oxygen required to chieve the idel wine. While the chromtic chnges during wine ging re well documented, little is known bout the romtic deteriortion in reltion to white wine oxidtion, even though it seems to tke plce prior to discolortion (Li et l., 28). Other thn tht, studies investigting oxidtive spoilge of wines were often crried out under conditions ccelerting oxidtion such s high oxygen exposure or high tempertures (Ferreir et l., 1997; Escudero et l., 2; Escudero et l., 22). Conversely, oxidtive processes tking plce during ging re usully rther mild nd the significnce of such levels of oxidtion on wide rnge of chemicl compounds nd sensory profile of the wines remin to be estblished. The interctions between rom compounds present t ny one time in the wine during bottle ging could lso significntly influence the wine s chrcter by chnging the perception of certin ttributes nd should be investigted. This review will focus on the vrious wine constituents of typicl Suvignon blnc wine nd the rections thereof towrds oxidtion. Sensory spects nd interctions between romtic compounds will lso be discussed. 9

24 2.2 WHITE WINE OXIDATION Oxygen during wine processing During winemking, the grpe must nd wine re exposed to different levels of oxygen. This hs been described s mcro, micro nd nno concentrtions. Figure 2.2 provides theoreticl illustrtion of the oxygen concentrtion during wine processing (Uglino et l., 21), however studies hve shown tht these circumstnces re not lwys pplicble nd lrge spikes of oxygen cn be introduced lte in the process. Figure 2.2 Representtion of the idel oxygen mngement (Uglino et l., 21). Reproduced with permission from The Austrlin Wine Industry Technicl Conference Inc. Any oenologicl prctices involving ir will undoubtedly cuse oxygen dissolution nd wines sturted with oxygen will contin bout 8 mg/l oxygen t cellr tempertures nd tmospheric pressure (Singleton et l., 1985). Wine is cpble of consuming considerble mount of dissolved oxygen which hs been scribed to the totl phenol content, which explins why red wine cn consume more oxygen thn white wine (Rossi & Singleton, 1966). The effect of single oenologicl processes on the level of dissolved oxygen in wines cn be clssified s high enrichment nd low enrichment tretments (Cstellri et l., 24). A study done on over 5 red nd white wines hs identified high enrichment prctices s rcking, centrifugtion, refrigertion, continuous trtric stbiliztion nd bottling (Cstellri et l., 24). Dissolved oxygen concentrtion fter these tretments rnged from 1 mg/l to bout 8.5 mg/l with refrigertion nd cold stbiliztion rnked s the two tretments cusing the highest oxygen pickup. Low enrichment tretments re prctices like pumping, filtrtion, het exchnge nd electrodilysis nd cused n uptke of up to.6 mg/l with filtrtion cusing the most oxygen pickup (Cstellri et l., 24). During nd fter bottling, both wine nd the gseous hedspce in the bottle will contin substntil mount of oxygen. The totl pckged oxygen s the sum of these two components cn typiclly rech 1 9 mg/l (Uglino et l., 213) depending on the specific hndling of the wine (cre tken to exclude oxygen). In the cse of 1

25 some closures such s corks or similr products, dditionl oxygen will be relesed into the bottle fter bottling. The solubility of oxygen in wine is influenced by the wine composition (e.g. ethnol content), but depends primrily on the temperture nd the prtil pressure of the gs, with greter solubility t lower tempertures nd when pure oxygen is used insted of ir (Wterhouse & Lurie, 26). The rte of the oxidtion rections, however, increses with incresed temperture (Mrglit, 1997; Vivs de Gulejc et l., 21; Ribéreu Gyon, 26). The contct of wine with oxygen cn be minimised by the use of inert gsses such s nitrogen, crbon dioxide nd even rgon gs, which cn displce the ir in tnk or brrel Oxygen nd phenolics Phenolic compounds re chrcterised by n romtic ring contining one or more hydroxyl groups ( OH), including their functionl derivtives. The concentrtion of phenolics in white wine will depend on the vriety, cultivtion conditions, climte, grpe mturity, winemking techniques s well s ging conditions. The polyphenol content of white wines is substntilly lower thn in red wines due to different winemking techniques fvouring phenolic extrction during red wine production. Phenolics re strong hydrogen donting species which mkes them idel oxidtion substrtes (Wilderndt & Singleton, 1974). Phenolic molecules originting from grpes cn be divided into non flvonoids nd flvonoids. The nonflvonoids re grpe derived nd consist mong others of hydroxybenzoic cids, hydroxycinnmic cids nd stilbenes nd re normlly the principl phenolic molecules in white wines. Compounds such s the cffeic cid, p coumric cid nd ferulic cid nd their trtric cid esters re exmples of nonflvonoids nd re the min phenolic molecules in white wine tht did not receive prolonged periods of skin contct. Of these, trns cftric cid is the predominnt hydroxycinnmte ester in grpe juice nd wine, together with smller quntities of coutric nd fertric cid (Singleton et l., 1978; Singleton et l., 1984; Vrhovšek, 1998). The nturlly occurring trtric esters re susceptible to hydrolysis, liberting the corresponding free hydroxycinnmic cids. The hydroxybenzoic cids represent minor clss of white wine polyphenols of which gllic cid, vnillic cid nd syringic cid re few exmples. The mjor clsses of flvonoids in grpes nd wine re flvn 3 ols, flvonols nd nthocynins. Anthocynins re pigments significnt only in red grpe vrieties nd re in generl bsent from white wines. The min flvn 3 ols found in grpes re (+) ctechin nd ( ) epictechin s well s the glte ester ( ) epictechin 3 O gllte (Tsi Su & Singleton, 1969), while the flvonols minly consist of quercetin, kempferol nd myricetin (Mongs et l., 25; Ribéreu Gyon, 26). In grpe must, enzymtic oxidtion of phenolic compounds tkes plce due to the presence of oxidtion enzymes (e.g. polyphenol oxidse) which ctlyse the oxidtion process, while in wine, non enzymtic 11

26 chemicl oxidtion is the predominnt oxidtion rection due to the bsence or inhibition of oxidtion enzymes. Although polyphenol oxidtion is extremely slow (Oszminski et l., 1985), it ffects wide rnge of polyphenols, depending on their individul redox potentils (Kilmrtin et l., 22). The oxidtion process will depend on number of fctors such s oxygen concentrtion, temperture, presence of ctlysts, the nture nd composition of polyphenols, ph, ethnol content nd presence of ntioxidnts (Berg & Akiyoshi, 1956; Cilliers & Singleton, 1989; Kilmrtin et l., 21; Wterhouse & Lurie, 26). Oxygen in wine is in the unrective triplet stte, nd its bility to rect directly with most wine components is low (Wterhouse & Lurie, 26). The presence of ctlyst (prticulrly iron nd copper) increses the rection speed by donting n electron to oxygen resulting in superoxide ion which exists s hydroperoxyl rdicl t wine ph (Figure 2.3). The rdicl hs reltively low rectivity in the wine environment nd will rect with strong hydrogen donting species such s phenolic molecules (Wilderndt & Singleton, 1974). The rection of the superoxide ion with o diphenols forms H 2 O 2 nd o quinones t wine ph (Figure 2.3). Figure 2.3 Formtion of o quinones nd hydrogen peroxide nd the consequent oxidtion of ethnol to form cetldehyde nd other rections in the wine mtrix. Reprinted with permission from Uglino, M., 213. Copyright 213 Americn Chemicl Society. The H 2 O 2 cn rect with ferrous ions vi the Fenton rection to produce hydroxyl rdicl which is extremely rective nd cn rect with vrious wine constituents (e.g. lcohols, orgnic cids nd sugrs) in proportion to their concentrtion, cusing the formtion of ldehydes nd ketones. In this cse, ethnol will be primrily oxidised to produce cetldehyde (Wterhouse & Lurie, 26). The o quinone 12

27 is unstble nd very rective nd cn rect further with other molecules with lower redox potentils such s other phenolic molecules, SO 2 nd thiol contining compounds including glutthione nd mino cids. This rry of rections cn cuse substntil chnge in the composition of wine The role of ntioxidnts The most common ntioxidnts (other thn phenols) present in wine re sulphur dioxide (SO 2 ), scorbic cid nd glutthione (GSH). These compounds interfere in the phenol oxidtion process, either by removing oxygen from the wine or by reversing or ltering the oxidtion process. Sulphur dioxide is n effective nd low cost dditive for the preservtion of wines nd other food products (Doyle & Beucht, 27). In wine, it serves s the min preservtive to prevent oxidtion nd decrese microbil ctivity. Even though SO 2 occurs nturlly in ll wines s by product of yest metbolism during fermenttion (Rnkine & Pocock, 1969), it is usully dded t severl stges in the process of conventionl winemking such s during crushing, settling or fter primry nd secondry fermenttion (Pul, 1975). Unfortuntely, excessive use of SO 2 cn be not only detrimentl to orgnoleptic qulity of wine, but poses helth risk for sensitive consumers (especilly sthmtics) nd its presence requires mndtory lbel wrning sttements in most jurisdictions (Kleinhns, 1982). The equilibrium estblished upon dissocition of SO 2 in wine is shown in Figure 2.4. Figure 2.4 Distribution of SO 2 species t different ph vlues (Sneyd et l., 1993; Herbst, 21). Adpted with permission from The Austrlin Wine Industry Technicl Conference Inc. 13

28 Sulphur dioxide exists in wine in both free nd bound form (the sum equlling totl SO 2 ). At wine ph (ph 3 to 4), free SO 2 cn exist in three forms: moleculr SO 2, bisulphite (HSO 3 ) nd sulphite (SO 2 3 ). The equilibrium is ph dependnt nd ffected by the presence of wine constituents tht bind the bisulphite s well s wine temperture (Usseglio Tomsset, 1992). The moleculr form is minly responsible for ntimicrobil properties due to its bility to penetrte the cellulr membrnes of microorgnisms (Beech et l., 1979). However, t wine ph only smll proportion of the free SO 2 is in moleculr form, the predominnt form being bisulphite (94 99%) which cn bind lrge rnge of wine components, consequently producing bound SO 2 (Zoecklein et l., 1995; Oliveir et l., 22). The sulphite cn rect directly with oxygen, but is present t extremely low concentrtions t wine ph. Direct rection of bisulphite with oxygen is slow nd the ntioxidnt ctivity lies in the bility to reduce the H 2 O 2 to wter, convert o quinones bck to o diphenols nd rect directly with o quinones to form sulphonic cids (Figure 2.5) (Dnilewicz, 27). Figure 2.5 Proposed mechnism illustrting SO 2 nd GSH ntioxidnt protection in wine. Reprinted with permission from Uglino, M., 213. Copyright 213 Americn Chemicl Society. A number of crbonyl compounds (minly cetldehyde, pyruvic cid nd α keto glutric cid) cn bind with free SO 2 individully to form complex compounds (Jckowetz & De Orduñ, 213). The bisulphitecetldehyde ddition product ccounts normlly for the mjority of the bound SO 2 in wine. Rections with crbonyl compounds such s cetldehyde leds to the formtion of bisulphite ddition products in reversible rections, however other rections such s those between bisulphite nd o quinones form 14

29 stble dducts nd re irreversible (Lurie et l., 212). Nevertheless, there is generl trend in the wine industry towrds minimising SO 2 content nd pproprite substitutes nd supplements re populr topic in reserch. To dte, no single replcement for SO 2 hs been found tht combines ntimicrobil nd ntioxidnt chrcteristics. Glutthione is sulphur contining tripeptide (L γ glutmyl L cysteinyl glycine) nd nturlly occurring ntioxidnt from the grpes nd yest metbolism (Figure 2.6). The concentrtion in the must is influenced by the nitrogen uptke of the vine (Choné et l., 26) nd it is ccumulted in the berry t the onset of vérison (Adms & Liynge, 1993). Concentrtions of GSH in 28 young Suvignon blnc wines verged t 12.5 mg/l (Jnes et l., 21) nd winemking conditions promoting oxygen exposure led to decrese in GSH concentrtions while higher concentrtions hve been observed in juices treted reductively (Du Toit et l., 27; Mggu et l., 27). GSH hs n electron rich nucleophilic mercpto group which cn be spontneously substituted by 1,4 Michel ddition into the electrophilic centre of the o quinone formed during oxidtion. The product is thioether, 2 S glutthionyl cftric cid or Grpe Rection Product (GRP) (Figure 2.6). The formtion of GRP trps the o quinone thus preventing ny further rections tking plce (Kritzinger et l., 213). Glutthione cn lso rect with oxygen species such s H 2 O 2 (Anderson, 1998) to be oxidised to glutthione disulphide (Figure 2.6). It hs been rgued tht the disulphide cn lso be formed by reducing the o quinone bck to the o diphenol (Cilliers & Singleton, 199). This my explin incresing trns cftric nd coutric cid concentrtions in must fter GSH dditions s reported by Penn et l. (21). Figure 2.6 Moleculr structures of glutthione (A), glutthione disulphide (B) nd grpe rection product (C). 15

30 Glutthione is lso cpble of performing nucleophilic rections with other compounds such s ldehydes (Cheynier et l., 1986; Cheynier & Vn Hulst, 1988; Sonni et l., 211) nd hs protective bilities towrd importnt rom compounds such s esters, monoterpenes nd voltile thiols (Ppdopoulou & Roussis, 21; Lvigne Cruège & Dubourdieu, 23; Roussis et l., 27; Ppdopoulou & Roussis, 28). However, GSH hs lso been shown to induce production of H 2 S during wine ging in low oxygen conditions (Uglino et l., 211). Glutthione hs been linked to the rom potentil of Suvignon blnc grpes due to the identifiction of glutthionylted precursors for 3MH nd 4MMP (Peyrot des Gnchos et l., 22). The ddition of GSH to must or wine is not leglly permitted. However, yest extrcts (which contins GSH) clim to preserve the freshness nd increse the mouthfeel s well s the romtic complexity of white wines (Kritzinger et l., 213b). These extrcts re commercilly vilble. Ascorbic cid is nturlly present in grpes, but disppers rpidly following crushing nd is therefore mostly dded just prior to bottling by wine producers. It is ble to rect rpidly with the moleculr oxygen present in wine, thereby preventing oxidtive browning nd prolonging shelf life of white wines (Kielhöfer & Würdig, 1958). This rection produces dehydroscorbic cid s well s H 2 O 2. It is thus importnt to hve sufficient SO 2 present to rect with the formed H 2 O 2 in order to prevent further oxidtion rections from tking plce (Peng et l., 1998). Consequently, the complete replcement of SO 2 with scorbic cid s n ntioxidnt in white wine is not recommended nd scorbic cid should only be used s supplement to scvenge oxygen. The potentil of scorbic cid to recycle o quinones bck to o diphenols hs lso been suggested (Cilliers & Singleton, 199; Iscs & Vn Eldik, 1997; Dnilewicz, 23). However, recent studies showed no indiction of rpid interction between scorbic cid nd o quinones on the cyclic voltmmogrms (Mkhotkin & Kilmrtin, 29). The function of scorbic cid s n ntioxidnt hs been questioned s scorbic cid cn led to incresed browning during storge together with n ccelerted consumption of free SO 2 (Kielhöfer & Würdig, 196). The pro oxidtive ctivity of scorbic cid in white wines under certin conditions hs been reserched since (Peng et l., 1998). Ascorbic cid initilly functions s n ntioxidnt, however over time the presence of scorbic cid cn led to enhnced SO 2 consumption nd oxidtion of phenolic compounds resulting in browning (Peng et l., 1998; Brdshw et l., 21; Brdshw et l., 23; Brdshw et l., 24; Clrk et l., 28) Effect of oxygen on white wine colour The colour of white wine is one of the importnt qulity prmeters. Drk yellow or brown colour usully indictes oxidtion or spoilge of white wine. A positive correltion between totl phenolic content nd browning hs been reported (Simpson, 1982). However the content of hydroxycinnmic 16

31 cids in the wines correlted poorly with browning (Fernndez Zurbno et l., 1995). The hydroxycinnmic cids my, however, contribute to the browning through coupled oxidtion rections (Simpson, 1982; Fernndez Zurbno et l., 1995). The monomeric flvn 3 ols nd dimeric procynidins plys n importnt role in white wine colourtion with ( ) epictechin more positively correlted to the brown colour compred to (+) ctechin (Simpson, 1982). Browning in white wine cn be due to different mechnisms. Firstly, the oxidtion of phenolic molecules to their corresponding o quinones leds to further rections with phenolic compounds to produce dimers which pper to be more susceptible to oxidtion, nd thus ccelerte phenol polymeristion nd utoctlytic oxidtion in wine (Singleton, 1987). The formtion these polymers cn result in the formtion of more intensely coloured yellowbrown compounds (Es Sfi et l., 1999). The second mechnism is the oxidtive degrdtion of trtric cid leding to the formtion glyoxylic cid which cn medite condenstion rections of flvn 3 ols, potentilly contributing to browning in white wines. The formtion of yellow pigments in n oxygented model wine medium contining the flvn 3 ol (+) ctechin, trtric cid nd ferrous ions (Fe 2+ ) hs been observed (Oszminski et l., 1996). The glyoxylic cid rects with flvn 3 ol units (vi crboxymethine bridge) generting colourless nd yellow compounds in model wine. The mjor colourless product hs been identified s (+) ctechin dimer (linked by crboxymethine bridge) (Fulcrnd et l., 1997). The yellow pigments hve been identified s xnthylium slts deriving from dehydrtion nd oxidtion of the colourless crboxymethine linked dimer (Es Sfi et l., 1999). The bsorbnce mxim of xnthylium ctions re close to 42 nm, the stndrd wvelength to mesure browning of white wine. Other crbonyl substnces such s cetldehyde could lso potentilly medite the condenstion rection (Es Sfi et l., 1999; Lopez Toledno et l., 24; Drinkine et l., 25; Mongs et l., 25). 2.3 SAUVIGNON BLANC AROMA AND EVOLUTION DURING AGING AND OXIDATION Vitis vinifer L. cv. Suvignon blnc is grpe vriety ntive to Bordeux (Grves) nd the Loire Vlley (Sncerre nd Pouilly Fumé) (Adms & Liynge, 1993). This vriety is now widely cultivted in mny other wine growing regions cross the globe including South Afric, Austrli, New Zelnd, Chile nd the United Sttes (Jckson, 28). The grpes ripen to moderte to high levels of cidity, producing fresh, crisp nd dry white wines with pungent rom nd flvour (Cooper, 28). Generlly, South Africn Suvignon blnc wines re cold fermented in stinless steel tnks nd not exposed to ok. In this wy, the terroir (often delivering fresh nd fruity roms) is reflected in the wine. Depending on the climte, Suvignon blnc wines cn offer rnge of wine styles. On the one hnd, the wine cn deliver fresh nd fruity chrcters reminiscent of guv, grpefruit, gooseberries nd pssion fruit which usully originte from grpes grown in wrmer climte. On the other hnd the 17

32 more green style Suvignon blnc cn be produced from cooler grpe growing regions delivering rom nunces such s green pepper, grssy nd sprgus (Lund et l., 29b). The rom of Suvignon blnc is known to chnge drmticlly over period of just yer in the bottle, mening tht the wine should be drunk erly to experience the intended romtic bouquet of the wine (Herbst, 21). Vrious chemicl compounds contribute to the romtic composition of wine. Suvignon blnc impct compounds such s the voltile thiols nd the methoxypyrzines cn contribute significntly to the typicl chrcter of the wine. Other rom groups such s ester, lcohols, cids nd monoterpenes cn lso contribute to the plesnt wine rom. Wines developing reductive rom will typiclly hve negtive rom descriptors due to low oxygen exposure, while vrious oxidtion relted compounds such s ldehydes, lctones nd cetls will contribute to unplesnt rom. In the following sections detils regrding the most importnt rom contributing groups will be discussed Voltile thiols Thiols (more trditionlly referred to s mercptns) re sulphur contining compounds possessing sulfhydryl group ( SH). Vrious sulphur contining compounds occur in wine nd cn contribute to rnge of romtic nunces, depending on the type of compound. Certin thiols typiclly contribute to the fruity chrcters of wine nd ttributes such s grpefruit, pssion fruit, box tree nd gooseberry hve been used to describe the odour (Drriet et l., 1995; Toming et l., 1998). These compounds will be referred to s the voltile thiols. Other sulphur contining compounds cn contribute to reductive rom due to low oxygen content in certin wines nd the rom reminiscent of rotten egg, grlic nd cbbge occurs (Ruhut, 1993; Brjkovich et l., 25). Compounds such s furfurl nd 5 hydroxymethylfurfurl (5 HMF) (lso sulphur contining compounds) cn lso imprt significnt rom to wine. The contribution of the sulphur contining compounds other thn the voltile thiols, will be discussed in Chpter 2, sections 2.4 nd 2.5. Although first identified in Suvignon blnc wine during the 199s, the presence of voltile thiols hs lso been reported in wines mde from other vrieties (both red nd white) such s Riesling, Colombrd, Semillon, Cbernet Suvignon nd Merlot (Bouchilloux et l., 1998; Toming et l., 2; Murt et l., 21b). Along with methoxypyrzines, voltile thiols re considered to be impct odournts for Suvignon blnc wines due to the fct tht they form prt of the typiclity of the vriety nd re highly sought fter by consumers (Gonik & Noble, 1987; Ruhut, 1993; Toming et l., 24; Lund et l., 29b; King et l., 211). The voltile thiols plying n importnt role in Suvignon blnc wine rom re minly 4 mercpto 4 methylpentn 2 one (4MMP), 3 mercptohexn 1 ol (3MH) nd 3 mercptohexn 1 ol cette (3MHA) (Tble 2.1) (Drriet et l., 1995; Toming et l., 1996; Toming et 18

33 l., 2; Toming et l., 24; Coetzee & Du Toit, 212). Other voltile thiols hve been identified, however the concentrtion of these compounds in wines re usully below the perception threshold (Toming et l., 1998). Tble 2.1 Voltile thiols involved in Suvignon blnc rom (Coetzee & Du Toit, 212). ) Rnge in wines from Frnce nd New Zelnd (Toming et l., 1998; Ribéreu Gyon et l., 26; Lund et l., 29b; Mteo Vivrcho et l., 21) b) Perception threshold is the minimum detectble concentrtion for 5% of the tsters; these vlues hve been tested in model wine solution (Toming et l., 1996; Toming et l., 1998b; Dubourdieu et l., 26) c) Drriet et l. (1995) d) Toming et l. (1998); Dubourdieu et l. (26) e) Toming et l. (1996); Dubourdieu et l. (26) 3MHA hs been described s sweet swety pssion fruit, grpefruit, box tree, gooseberry nd guv (Toming et l., 1996; Nicolu et l., 26) nd hs very low perception threshold of 4.2 ng/l in model wine solution (Toming et l., 1996). 3MH hs been described s pssion fruit, grpefruit, gooseberry nd guv (Swiegers et l., 25; Nicolu et l., 26; Vn Wyngrd, 213) with perception threshold of 6 ng/l (Toming et l., 1998), while 4MMP hs been described s box tree, pssion fruit, broom nd blck current bud (Drriet et l., 1995) with perception threshold of.8 ng/l in model wine solution (Toming et l., 1998). The low perception thresholds reported mkes these compounds potent romtic contributors in wine. In survey done on 24 South Africn Suvignon blnc wines (211 vintge) the verge thiol concentrtions were 1.1 ng/l, ng/l nd 969 ng/l for 4MMP, 3MHA nd 3MH respectively (Vn Wyngrd, 213). These concentrtions were in line with those found in other studies investigting Suvignon blnc wines from ll over the world including New Zelnd, Austrli, Frnce nd Chile (Benkwitz et l., 212b), with the exception of 3MH for which the concentrtions were lower (Lund et l., 29b; Benkwitz et l., 212b). 19

34 Unlike the methoxypyrzines which re present s such in the grpes, the voltile thiols (4MMP nd 3MH) re thought to be in prt relesed by the yest from odourless, non voltile precursors during fermenttion (Drriet et l., 1995; Toming et l., 1996). Some precursors in the juice hve been identified s being cysteinylted or glutthionylted precursors. However, studies hve shown tht these precursors only ccount for frction of the totl mount of thiols present in the wine (Drriet et l., 1995; Toming et l., 1998c; Peyrot des Gnchos et l., 22; Subileu et l., 28; Fedrizzi et l., 29), moreover no direct correltion between precursor concentrtion nd the mount of free voltile thiols in the resulting wines hve been observed (Cpone et l., 21; Rolnd et l., 211). Other thn the relese from precursors, the biogenesis of voltile thiols hs been proposed. The direct ddition of H 2 S or nother sulphur donor to conjugted crbonyl compounds such s (E) 2 hexenl nd mesityl oxide followed by reduction step could be nother source of voltile thiol formtion (Schneider et l., 26). The formtion of the voltile thiols is not fully understood, since the min precursor hs yet to be elucidted nd the production of these voltile thiols needs further investigtion. 3MHA is formed by the esterifiction of 3MH with cetic cid during fermenttion. The finl concentrtion of 3MHA (nd other fermenttive esters) depends on the blnce of ctivities of lcohol cetyltrnsferse (promoting estrifiction of the corresponding lcohol) nd esterse (promoting hydrolysis). Yest strins differ in their bility to relese the voltile thiols from their precursors nd lso in their bility to convert 3MH to 3MHA (Murt et l., 21; Swiegers et l., 25; Swiegers et l., 26; Anfng et l., 29). 3MHA typiclly mounts to up to 1% of the concentrtion of 3MH (Toming et l., 2; Dubourdieu et l., 26). The formtion of voltile thiols cn be mnipulted through both viticulturl nd oenologicl opertions. Reserch on the impct of viticulturl prctices on the production of voltile thiol precursors is limited nd under question due to the min mechnisms of voltile thiol formtion from precursors not yet being identified. The effect of fctors such s nitrogen fertiliztion, wter vilbility nd Botrytis cinere infection on certin precursors hs been reported (Choné, 21; Peyrot des Gnchos et l., 25; Choné et l., 26; Srrzin et l., 27; Thibon et l., 29). During winemking, vrious processes cn lso be dpted to mximise the precursor extrction nd/or formtion s well the libertion of the thiol from the precursor. Prctices such s mechnicl hrvesting, skin contct, pressure during pressing nd oxygen exposure could significntly increse the precursor content in the juice s well s the voltile thiol concentrtion in the corresponding wines (Mggu et l., 27; Ptel et l., 21; Rolnd et l., 21; Cpone & Jeffery, 211). The libertion of the voltile thiols from the respective precursors depends on conditions such s yest strin, nitrogen vilbility nd fermenttion temperture (Murt et l., 21; Msneuf Pomrède et l., 26; Swiegers et l., 26; Subileu et l., 28b; Anfng et l., 29; Hrsch et l., 213). More detil regrding the influence of these processes on the voltile thiol concentrtions cn be obtined in review rticle published in 212 (Coetzee & Du Toit, 212). 2

35 The voltile thiols re prticulrly susceptible to oxidtion during ging (Murt et l., 23; Blnchrd et l., 24) nd vrious pckging nd ging studies hve reported the decrese of the voltile thiols during storge (Murt, 25; Lopes et l., 29; Herbst Johnstone et l., 211; Ghidossi et l., 212). Three mechnisms hve been identified vi which the thiol content cn decrese (Figure 2.7). Voltile thiols cn oxidise esily in the presence of oxygen nd iron to form the corresponding disulphides (rection 1) (Jocelyn, 1972; Kotserides et l., 2). Furthermore, these thiols re nucleophilic nd cpble of ddition rections with electrophiles such s polymeric phenolic compounds (Ribéreu Gyon, 1998, 24b) nd prticipte in chemicl rections (1,4 Michel type ddition) with products of phenolic oxidtion such s o quinones (rection 8) (Herbst et l., 28; Nikolntonki et l., 21). The formed dducts re non voltile nd would cuse loss in vrietl chrcter. Recent observtions reported the o quinone trpping s the min mechnism ccounting for 3MH loss in wine under oxidtive conditions, while other rections seems to contribute mrginlly (Kreitmn et l., 213). Figure 2.7 Possible mechnistic pthwys for the the degrdtion of the voltile thiols in wine under oxidtive conditions. Reprinted with permission from Nikolntonki et l., 21. Copyright 21 Americn Chemicl Society. The rte of these rections re ph dependnt, since t wine ph the concentrtion of the thiolte ion (RS ), which is more rective thn its protonted form (RSH), re low with pk vlues for thiols being typiclly between 9 nd 12. The greter proportion of thiolte nions t higher ph my explin the consistently lower levels of 3MH in these conditions (Blnchrd et l., 24). 21

36 The decrese in especilly 3MHA cn lso be ttributed to cid ctlysed hydrolysis t wine ph (Figure 2.8) (Toming et l., 24). The ester structure of 3MHA mkes it susceptble to this type of rection. Voltile cette esters re introduced enzymticlly to the wine during fermenttion by the ction of yest vi combintion of cetyl CoA with n lcohol which is ctlysed by lcohol cetyltrnsferse. The products of this rection re 3MH nd cetic cid. The hydrolysis rection is expected to be ccelerted by higher temperture, but not directly ffected by oxidtive conditions (Mkhotkin et l., 212). Figure 2.8 Hydrolysis of 3 mercptohexn 1 ol cette to 3 mercptohexn 1 ol nd cetic cid (Herbst Johnstone et l., 211). Acid hydrolysis ws found to hve stronger influence thn o quinone trpping on the decrese of 3MHA concentrtions during bottle ging of Suvignon blnc wine under highly hermetic conditions such s screw cps (Herbst et l., 28; Herbst Johnstone et l., 211). As 3MH hs higher perception threshold thn 3MHA, the hydrolysis is expected to result in lower romtic intensity nd possibly different rom. The presence of SO 2 ws shown to prevent the loss of voltile thiols due to its bility to recycle the o quinones to the originl phenol, bind them directly, or reduce H 2 O 2 to wter (Dnilewicz et l., 28; Nikolntonki et l., 21; Lurie et l., 212). However, the formtion of the o quinone dducts ws shown to occur even in the presence of SO 2, lthough to lower extent (Nikolntonki et l., 212) Methoxypyrzines Methoxypyrzines re nitrogen contining ring structures nd secondry plnt metbolites which re responsible for rom descriptors such s green pepper, sprgus, grssy, herbceous nd vegettive. The romticlly potent lkyl methoxypyrzines re the min compounds thought to be responsible for these flvours in especilly Suvignon blnc wines. Three min methoxypyrzines exist in wines, nmely 3 isobutyl 2 methoxypyrzine (IBMP), 3 isopropyl 2 methoxypyrzine (IPMP) nd 3 secbutyl 2 methoxypyrzine (SBMP) (Figure 2.9) (Allen et l., 1991; Lcey et l., 1991; Mris, 1994). These compounds hve very low odour threshold vlue of 1 2 ng/l in wter (Buttery et l., 1969; Allen et l., 1991; Lcey et l., 1991; Mris, 1994) nd surveys done on South Africn Suvignon blnc wines 22

37 reported the IBMP (the most importnt methoxypyrzine) concentrtions to be in the rnge of 1.2 to 4 ng/l (Alberts et l., 29; Vn Wyngrd, 213). SBMP is rrely detected in Suvignon blnc wines, wheres IPMP cn occur t concentrtions bove its perception threshold (up to 13.7 ng/l) (Lubbers et l., 1994; Guyot et l., 1995; Kwitkowski et l., 27). Figure Isobutyl 2 methoxypyrzine (IBMP) (1), 3 isopropyl 2 methoxypyrzine (IPMP) (2) nd 3 sec butyl 2 methoxypyrzine (SBMP) (3). The herbceous notes of methoxypyrzines re usully ssocited with under ripe grpes nd re undesirble for most vrieties, however Suvignon blnc seems to be the exception. The formtion of these compounds in the vine hs not been fully elucidted (Ebeler & Thorngte, 29). The structure would suggest their origin s secondry product during the metbolism of mino cids (Mris, 1994; Swiegers et l., 26; Ebeler & Thorngte, 29). The methoxypyrzine concentrtion in the berry evolves during the mturtion with the mximum concentrtion found t vérison. After this, the concentrtions decrese with ripening (Lcey et l., 1991). Elevted levels of these compounds re normlly found in cooler climtes (Lcey et l., 1988; Green et l., 211). Wrmer climtes usully deliver lower concentrtions which is thought to be due to incresed light exposure during ripening (methoxypyrzines re light sensitive) (Hshizume & Smut, 1999; Ryon et l., 28) in these res s well s the elevted verge tempertures (Flco et l., 27). In generl, the methoxypyrzine content is more ffected by the viticulturl prctices thn by the vinifiction processes (Mris, 1994; De Boubee et l., 22; Hunter et l., 24; Mggu et l., 27). The methoxypyrzines re highly extrctble from Suvignon blnc grpes. Skin contct hs greter impct on the IBMP concentrtion compred to the pressure pplied during pressing (Mg, 1989). However, settling of the juice before fermenttion cn led to decrese in methoxypyrzine concentrtions (Kotserides et l., 28), while concentrtions hve been reported to remin stble during fermenttion (Sl et l., 24). The compounds re generlly resilient to stndrd wine fining prctices (Pickering et l., 26) nd during ging concentrtions did not differ when exposed to vrious ging conditions such s light exposure nd temperture vritions (Blke et l., 21). Even 23

38 hyperoxidtion of musts nd wine did not lter the methoxypyrzine content (Mris, 1998; Coetzee et l., 213). Surprisingly, poor correltion ws obtined when compring methoxypyrzine concentrtion with the romtic intensity of the green pepper ttribute. The contribution of other romtic compounds such s the vrietl thiols nd dimethyl sulphide to these types of ttributes cn be significnt in Suvignon blnc wines nd help explin the wek correltion observed previously (Prk et l., 1994; Mris et l., 1998; Lund et l., 29b; King et l., 211; Vn Wyngrd, 213) Esters, higher lcohols, ftty cids Esters constitute one of the most importnt clsses of rom compounds nd re lrgely responsible for the fruity roms ssocited with wine (Lilly et l., 2; Swiegers et l., 26). Esters re minly produced during fermenttion by the condenstion of n lcohol nd coenzyme A ctivted cid, by the ction of lcohol cetyl trnsferse (Lmbrechts & Pretorius, 2). A lrge vriety of esters cn be formed s ll of the lcohols (especilly ethnol) nd ftty cids my rect to form esters. Acette esters of the higher lcohols nd the ethyl esters of stright chin, sturted ftty cids re the most significnt esters produced in wine (Lmbrechts & Pretorius, 2). Some of the most quntittively significnt esters in wine hve been identified to be isomyl cette, ethyl hexnote nd 2 phenylethyl cette (Thurston et l., 1981). Ester concentrtion cn decrese during ging due to chemicl hydrolysis (Mris, 1978; Ferreir et l., 1997; Lmbropoulos & Roussis, 27) or oxidtion by hydroxyl rdicl oxidtion relted processes (Litchev, 1989; Escudero et l., 2) which cn both led to loss in the fruity chrcter of wine. The hydrolysis is fvoured t elevted tempertures nd low ph vlues (Rmey & Ough, 198). The decline in especilly the cette esters contributes to the loss of freshness nd fruitiness in white wines during bottle ging. Acette esters of higher lcohols tend to diminish more rpidly during ging compred to ethyl esters of ftty cids. Ethyl esters re thought to be close to their chemicl equilibrium in young wines (Simpson, 1978) nd the hydrolysis tht occurs during storge hppens reltively slowly (Rmey & Ough, 198). Higher lcohols re lso product of lcoholic fermenttion nd cn be importnt precursors for ester formtion (Soles et l., 1982). At concentrtions below 3 mg/l they generlly contribute to the complexity of the wine rom (Rpp & Mndery, 1986). However, t higher concentrtions the rom cn become too intense nd cn contribute to strong pungent smell nd tste (Nykänen, 1986), thus possibly msking other rom contributors. Higher lcohols cn be nboliclly synthesized from intermedites of the sugr metbolism or ctboliclly synthesized from brnch chin mino cids, through the Ehrlich pthwy (Nykänen, 1986; Boulton et l., 1996; Dickinson et l., 1997; Dickinson et l., 23). During ging, lcohols cn be oxidised to form ldehydes (Mris & Pool, 198), cusing the concentrtion to decrese, however mny studies report stble lcohol concentrtions during the ging 24

39 of wines (Mris, 1978; Roussis et l., 25; Roussis et l., 27; Blke et l., 29). Unlike other higher lcohols, hexnol concentrtion cn increse during storge, which is probbly due to the oxidtion of linoleic nd linolenic cids (Oliveir et l., 26). Ftty cids, of which the most bundnt re cetic, hexnoic, octnoic nd decnoic cid contribute to the fresh flvour of wine. However, t too high concentrtions ftty cids it cn led to unwnted flvours ssocited with rncid, cheesy nd vinegr rom (Schreier, 1979; Lmbrechts & Pretorius, 2). Medium chin ftty cids, such s hexnoic, octnoic nd decnoic cid re produced by the yest s intermedites in the biosynthesis of long chin ftty cids. The hydrolysis of ethyl esters during ging cn led to n increse in the corresponding cid. However, n increse in these cids hs not lwys been observed s studies hve shown tht the voltile ftty cid stbility is not uniform s some compounds increse while others decrese or remin stble during ging (Roussis et l., 25; Câmr et l., 26; Blke et l., 29; Lee et l., 211). Mny fctors cn influence the formtion nd degrdtion of esters, higher lcohols nd ftty cids during fermenttion, ging or oxidtion. Fermenttion conditions such s temperture, juice clrifiction, yest strin s well s other prmeters such s oxygen exposure nd SO 2 dditions re some of the most importnt fctors (Bertrnd, 1968; Dudt & Ough, 1973; Mris, 21; Grde Cerdán & Ancín Azpilicuet, 27; Coetzee et l., 213) Monoterpenes Monoterpenes re known for their florl, fruity, citrus nd perfume odours usully expressed by gerniol, linlool, nerol nd α terpineol (Mris, 1983). Suvignon blnc cn be clssified s member of intermedite clss between monoterpene dependnt florl grpes nd monoterpene deficient nonflorl grpes (Benkwitz et l., 212). Most terpenes increse during ripening (Byonove & Cordonnier, 1971) with some studies reporting decrese t the overripe stge (Versini et l., 1981). A considerble proportion of these compounds is in the bound form in the juice nd is relesed during fermenttion by the yest, however some cn become romtic by chemicl rerrngement (Willims et l., 198; Loscos et l., 27). During ging, wines re known to lose some of the florl roms ssocited with monoterpenes (Rpp & Mndery, 1986; Rpp, 1988). Linlool specificlly is known to decrese during storge (Ferreir et l., 1997; Lmbropoulos & Roussis, 27) while α terpineol initilly incresed (probbly due to the oxidtion of other terpenols) nd then decresed t lter stge (Ferreir et l., 1997). Terpenes re sensitive to cidic conditions, storge time nd temperture nd cn be trnsformed into other compounds which could contribute to different rom nd hve different perception thresholds (Ppdopoulou & Roussis, 21; Roussis et l., 25; Roussis et l., 27). The importnce of monoterpenes in wines is ccentuted by the fct tht 25

40 they ct synergisticlly in wine medium nd cn thus influence the romtic composition of wine (Ribéreu Gyon et l., 1975). 2.4 REDUCTIVE AROMA Reductive rom properties hve been described s rotten egg, cbbge nd grlic. The compounds responsible for these rom ttributes usully form fter period in the bottle. The occurrence of these odours hs been ttributed to the presence of low moleculr weight sulphur compounds. Hydrogen sulphide (H 2 S) nd methyl mercptn (MeSH) hve been identified to be primrily responsible for the post bottling reduction (Lopes et l., 29; O'Brien et l., 29; Uglino et l., 211; Uglino et l., 212). H 2 S cn ccumulte during bottle ging nd concentrtions 3 4 times higher thn its perception threshold hs been reported fter 6 months in the bottle (Uglino et l., 211). The mechnisms for the formtion of the reductive compounds re not fully understood, however the formtion of H 2 S from cysteine precursor in the presence of dicrbonyl compound s well s the direct reduction of sulphte or sulphite hve been suggested (Pripis Nicolu et l., 2; Lopes et l., 29). Interestingly, studies hve shown n increse in the formtion of reductive odours in wines tht were treted with copper sulphite t bottling, which is common prctice mongst winemkers to prevent the formtion of these odours (Uglino et l., 211). H 2 S cn lso prticipte in other rections such s the rection with benzldehyde to form methyl mercptn which cn contribute to the reductive rom of the wine with smokey/empyreumtic, cbbge nd sewge nunces, however this hs not been conclusively demonstrted (Toming, 23; Toming et l., 23; Uglino, 213). Low levels of oxygen exposure during bottle ging correspond to lrger ccumultion of H 2 S nd MeSH (Lopes et l., 29; Uglino et l., 211; Uglino et l., 212) with the ccumultion of MeSH minly occurring between the first 6 to 12 months of bottle storge (Uglino et l., 212). Dimethyl sulphide (DMS) lso increses during bottle ging nd cn hve n importnt contribution to wine rom (Segurel et l., 25; Escudero et l., 27). The formtion of DMS does not seem to be ffected by oxygen exposure (Uglino et l., 212), lthough decrese in DMS hs been reported in the presence of excess oxygen (Silv Ferreir et l., 23d; Fedrizzi et l., 211). Oxygen exposure will lso led to decrese in H 2 S nd MeSH concentrtions. Oxidtion of mercptns to the corresponding disulphides hs been proposed, but not proven (Limmer, 25) nd recent dt did not find direct reltionship between the mercptn nd the disulphides (Nguyen et l., 21; Uglino et l., 212). The rection of reductive compounds with o quinones resulting from phenolic oxidtion could explin the decrese of especilly H 2 S nd MeSH in n oxygented wine due to the high rectivity of H 2 S towrd o quinones in wine like 26

41 solutions (Nikolntonki & Wterhouse, 212). The presence of oxygen t the during the erly stges of bottle ging will thus prevent the formtion of these compounds, possibly t precursor level, however this needs further investigtion (Uglino, 213). 2.5 OXIDATION AROMA Erly work on oxidtive spoilge of white wines indicted tht oxidtion brought bout sensory chrcters described s honey, frmfeed, woody nd cooked vegetbles (Escudero et l., 22; Silv Ferreir et l., 23c). A number of trce rom compounds hve been identified s mjor contributors to these rom ttributes, including vrious ldehydes, lctones nd cetls (Escudero et l., 2; Escudero et l., 2b; Escudero et l., 22; Silv Ferreir et l., 22b; Silv Ferreir et l., 23c). The ldehydes re especilly importnt due to their possible impct on wine romtic composition (Escudero et l., 2b; Silv Ferreir et l., 23c; Culleré et l., 27). The vrious chemicl compounds potentilly contributing to wine oxidtion roms will be discussed in the following sections Aldehydes Acetldehyde is one of the most importnt crbonyl compounds in wine nd constitutes up to 9 % of the totl mount of ldehydes found in wine (Nykänen, 1986). Its orgnoleptic influence nd its bility to combine rpidly with SO 2 even t low tempertures mkes this compound one of the criticl mrkers during winemking (Burroughs & Sprks, 1973). Acetldehyde is formed by the yest during lcoholic fermenttion (Mrglit, 1997) nd cn lso originte from microbil ctivity of other microbes such s lctic cid bcteri nd cetic cid bcteri (Drysdle & Fleet, 1988; Liu & Pilone, 2). However, winemking prctices tht enhnce cetldehyde formtion re usully post fermenttion nd cn led to moderte to importnt cetldehyde increses (Jckowetz & De Orduñ, 213). The most importnt production of cetldehyde is during the oxidtion of ethnol (Wilderndt & Singleton, 1974; Ribéreu Gyon, 1998, 24b). This rection is not direct, but rther vi coupled utooxidtion of certin phenolic compounds (Figure 2.3) (Wilderndt & Singleton, 1974) nd the formtion will depend on the mount of oxygen present (Kielhöfer & Würdig, 196; Schneider, 23, 25). Free SO 2 present in the wine will prevent this oxidtion by recting with intermedite oxidtion products s well s the formed cetldehyde resulting in n odourless sulphite combintion, known s hydroxysulphonte, which is stble in the cid medium (Figure 2.1) (Wterhouse & Lurie, 26). 27

42 CH 3 CHO + HSO 3 CH 3 CHOH SO 3 Figure 2.1 Rection of cetldehyde with bisulphite The rection between cetldehyde nd bisulphite is rpid nd t ph of 3.3, 98% of the cetldehyde will be combined with the sulphite within 9 minutes. It hs been estimted tht only.4% of cetldehyde is in the free form in the presence of 3 mg/l free SO 2 (Blouin, 1965). Thus only when ll the free cetldehyde hs been consumed cn SO 2 exist in the free form in wine. Acetldehyde hs been identified s the most importnt SO 2 binder in study screening 237 tble wines nd ccounted for 55% to over 7% of the estimted bound SO 2 for red nd white wines respectively (Jckowetz & De Orduñ, 213). The production of cetldehyde will thus effectively cuse decrese in free SO 2 concentrtion with the subsequent increse in bound SO 2. Due to the decresed preservtive ctivity of the bound form (Rnkine, 1968), the ddition of further SO 2 will be required to efficiently mintin ntimicrobil nd ntioxidnt properties. Acetldehyde lso redily binds to proteins nd especilly to GSH or individul mino cids (Liu & Pilone, 2). Cysteine dditions reduced cetldehyde concentrtions in beer (Suovniemi et l., 26) nd the use of other sulphur contining compounds s dditives to bind cetldehyde in beverges should be investigted. Acetldehyde concentrtion (free nd bound) in dry white wines hs been reported to rnge between 7 nd 24 mg/l verging t 4 mg/l (Lopes et l., 29; Jckowetz & De Orduñ, 213), while the highest concentrtions were reported in fortified wines rnging from 12 8 mg/l (Lchenmeier & Sohnuis, 28). Acetldehyde in the free form is key impct rom compound for sherry wines nd cn rech high concentrtions in these types of wines (Mrtínez et l., 1998) due to the bsence of free SO 2. Odours ssocited with the presence of free cetldehyde hve been described s green pple, overripe bruised pple, grssy, pungent, nutty nd sherry (Mrglith, 1981; Henschke & Jirnek, 1993; Miyke & Shibmoto, 1993; Frivik & Ebeler, 23). At low concentrtions, the presence of free cetldehyde could contribute to the plesnt fruity rom of wine, however t higher concentrtions the typicl oxidtion relted nunces will develop (Miyke & Shibmoto, 1993; Ze et l., 21). The bound form reduces the sensory effect of cetldehyde (Jckowetz et l., 211) nd it is thus recommended to mintin positive level of free SO 2 to ensure fixtion of cetldehyde (Somers & Wescombe, 1987). Other ldehydes cn contribute to honey, boiled vegetble nd rotten potto nunces in oxidised wines (Escudero et l., 2b; Culleré et l., 27). Of these ldehydes, the Strecker ldehydes, 3 (methylthio) propionldehyde (methionl) nd phenylcetldehyde, hve been identified s mjor contributors (Figure 2.11) (Escudero et l., 2; Escudero et l., 2b; Escudero et l., 22; Silv Ferreir et l., 22b; Silv Ferreir et l., 23c). The formtion of these ldehydes hs been proposed 28

43 to occur vi the Strecker rection of dicrbonyl compounds with mino cids, methionine nd phenyllnine to form methionl nd phenylcetldehyde respectively (Figure 2.12) (Escudero et l., 2b; Pripis Nicolu et l., 2; Silv Ferreir et l., 22b; Rizzi, 26). Wine contins severl types of dicrbonyl compounds. Dicrbonyls (such s o quinones) cn be formed vi oxidtion of phenolic molecule, while other dicrbonyls such s dicetyl, glyoxl nd methylglyoxl originte from microbil origin (Pripis Nicolu et l., 2). The formtion of ldehydes vi the Strecker rection in wine hs been brought under question s the rection rte between 4 methyl 1,2 benzoquinone nd the α mino cids in synthetic wine medium ws essentilly zero (Nikolntonki & Wterhouse, 212), compred to neutrl queous solution without dded ethnol t 22 C, conditions tht re different from the cidic wine mtrix (Rizzi, 26). The ph of the medium could hve n importnt role in controlling the rte of α mino ddition (Youngblood, 1986; Modic et l., 21). A B Figure 2.11 Strecker ldehydes. A) Methionl; B) Phenylcetldehyde 29

44 Figure 2.12 Mechnism of formtion of brnched chin ldehydes in wines vi (A) Strecker degrdtion of mino cids involving o quinones derived from phenolic oxidtion, (B) Strecker degrdtion involving dicrbonyl compounds of microbil origin, or (C) oxidtion of higher lcohols by hydroxyl rdicl. Reprinted with permission from Uglino, M., 213. Copyright 213 Americn Chemicl Society. The formtion of ldehydes from the oxidtion of relted fermenttion derived lcohols (phenylethnol nd methionol) hs lso been suggested (Escudero et l., 2; Escudero et l., 2b; Pripis Nicolu et l., 2; Silv Ferreir et l., 22b; Rizzi, 26; Loscos et l., 21; Fedrizzi et l., 211; Nikolntonki & Wterhouse, 212) nd good correltion between the lcohol nd corresponding ldehyde hs been observed in ged wines (Sn Jun et l., 212). Such mechnism would be nlogous to the one generting cetldehyde from ethnol. Methionl nd phenylcetldehyde hve perception thresholds of.5 µg/l nd 1 µg/l in synthetic wine medium respectively (Escudero et l., 2b; Culleré et l., 27) nd concentrtions in young Suvignon blnc wine just fter bottling were found to be below.5 nd 5 µg/l respectively. After 18 months ging t 2 C, the concentrtions incresed to bout 5.5 nd 34 µg/l respectively (Ghidossi et l., 212). Higher concentrtions of these compounds hve been reported in more extreme storge conditions such s higher dissolved oxygen concentrtions nd higher storge tempertures (Silv Ferreir et l., 23c). 3

45 The formtion of methionl nd phenylcetldehyde re gretly ffected by temperture nd the dissolved oxygen concentrtion (Silv Ferreir et l., 22b). However, both fctors re not required for the formtion of the ldehydes s it ws found in beer tht the ldehydes formed in the bsence of oxygen (<.2 mg/l) (Sores d Cost et l., 24). Oxidtion of the lcohols s well s the Strecker degrdtion rection will be prevented by the presence of SO 2 (Silv Ferreir et l., 23c), probbly due to its bility to reduce H 2 O 2 to wter, convert the o quinone bck to the o diphenol nd lso reversibly recting with the microbil derived dicrbonyls. Thus SO 2 should effectively be ble to prevent the formtion of these ldehydes. Furfurl nd other furnic ldehydes such s 5 methylfurfurl (5 MF) nd 5 hydroxymethylfurfurl (5 HMF) usully originte from the heting process of ok during brrel tosting (Moutounet et l., 1989). The therml degrdtion of polyscchrides produces these furnic ldehydes from crbohydrte polymers. The formtion of furfurl due to the non oxidtive decy of scorbic cid hs lso been reported (Buernfeind & Pinkert, 197; Yun & Chen, 1998; Wllington et l., 213), however in the cse of ged wines (especilly those not contining scorbic cid), the occurrence of these compounds is probbly due to the degrdtion of crbohydrtes (Câmr et l., 24). A good correltion between ging time nd the concentrtion of these furnic compounds were found in Port nd Mdeir wines (Silv Ferreir et l., 23; Câmr et l., 24). Ferreir et l. (1997) lso reported correltion between furfurl concentrtion nd wine browning during ging. The concentrtion of furfurl nd 5 MF in rnge of white wines oxidised for 1 week ws found to be between µg/l nd 1 6 µg/l respectively (Escudero et l., 22). Perception thresholds of 15 mg/l, 2 mg/l nd 1 mg/l hve been reported for furfurl, 5 MF nd 5 HMF respectively (Meilgrd, 1975; Câmr et l., 24) which would suggest minor orgnoleptic contribution due to the concentrtions being below the perception threshold (Ribéreu Gyon, 26). However, the contribution of these compounds to woody or mderized nunces during the oxidtion of white wines hs been reported nd could explin the perception of these odours in wines tht did not receive ny wood contct (Escudero et l., 22; Cmpo et l., 28). Benzldehyde is lwys present in wine nd is produced by lcoholic fermenttion in concentrtions up to.5 mg/l. During wine ging, benzldehyde formtion hs been ttributed to phenyllnine oxidtion (Loyux et l., 1981), while other uthors reported its origin from mygdlin (Nykänen & Suomlinen, 1982). The concentrtion of benzldehyde in rnge of white wines oxidised for 1 week ws found to be between µg/l (Escudero et l., 22) nd the compound correlted well with the term liquor (Escudero et l., 22). The odour detection threshold for benzldehyde is.35 mg/l in wter (Belitz et l., 29). 31

46 2.5.2 Other crbonyl compounds nd cetls Sotolon is chirl furnone (lctone) (Figure 2.13) nd very powerful odournt tht hs been described to hve the odour of curry nd myrrh t high concentrtions with rosting, mple syrup, burnt sugr nd crmel t lower concentrtions (Escudero et l., 2; Ghidossi et l., 212). Sotolon hs been demonstrted to be key odournt in sherry, port, vin june, botrytized nd Toki wines (Msud et l., 1984; Mrtin & Etiévnt, 1991; Guichrd et l., 1992; Mrtin et l., 1992; Silv Ferreir et l., 23; Collin et l., 212), but cn hve detrimentl effect when present in dry white wines (Escudero et l., 22; Silv Ferreir et l., 23c; Lvigne et l., 28). Figure 2.13 Chemicl structure of sotolon. The mechnism for the formtion of this compound is still not fully understood, however few pthwys hve been suggested nd the connection between oxidtion nd sotolon formtion seems to be evident (Phm et l., 1995; Cutzch et l., 1999; Silv Ferreir et l., 23; Silv Ferreir et l., 23c; Escudero et l., 211). The formtion of sotolon cn occur vi the enzymtic or chemicl demintion of threonine followed by ldol condenstion with cetldehyde in vrious wine medi (Tkhshi et l., 1976; Phm et l., 1995; Cutzch et l., 1998). Threonine is deminted by threonine deminse, forming α ketobutyric cid. The condenstion rection between α ketobutyric cid nd cetldehyde followed by lctonistion cn then form sotolon. Sotolon cn lso be formed from the oxidtive degrdtion of scorbic cid in the presence of ethnol (König et l., 1999; Pons et l., 21). This mechnism is most likely to occur in dry white wines, especilly considering the common prctice of dding scorbic cid to Suvignon blnc must/wine during winemking. During the oxidtive decomposition of the scorbic cid, α ketobutyric cid is formed s degrdtion intermedite. Millrd rections in severl combintions of binry mixtures of cysteine nd three sugrs, ribose, glucose nd rhmnose lso resulted in the formtion of sotolon (Hofmnn & Schieberle, 1995, 1997). This seems to be less likely mechnism in the cse of white wines, s these rections occur under much hrsher conditions thn those found during ging (Pons et l., 21). The concentrtions of this compound in dry white wines re much less compred to other lcoholic beverges (fortified wine, port, sherry) (Slmon et l., 1999). Concentrtion in young Suvignon blnc 32

47 wines just fter bottling were found to be below.5 µg/l while Suvignon blnc wine stored under vrious types of closures during 24 month period hd sotolon content rnging from µg/l (Lopes et l., 29; Ghidossi et l., 212). Another study mesured concentrtions of up to 1 µg/l for Suvignon blnc wines ged between 2 nd 32 yers (Lvigne et l., 28). The formtion of this compound in white wines seems to be somewht dependnt on temperture nd concentrtion of round 8 µg/l were obtined during forced ged experiment conducted t 6 C (Silv Ferreir et l., 23b). The odour threshold for this compound hs been determined to be 15 µg/l in flor sherry (Mrtin et l., 1992), 2 µg/l in model wine (Pons et l., 21) nd 8 µg/l in dry white wines (Lvigne et l., 28). Acetls re formed by the condenstion rection between glycerol nd cetldehyde in n cid medium. Four isomers cn be formed vi this rection: cis nd trns 5 hydroxy 2 methyl 1,3 dioxne nd cis nd trns 4 hydroxymethyl 2 methyl 1,3 dioxolne (Figure 2.14). In Port wine, isomers of glycerol nd cetldehyde cetls hve been found t totl concentrtions rnging from 9.4 to mg/l (Silv Ferreir et l., 22). During oxidtion of wine lrge mounts of the dioxnes nd dioxolnes re expected to be produced due to the fct tht mple quntity of substrtes is vilble. The romtic impct of these compounds could contribute significntly to the oxidised odour perceived in some wines. Odour evlution by sniffing in gs chromtogrphy rted the trns 5 hydroxy 2 methyl 1,3 dioxne s the compound with the highest intensity rom contributing to sweet nd old portlike rom (Silv Ferreir et l., 22). A B C D Figure 2.14 Isomers of cetldehyde nd glycerol cetls. A) cis 5 hydroxy 2 methyl 1,3 dioxne ; B) trns 5 hydroxy 2 methyl 1,3 dioxne; C) cis 4 hydroxymethyl 2 methyl 1,3 dioxolne; D) trns 4 hydroxymethyl 2 methyl 1,3 dioxolne. 2.6 SENSORY SCIENCE Evluting the chemicl composition of wine cn deliver vluble informtion such s the type nd quntity of specific romtic compounds present, however the use of the chemicl composition only is not dequte to predict the rom profile nd wine qulity. This is importnt in complex medium such 33

48 s wine where interctions between ll compounds present could influence the sensory perception of the rom. Sensory nlysis (especilly descriptive nlysis) of wines is thus of criticl importnce for the evlution of wine rom (Stone & Sidel, 1993; Lwless & Heymnn, 1998). The combintion of the chemicl nd sensory dt could deliver vluble informtion concerning the rom profile of the wine nd how vrious compounds interct Descriptive nlysis Descriptive nlysis is stndrd profiling method tht hs been used since 1974 to define the sensory ttributes for specific product set (Chollet et l., 211). According to Lwless & Heymnn (21), descriptive nlysis methods such s Quntittive Descriptive Anlysis (QDA) re considered to be the most sophisticted tools in sensory science. These tests identify, describe nd quntify the difference in sensory ttribute intensities between smples within smple set by using smll pnel of usully 8 to 12 people (Lwless & Heymnn, 21) whom re extensively trined for specific product ctegory nd re expected to then work s n nlyticl instrument (Nestrud & Lwless, 28; Næs et l., 21). With proper trining, the resercher cn expect little vrince between individul results which justifies the use of smller pnel. The error in vrince cn thus be lowered by intensive trining, whilst still mintining sttisticl power nd test sensitivity (Lwless & Heymnn, 21). There re vrious descriptive nlysis tests such s the Flvor Profile method (Cul, 1957); the Texture Profile method (Brndt et l., 1963; Szczesnik, 1975); Quntittive Descriptive Anlysis (Stone & Sidel, 24); the Spectrum Method (Meilgrd et l., 26) nd other hybrid versions of these techniques tht cn be used by reserchers (Einstein, 1991). These techniques re clssified s the most comprehensive nd informtive sensory methods in the history of sensory science (Lwless & Heymnn, 21). The procedure of QDA strts with the genertion of ttributes to describe the products. Therefter physicl or chemicl stndrds re selected which ct s references for the selected ttribute terms during the trining period. Assocition of n ttribute with single word of phrse is criticl step during QDA. After concept formtion nd verblistion trining, ssessors re trined to rte ll the pplicble ttribute intensities on n intensity line scle (Chollet et l., 211). Intensity scles cn be presented s vrious numberings with n upper nd lower limit or s structured or unstructured line scle (Næs et l., 21). Repetitive trining using the line scle is then done to profile the specific smple set ccording to the ttribute list. The finl profiling test does not include discussion nd the smples re rndomised per pnellist in order to prevent ny first order effect to influence the pnellist s perception of specific smple. Due to the fct tht pnellists re intensively trined to identify differences in specific sensoril ttribute intensities, one cn use QDA on product smple sets tht re similr to one nother. Dt collected from QDA tests cn be nlysed sttisticlly by using prmetric sttistics such s nlysis of vrince 34

49 (ANOVA), wheres the reltionships between the ttributes re often described using multivrite methods such s principle component nlysis (PCA). QDA does hve few disdvntges. This technique is time consuming. Pnel members must thus commit to the trining procedures which cn tke up to months due to criticl lnguge development nd clibrtion (Chollet et l., 211). The long periods of time needed for trining nd testing s well s lrge volumes of wine needed, mke this test procedure quite expensive. Another disdvntge is the fct tht the ttributes generted for specific product cnnot be used s generic for other products. This would effectively men tht pnel needs to be trined new for ech new product set dding more hours to the nlysis which cn become finncilly too strenuous for mny compnies nd reserchers (Bitnes et l., 27). Consequently most industries cnnot routinely use this technique (Stone & Sidel, 24; Meilgrd et l., 26; Kemp et l., 29). More cost efficient methods such s free choice profiling (Willims & Lngron, 1984), flsh profiling (Dirou & Sieffermnn, 22) nd projective mpping (Npping ) (Risvik et l., 1994; Risvik et l., 1997) hve been developed in the pst few yers in n ttempt to replce QDA (Guinrd et l., 21). QDA could lso potentilly crete comprehension nd greement problems within the pnel due to the importnce of lnguge (Chollet et l., 211). The difference in the use of the intensity scles by the vrious pnellists is lso disdvntge. However, this cn be corrected by continuously checking the pnel for consensus nd consistency using specific softwre pckges such s PnelCheck (V1.4., NOFIMA, Norwy) which is freewre R bsed progrm (Lwless & Heymnn, 21). Methods to evlute the pnel such s Tucker plots nd the use of ANOVA hve proven to deliver best results, however the combintion of the two methods would be considered idel. Tucker plots re used to ssess ech pnel member s performnce in reltion to specific ttribute or the whole pnel. It will identify which pnellists struggle to differentite between different smples, while ANOVA is used to elucidte the importnce of n ttribute nd the significnce of tht ttribute in the wines tested (Tomic et l., 27; Tomic et l., 21). QDA describes the wine ttributes nd the intensities t which these ttributes occur in ech smple. Intensive trining would lso llow the identifiction of smll but significnt differences between the smples even in cses where the smples re sensorilly similr. Descriptive nlysis hs successfully been used s tool to differentite between Suvignon blnc wines from different vintges nd countries (Lund et l., 29b; Green et l., 211). These techniques hve lso been used to investigte interctions between vrious compounds in Suvignon blnc wine (King et l., 211; Benkwitz et l., 212) s well s to identify off flvours in oxidtion spoiled white wines (Silv Ferreir et l., 23c). 35

50 2.6.2 Sensory interction studies Over the lst two decdes rom reconstitution pproches hve dded new dimension to sensory investigtions. In this wy, sensory interctions of chemicl compounds t different concentrtions nd in vrious medi re investigted (Guth, 1997; Lee, 23; Ferreir et l., 26). Studies indicted tht wine rom is influenced by complex interctions between vrious wine constituents nd rrely dominted by single component (Escudero et l., 27). In bsic or neutrl wine the mixture of ethnol, esters, lcohols nd cids will crete n rom buffer (Ferreir et l., 28). This is considered to form the bse of the wine rom nd consists of vrious romtic compounds usully present in concentrtions bove their perception threshold. This buffering property cn ttenute the ddition or omission effect of rom compounds. The compounds re not perceived s individul entities but rther s n integrted complex medium (Ferreir et l., 22; Escudero et l., 24). The bility of n odournt to brek the buffer nd chnge the romtic composition of the medium will define or clssify the odournt (Ferreir et l., 28). One or more of the following mechnisms could brek the buffer: 1) single type of molecule present t high enough concentrtions 2) group of compounds closely relted in terms of chemicl or romtic properties 3) lrge group of rom compounds with similr descriptors 4) n ssocition between n rom enhncer nd molecule unble to brek the buffer itself (Ferreir et l., 28). Impct compounds re compounds tht cn brek the buffer when present t norml concentrtions. These compounds cn trnsmit the specific rom without the support of other chemicls in wine. Arom compounds with the bility to do this include voltile thiols, methoxypyrzines nd oxidtion compounds such s sotolon, phenylcetldehyde nd methionl. Wine is complex medium contining vrious types of rom compounds tht cn ll hve n effect (differing degrees of suppression nd msking) on ech other (Frncis & Newton, 25). This complictes the investigtion of interctive effects between compounds. Over the yers, reserchers hve been trying to understnd the interctions of rom compounds in vrious medi s different compounds cn mnifest differently in different medi (Mg, 1989; Mris & Swrt, 1999; Cmpo et l., 25; Escudero et l., 27). Severl studies hve indicted difficulties when working in wine mtrix resulting in conflicting results when compring mtrices such s wter, model wine nd rel or deromtized wines (Aronson & Ebeler, 24). Even deromtized wines still hd sufficient chemicl compounds present to influence interction studies. In these deromtized wines most of the rom compounds would be removed by using resins such s mberlite or chrcol. The remining mtrix would still contin n unknown mount of unidentified chemicl compounds tht could potentilly influence the results of the romtic interction study. Lund et l. (29) proved the significnt effect of non voltile polyphenols on the perception of some romtic compounds, while other studies lso noticed the effect of compounds 36

51 present t their perception thresholds influencing the perception of other romtic compounds (Culleré et l., 27; Vn Wyngrd, 213). This indictes tht romtic compounds could still hve n enhncing or suppressing effect on other compounds even when it is not voltile or not detectble by humns. Vrious studies hve undertken the tsk to elucidte interctive effects between compounds in wines. Ethnol hs been found to ply very importnt role s it not only influences the solubility of odournts but it cn hve n impct on the ctul chemoreceptor system. A higher ethnol concentrtion will increse the solubility of the odournt cusing decrese in the mount of voltiles reching the pituitry glnd. Even more, ethnol hs proven to msk or suppress the fruity notes of esters (Escudero et l., 27). The ddition of β dmscenone nd β ionone to the mixture of esters brought cler increse of the fruity note (Pineu et l., 27). Dimethyl sulphide lso intensified the fruity rom especilly when β dmscenone ws lso present. The study concluded tht the fruity notes of red wines re the result of complex interction between esters, ethnol, norisoprenoids, dimethyl sulphide nd other voltiles (Escudero et l., 27). Other studies investigted interctive effects between rom impct compounds. Combintions contining high concentrtions of 4MMP, 3MH nd 3MHA not only delivered high intensity rtings for the ct urine/swety ttribute but lso the cooked green vegetl descriptor thus demonstrting the contribution of the thiols to the green chrcteristics of wine (King et l., 211). Synergism between IBMP nd 4MMP hs lso been reported. When dded individully t low concentrtions (2 ng/l for IBMP;.2 ng/l for 4MMP), dusty ws the only ttribute generted, however when dded together t the sme concentrtion, the mount of ttributes generted by the pnel incresed to dusty, grssy nd herbceous (Mris & Swrt, 1999). When dded t higher concentrtions, mutul suppression between the voltile thiols nd the methoxypyrzines ws observed (Mris & Swrt, 1999; Cmpo et l., 25; King et l., 211; Vn Wyngrd, 213). The ddition of esters to the bse wine incresed the perception of the confectionry ttribute nd lso incresed the rtings of overll fruit rom, tropicl nd cooked green vegetl ttributes (ttributes lso ssocited with thiols). Esters re known to be prticulrly importnt in the bouquet of young white wines nd interct in n dditive mnner (Vn der Merwe & Vn Wyk, 1981; Cmpo et l., 25). High thiol concentrtions in the presence of n ester combintion decresed the rtings for confectionry, indicting suppressive effect of the thiols on the esters when present t high levels. To the contrry, in the presence of moderte to high thiol concentrtions, the esters ctully enhnced some thiol relted sensory ttributes (Ferreir et l., 22; Cmpo et l., 25; King et l., 211; Benkwitz et l., 212). Overll, the ddition of IBMP hd greter impct on the rom of wine when compred to the ddition of the thiols. The effect of IBMP hs been described s being dominnt nd in wines contining both thiols nd IBMP, the rom is often driven by IBMP (Hein et l., 29; King et l., 211; Vn Wyngrd, 213). Lund et l. (29) investigted the interction of non voltile wine compounds such s polyphenols nd voltile rom compounds in diluted wine medium. They found tht the perception of IBMP, 3MH nd 37

52 ethyl decnote were lrgely suppressed by the ddition of polyphenols ((+) ctechin, quercetin nd cffeic cid). On the other hnd the ddition of specificlly cffeic cid enhnced the perception of 3MH. This could be due to the suppression of other rom compounds tht initilly msked the 3MH rom (Lund et l., 29). The rom perception of 3MHA ws lest ffected by the polyphenol dditions when compred to the other rom compounds tested in the study. The presence of the ester functionl group is thought to mke it less susceptible for interction with the polyphenol when compred to the lcohol group of 3MH (Lund et l., 29). Very few studies hve investigted the interctive effect of unplesnt rom compounds such s compounds relted to oxidtion. The ddition of brnched liphtic ldehydes nd (E) 2 lkenls generted ttributes such s sweet ornge, dried fruit, fusel, closed room, dirty, mouldy nd rncid nd n dditive effect between these ldehydes were observed (Culleré et l., 27). There seems to be lck of informtion regrding these types of studies, especilly considering tht no reserch hs been done to dte to investigte the interctions between chrcter impct compounds nd potent compounds originting due to oxidtion. 2.7 CONCLUSION Oxygen mngement relted issues re estimted to compromise the qulity of pproximtely 3% of ll wines before the bottles rech the retiler (Nygrd, 21). Consumers re sensitive to wine fults, such s reductive nd oxidtive roms nd this will decrese the liking scores nd purchse intent (Nygrd, 21). The mngement of oxygen during winemking is thus generting considerble interest in the wine world. If oxygen exposure is too high or too low, wines cn develop defects tht cn compromise their sensory qulity. Too little oxygen exposure could led to wine with reductive odours, while too much oxygen cn result in n oxidised wine displying other unplesnt roms together with the decrese in plesnt fruity roms. The chllenge is thus to identify the right mount of oxygen needed for specific wine to mintin the perfect blnce between chemicl composition nd sensory perception. Other thn tht, reserch focussing on the evolution of the compounds during oxidtive ging of wines is needed to better understnd the reltionships between compounds nd mechnisms occurring. Such reserch should lso be conducted under conditions which the wine might encounter in commercil cellr sitution nd not only be performed under conditions of enhnced oxidtion. The complexity of wine s medium gretly complictes reserch ttempts. Suvignon blnc wine especilly is known for the lrge diversity of flvours which will dd to the complexity. The voltile thiols nd the methoxypyrzines re the min compounds involved in the unique chrcter of Suvignon blnc wines, however by themselves these key molecules do not ccount for the romtic complexity of high 38

53 qulity wines. Vrious interctions occur between both voltile nd non voltile wine constituents nd these interctions cn chnge s the concentrtions chnge. Studies ttempted to identify some of these interctive effects, however much more reserch is needed to successfully ddress this issue. 2.8 ABBREVIATIONS USED 3 mercptohexn 1 ol (3MH); 3 mercptohexn 1 ol cette (3MHA); 4 mercpto 4 methylpentn 2 one (4MMP); 3 isobutyl 2 methoxypyrzine (IBMP); 3 isopropyl 2 methoxypyrzine (IPMP); 3 sec butyl 2 methoxypyrzine (SBMP); glutthione (GSH); oxidised glutthione (GSSG); grpe rection product (GRP); methyl mercptn (MeSH); dimethyl sulphide (DMS); 5 hydroxymethylfurfurl (5 HMF); 5 methylfurfurl (5 MF); nlysis of vrince (ANOVA); principle component nlysis (PCA); quntittive descriptive nlysis (QDA). 2.9 REFERENCES Adms, D.O. & Liynge, C., Glutthione increses in grpe berries t the onset of ripening. Americn Journl of Enology nd Viticulture. 44, Alberts, P., Stnder, M.A., Pul, S.O. & De Villiers, A., 29. Survey of 3 lkyl 2 methoxypyrzine content of South Africn Suvignon blnc wines using novel LC APCI MS/MS method. Journl of Agriculturl nd Food Chemistry. 57, Allen, M.S., Lcey, M.J., Hrris, R.L.N. & Brown, W.V., Contribution of methoxypyrzines to Suvignon blnc wine rom. Americn Journl of Enology nd Viticulture. 42, Anderson, M.E., Glutthione: An overview of biosynthesis nd modultion. Chemico Biologicl Interctions. 112, Anfng, N., Brjkovich, M. & Goddrd, M.R., 29. Co fermenttion with Pichi kluyveri increses vrietl thiol concentrtions in Suvignon blnc. Austrlin Journl of Grpe nd Wine Reserch. 15, 1 8. Aronson, J. & Ebeler, S.E., 24. Effect of polyphenol compounds on the hedspce voltility of flvors. Americn Journl of Enology nd Viticulture. 55, Arrehnius, S.P., McCloskey, L.P. & Sylvn, M., Chemicl mrkers for rom of Vitis vinifer vr. Chrdonny regionl wines. Journl of Agriculturl nd Food Chemistry. 44, Buernfeind, J.C. & Pinkert, D.M., 197. Food processing with dded scorbic cid. In Advnces in food reserch. C.O. Chichester, E.M.M.G.F.S. (ed.), pp Acdemic Press, London. Byonove, C. & Cordonnier, R., Recherches sur l rôme du musct. III. Étude de l frction terpénique. Ann. Technol. Agric. 2,

54 Beech, F.W., Burroughs, L.F., Timberlke, C.F. & Whiting, G.C., Progrès récents sur l'spect chimique et l'ction ntimicrobienne de l'nhydride sulfureux (SO 2 ). Bulletin de L'O.I.V. 52, Belitz, H. D., Grosch, W. & Schieberle, P., 29. Food Chemistry. Springer, Berlin Heidelberg. Benkwitz, F., Nicolu, L., Lund, C., Beresford, M., Wohlers, M. & Kilmrtin, P.A., 212. Evlution of key odornts in Suvignon blnc wines using three different methodologies. Journl of Agriculturl nd Food Chemistry. 6, Benkwitz, F., Toming, T., Kilmrtin, P.A., Lund, C., Wohler, M. & Nicolu, L., 212b. Identifying the chemicl composition relted to the distinct rom chrcteristics of New Zelnd Suvignon blnc wines. Americn Journl of Enology nd Viticulture. 63, Berg, H.W. & Akiyoshi, M., Some fctors involved in browning of white wine. Americn Journl of Enology nd Viticulture. 7, 1 7. Bertrnd, A., Utilistion de l chromtogrphie en phse gzeuse pour le dosge des constitunts voltils du vin. Thesis, University of Bordeux. Bitnes, J., Rødbotten, M., Le, P., Uelnd, Ø. & Mrtens, M., 27. Effect of product knowledge on profiling performnces compring vrious sensory lbortories. Journl of Sensory Science. 22, Blke, A., Kotserides, Y., Brindle, I.D., Inglis, D. & Pickering, G.J., 21. Effect of light nd temperture on 3 lkyl 2 methoxypyrzine concentrtion nd other impct odournts of Riesling nd Cbernet frnc wines during bottle geing. Food Chemistry. 119, Blke, A., Kotseridis, Y., Brindle, I.D., Inglis, D., Sers, M. & Pickering, G.J., 29. Effect of closure nd pckging type on 3 lkyl 2 methoxypyrzines nd other impct odornts of Riesling d Cbernet frns wines. Journl of Agriculturl nd Food Chemistry. 57, Blnchrd, L., Drriet, P. & Dubourdieu, D., 24. Rectivity of 3 mercptohexnol in red wine: Impct of oxygen, phenolic frctions, nd sulfur dioxide. Americn Journl of Enology nd Viticulture. 55, Blouin, J., Contribution L Étude des combinisons de l nhydride sulfureux dns les moûts et les vins. Thesis, L'Université de Bordeux, Bordeux, Frnce. Bouchilloux, P., Drriet, P., Henry, R., Lvigne Cruege, V. & Duboudieu, D., Identifiction of voltile nd powerful odorous thiols nd Bordeux red wine vrieties. Journl of Agriculturl nd Food Chemistry. 46, Boulton, R.B., Singleton, V.L., Bisson, L.F. & Kunkee, R.E., Principles nd prctices of winemking. Chpmn & Hll, New York. Brdshw, M.P., Cheynier, V., Scollry, G.R. & Prenzler, P.D., 23. Defining the scorbic cid crossover from nti oxidnt to pro oxidnt in model wine mtrix contining (+) ctechin. Journl of Agriculturl nd Food Chemistry. 51,

55 Brdshw, M.P., Scollry, G.R. & Prenzler, P.D., 21. Ascorbic cid induced browning of (+) ctechin in model wine system. Journl of Agriculturl nd Food Chemistry. 49, Brdshw, M.P., Scollry, G.R. & Prenzler, P.D., 24. Exmintion of the sulfur dioxide scorbic cid nti oxidnt system in model white wine mtrix. Journl of the Science of Food nd Agriculture. 84, Brjkovich, M., Tibbits, N., Peron, G., Lund, C.M., Dykes, S.I., Kilmrtin, P.A. & Nicolu, L., 25. Effect of screwcp nd cork closures on SO 2 levels nd roms in Suvignon blnc wine. Journl of Agriculturl nd Food Chemistry. 53, Brndt, M.A., Skinner, E.Z. & Colemn, J.A., The texture profile method. Journl of Food Science. 28, Burroughs, L.F. & Sprks, A.H., Sulphite binding power of wines nd ciders. I. Equilibrium constnts for the dissocition of crbonyl bisulphite compounds. Journl of the Science of Food nd Agriculture. 24, Buttery, R.G., Seifert, R.M., Gudgni, D.G. & Ling, L.C., Chrcteriztion of some voltile constituents of bell peppers. Journl of Agriculturl nd Food Chemistry. 17, Câmr, J.S., Alves, M.A. & Mrques, J.C., 26. Chnges in voltile composition of Mdeir wines during their oxidtive ging. Anlytic Chimic Act. 563, Câmr, J.S., Mrques, J.C., Alves, M.A. & Silv Ferreir, A.C., Hydroxy 4,5 dimethyl 2(5H) furnone levels in fortified Mdeir wines: Reltionship to sugr content. Journl of Agriculturl nd Food Chemistry. 52, Cmpo, E., Ccho, J. & Ferreir, V., 28. The chemicl chrcteriztion of the rom of dessert nd sprkling white wines (Pedro Ximénez, Fino, Suternes nd Cv) by gs chromtogrphyolfctometry nd chemicl quntittive nlysis. Journl of Agriculturl nd Food Chemistry. 56, Cmpo, E., Ferreir, V., Escudero, A. & Ccho, J., 25. Prediction of wine sensory properties relted to grpe vriety from dynmic hedspce gs chromtogrphy olfctory dt. Journl of Agriculturl nd Food Chemistry. 53, Cpone, D.L. & Jeffery, D.W., 211. Effects of trnsporting nd processing Suvignon blnc grpes on 3 mercptohexn 1 ol precursor concentrtions Journl of Agriculturl nd Food Chemistry. 59, Cpone, D.L., Sefton, M.A., Hysk, Y. & Jeffery, D.W., 21. Anlysis of precursors to wine odornt 3 mercptohexn 1 ol using HPLC MS/MS: Resolution nd quntittion of distereomers of 3 Scysteinylhexn 1 ol nd 3 S glutthionylhexn 1 ol. Journl of Agriculturl nd Food Chemistry. 58, Cstellri, M., Simonto, B., Tornielli, G.B., Spinelli, P. & Ferrrini, R., 24. Effects of different enologicl tretments on dissolved oxygen in wines. Itlin Journl of Food Science. 16,

56 Cul, J.F., The profile method of flvor nlysis. Advnces in Food Reserch. 7, 1 4. Cheynier, V., Trousdle, E.K., Singleton, V.L., Slgeus, M.J. & Wylde, R., Chrcteriztion of 2 Sglutthionylcftric cid nd its hydrolysis in reltion to grpe wines. Journl of Agriculturl nd Food Chemistry. 34, Cheynier, V. & Vn Hulst, M.W.J., Oxidtion of trns cftric cid nd 2 S glutthionyl cftric cid in model solutions. Journl of Agriculturl nd Food Chemistry. 36, Chollet, S., Lelièvre, M., Abdi, H. & Vlentin, D., 211. Sort nd beer: Everything you wnted to know bout the instructed sorting but did not dre to sk. Food Qulity nd Preference. 22, Choné, X., 21. Contribution à l'étude des terroirs de Bordeux: étude des déficits hydriques modérés, de l'limenttion en zote et de leurs effets sur le potentiel romtique des risins de Vitis vinifer L. cv. Suvignon blnc. Thesis, University of Bordeux. Choné, X., Lvigne Cruege, V., Toming, T., Vn Leewin, C., Cstgnede, C., Sucier, C. & Duboudieu, D., 26. Effect of vine nitrogen sttus on grpe romtic potentil: flvour precursors (S cysteine conjugtes), glutthione nd phenolic content in Vitis vinifer L. cv. Suvignon blnc grpe juice. Journl Interntionl des Sciences de l Vigne et du Vin. 4, 1 6. Chrisholm, M.G., Guiher, L.A. & Zczkiewicz, S.M., Arom chrcteristics of ged Vidl blnc wine. Americn Journl of Enology nd Viticulture. 46, Cilliers, J.J.L. & Singleton, V.L., Nonenzymtic utooxidtive phenolic browning rections in cffeic cid model system. Journl of Agriculturl nd Food Chemistry. 37, Cilliers, J.J.L. & Singleton, V.L., 199. Cffeic cid utoxidtion nd the effects of thiols. Journl of Agriculturl nd Food Chemistry. 38, Clrk, A.C., Pedretti, F., Prenzler, P.D. & Scollry, G.R., 28. Impct of scorbic cid on the oxidtive colourtion nd ssocited rections of model wine solution contining (+) ctechin, cffeic cid nd iron. Austrlin Journl of Grpe nd Wine Reserch. 14, Coetzee, C. & Du Toit, W.J., 212. A comprehensive review on Suvignon blnc rom with focus on certin positive voltile thiols. Food Reserch Interntionl. 45, Coetzee, C., Lisjk, K., Nicolu, L., Kilmrtin, P.A. & Du Toit, W.J., 213. Oxygen nd sulfur dioxide dditions to Suvignon blnc must: effect on must nd wine composition. Flvour nd Frgrnce Journl. 28, Collin, S., Nizet, S., Bouuert, T.C. & Desptures, P. M., 212. Min odornts in jur Flor sherry wines. Reltive contributions of sotolon, bhexon, nd thespirne derived compounds. Journl of Agriculturl nd Food Chemistry. 6, Cooper, M., 28. Wine tls of New Zelnd. Hodder, Mo, Aucklnd, New Zelnd. Culleré, L., Ccho, J. & Ferreir, V., 27. An ssessment of the role plyed by some oxidtion relted ldehydes in wine rom. Journl of Agriculturl nd Food Chemistry. 55,

57 Cutzch, I., Chtonnet, P. & Duboudieu, D., Study the mechnisms of some voltile compounds during ging of sweet fortified wines. Journl of Agriculturl nd Food Chemistry. 47, Cutzch, I., Chtonnet, P. & Dubourdieu, D., Rôle du sotolon dns l rôme des vins doux nturels, influence des conditions d élevge et de vieillissement. Journl Interntionl des Sciences de l Vigne et du Vin. 32, Dirou, V. & Sieffermnn, J.M., 22. A comprison of 14 jms chrcterized by conventionl profile nd quick originl method, flsh profile. Journl of Food Science. 67, Dnilewicz, J.C., 23. Review of rection mechnisms of oxygen nd proposed intermedite reduction products in wine: centrl role of iron nd copper. Americn Journl of Enology nd Viticulture. 54, Dnilewicz, J.C., 27. Interction of sulfur dioxide, polyphenols, nd oxygen in wine model system: Centrl role of iron nd copper. Americn Journl of Enology nd Viticulture. 58, Dnilewicz, J.C., Seccombe, J.T. & Wheln, J., 28. Mechnism of interction of polyphenols, oxygen, nd sulphur dioxide in model wine nd wine. Americn Journl of Enology nd Viticulture. 59, Drriet, P., Toming, T., Lvigne, V., Boidron, J. & Dubourdieu, D., Identifiction of powerful romtic compound of Vitis vinifer L. vr. Suvignon wines: 4 Mercpto 4 methylpentn 2 one. Flvour nd Frgrnce Journl. 1, Dudt, C.E. & Ough, C.S., Vritions in some voltile cette esters formed during grpe juice fermenttion temperture, SO 2, yest strin, nd grpe vriety. Americn Journl of Enology nd Viticulture. 24, De Boubee, D.R., Cumsille, A.M., Pons, M. & Dubourdieu, D., 22. Loction of 2 methoxy 3 isobutylpyrzine in Cbernet Suvignon grpe bunches nd its extrctibility during vinifiction. Americn Journl of Enology nd Viticulture. 53, 1 5. Dickinson, J.R., Lntermn, M.M., Dnner, D.J., Person, B.M., Snz, P., Hrrison, S.J. & Hewlins, J.E., A 13C nucler mgnetic resonnce investigtion of the metbolism of leucine to isomyl lcohol in Scchromyces cerevisie. The Journl of Biologicl Chemistry. 272, Dickinson, J.R., Slgdo, L.E. & Hewlins, J.E., 23. The ctbolism of mino cids to long chin nd complex lcohols in Scchromyces cerevisie. The Journl of Biologicl Chemistry. 278, Doyle, M. & Beucht, L., 27. Food microbiology: Fundmentls nd frontiers. ASM Press, Wshington, DC. Drinkine, J., Glories, Y. & Sucier, C., 25. (+) Ctechin ldehyde condenstions: Competition between cetldehyde nd glyoxylic cid. Journl of Agriculturl nd Food Chemistry. 53, Drysdle, G.S. & Fleet, G.H., Acetic cid bcteri in winemking: review. Americn Journl of Enology nd Viticulture. 39,

58 Du Toit, W.J., Lisjk, K., Stnder, M. & Prevoo, D., 27. Using LC MSMS to ssess glutthione levels in South Africn white grpe juices nd wines mde with different levels of oxygen. Journl of Agriculturl nd Food Chemistry. 55, Dubourdieu, D., Toming, T., Msneuf, I., Peyrot des Gchons, C. & Murt, M.L., 26. The role of yest in grpe flvour development during fermenttion: The exmple of Suvignon blnc. Americn Journl of Enology nd Viticulture. 57, Ebeler, S.E. & Thorngte, J.H., 29. Wine chemistry nd flvour: Looking into the crystl glss. Journl of Agriculturl nd Food Chemistry. 57, Einstein, M.A., Descriptive techniques nd their hybridiztion. In Sensory science theory nd pplictions in foods. Lwless, H.T. nd Klein, B.P. (eds.). Mrcel Dekker, New York. Es Sfi, N., Le Guernevé, C., Lbrbe, B., Fulcrnd, H., Cheynier, V. & Moutounet, M., Structure of new xnthylium slt derivtive. Tetrhedron Letters. 4, Escudero, A., Asencio, E., Ccho, J. & Ferreir, V., 22. Sensory nd chemicl chnges of young white wines stored under oxygen. An ssessment of the role plyed by ldehydes nd some other importnt odornts. Food Chemistry. 77, Escudero, A., Ccho, J. & Ferreir, V., 2. Isoltion nd identifiction of odournts generted in wine during its oxidtion: gs chromtogrphy olfctometry study. Europen Food Reserch nd Technology. 211, Escudero, A., Ccho, J. & Ferreir, V., 211. Isoltion nd identifiction of odornts generted in wine during its oxidtion: gs chromtogrphy olfctometric study. Europen Food Reserch nd Technology Escudero, A., Cmpo, E., Frin, L., Ccho, J. & Ferreir, V., 27. Anlyticl chrcteriztion of the rom of five premium red wines. Insights into the role of odor fmilies nd the concept of fruitiness of wines. Journl of Agriculturl nd Food Chemistry. 55, Escudero, A., Gororz, B., Melus, M.A., Ortin, N., Ccho, J. & Ferreir, V., 24. Chrcteriztion of the rom of wine from Mccbeo. Key role plyed by compounds with low odor ctivity vlues. Journl of Agriculturl nd Food Chemistry. 52, Escudero, A., Hernndez Orte, P., Ccho, J. & Ferreir, V., 2b. Clues bout the role of methionl s chrcter impct odornt of some oxidized wines. Journl of Agriculturl nd Food Chemistry. 48, Flco, L.D., De Revel, G., Perello, M.C., Moutsiou, A., Znus, M.C. & Bordignon Luiz, M.T., 27. A survey of sesonl temperture nd vineyrd ltitude influences on 2 methoxypyrzine, C 13 norisoprenoids, nd their sensory profile of Brzilin Cbernet Suvignon wines. Journl of Agriculturl nd Food Chemistry. 55,

59 Fedrizzi, B., Prdon, K.H., Sefton, M.A., Elsey, G.M. & Jeffery, D.W., 29. First identifiction of 4 Sglutthionyl 4 methylpentn 2 one, potentil precursor of 4 mercpto 4 methylpentn 2 one, in Suvignon blnc juice. Journl of Agriculturl nd Food Chemistry. 57 (3), Fedrizzi, B., Zpproli, G., Finto, F., Tosi, E., Turri, A., Azzolini, M. & Versini, G., 211. Model ging nd oxidtion effects on vrietl, fermenttive, nd sulphur compiounds in dry botrytized red wine. Journl of Agriculturl nd Food Chemistry. 59, Fernndez Zurbno, P., Ferreir, V., Pen, C., Escudero, A., Serrno, F. & Ccho, J., Prediction of oxidtive browning in white wines s function of their chemicl composition. Journl of Agriculturl nd Food Chemistry. 43, Ferreir, V., Escudero, A., Cmpo, E. & Ccho, J., 28. Understnding the role plyed by the odornts nd their interctions on wine flvour. Proc. Wine Active Compounds. Beune, Frnce. Ferreir, V., Escudero, A., Fernnádez, P.E. & Ccho, J., Chnges in the profile of voltile compounds in wines stored under oxygen nd their reltionship with the browning process. Zeitschrift für Lebensmittel Untersuchung und Forschung A. 25, Ferreir, V., Ortin, N., Escudero, A., Lopez, R. & Ccho, J., 22. Chemicl chrcteriztion of the rom of Grenche rosé wines: Arom extrct dilution nlysis, quntittive determintion, nd sensory reconstitution studies. Journl of Agriculturl nd Food Chemistry. 5, Ferreir, V., Pet'k, A. & Ccho, J., 26. Intensity nd persistence profiles of flvor compounds in synthetic solutions. Simple model for explining the intensity nd persistence of their ftersmell. Journl of Agriculturl nd Food Chemistry. 54, Frncis, I.L. & Newton, J.L., 25. Determining wine rom from compositionl dt. Austrlin Journl of Grpe nd Wine Reserch. 11, Frivik, S.K. & Ebeler, S.E., 23. Influence of sulfur dioxide on the formtion of ldehydes in white wine. Americn Journl of Enology nd Viticulture. 54, Fulcrnd, H., Cheynier, V., Oszminski, J. & Moutounet, M., An oxidized trtric cid residue s new bridge potentilly competing with cetldehyde in flvn 3 il condenstion. Phytochemistry. 46, Grde Cerdán, T. & Ancín Azpilicuet, C., 27. Effect of SO 2 on the formtion nd evolution of voltile compounds in wines. Food Control. 18, Ghidossi, R., Poupot, C., Thibon, C., Pons, A., Drriet, L., Riquier, G., De Revel, G. & Mietton Peuchot, M., 212. The influence of pckging on wine conservtion. Food Control. 23, Godden, P., Leigh, F., Field, J., Gishen, M., Coulter, A., Vlente, P., HoJ, P. & Robinson, E., 21. Wine bottle closures: physicl chrcteristics nd effect on composition nd sensory properties of Semillon wine. 1. Performnce up to 2 months post bottling. Austrlin Journl of Grpe nd Wine Reserch. 7,

60 Gonik, O.J. & Noble, A.C., Sensory study of selected voltile sulfur compounds in white wine. Americn Journl of Enology nd Viticulture. 38, Green, J.A., Prr, W.V., Breitmeyer, J., Vlentin, D. & Sherlock, R., 211. Sensory nd chemicl chrcteristion of Suvignon blnc wine: Influence of source of origin. Food Reserch Interntionl. 44, Guichrd, E., Etiévnt, P.X., Henry, R. & Mosndl, A., Enntiomeric rtios of pntolctone, solerone, 4 crboethoxy 4 hydroxy butyrolctone, nd of sotolon, flvour impct compound of flor sherry nd botrytized wines. Zeitschrift für Lebensmittel Untersuchung und Forschung A. 195, Guinrd, J.X., Uotni, B. & Schlich, P., 21. Internl nd externl mpping of preferences of commercil lger beers: Comprison of hedonic rtings by consumers blind versus with knowledge of brnd nd price. Food Qulity nd Preference. 12, Guth, H., Quntittion nd sensory studies of chrcter impct odornts of different white wine vrieties. Journl of Agriculturl nd Food Chemistry. 45, Guyot, S., Cheynier, V., Souquet, J.M. & Moutounet, M., Influence of ph on the enzymtic oxidtion of (+) ctechin in model systems. Journl of Agriculturl nd Food Chemistry. 43, Hllidy, J. & Johnson, H., The rt nd science of wine. Mitchell Bezley, London. Hrsch, M.J., Benkwitz, F., Frost, A., Colonn Ceccldi, B. & Grdner, R.C., 213. New precursor of 3 mercptohexn 1 ol in grpe juice: Thiol forming potentil nd kinetics during erly stges of must fermenttion. Journl of Agriculturl nd Food Chemistry. 61, Hshizume, K. & Smut, T., Grpe mturity nd light exposure ffect berry methoxypyrzine concentrtion. Americn Journl of Enology nd Viticulture. 5, Hein, K., Ebeler, S.E. & Heymnn, H., 29. Perception of fruity nd vegetl roms in red wine. Journl of Sensory Studies. 24, Henschke, P.A. & Jirnek, V., Yests: Growth during fermenttion. In Wine Microbiology nd biotechnology. Fleet, G.H. (ed.), pp Hrdwood cdemic publishers, Chur, Switzerlnd. Herbst Johnstone, M., Nicolu, L. & Kilmrtin, P.A., 211. Stbility of vrietl thiols in commercil Suvignon blnc wines. Americn Journl of Enology nd Viticulture. 62, Herbst, M., 21. Investigtion into the rom stbility of New Zelnd Suvignon blnc. Thesis, University of Aucklnd Aucklnd, New Zelnd. Herbst, M., Kilmrtin, P.A. & Nicolu, L., 28. Arom stbility in Suvignon blnc wines. The Austrlin nd New Zelnd Grpegrower nd Winemker. 6, Hofmnn, T. & Schieberle, P., Evlution of the key odornts in thermlly treted solution of ribose nd cysteine by rom extrct dilution techniques. Journl of Agriculturl nd Food Chemistry. 43,

61 Hofmnn, T. & Schieberle, P., Identifiction of potent rom compounds in thermlly treted mixtures of glucose/cysteine nd rhmnose/cysteine using rom extrct dilution technique. Journl of Agriculturl nd Food Chemistry. 45, Hunter, J.J., Volschenk, C.G. & Fouche, G.W., 24. Composition of Suvignon blnc grpes s ffected by pre vérison cnopy mnupiltion nd ripeness level. South Africn Journl of Enology nd Viticulture. 25, Iscs, N. & Vn Eldik, R.A., A mechnistic study of the reduction of quinones by scorbic cid. Journl of the Chemicl Society, Perkin Trnsctions Jckowetz, J.N. & De Orduñ, M., 213. Survey of SO 2 binding crbonyls in 237 red nd white tble wines. Food Control. 32, Jckowetz, J.N., Dierschke, S.E. & Mir de Orduñ, R., 211. Multifctoril nlysis of cetldehyde kinetics during lcoholic fermenttion by Scchromyces cerevisie. Food Reserch Interntionl. 44, Jckson, R.S., 28. Wine Science: Principles nd pplictions. Elsevier/Acdemic Press, Amsterdm. Jnes, L., Lisjk, K. & Vnzo, A., 21. Determintion of glutthione content in grpe juice nd wine by high performnce liquid chromtogrphy with fluorescence detection. Anlytic Chimic Act. 674, Jocelyn, P.C., Biochemistry of the SH group; the occurrence, chemicl properties, metbolism nd biologicl function of thiols nd disulphides. Acdemic Press, London, UK. Kemp, S.E., Hollowood, T. & Hort, J., 29. Sensory Evlution. A Prcticl Hndbook. Wiley Blckwell, Oxford. Kielhöfer, E. & Würdig, G., Die bindung des suerstoffs durch schweflige säure und scorbinsäure. Weinberg und Keller. 5, Kielhöfer, E. & Würdig, G., 196. Die n ldehyd gebundene schweflige säure im wein. I. Mitteilung: Aldehydbildung und enzymtische und nichtenzymtische lkohol oxidtion. II. Acetldehydbildung bei der gärung. Weinberg und Keller. 7, 16 22; Kilmrtin, P.A., Zou, H. & Wterhouse, A.L., 21. A cyclic voltmmetry method suitble for chrcterizing ntioxidnt properties of wine nd wine phenolics. Journl of Agriculturl nd Food Chemistry. 49, Kilmrtin, P.A., Zou, H. & Wterhouse, A.L., 22. Correltion of wine phenolic composition versus cyclic voltmmetry response. Americn Journl of Enology nd Viticulture. 53, King, E.S., Osidcz, P., Curtin, C., Bstin, S.E.P. & Frncis, I.L., 211. Assessing desirble levels of sensory properties in Suvignon Blnc wines consumer preferences nd contribution of key rom compounds. Austrlin Journl of Grpe nd Wine Reserch. 17, Kleinhns, D., Sensitivity of sthmtics ginst sulphur compounds in food, wine nd injection solutions. Deutsche Medizinische Wochenschrift. 17,

62 König, T., Gutsche, B., Hrtl, M., Hübscher, R., Schreier, P. & Schwb, W., Hydroxy 4,5 dimethyl 2(5H) furnone (sotolon) cusing off flvor: Elucidtion of its formtion pthwys during storge of citrus soft drinks. Journl of Agriculturl nd Food Chemistry. 47, Kotserides, Y., Ry, J., Augier, C. & Bumes, R., 2. Quntittive determintion of sulfur contining wine odornts t sub ppb levels. 1. Synthesis of the deuterted nlogues. Journl of Agriculturl nd Food Chemistry. 48, Kotserides, Y., Spink, M., Brindle, I.D., Blke, A., Sers, M., Chen, X., Soles, G., Inglis, D. & Pickering, G.J., 28. Quntittive nlyses of 3 lkyl 2 methoxypyrzines in juice nd wine using stble isotope lbelled internl stndrd ssy. Journl of Chromtogrphy A. 119, Kreitmn, G.Y., Lurie, F. & Elis, R., 213. Investigtion of ethyl rdicl quenching by phenolics nd thiols in model wine. Americn Journl of Enology nd Viticulture. 61, Kritzinger, E.C., Buer, F.F. & Du Toit, W.J., 213. Role of Glutthione in Winemking: A Review. Journl of Agriculturl nd Food Chemistry. 61, Kritzinger, E.C., Stnder, M. & Du Toit, W.J., 213b. Assessment of glutthione levels in model solution nd grpe ferments supplemented with glutthione enriched inctive dry yest preprtions using novel UPLC MS/MS method. Food Additives & Contminnts: Prt A. 3, Kwitkowski, M., Skouroumounis, G.K., Lttey, K.A. & Wters, E.J., 27. The impct of closures, including screw cp with three different hedspce volumes, on the composition, colour nd sensory properties of Cbernet Suvignon wine during two yers' storge. Austrlin Journl of Grpe nd Wine Reserch. 13, Lcey, M.J., Allen, M.S., Hrris, R.L.N. & Brown, W.V., Methoxypyrzines in Suvignon blnc grpes nd wines. Americn Journl of Enology nd Viticulture. 42, Lcey, M.J., Brown, W.V., Allen, M.S. & Hrris, R.L.N., Alkyl methoxypyrzines nd Suvignon blnc chrcter. Proc. Cool Climte Symposium for Viticulture nd Oenology. Aucklnd, New Zelnd. Lchenmeier, D.W. & Sohnuis, E. M., 28. The role of cetldehyde outside ethnol metbolism in the crcinogenicity of lcoholic beverges: Evidence from lrge chemicl survey. Food nd Chemicl Toxicology. 46, Lmbrechts, M.G. & Pretorius, I.S., 2. Yest nd its importnce to wine rom A review. South Africn Journl of Enology nd Viticulture. 21, Lmbropoulos, I. & Roussis, I., 27. Inhibition of decrese of voltile esters nd terpenes during storge of white wine nd model white wine medium by cffeic cid nd gllic cid. Food Reserch Interntionl. 4, Lurie, F., Zúñig, M.C., Crrsco Sánchez, V., Sntos, L., Cñete, A., Ole Azr, C., Uglino, M. & Agosin, E., 212. Rectivily of 3 sulfnyl 1 hexnol nd ctechol conting phenolics in vitro. Food Chemistry. 131,

63 Lvigne Cruège, V. & Dubourdieu, D., 23. Rôle du glutthion sur l évolution romtique des vins blncs secs. Presented t the 7e Symposium Interntionl d Oenologie, Arcchon, Frnce. Lvigne, V., Pons, A., Drriet, P. & Duboudieu, D., 28. Chnges in sotolon content of dry white wines during brrel nd bottle ging. Journl of Agriculturl nd Food Chemistry. 56, Lwless, H.T. & Heymnn, H., Sensory evlution food, principles nd prctice. Chpmn nd Hll, New York, USA. Lwless, H.T. & Heymnn, H., 21. Sensory evlution of food: principles nd prctices. Springer, New York. Lee, D. H., Kng, B. S. & Prk, H. J., 211. Effect of oxygen on voltile nd sensory chrcteristics of Cbernet Suvignon during secondry shelf life. Journl of Agriculturl nd Food Chemistry. 59, Lee, S.J., 23. Finding key odornts in foods: Gs chromtogrphy olfctometry (GC/O). Food Science nd Biotechnology. 12, Li, H., Guo, A. & Wng, H., 28. Mechnisms of oxidtive browning of wine. Food Chemistry. 18, Lilly, M., Lmbrechts, M.G. & Pretorius, I.S., 2. Effect of incresed yest lcohol cetyltrnsferse ctivity on flvor profiles of wine nd distilltes. Applied nd Environmentl Microbiology. 66, Limmer, A., 25. Suggestions for deling with post bottling sulfides. Austrlin nd New Zelnd grpegrower nd winemker. 476, Litchev, V., Influence of oxidtion processes on the development of the tste nd flvor of wine distilltes. Americn Journl of Enology nd Viticulture. 4, Liu, S. Q. & Pilone, G.J., 2. An overview of formtion nd roles of cetldehyde in winemking with emphsis on microbil implictions. Interntionl Journl of Food Science nd Technology. 35, Lopes, P., Sucier, C. & Glories, Y., 25. Nondestructive colorimetric method to determine the oxygen diffusion rte through closures used in winemking. Journl of Agriculturl nd Food Chemistry. 53, Lopes, P., Silv, M.A., Pons, A., Toming, T., Lvigne, V., Sucier, C., Drriet, P., Teissedre, P.L. & Dubourdieu, D., 29. Impct of oxygen dissolved t bottling nd trnsmitted through closures on the composition nd sensory properties of Suvignon blnc wine during bottle storge. Journl of Agriculturl nd Food Chemistry. 57, Lopez Toledno, A., Villno Vlenci, D., Myen, M., Merid, J. & Medin, M., 24. Interction of yest with the products resulting from the condenstion rection between (+) ctechin nd cetldehyde. Journl of Agriculturl nd Food Chemistry. 52,

64 Loscos, N., Hernndez Orte, P., Ccho, J. & Ferreir, V., 27. Relese nd formtion of vrietl rom compounds during lcoholic fermenttion from nonflorl grpe odorless flvor precursors frctions. Journl of Agriculturl nd Food Chemistry. 55, Loscos, N., Hernndez Orte, P., Ccho, J. & Ferreir, V., 21. Evolution of the rom composition of wines supplemented with grpe flvour precursors from different vrietls during ccelerted wine ging. Food Chemistry. 12, Loyux, D., Roger, S. & Add, J., The evolution of Chmpgne voltiles during geing. Journl of the Science of Food nd Agriculture. 32, Lubbers, S., Chrpentier, C., Feuillt, M. & Voilley, A., Influence of yest wlls on the behviour of rom compounds in model wine. Americn Journl of Enology nd Viticulture. 45, Lund, C.M., Nicolu, L., Grdner, R.C. & Kilmrtin, P.A., 29. Effect of polyphenols on the perception of key rom compounds from Suvignon blnc wines. Austrlin Journl of Grpe nd Wine Reserch. 15, Lund, C.M., Thompson, M.K., Benkwitz, F., Wohler, M.W., Triggs, C.M., Grdner, R., Heymnn, H. & Nicolu, L., 29b. New Zelnd Suvignon blnc distinct flvor chrcteristics: Sensory, chemicl, nd consumer spects. Americn Journl of Enology nd Viticulture. 6, Mg, J.A., Sensory nd stbility properties of dded methoxypyrzines to model nd uthentic wines. Proc. 6th Interntionl Flvor Conference: Flvors nd Off flvors.. Rethymnon, Grete, Greece. pp Mggu, M., Winz, R., Kilmrtin, P.A., Trought, M.C.T. & Nicolu, L., 27. Effect of skin contct nd pressure on the composition of Suvignon blnc must. Journl of Agriculturl nd Food Chemistry. 55, Mkhotkin, O. & Kilmrtin, P.A., 29. Uncovering the influence of ntioxidnts on polyphenol oxidtion in wines using n electrochemicl method: Cyclic voltmmetry. Journl of Electronlyticl Chemistry. 633, Mkhotkin, O. & Kilmrtin, P.A., 212. Hydrolysis nd formtion of voltile esters in New Zelnd Suvignon blnc wine. Food Chemistry. 135, Mris, J., The effect of ph on esters nd qulity of Colombr wine during mturtion. Vitis. 19, Mris, J., Terpenes in the rom of grpes nd wines: A review. South Africn Journl of Enology nd Viticulture. 4, Mris, J., Suvignon blnc cultivr rom review. South Africn Journl of Enology nd Viticulture. 15, Mris, J., Effect of grpe temperture, oxidtion nd skin contct on Suvignon blnc juice nd wine composition nd wine qulity. South Africn Journl of Enology nd Viticulture. 19,

65 Mris, J., 21. Effect of grpe temperture nd yest strin on Suvignon blnc wine rom composition nd qulity. South Africn Journl of Enology nd Viticulture. 22, Mris, J., Minnr, P. & October, F., Methoxy 3 isobutylpyrzine levels in spectrum of South Africn Suvignon blnc wines. South Africn Journl of Enology nd Viticulture. 19, Mris, J., Minnr, P. & October, F., Methoxy 3 isobutylpyrzine levels in spectrum of South Africn Suvginon blnc wines. In Wynboer (Februry). Mris, J. & Pool, H.J., 198. Effect of storge time nd temperture on the voltile composition nd qulity of dry white tble wines. Vitis. 19, Mris, J. & Swrt, E., Sensory impct of 2 methoxy 3 isobutylpyrzine nd 4 mercpto 4 methylpentn 2 one dded to neutrl Suvignon blnc wine. South Africn Journl of Enology nd Viticulture. 2, Mrglit, Y., Concepts in wine chemistry. The wine pprecition guild, Sn Frncisco. Mrglith, P.Z., Flvour Microbiology. Chrles C. Thoms Publishers, Springfield, Illinois. Mrtin, B. & Etiévnt, P.X., Quntittive determintion of solerone nd sotolon in flor sherries by two dimensionl cpillry GC. Journl of High Resolution Chromtogrphy & Chromtogrphy Communictions. 14, Mrtin, B., Etiévnt, P.X., Le Queré, L. & Schlich, P., More clues bout the sensory impct of sotolon in flor sherry wines. Journl of Agriculturl nd Food Chemistry. 4, Mrtínez, P., Vlcárcel, M.J., Piérez, L. & Benítez, T., Metbolism of Scchromyces cerevisie flor yests during fermenttion nd biologicl ging of fino sherry: By products nd rom compounds. Americn Journl of Enology nd Viticulture. 49, Msneuf Pomrède, I., Mnsour, C., Murt, M.L., Toming, T. & Dubourdieu, D., 26. Influence of fermenttion temperture on voltile thiols concentrtions in Suvignon blnc wines. Interntionl Journl of Food Microbiology. 18, Msud, M., Okw, E., Nishimur, K. & Yunome, H., Identifiction of 4,5 dimethyl 3 hydroxy 2(5H) furnone (sotolon) nd ethyl 9 hydroxynonnote in botrytised wine nd evlution of the roles of compounds chrcteristic. Agriculturl nd Biologicl Chemistry. 48, Mteo Vivrcho, L., Zpt, J., Ccho, J. & Ferreir, V., 21. Anlysis, occurrence, nd potentil sensory significnce of five polyfunctionl mercptns in white wines. Journl of Agriculturl nd Food Chemistry. 58, Meilgrd, M., Civille, C.V. & Crr, B.T., 26. Sensory evlution techniques. CRC, Boc Rton, FL. Meilgrd, M.C., Flvour chemistry of beer. Prt II: Flvour nd threshold of 239 rom voltiles. Mster Brewers Assocition of the Americs, Technicl Qurterly. 12, Miyke, T. & Shibmoto, T., Quntittive nlysis of cetldehyde in foods nd beverges. Journl of Agriculturl nd Food Chemistry. 41,

66 Modic, E., Znletti, R., Freccero, M. & Mell, M., 21. Alkyltion of mino cids nd glutthione in wter by o quinone methide. Rectivity nd selectivity. The Journl of Orgnic Chemistry. 66, Mongs, M., Brtolomé, B. & Gómez Cordovés, C., 25. Updted knowledge bout the presence of phenolic compounds in wine. Criticl Reviews in Food Science nd Nutrition. 45, Moutounet, M., Rbier, P., Puench, J.L., Verette, E. & Brillére, J.M., Anlysis of HPLC of extrctble substnces of ok wood ppliction to Chrdonny wine. Sciences des Aliments. 9, Murt, M.L., 25. Recent findings on rosé roms. Prt 1: Identifying roms studying the romtic potentil of grpes nd juice. Austrlin nd New Zelnd grpegrower nd winemker. 497, Murt, M.L., Msneuf, I., Drriet, I., Lvigne, V., Toming, T. & Dubourdieu, D., 21. Effect of Scchromyces cerevisie yest strins on the libertion of voltile thiols in Suvignon blnc wine. Americn Journl of Enology nd Viticulture. 52, Murt, M.L., Toming, T. & Dubourdieu, D., 21b. Assessing the romtic potentil of Cbernet Suvignon nd Merlot musts used to produce rosé wine by ssying the cysteinylted precursor of 3 mercptohexn 1 ol. Journl of Agriculturl nd Food Chemistry. 49, Murt, M.L., Toming, T., Sucier, C., Glories, Y. & Dubourdieu, D., 23. Effect of nthocynins on stbility of key odourous compound, 3 mercptohexn 1 ol, in Bordeux rosé wines. Americn Journl of Enology nd Viticulture. 54, Næs, T., Bruckhoff, P.B. & Tomic, O., 21. Qulity control of sensory profile dt. John Wiley nd Sons, Ltd, United Kingdom. Nestrud, M.A. & Lwless, H.T., 28. Perceptul mpping of citrus juices using projective mpping nd profiling dt from culinry professionls nd consumers. Food Qulity nd Preference. 19, Nguyen, D. D., Nicolu, L., Dykes, S.I. & Kilmrtin, P.A., 21. Influence of microoxygention on reductive sulfur off odours nd color development in Cbernet Suvignon wine. Americn Journl of Enology nd Viticulture. 61, Nicolu, L., Benkwitz, F. & Toming, T., 26. Chrcterising the rom of New Zelnd Suvignon blnc. Austrlin nd New Zelnd grpegrower nd winemker. 59, Nikolntonki, M., Chichuc, I., Teissedre., P.L. & Drriet, P., 21. Rectivity of voltile thiols with polyphenols in wine model medium: Impct of oxygen, iron, nd sulfur dioxide. Anlytic Chimic Act. 66, Nikolntonki, M., Jourdes, M., Shinod, K., Teissedre, P.L., Quideu, S. & Drriet, P., 212. Identifiction of dducts between n odoriferous voltile thiol nd oxidized grpe phenolic compounds: 52

67 Kinetic study of dduct formtion under chemicl nd enzymtic oxidtion conditions. Journl of Agriculturl nd Food Chemistry. 6, Nikolntonki, M. & Wterhouse, A.L., 212. A method to quntify quinone rection rtes with wine relevnt nucleophiles: A key to the understnding of oxidtive loss of vrietl thiols. Journl of Agriculturl nd Food Chemistry. 6, Noble, A.C., Arnold, R.A., Buechsenstein, J., Lech, E.J., Schmidt, J.O. & Stern, P.M., Modifiction of stndrdized system of wine rom terminology. Americn Journl of Enology nd Viticulture. 38, Nygrd, M., 21. Oxygen mngement from grpes to glss. Wine Industry Journl. 25, Nygrd, M., Osidcz, P., Roget, W., Frncis, L., Vidl, S. & Agrd, O., 21. The effect of closure choice on consumer preference rting of wines. Austrlin nd New Zelnd grpegrower nd winemker. 563, Nykänen, L., Formtion nd occurrence of flvor compounds in wine nd distilled lcoholic beverges. Americn Journl of Enology nd Viticulture. 37, Nykänen, L. & Suomlinen, H., Arom of beer, wine nd distilled lcoholic beverges. Reidel, Dordrecht. O'Brien, V., Frncis, L. & Osidcz, P., 29. Pckging choices ffect consumer enjoyment of wines. Wine Industry Journl. 24, Oliveir, A.C., Silv Ferreir, A.C., Guedes de Pinho, P. & Hogg, T.A., 22. Development of potentiometric method to mesure the resistnce to oxidtion of white wines nd the ntioxidnt power of their constituents. Journl of Agriculturl nd Food Chemistry. 5, Oliveir, J.M., Fri, M., Sá, F., Brros, F. & Arújo, I.M., 26. C 6 lcohols s vrietl mrkers for ssessment of wine origin. Anlytic Chimic Act. 563, Oszminski, J., Cheynier, V. & Moutounet, M., Iron ctlyzed oxidtion of (+) ctechin in model systems. Journl of Agriculturl nd Food Chemistry. 44, Oszminski, J., Spis, J.C. & Mcheix, J.J., Chnges in grpe seed phenols s ffected by enzymic nd chemicl oxidtion in vitro. Journl of Food Science. 5, 155. Ppdopoulou, D. & Roussis, I.G., 21. Inhibition of the decline of linlool nd α terpineol in Musct wines by glutthione nd N cetyl cysteine. Itlin Journl of Food Science. 13, Ppdopoulou, D. & Roussis, I.G., 28. Inhibition of the decrese of voltile esters nd terpenes during storge of white wine nd model wine medium by glutthione nd N cetylcysteine. Interntionl Journl of Food Science nd Technology. 43, Prk, S.K., Boulton, R.B., Brtr, E. & Noble, A.C., Incidence of voltile sulfur compounds in Cliforni wines. A preliminry survey. Americn Journl of Enology nd Viticulture. 45,

68 Prr, W.V., Green, J.A., White, K.G. & Sherlock, R.R., 27. The distinctive flvour of New Zelnd Suvignon blnc: Sensory chrcteristion by wine professionls. Food Qulity nd Preference. 18, Ptel, P., Herbst Johnstone, M., Lee, S.A., Grdner, R.C., Wever, R., Nicolu, L. & Kilmrtin, P.A., 21. Influence of juice pressing conditions on polyphenols, ntioxidnts, nd vrietl rom of Suvignon blnc microferments. Journl of Agriculturl nd Food Chemistry. 58, Pul, F., SO 2 reltionships in fermenting red wine mshes nd musts. Proc. Fourth Interntionl Enologicl Symposium. Vlenci/Spin. pp Peng, Z., Duncn, B., Pocock, K.F. & Sefton, M.A., The effect of scorbic cid on oxidtive browning of white wines nd model wines. Austrlin Journl of Grpe nd Wine Reserch. 4, Penn, N.C., Dudt, C.E., Brendel, M. & Henriques, J.A.P., 21. Evolution of glutthione levels nd hydroxycinnmic cid during fermenttion of Suvignon blnc, Chenin blnc, Niágr nd Tnnt. Alimentri. 326, Peyrot des Gnchos, C., Toming, T. & Dubourdieu, D., 22. Sulfur rom precursor present in S glutthione conjugte form: identifiction of S 3 (hexn 1 ol) glutthione in must from Vitis vinifer L. cv. Suvignon blnc. Journl of Agriculturl nd Food Chemistry. 5, Peyrot des Gnchos, C., Vn Leewin, C., Toming, T., Soyer, J. P., Gudillere, J. P. & Dubourdieu, D., 25. Influence of wter nd nitrogen deficit on fruit ripening nd rom potentil of Vitis vinifer L. cv. Suvignon blnc in field conditions. Journl of Agriculturl nd Food Chemistry. 85, Phm, T.T., Guichrd, E., Pscl, S. & Chrpentier, C., Optiml conditions for the formtion of Sotolon from α ketobutyric cid in the French "Vin June". Journl of Agriculturl nd Food Chemistry. 43, Pickering, G.J., Lin, J., Reynolds, A., Soles, G. & Riesen, R., 26. The evlution of remedil tretments for wine ffected by Hrmoni xyridis. Interntionl Journl of Food Science nd Technology. 41, Pineu, B., Brbe, J. C., Vn Leewin, C. & Duboudieu, D., 27. Which impct for β dmscenone on red wine rom? Journl of Agriculturl nd Food Chemistry. 55, Pons, A., Lvigne, V., Lndis, Y., Drriet, P. & Duboudieu, D., 21. Identifiction of sotolon pthwy in dry white wines. Journl of Agriculturl nd Food Chemistry. 58, Pripis Nicolu, L., de Revel, G., Bertrnd, A. & Mujen, A., 2. Formtion of flvour components by the rection of mino cid nd crbonyl compounds in mild conditions. Journl of Agriculturl nd Food Chemistry. 48, Rmey, D.D. & Ough, C.S., 198. Voltile ester hydrolysis or formtion during storge of model solutions nd wines. Journl of Agriculturl nd Food Chemistry. 28,

69 Rnkine, B.C., The importnce of yest in determining the composition nd qulity of wines. Vitis. 7, Rnkine, B.C. & Pocock, K.F., Influence of yest strins on binding of sulphur dioxide in wines, nd on its formtion during fermenttion. Journl of the Science of Food nd Agriculture. 2, Rpp, A., Studies on terpene compounds in wines. In Frontiers of flvor. Chrlmbous, G. (ed.), pp Elsevier, Amsterdm. Rpp, A. & Mndery, H., Wine Arom. Experimenti. 42, Ruhut, D., Yests production of sulfur compounds. In Wine microbiology nd biotechnology. Fleet, G.H. (ed.), pp Hrdwood Acdemic, Chur, Switzerlnd. Renouil, Y., Dictionire du vin. Tours: Sézme. Ribéreu Gyon, P., Boidron, J.N. & Terrier, A., Arom of Musct grpe vrieties. Journl of Agriculturl nd Food Chemistry. 23, Ribéreu Gyon, P., Dubourdieu, D., Doneche, B. & Lonvud, A., 26. Hndbook of Enology. John Wiley & Sons Ltd, Chichester. Ribéreu Gyon, P., Glories, Y., Dubourdieu, D., Mujen, A., 1998, 24b. Trite d'oenologie. Chimie du vin. Stbilistion et tritements., Dunod Pris. Ribéreu Gyon, P., Glories, Y., Mujen, A., Dubourdieu, D., 26. Hndbook of Enology: The chemistry of wine stbiliztion nd tretments. John Wiley & Sons Ltd, Chichester. Risvik, E., McEvn, J.A., Colwill, J.S., Rogers, R. & Lyon, D.H., Projective mpping: A tool for sensory nlysis nd consumer reserch. Food Qulity nd Preference. 5, Risvik, E., McEwn, J.A. & Rødbotten, M., Evlution of sensory profiling nd projective mpping dt. Food Qulity nd Preference. 8, Rizzi, G.P., 26. Formtion of Strecker ldehydes from polyphenol derived quinones nd α mino cids in nonenzymtic model system. Journl of Agriculturl nd Food Chemistry. 54, Rolnd, A., Cvelier, F. & Schneider, R., 211. New insights in vrietl thiols biogenesis in wine. Proc. Interntionl conference series on wine ctive compounds. Beune, Frnce. pp Rolnd, A., Vilret, J., Rzungles, A., Rigou, P. & Schneider, R., 21. Evolution of S Cysteinylted nd S glutthionylted thiol precursors during oxidtion of Melon B. nd Suvignon blnc musts. Journl of Agriculturl nd Food Chemistry. 58, Rossi, J.A. & Singleton, V.L., Contribution of grpe phenols to oxygen bsorption nd browning of white wine. Americn Journl of Enology nd Viticulture. 17, Roussis, I.G., Lmbropoulos, I. & Ppdopoulou, D., 25. Inhibition of the decline of voltile esters nd terpenols during oxidtive storge of Musct white nd Xinomvro red wine by cffeic cid nd N cetyl cysteine. Journl of Agriculturl nd Food Chemistry. 93,

70 Roussis, I.G., Lmbropoulos, I. & Tzims, P., 27. Protection of voltiles in wine with low sulfur dioxide by cffeic cid or glutthione. Americn Journl of Enology nd Viticulture. 58, Ryon, I., Pn, B., INtrigliolo, D., Lkso, A. & Scks, G., 28. Effect of cluster light exposure on 3 isobutyl 2 methoxypyrzine ccumultion nd degrdtion ptterns in red wine grpes (Vitis vinifer L. Cv. Cbernet frnc). Journl of Agriculturl nd Food Chemistry. 56, Sl, C., Busto, O., Gusch, J. & Zmor, F., 24. Influence of vine trining nd sunlight exposure on the 3 lkyl 2 methoxypyrzine content in must nd wines from Vitis vinifer vriety Cbernet Suvignon. Journl of Agriculturl nd Food Chemistry. 52, Slmon, J.M., Fornison, C. & Moutounet, M., Modifiction chimique des lies consécutive à l pport d oxygène pendnt l élevge des vins sur lies. Proc. Oenologie 99, 6eme Symposium Interntionl d'oenologie. Pris, Frnce. pp Sn Jun, F., Ccho, J., Ferreir, V. & Escudero, A., 212. Arom chemicl composition of red wines from different price ctegories nd its reltionship to qulity. Journl of Agriculturl nd Food Chemistry. 6, Srrzin, E., Dubourdieu, D. & Drriet, D., 27. Chrcteriztion of key rom compounds of botrytized wines, influence of grpe botrytiztion. Food Chemistry. 13, Schneider, R., Chrrier, F., Rzungles, A. & Bumes, R., 26. Evidence for n lterntive biogenetic pthwy leding to 3 mercptohexnol nd 4 mercpto 4 methylpentn 2 one in wines. Anlytic Chimic Act. 563, Schneider, V., 23. Alterung von Weißwein, I: Die Reduktionskrft der Hefe. Die Winzer Zeitung. 5, Schneider, V., 25. Postfermenttive Phse: Die Hefe nch der Gärung. Der Winzer. 11, Schreier, P., Flvour composition of wines: review. CRC Criticl Reviews in Food Science & Nutrition. 12, Segurel, M.A., Rzungles, A.J., Riou, C., Triguiero, M.G.L. & Bumes, R., 25. Ability of possible DMS precursors to relese DMS during wine ging in the conditions of het lkline tretment. Journl of Agriculturl nd Food Chemistry. 53, Silv Ferreir, A.C., Brbe, J. C. & Bertrnd, A., 22. Heterocyclic cetls from glycerol nd cetldehyde in Port wines: Evolution with ging. Journl of Agriculturl nd Food Chemistry. 5, Silv Ferreir, A.C., Brbe, J. C. & Bertrnd, A., Hydroxy 4,5 dimethyl 2(5H) furnone: A key odornt of the typicl rom of oxidtive ged Port wine. Journl of Agriculturl nd Food Chemistry. 51, Silv Ferreir, A.C., Guedes de Pinho, P., Rodrigues, P. & Hogg, T., 22b. Kinetics of oxidtive degrdtion of white wines nd how they re ffected by selected technologicl prmeters. Journl of Agriculturl nd Food Chemistry. 5,

71 Silv Ferreir, A.C., Hogg, T. & De Pinho, P.G., 23b. Identifiction of key odornts relted to the typicl rom of oxidtion spoiled white wines. Journl of Agriculturl nd Food Chemistry. 51, Silv Ferreir, A.C., Hogg, T. & Guedes de Pinho, P., 23c. Identifiction of key odornts relted to the typicl rom of oxidtion spoiled white wines. Journl of Agriculturl nd Food Chemistry. 51, Silv Ferreir, A.C., Rodrigues, P., Hogg, T. & De Pinho, P.G., 23d. Influence of some technologicl prmeters on the formtion of dimethyl sulfide, 2 mercptoethnol, methionol, nd dimethyl sulfone in port wines. Journl of Agriculturl nd Food Chemistry. 51, Simpson, R.F., Arom nd compositionl chnges in wine with oxidtion, storge nd geing. Vitis. 17. Simpson, R.F., Fctors ffecting oxidtive browning of white wine. Vitis. 21, Singleton, V.L., Oxygen with phenols nd relted rections in must, wines nd model systems: observtions nd prcticl implictions. Americn Journl of Enology nd Viticulture. 38, Singleton, V.L., Slgues, J., Zy, J. & Trousdle, E., Cftric cid disppernce nd conversion to products of enzymtic oxidtion in grpe must nd wine. Americn Journl of Enology nd Viticulture. 36, Singleton, V.L., Timberlke, C.F. & Le, A.G.H., The phenolic cinnmtes of white grpes nd wines. Journl of Sciences nd Food Agriculture. 29, Singleton, V.L., Trousdle, E. & Zy, J., Oxidtion of wines I. Young white wines periodiclly exposed to ir. Americn Journl of Enology nd Viticulture. 3, Singleton, V.L., Zy, J., Trousdle, E. & Slgues, M., Cftric cid in grpes nd conversion to rection product during processing. Vitis. 23, Skouroumounis, G.K., Kwitkowski, M., Frncis, I.L., Okey, H., Cpone, D., Duncn, B., Sefton, M.A. & Wters, E.J., 25. The impct of closure type nd storge conditions on the composition, colour nd flvour properties of Riesling nd wooded Chrdonny wine during five yers' storge. Austrlin Journl of Grpe nd Wine Reserch. 11, Sneyd, T.N., Leske, P.A. & Dunsford, P.A., How Much Sulfur? Proc. 8th Austrlin Wine Industry Technicl Conference. Melbourne, Victori. pp Sores d Cost, M., Gonçlves, C.F., A., Ibsen, C., Guedes de Pinho, P. & Silv Ferreir, A.C., 24. Further insights into the role of methionl nd phenylcetldehyde in lger beer flvour tbility. Journl of Agriculturl nd Food Chemistry. 52, Soles, R.M., Ough, C.S. & Kunkee, R.E., Ester concentrtion differences in wine fermented by vrious species nd strins of yests. Americn Journl of Enology nd Viticulture. 33, Somers, T.C. & Wescombe, L.G., Evolution of red wines. I. An ssessment of the role of cetldehyde. Vitis. 26,

72 Sonni, F., Clrk, A.C., Prenzler, P.D., Riponi, C. & Scollry, G.R., 211. Antioxidnt ction of glutthione nd the scorbic cid/glutthione pir in model white wine. Journl of Agriculturl nd Food Chemistry. 59, Stone, H. & Sidel, J.L., Sensory evlution prctices. Cliforni, USA. Stone, H. & Sidel, J.L., 24. Sensory evlution prctices. Acdemic, Orlndo, FL. Subileu, M., Schneider, R., Slmon, J.M. & Degryse, E., 28. New insights on 3 mercptohexnol (3MH) biogenesis in Suvignon blnc wines: Cys 3MH nd (E) Hexen 2 l re not the mjor precursor. Journl of Agriculturl nd Food Chemistry. 56, Subileu, M., Schneider, R., Slmon, J.M. & Degryse, E., 28b. Nitrogen ctbolite repression modultes the production of romtic thiols chrcteristic of Suvignon blnc t the level of precursor trnsport. FEMS Yest Reserch. 8, Suovniemi, O., Slspuro, M., Slspuro, V. & Mrvol, M., 26. Food composition for binding cetldehyde in mouth nd digestive trct, nd method for the preprtion of the composition. Interntionl Ptent Appliction WO 26/13316 A Swiegers, J.H., Frncis, I.L., Herderich, M.J. & Pretorius, I.S., 26. Meeting consumer expecttions through mngement in vineyrd nd winery: the choice of yest for fermenttion offers gret potentil to djust the rom of Suvignon blnc wine. Austrlin nd New Zelnd Wine Industry Journl. 21, Swiegers, J.H., Willmott, R., Hill Ling, A., Cpone, D.L., Prdon, K.H., Elsey, G.M., Howell, K.S., de Brros Lopes, M.A., Sefton, M.A., Lilly, M. & Pretorius, I.S., 25. Modultion of voltile thiol nd ester roms by modified wine yest. Proc. Weurmn flvour reserch symposium. Roskilde, Denmrk. Szczesnik, A.S., Generl foods texture profile revisited ten yers perspective. Journl of Texture Studies. 6, Tkhshi, T., Tdenum, M. & Sto, S., Hydroxy 4,5 dimethyl 2(5H) furnone, burnt flvouring compound from ged ske. Agriculturl nd Biologicl Chemistry. 4, Thibon, C., Dubourdieu, D., Drriet, P. & Toming, T., 29. Impct of noble rot on the rom precursor of sulfnylhexnol content in Vitis vinifer L. Cv Suvignon blnc nd Semillon grpe juice. Food Chemistry. 114, Thurston, P.A., Tylor, R. & Ahveninen, J., Effects of linoleic cid supplements on the synthesis by yest of lipids nd cette esters. Journl of the Institute of Brewing. 87, Tomic, O., Lucino, G., Nilsen, A., Hyldig, G., Lorensen, K. & Næs, T., 21. Anlysing sensory pnel performnce in proficiency tests using the PnelCheck softwre. Europen Food Reserch nd Technology. 23, Tomic, O., Nilsen, A., Mrtens, M. & Næs, T., 27. Visuliztion of sensory profiling dt for performnce monitoring. LWT Food Science nd Technology. 4,

73 Toming, T., 23. Contribution of benzenemethnethiol to the smoky roms of certin Vitis vinifer wines. Journl of Agriculturl nd Food Chemistry. 51, Toming, T., Bltenweck Guyot, R., Peyrot des Gchons, C. & Dubourdieu, D., 2. Contribution of voltile thiols to the roms of white wines mde from severl Vitis vinifer grpe vrieties. Americn Journl of Enology nd Viticulture. 51, Toming, T., Drriet, P. & Dubourdieu, D., Identifiction of 3 mercptohexyl cette in Suvignon wine, powerful romtic compound exhibiting box tree odor. Vitis. 35, Toming, T., Furrer, A., Henry, R. & Dubourdieu, D., Identifiction of new voltile thiols in the rom of Vitis vinifer L. vr. Suvignon blnc wines. Flvour nd Frgrnce Journl. 13, Toming, T., Guimbertu, G. & Dubourdieu, D., 23. Role of certin voltile thiols in the bouquet of ged Chmpgne wines. Journl of Agriculturl nd Food Chemistry. 51, Toming, T., Msneuf, I. & Dubourdieu, D., 24. Powerful romtic voltile thiols in wines mde from severl Vitis vinifer L. cv. Suvignon blnc. Proc. ACS Symposium Series. pp Toming, T., Murt, M.L. & Dubourdieu, D., 1998b. Development of method nlyzing the voltile thiols involved in the chrcteristic rom of wines mde from Vitis vinifer L. cv. Suvignon blnc. Journl of Agriculturl nd Food Chemistry. 46, Toming, T., Peyrot des Gchons, C. & Dubourdieu, D., 1998c. A new type of flvour precursors in Vitis vinifer L. cv. Suvignon blnc: S Cysteine conjugtes. Journl of Agriculturl nd Food Chemistry. 46, Toukis, G., Chemistry of wine stbiliztion. A review. In Chemistry of winemking. Webb, A.D. (ed.), pp Advn. Chem. Americn Society, Wshington. Tsi Su, C. & Singleton, V.L., Identifiction of three flvn 3 ols from grpes. Phytochemistry. 8, Uglino, M., 213. Oxygen contribution to wine rom evolution during bottle ging. Journl of Agriculturl nd Food Chemistry. 61, Uglino, M., Dievl, J., Siebert, T., Kwitkowski, M., Agrd, O., Vidl, S. & Wters, E.J., 212. Oxygen consumption nd development of voltile sulfur compounds during bottle ging of two Shirz wines. Influence of pre nd postbottling controlled oxygen exposure. Journl of Agriculturl nd Food Chemistry. 6, Uglino, M., Dievl, J.B., Dimkou, E., Wirth, J., Cheynier, V., Jung, R. & Vidl, S., 213. Controlling oxygen t bottling to optimize post bottling development of wine. Prcticl Winery & Vineyrd Journl. 34, Uglino, M., Kwitkowski, M., Trvis, B., Frncis, I.L., Wters, E.J., Herderich, M. & Pretorius, I.S., 29. Post bottling mngement of oxygen to reduce off flvour formtion nd optimize wine style. Austrlin nd New Zelnd Wine Industry Journl. 24,

74 Uglino, M., Kwitkowski, M., Vidl, S., Cpone, D., Siebert, T., Dievl, J., Agrd, O. & Wters, E.J., 211. Evolution of 3 mercptohexnol, hydrogen sulfide, nd methyl mercptn during bottle storge of Suvignon blnc wines. Effect of glutthione, copper, oxygen exposure, nd closurederived oxygen. Journl of Agriculturl nd Food Chemistry. 59, Uglino, M., Kwitkowski, M., Vidl, S., Cpone, D., Solomon, M., Dievl, J.B., Agrd, O. & Wters, E.J., 21. The role of copper nd glutthione ddition nd oxygen exposure in the evlution of key rom compounds of Suvignon Blnc. Poster presented t 14th AWITC, Adelide. Usseglio Tomsset, L., Properties nd use of sulphur dioxide. Food Additives & Contminnts. 9, Vn der Merwe, C.A. & Vn Wyk, C.J., The contribution of some fermenttion products to the odor of dry white wines. Americn Journl of Enology nd Viticulture. 32, Vn Wyngrd, E., 213. Voltiles plying n importnt role in South Africn Suvignon blnc wines. Thesis, University of Stellenbosch, Stellenbosch. Versini, G., Inm, S. & Srtori, G., A cpillry column gs chromtogrphic reserch into the terpene constituents of "Riesling Romno" (Rhine Riesling) wine from Trentino Alto Adige: Their distribution within berries, their pssge into the must nd their presence in the wine ccording to different wine mking procedures. Orgnoleptic considertions. Vini d'itli. 23, Vivs de Gulejc, N., Vivs, N., Nonier, M., Abslon, C. & Bourgeois, G., 21. Study nd quntifiction of monomeric flvn 3 ol nd dimeric procynidin quinonic forms by HPLC/ESI MS. Applictions to red wine oxidtion. Journl of the Science of Food nd Agriculture. 81, Vrhovšek, U., Extrction of hydroxycinnmoyltrtric cids from berries of different grpe vrieties. Journl of Agriculturl nd Food Chemistry. 46, Wllington, N., Clrk, A.C., Prenzler, P.D., Brril, C. & Scollry, G.R., 213. The decy of scorbic cid in model wine system t low oxygen concentrtion. Food Chemistry. 141, Wterhouse, A.L. & Lurie, V.F., 26. Oxidtion of wine phenolics: A criticl evlution nd hypotheses. Americn Journl of Enology nd Viticulture. 57, Wilderndt, H.L. & Singleton, V.L., The production of ldehydes s result of oxidtion of polyphenolic compounds nd its reltion to wine ging. Americn Journl of Enology nd Viticulture. 25, Willims, A.A. & Lngron, S.P., The use of free choice profiling for the exmintion of commercil ports. Journl of the Science of Food nd Agriculture. 35, Willims, P.J., Struss, C.R. & Wilson, B., 198. Hydroxylted linlool derivtives s precursors of voltile monoterpenes of Musct grpes. Journl of Agriculturl nd Food Chemistry. 28, Youngblood, M.P., Kinetics nd mechnism of the ddition of sulfite to p benzoquinone. The Journl of Orgnic Chemistry. 51,

75 Yun, J. P. & Chen, F., Degrdtion of scorbic cid in queous solution. Journl of Agriculturl nd Food Chemistry. 46, Ze, L., Moyno, L., Ruiz, M.J. & Medin, M., 21. Chromtogrphy olfctometry study of the rom of fino sherry wines. Interntionl Journl of Anlyticl Chemistry. Article ID doi:1.1155/21/ Zoecklein, B.W., Fugelsng, K.C., Gump, B.H. & Nury, F.S., Wine nlysis nd production. Chpmn & Hll, London. 61

76 Chpter 3 Reserch results Sensory interction studies combining vrietl nd oxidtion relted compounds tht occur in Suvignon blnc wines 62

77 Chpter 3: Reserch Results Sensory interction studies combining vrietl nd oxidtion relted compounds tht occur in Suvignon blnc wines 3.1 INTRODUCTION MATERIALS AND METHODS Medium Chemicls nd spiking Experimentl design Sensory nlysis Dt Anlysis RESULTS AND DISCUSSION Pnel evlution Individul compound evlution Multiple compound evlution CONCLUSION ABBREVIATIONS USED REFERENCES 92 63

78 3.1 INTRODUCTION Wine rom is not the result of single rom ctive compounds, but rther due to the vrious complex interctions between specific compounds (Escudero et l., 27; Fischer, 27). These interctions give rise to wine s tste nd rom (Dll'Ast et l., 211). Over the lst two decdes rom reconstitution pproches hve dded new dimension to sensory investigtions. Sensory interctions of chemicl compounds t different concentrtions nd in vrious medi hs been investigted (Ribéreu Gyon et l., 1975; Vn der Merwe & Vn Wyk, 1981; Mg, 1989; Guth, 1997; Mris & Swrt, 1999; Ferreir et l., 22; Lee, 23; Aronson & Ebeler, 24; Cmpo et l., 25; Frncis & Newton, 25; Ferreir et l., 26; Culleré et l., 27; Escudero et l., 27; Pineu et l., 27; Lund et l., 29; King et l., 211; Benkwitz et l., 212; Vn Wyngrd, 213). Studies performed by dding rnge of rom compounds to vrious medi show complex interctions between groups of rom compounds such s esters, thiols, lcohols, pyrzines, ketones nd terpenes (Ribéreu Gyon et l., 1975; Ferreir et l., 22; Cmpo et l., 25; Escudero et l., 27; Pineu et l., 27; King et l., 211). Interctive studies between thiols, esters nd 3 isobutyl 2 methoxypyrzine (IBMP) demonstrted the complex sensory interctions (synergism, suppression nd enhncement) of the compounds nd unexpected contributions of certin thiols nd combintion of esters nd thiols to the green chrcteristics (Mris & Swrt, 1999; King et l., 211). Furthermore, it hs previously been shown tht non voltile polyphenols cn ffect the perception of vrious rom compounds including thiols, methoxypyrzines nd esters (Lund et l., 29). The bove mentioned studies lmost exclusively investigte the effect of positively relted compounds on ech other. Limited reserch hs been performed on the intensity of sensory ttributes generted or the modifiction of these ttributes by vrying concentrtions of individul rom compounds or combintions of compounds. The suppression/msking/enhncing effect of off odours or fulty odours should be investigted s it cn develop in wine during ging. Suvignon blnc is complex cultivr nd cn yield very different wine styles. The voltile thiols (4 mercpto 4 methylpentn 2 one (4MMP), 3 mercptohexn 1 ol (3MH) nd 3 mercptohexn 1 ol cette (3MHA)) cn led to tropicl nd fruity style wines, while the methoxypyrzines (3 isobutyl 2 methoxypyrzine (IBMP), 3 isopropyl 2 methoxypyrzine (IPMP) nd 3 sec butyl 2 methoxypyrzine (SBMP)) cn contribute green pepper, grssy nd sprgus rom of the wine (Lcey et l., 1991; Toming et l., 1998; Swiegers et l., 26; Coetzee & Du Toit, 212). Both of these odour groups re often well sought fter for the production of qulity wines (Lund et l., 29b; King et l., 211). 64

79 The voltile thiols, 3MH nd 4MMP, re relesed from their precursors by yest during lcoholic fermenttion (Subileu et l., 28). 3MHA results from the esterifiction of 3MH by yest (Swiegers et l., 29). The voltile thiols re known to be sensitive to oxidtion nd cid hydrolysis nd their concentrtions decrese during ging (Toming et l., 24; Herbst et l., 28). The methoxypyrzines re, however, not sensitive to oxidtion nd studies hve shown methoxypyrzine concentrtions to remin constnt during oxidtive hndling of juice nd wine (Mris, 1998; Coetzee et l., 213). Methoxypyrzines re present in grpes nd their concentrtion decreses during grpe mturtion (Lcey et l., 1991). A comprehensive review on Suvignon blnc rom (with focus on voltile thiols) nd the rection of vrious rom groups to oxidtion hs been published previously (Coetzee & Du Toit, 212). During ging, new rom compounds form which cn chnge the romtic composition of wine. In white wine, compounds such s cetldehyde, 3 (methylthio) propionldehyde (methionl), phenylcetldehyde nd sotolon hve been shown to ply key role in the developed ging chrcter of white wines. These compounds contribute to the typicl ging chrcter by lending rom ttributes such s green pple skin, cooked potto, flower, honey nd curry. (Silv Ferreir et l., 22; Frivik & Ebeler, 23; Silv Ferreir et l., 23; Jckowetz & De Orduñ, 213). A study done on oxidtion relted ldehydes nd how they ffect ech other showed the dditive/enhncing effect of the ldehydes on ech other s well s the importnce of the compounds even when present t concentrtions below their perception thresholds (Culleré et l., 27). The study lso demonstrted the effect of groups of compounds (brnched liphtic ldehydes vs (E) 2 lkenls) on ech other. However, to dte very little reserch hs focussed on mixture of fresh nd oxidtive or ged roms nd the question still remins how these rom compounds will interct when combined in wine like medium. The sensory impct of the compounds t wide rnge of concentrtions (including below nd t the perception threshold) would lso be importnt to investigte. In this study, the sensory interctions between the vrietl chrcters of Suvignon blnc nd typicl white wine oxidtion compounds were investigted. Vrious compounds were dded singulrly nd in combintion to model wine medium nd evluted by trined sensory pnel using descriptive nlysis. Testing the interctions should clrify the evolution of white wine rom during bottle ging nd possibly serve s model to evlute t wht concentrtions interctions become detrimentl to the wine romtic composition. 65

80 3.2 MATERIALS AND METHODS Medium Model wine ws the preferred choice of spiking medium s the complexity of wine could interfere with the interction study. Model wine consisted of distilled wter, 5 g/l trtric cid nd 12% ethnol nd ph djusted to 3.5 using sodium hydroxide. After prepring the model wine, the composition ws confirmed using WineScn FT 12 instrument (FOSS Anlyticl, Denmrk) Chemicls nd spiking The compounds used in this study were 3MH, IBMP, methionl nd phenylcetldehyde. Solutions of mg/l 3MH (Interchim), 25 mg/l IBMP (Sigm), 1585 mg/l methionl (Sigm) nd 1561 mg/l phenylcetldehyde (Sigm) were prepred in 99.5% ethnol (Merck Chemicls, South Afric). The exct concentrtion of 3MH ws lso confirmed by Ellmn s regent (Ellmn, 1959). 3MH nd IBMP solutions were stored in the drk t 8 C, while methionl nd phenylcetldehyde solutions were stored t 4 C. These solutions were used to spike the model wine to the desired concentrtions one hour prior to tsting Experimentl design Initilly ech compound ws spiked individully in order to ssess the effect of concentrtion chnges on the romtic perception of ech compound. The concentrtions used throughout this study were kept constnt nd cn be seen in Tble 3.1. This ws done to compre profiles of both trils (singulr dditions nd multiple dditions). Smples re numbered from 1 to 5 ccording to the level of ech of the compounds. Using the level s the indictor of the concentrtion simplifies the interprettion of the vrious figures in this study. Refer to Tble 3.1 for the concentrtion of ech compound level. A centrl composite design ws used for the sensory nlysis of the smples contining multiple compounds (to ssess interctions). The suitbility of the centrl composite design lies in the fct tht it cn be used to test interctions between different fctors. Idelly fctoril design could be used to test ll combintions, however this design would result in very lrge number of combintions to be ssessed. The centrl composite design only selects certin mount of combintions specificlly chosen by the design. The use of this method led to design with eight str 66

81 points, sixteen cube points nd one centre point which resulted in 25 smples in totl to be tested (Tble 3.2). Tble 3.1 Concentrtions tested. Level nd concentrtion Compound Mercptohexn 1 ol (ng/l) Isobutyl 2 methoxypyrzine (ng/l) Methionl (µg/l) Phenylcetldehyde (µg/l) Tble 3.2 Composition (concentrtion levels) of vrious smples in the centrl composite design. Compound Smple 3 Mercptohexn 1 ol methoxypyrzine 3 Isobutyl 2 Methionl Phenylcetldehyde Centre Cube Cube Cube Cube Cube Cube Cube Cube Cube Cube Cube Cube Cube Cube Cube Cube Str Str Str Str Str Str Str Str

82 3.2.4 Sensory nlysis The sensory pnel consisted of 11 judges (ll femle between the ge of 25 nd 54). Intensity rting ws done using 1 mm unstructured line scle, rnking intensity from none to intense. Testing ws done in booths using stndrd ISO wine tsting glsses. The booths hve stndrd rtificil dylight lighting nd the temperture ws controlled t 2±2 C. Smple glsses were mrked with rndom three digit codes unique for ech judge nd glsses were covered with plstic lid prior to sensory ssessment to prevent the rom contminting the lbortory environment. The order of the smples ws rndomised nd blnced cross the ssessors. Along with the set of smples contining the spiked compound, blnk glss contining the unspiked model wine only, ws lso provided for comprison. Pnellists evluted the smples orthonslly only nd the dt ws collected on pper bllot. For resons of humn ethics, brief explntion of the ddition of flvour to the smples ws given prior to testing, lthough cre ws tken to exclude ny informtion tht could hve cused bis. Compounds were first nlysed individully. Eight trining sessions (one hour ech) were conducted in n ttempt to rech consensus regrding different smples nd terminology used. In this wy, the pnellists were exposed to ech compound twice. At ech trining session, the pnel received ll five levels of compound s well s blnk smple consisting only of model wine. A rnge of reference stndrds ws lso given. During the first discussion session, the pnel generted descriptors nd brief line scling exercise ws done. During the second trining session the descriptors nd the smple position on the line scle ws finlised. Ech individul compound ws nlysed during forml test session in triplicte. The profiling of the smples contining multiple compounds ws done fter the evlution of the singulr compounds. At this stge the judges were lredy fmilir with the compounds. During profiling, the smples were divided into three subsets s ll smples could not be nlysed in one session. Subsets were trined nd tested on individul dys to void ftigue. The first subset ws cube points 1 8, the second subset ws cube points 9 16 nd the third subset ws the str points 1 8 s suggested by Esbensen, (22). The experimentl lyout nd the combintions of compounds for ech smple nd subset cn be seen in Tble 3.2. Two trining sessions consisting of two hours ech were used for ech subset (4 hours per subset). The centre smple (ll compounds present t level 3) ccompnied ech set during trining nd testing. During trining, the pnellists were not informed of the composition of ech smple nd reference stndrds were vilble for the pnellists during the tril (section 3.3.2). 68

83 Smples were scled on 1 mm unstructured line scle using the ttributes generted during the profiling of the singulr compounds. Anlysis were performed s 3 different tests (ech subset seprtely) with 3 replictions ech (three two hour sessions). With the set of eight spiked smples (long with the centre smple), blnk glss contining model wine only, ws lso provided for comprison. Cube points 1 8 were tested first, followed by cube points 9 16 nd str points 1 8. The judges took regulr breks between smples to prevent ftigue Dt Anlysis Assessor performnce ws evluted using PnelCheck (Version V1.4., Nofim, Tromsø, Norwy) ccording to the workflow s described by Tomic et l. (21). Sensory descriptive dt for ech ttribute ws nlysed using mixed model nlysis of vrince (ANOVA). Principle component nlysis (PCA) biplots were creted using the correltion mtrix of the men dt. Post hoc Fisher s LSD tests were used to test for significnce of sensoril differences between compound combintions nd the significnt sttisticl reltionships were modelled s network nd visulised with Cytoscpe (Shnnon et l., 23). In custom built python progrm using networkx nd scipy librries, t tests were used to test for the significnce of the sensoril differences between compound combintions, nd the significnt sttisticl reltionships were modelled s network in which coloured nodes indicted the intensity of sensory ttribute compred to the control levels (tht were indicted by n uncoloured node locted t the centre of the network). Response surfce plots were constructed (Sttistic 1, Sttsoft Inc., Tuls, USA) to investigte interctions. A p vlue threshold of.5 (p<.5) ws used for the determintion of sttisticl significnce. 3.3 RESULTS AND DISCUSSION Pnel evlution Pnel performnce ws monitored s suggested by Tomic et l. (21). 3 Wy ANOVA ws used to determine significnt ttributes nd Tucker plots were used to ssess the consensus mongst judges (Figure 3.1). For ll the significnt ttributes generted there ws definite grouping of the pnellists. For most of the ttributes this grouping ws situted between the inner nd outer ellipses. This shows good pnel consensus with 5 1% of the vrince explined by the judges. 69

84 Stellenbosch University Figure 3.1 Tucker plot showing significnce of ll the ttributes generted by the pnel Individul compound evlution Model wine ws used s the testing medium s even neutrl or deromtized wine hs sensory properties due to certin romtic compounds tht remin in the wine. For the purpose of this study, the bsence of ny other romtic nd non romtic compounds (tht could influence the perception of the rom) ws criticl (Lund et l., 29). The concentrtions were chosen bsed on concentrtions of the rom compounds found in white wines s reported in literture, s well s preliminry sensory ssessment of the individul rom compounds nd mixtures of the compounds. Pre screenings by experienced wine tsters (ll who completed certificte course in wine evlution) were done before finlizing the concentrtions to ensure tht the levels chosen fulfil the requirements for the im of this study. The concentrtions (levels) for this study ws chosen s follows: the first level ws chosen to be below the perception threshold, s this would give n indiction of enhncing or suppressive effects of these compounds even when they re not present in concentrtions high enough to be 7

85 individully perceived. The second level ws t the perception threshold s reported in literture. A complete detection threshold test ws not done for ech compound s concentrtions reported were deemed sufficient (especilly when reported in model wine) (Buttery et l., 1969; Escudero et l., 2; Dubourdieu et l., 26; Culleré et l., 27). The third level ws t low to medium concentrtion while the fourth level ws t medium to high concentrtion found in wines ccording to literture. The fifth level ws considered to be very high ccording to levels found in white wines in generl (Silv Ferreir et l., 23; Silv Ferreir et l., 23b; Dubourdieu et l., 26; Culleré et l., 27; Alberts et l., 29; Swiegers et l., 29). This wy the effect of these compounds below, t nd bove their perception threshold, could be investigted. The ttributes generted by the pnel for ech compound re shown in Tble 3.3 nd Figure 3.2. In Tble 3.3 the definition of ech descriptor is shown s defined by the pnel. The reference stndrds used for ech ttribute re lso defined (Tble 3.3). The list of descriptors did not chnge between the two trils (individul or combintion of compounds), however not ll the ttributes were used by the judges. For exmple, the descriptor erthy ws not used by the pnel when evluting the smples ccording to the centrl composite design s this ttribute did not describe the odour of the smples when the compounds were present in combintion. A correltion network ws used in order to visulise the (Person) correltions between compounds nd the sensory ttributes ssocited with them. Figure 3.2 shows the correltion network constructed from dt obtined from nlysing the compounds individully. 71

86 Tble 3.3 Attributes nd reference stndrds used for descriptive nlysis of spiked model wines t vrious concentrtions Descriptor Definition Reference stndrd Erthy Smell ssocited with soil none Dusty Smell ssocited with closed bsement or cupbord none Grpefruit Typicl grpefuit Grpefuit pieces in model wine Guv green Unripe guv (green skin) Unripe green guv pieces in model wine Guv ripe Very ripe guv (pink skin) Ripe guv pieces in model wine Pssion fruit Typicl pssion fruit Pssion fruit pieces in model wine Tomto lef/stlk Fresh tomto plnt stlk Fresh cherry tomto on the vine with long stlk Cooked potto Thoroughly cooked potto Stemed potto pieces in model wine Cooked bens Cooked green bens Cnned cooked bens in model wine Green pepper Typicl green pepper Fresh green pepper Grss Freshly cut grss Freshly cut grss Florl Bouquet of fresh flowers none Honey Potpourri Typicl honey Dried flower leves (mostly rose petls) 2 tespoons honey dissolved in model wine Dried rose petls (vrious colours) 72

87 Figure 3.2 A correltion network showing the ttributes generted by dding individul compounds to model wine medium t the vrious levels. Vlues displyed on the edge re the correltion between the compound concentrtion nd intensity of the ttribute. Thickness of the edges indictes the strength of the correltion. Broken line indictes negtive correltion. Correltions between the compound concentrtion nd ttribute intensity cn be seen on the edges between the compound nd the ttribute. Dusty ws used s descriptor for IBMP, 3MH nd phenylcetldehyde. The correltion between dusty nd IBMP ws high (.839) unlike the correltion with the other two compounds which were wekly negtive. The other descriptor used for more thn one compound ws cooked bens nd ws used s n ttribute for both methionl nd IBMP. Both of these compounds contributed to this descriptor with strong correltions, with methionl (.891) hving lower correltion thn IBMP (.973). This could indicte possible enhncement effect when these two compounds re present in the sme solution. Guv ws divided into two clsses. Guv green ws described s the odour from the unripe fruit while guv 73

88 ripe refers to the fully ripened or even overripe fruit. Guv ripe hd very strong correltion (.936) with 3MH, while guv green showed wek negtive correltion with 3MH (.44). The use of the guv descriptor hs rrely been reported in literture (Swiegers et l., 26). However, Vn Wyngrd (213) showed guv to lso be one of the dominnt ttributes for 3MH. This could be due to the South Africn pnel being more fmilir with the odour of this fruit when compred to French pnel (Vlentin, 1997). With the exception of grss (.622), ll other ttributes (cooked potto, erthy, green pepper, tomto lef/stlk, pssion fruit, grpefruit, florl, potpourri nd honey) hd high positive correltions (>.875) with the corresponding compound. The use of most of these ttributes to describe the vrious compounds hve previously been reported in literture (Vn Strten et l., 1981; Augustyn et l., 1982; Toming et l., 1996; Toming, 1998; Toming et l., 1998; Escudero et l., 22; Silv Ferreir et l., 23; Dubourdieu et l., 26; Vn Wyngrd, 213). The use of tomto lef s n ttribute for Suvignon blnc is not uncommon nd the presence of 3MH hs been identified in tomto lef s well s rhubrb petiole (Toming, 1998). The effect of vrying 3MH levels on the ttribute intensities cn be seen in Figure 3.3. No specific descriptor ws used to describe the fint odour observed t 4 nd 6 ng/l, however dusty did occur t low intensities (<1 intensity units). Mteo Vivrcho et l. (21) lso found 3MH concentrtions close to the perception threshold (21 or 81 ng/l) not to contribute significntly to the flvour. 74

89 b Attribute intensity rting g h f c d g c e 1 i j jkl jk jklm jk jkl npm npl nk nko npnpnp np np npm po np MH (ng/l) Dusty Grpefruit Guv green Guv ripe Pssion fruit Tomto lef/stlk Figure 3.3 Effect of 3MH concentrtion on the vrious ttribute intensities. Different letters indicte significnt differences t p<.5. Dusty remined low nd decresed slightly during the increse of 3MH concentrtions. At 5 ng/l, significnt increse in grpefruit nd guv green intensities ws observed. Tomto lef/stlk lso incresed, but the intensity ws still reltively low. At 2 ng/l guv green decresed significntly to below 1 units nd remined low even t 6 ng/l of 3MH. This is good indiction of nonliner reltionship between the 3MH concentrtion nd the ttribute perception nd could pply to other descriptors nd compounds s well. This phenomenon should be kept in mind when generting ttributes for compound without testing vrious concentrtions s often liner reltionship between concentrtion nd n increse in sensory intensity is ssumed. Pssion fruit nd guv ripe incresed to n intensity of bove 4 units nd surpssed the other descriptor intensities t 2 ng/l. The two descriptors remined dominnt t 6 ng/l with guv ripe reching n intensity of bove 7. Grpefruit nd tomto lef/stlk lso significntly incresed from 5 to 6 ng/l. Mteo Vivrcho et l. (21) reported the contribution of 3MH to the fruitiness of wine when present in concentrtions bove 1497 ng/l, while King et l. (211) lso reported 75

90 increses in the cooked green vegetl ttribute with n increse in thiol concentrtions which shows the contribution of 3MH to green ttributes. Dusty reched higher intensities when IBMP ws dded to the medium compred to 3MH (Figure 3.4). The intensity incresed significntly from 2 to 1 ng/l fter which no further intensity increses were observed. Green pepper ws identified s the min ttribute used to describe the odour of IBMP when dded t 1 to 4 ng/l Attribute intensity rting fge de fg b d fe df de c g 1 hij hkhk kli h hij kli klj h IBMP (ng/l) Dusty Cooked bens Green pepper Grss Figure 3.4 Effect of IBMP concentrtion on the vrious ttribute intensities. Different letters indicte significnt differences t p<.5. The intensity of green pepper lso incresed with lrge intervls reching n intensity of bove 5 units t 4 ng/l. Cooked bens lso significntly incresed with intensity from 2 ng/l s IBMP concentrtions incresed. Grss did not lwys increse with incresed IBMP concentrtions nd the intensity ws significntly lower t 4 ng/l compred to 2 ng/l. This is contrdictory to results reported by Vn Wyngrd, (213) where further increse in the intensity of the grss ttribute ws reported from 15 (bout 2 intensity units) to 4 ng/l (bout 3 intensity units). The use of different pnel nd mrix could explin the differences observed. 76

91 The three descriptors used to describe the odour of methionl ll incresed linerly from.5 µg/l with incresed methionl concentrtions (Figure 3.5). Cooked potto ws used s the min ttribute t ll concentrtions nd reched the intensity level of bout 6 units t 15 µg/l methionl. After cooked potto; cooked bens nd erthy followed in descending order with intensities verging t 38 nd 24 units t 15 µg/l methionl respectively. Even though the use of these ttributes hve been used to describe the compound (Escudero et l., 2), the chnge in intensity s the concentrtion is ltered hs not been reported b Attribute intensity rting d f ef d e c 1 gh jhi ji jh gi j g Methionl (µg/l) Erthy Cooked potto Cooked bens Figure 3.5 Effect of methionl concentrtion on the vrious ttribute intensities. Different letters indicte significnt differences t p<.5. Florl ws minly used to describe the odour cused by phenylcetldehyde (Figure 3.6). The intensity significntly incresed from.5 to 13 µg/l phenylcetldehyde. Only t 13 µg/l did the potpourri ttribute surpss the intensity of florl. Potpourri did not increse initilly nd only occurred t 8 nd 13 µg/l where the intensities were rted reltively high (3.5 nd 55.9 intensity units respectively). Honey incresed with incresed phenylcetldehyde concentrtions, but reched plteu t 8 µg/l. The use of the honey ttribute hs been reported (Silv Ferreir et l., 23), 77

92 however the ttributes florl nd especilly potpourri in reltion to wine hve very little to no references in literture b Attribute intensity rting e c e d e 1 f fg hg ki hij ki kj kj kj k kj kj kj Phenylcetldehyde (µg/l) Dusty Florl Honey Potpourri Figure 3.6 Effect of phenylcetldehyde concentrtion on the vrious ttribute intensities. Different letters indicte significnt differences t p< Multiple compound evlution Smples contining multiple compounds will be referred to s four digit code in the text. The codes re n indiction of the level (concentrtion) of the specific compound s reported in Tble 3.1. For exmple, code would represent the smple contining level 1 of 3MH (4 ng/l), level 3 of IBMP (1 ng/l), level 5 of methionl (15 µg/l) nd level 2 of phenylcetldehyde (1. µg/l). PCA biplots were constructed to investigte the interctions tht occurred when multiple compounds were dded to the sme smple. In PCA biplot, the correltion of vribles with objects (nd between vribles themselves) depends on mny fctors. The plot my be interpreted using different pproches, including trends in the mgnitude of the vribles, ngles nd distnces between vribles s well s the distnce between the dt points. 78

93 Sixty percent of the vrition cn be explined by the first two PCs while nother 17% ws explined by PC3 (Figures 3.7 nd 3.8). In Figure 3.7 PC1 seprtes the smples ccording to ttributes: Florl, honey nd potpourri grouped on the left hnd side of the PC with green pepper to the middle nd guv green, tomto lef/stlk, grpefruit, guv ripe nd pssion fruit on the right. Green pepper 3MH IBMP Meth Phencet PC 2(25%) Guv green Tomto lef/stlk Grpefruit Guv ripe Pssion fruit Florl Honey Potpourri PC 1(35%) Figure 3.7 PCA biplot showing PC1 nd PC2. The selected re is enlrged to better evlute the clusters in the centre of the PCA. As expected, smple correlted strongly with florl, honey nd potpourri. Smples nd correlted strongly with green pepper. Smple strongly correlted with the thiol ssocited ttributes while showed good correltion with cooked potto nd cooked bens in PC3 (Figure 3.8). Of ll the smples tested, the smple combintions contining compound t level 5 correlted very strongly with the ssocited descriptors s indicted in results from the individul compound tests (Figures 3.3 to 3.6) with the exception of IBMP where the smple lso hd strong correltion with green pepper. None of the other 3 compounds, when present t level 4 ( ; ; ), elicited the sme strength in correltion with the ssocited 79

94 ttributes. This could be n indiction of IBMP being dominting compound when in combintion with the other rom compounds. This tendency ws lso found in other studies (Prr et l., 27; Hein et l., 29; King et l., 211; Vn Wyngrd, 213). Smples with combintions of levels 3 nd 1 were mostly clustered to the middle of the PC (Figures 3.7 nd 3.8). The low level of one compound (level 1) when in combintion with ll of the other compounds present t level 3, seems to not hve big influence on the perception of the rom compounds nd clustered together with smple which ws considered to be the centre smple. Interestingly, smples nd lso clustered in this group nd seem to correspond closely to smple probbly due to the fct tht no singulr ttribute defined those smples. Smples , , nd correlted better with the thiol relted ttributes compred to the other smples (Figures 3.7 nd 3.8). In these smples 3MH ws present t level 4 while methionl ws present t level 2 (perception threshold). This could be n indiction of methionl suppressing the perception of the thiol relted descriptors especilly when looking t smples , nd which were locted in the centrl cluster. These smples still contined 3MH t level 4 but with methionl lso t reltively high concentrtions. The only other smple with 3MH t level 4 ws This smple ws locted more towrds the centre of the biplot which could indicte suppression of 3MH by phenylcetldehyde s well, however this suppression ws not s severe s with methionl. A slight suppression of the thiol relted ttributes by IBMP cn be seen s the smples contining IBMP t level 4 ( nd ) did not correlte s strongly with the thiol ssocited chrcteristics when compred to With the exception of nd , ll other smples contining phenylcetldehyde t level 4 correlted better to florl, honey nd potpourri compred to the other smples. The presence of IBMP t level 4 could suppress phenylcetldehyde s ttribute intensities seeing tht smples , nd correlted best with the ttributes ssocited with phenylcetldehyde. Interestingly, smple which contined higher concentrtions of methionl showed similr correltion to the phenylcetldehyde ttributes thn smple This could indicte the nonsuppressive effect of methionl on phenylcetldehyde, lso considering smple which hd weker correltion with florl, honey nd potpourri when compred to smple In PC3 smple seemed to correlte well with cooked bens nd cooked potto (Figure 3.8). As mentioned before, the cooked bens ttribute ws used for describing both IBMP nd methionl. The presence of IBMP s well s methionl t reltively high concentrtions in the sme smple could enhnce ech other nd intensify the cooked bens descriptor. This is especilly interesting 8

95 when locting smple , which did not correlte s strongly to the cooked bens nd cooked potto ttributes s Dusty 3MH IBMP Meth Phencet PC 3(17%) Tomto lef/stlk Guv green Grpefruit Guv ripe Pssion fruit Cooked bens Cooked potto PC 1(35%) Figure 3.8 PCA biplot PC1 nd PC3. The selected re is enlrged to better evlute the clusters in the centre of the PCA. The PCA biplots effectively show the correltions between the ttributes nd therefter the reltionship with the vrious smples (dt points). The grouping of the ttributes in this nlysis mkes it difficult to drw conclusions regrding individul descriptors nd their reltionship with the smples. Sttisticl networks were constructed to investigte ech ttribute seprtely. Node colours were scled ccording to the differences in the verge intensity units between the smple in question nd the centre smple. The networks provide visul indictions of how the vrious smples influence the intensity of the ttribute. Only the ttributes tht re ffected by the smples nd the smples tht hve sttisticlly significnt difference from the centre smple re shown in the network. 81

96 As indicted by the biplots, smples , nd cn ech be considered centre points (Figure 3.7 nd 3.8). Sttisticl networks were constructed with these three smples s centre smples to investigte the effect of the chnge in concentrtions on the ttribute intensities. Figure 3.9 shows the network contining smples with compound combintions of levels 1, 3 nd 5 with smple s centre smple. As expected, the red nodes (increse in ttribute intensity) only occurred when the corresponding compound ws present t level 5. These increses in intensities were lrge with guv ripe, potpourri nd grpefruit incresing with 58.6, 42.6 nd 4.5 units respectively. Other increses in green pepper (31.5 units), cooked potto (3.8 units), florl (36.1 units) nd pssion fruit (35.9 units) were lso high. Cooked bens seems to be more influenced by the presence of methionl (25.2 units for ) when compred to IBMP (6.9 units for ) for the specific concentrtions tested. 3MH IBMP Meth Phencet Figure 3.9 A sttisticl network (centre smple ) showing the effect of smple combintions on ttribute intensity. Edges between ttributes nd coloured nodes indicte sttisticlly significnt difference from the centre smple. Red nd underlined smple nodes indicte ttribute intensity increse; blue nodes indicte ttribute intensity decrese. Drker colour shows lrger increse/decrese. 82

97 Smple consisted of ll four compounds present t their perception thresholds. Figure 3.1 shows the network constructed with this smple s the centre smple. The network is dominted by red nodes which would indicte generl increse in ttribute intensities s the smple composition chnged. Decreses in dusty were evident with these chnges. With the testing of the single compounds, negtive correltion between dusty nd both 3MH nd phenylcetldehyde ws observed, while positive correltion ws observed between dusty nd IBMP (Figure 3.2). This effect is repeted in combintion with the other compounds s there ws n increse in intensity (8.1 units) for smple MH IBMP Meth Phencet Figure 3.1 A sttisticl network (centre smple ) showing the effect of smple combintions on ttribute intensity. Edges between ttributes nd coloured nodes indicte sttisticlly significnt difference from the centre smple. Red nd underlined smple nodes indicte ttribute intensity increse; blue nodes indicte ttribute intensity decrese. Drker colour shows lrger increse/decrese. 83

98 In ll the smples contining 3MH t level 4 there ws n increse in grpefruit, however, in the presence of methionl nd 3MH t level 4 ( ), the increse ws lower (7.7 units) compred to other smples ( units). Grpefruit intensity ctully decresed ( units) when methionl ws present t level 4 with 3MH t level 2 when compred to smple This could show the suppressive effect of methionl on 3MH nd the grpefruit ttribute specificlly. Tomto lef/stlk ws used to describe the odour of 3MH when testing the individul compounds s previously mentioned, however the ddition of IBMP t level 4 in combintion with 3MH ( ) or without 3MH t level 4 ( ) lso resulted in n increse in the tomto lef/stlk ttribute. In this cse the tomto lef/stlk intensity does not seem to be enhnced by these increses s the intensity ws bout the sme for the three smples (5.8, 5.8 nd 6.2 units for , nd respectively). Guv green lso incresed with incresed 3MH concentrtions (6. units), however, this increse ws the sme or slightly higher when IBMP ws lso present t level 4 (7.2 units). The perception of the guv green nd tomto lef/stlk cn possibly be enhnced by the green odours from IBMP. Guv ripe incresed with incresed 3MH concentrtions. The increse ws less in the presence of phenylcetldehyde (1. units) nd IBMP (7.5 units) compred to 3MH only (12.5 units). A slight suppression might thus occur. Interestingly, green pepper not only incresed with incresed IBMP concentrtions, but it lso incresed when 3MH ws t level 4. The intensity of green pepper incresed with 18.4 units in smple compred to 13.3 units in smple This could show the enhncing effect of 3MH on certin chrcteristics ssocited with IBMP. Furthermore, n increse in 3MH ( ) lso incresed the green pepper intensity with 11.9 units. The contribution of thiols t certin concentrtions to the cooked green vegetl ttribute hs been reported in erlier studies, however there ws no contribution to the fresh green ttribute by these compounds (King et l., 211). Cooked bens incresed whenever one or both IBMP nd methionl were present t level 4. The smllest increse ws observed when 3MH were lso t level 4 (4.6 units for ). The lrgest increse ws when IBMP nd methionl both were t level 4 in the sme smple (27.2 units for smple ). The presence of phenylcetldehyde did not seem to hve n influence s the increses for nd were 19.4 nd 18.7 units respectively. As mentioned before, the cooked bens ttribute ws used for both IBMP nd methionl when evluted singulrly nd the increse in this ttribute could be n indiction of n enhncing effect between these two compounds. Cooked potto ws significntly incresed by incresed concentrtions of methionl. With methionl present t level 4 the increses were 28.6, 28.7 nd 29.8 units for , nd respectively. The presence of phenylcetldehyde nd 3MH t level 4 thus did not seem 84

99 to suppress the cooked potto ttribute. To the contrry, when IBMP nd methionl were present t level 4 in the sme smple ( ), the increse in cooked potto reched 41. units. It could be concluded tht IBMP could enhnce the cooked potto descriptor when in the presence of methionl. The florl nd potpourri descriptors incresed with phenylcetldehyde concentrtions. Florl did not seem to be drsticlly influenced by the presence of other compounds s the increses rnged from units for smples , , nd Potpourri showed the highest increse when phenylcetldehyde ws the only compound present t level 4 (16.2 units for ). In the other smples ( , nd ), the increse ws between 8.1 nd 1.1 units. The perception of this ttribute might be suppressed by the presence of other compounds. The sttisticl network constructed with s the centre smple (Figure 3.11) cn be seen s complimentry to the network with s centre smple, however, the network still provides dditionl informtion not displyed in the other networks. Grpefruit incresed when methionl concentrtions decresed to level 2 (compred to smple ). This phenomenon does not seem to be influenced by decreses of the other rom compounds s the increse in intensity ws between 6.8 nd 9.1 units for smples , nd The decrese in grpefruit lso seemed to remin the sme (between 14.1 nd 15.9 units) when 3MH ws lowered to level 2 despite the concentrtion chnges of the other compounds. There ws slight increse in the tomto lef/stlk intensity (4.9 units) when methionl nd phenylcetldehyde levels were lowered to level 2. Guv ripe decresed with decresed 3MH concentrtions with n increse (7.2 units) when IBMP nd methionl levels decresed to level 2 ( ). Guv green intensities decresed ( units) with lowered 3MH levels nd incresed ( units) when methionl ws present t level 2. The effect of 3MH on the perception of green pepper ws observed once gin. With decresed levels of 3MH or both 3MH nd IBMP, the green pepper intensity lso decresed by 8.4 units ( ) nd 1.3 units ( ) respectively. The decrese in methionl nd phenylcetldehyde concentrtions led to n increse in green pepper intensity (9.2 units for ) displying the possible suppression effect of these two compounds on the perception of green pepper. Cooked bens nd cooked potto intensity incresed with 12.1 nd 15.2 units respectively with IBMP nd methionl t level 4 nd 3MH nd phenylcetldehyde t level 2 ( ). The dditive effect of IBMP nd methionl (in the presence of low concentrtions of the other compounds) might ply role in this phenomenon. Cooked bens nd cooked potto intensities decresed when either IBMP or methionl concentrtions decresed. Potpourri might be suppressed by the presence of IBMP s 85

100 smples contining IBMP t level 2 (with phenylcetldehyde remining t level 4) hd n increse in potpourri intensity ( units). 3MH IBMP Meth Phencet Figure 3.11 A sttisticl network (centre smple ) showing the effect of smple combintions on ttribute intensity. Edges between ttributes nd coloured nodes indicte sttisticlly significnt difference from the centre smple. Red nd underlined smple nodes indicte ttribute intensity increse; blue nodes indicte ttribute intensity decrese. Drker colour shows lrger increse/decrese. To show the effect of some of these interctions, response surfce grphs were constructed. Response surfce grphs re three dimensionl representtions of the dt. Three vribles re shown in the grph, two rom compounds together with the descriptor. The grph thus shows how the intensity of the descriptor chnges with the chnging concentrtion of the two rom compounds. The response surfce plot for grpefruit (Figure 3.12) clerly showed the suppressive effect of methionl. At low methionl concentrtion, the increse in 3MH concentrtion will led to n increse in grpefruit intensity. However, s the concentrtion of methionl increses, the grpefruit intensity will decrese even t high 3MH concentrtions. In Figure 3.13 the suppressive 86

101 effect of methionl on 3MH cn gin be observed s the incresing concentrtion of methionl suppressed the guv ripe ttribute. Overll, methionl hs dominnt effect over 3MH. This could hve importnt implictions during wine ging. Not only re the thiols oxidtion sensitive (Blnchrd et l., 24; Nikolntonki et l., 21), but the remining thiols will be suppressed s soon s methionl is formed in the wine. 3MH present t reltively low concentrtions enhnced the green pepper ttribute in the presence of IBMP (Figure 3.14). This ws seen throughout the IBMP rnge. However, t higher 3MH levels, the ttribute intensity decresed. The suppression of IBMP odours by 3MH hs lso been observed in other studies (King et l., 211; Vn Wyngrd, 213). The present study highlighted the importnce of testing intensities t vrious concentrtions s the some interctions between compounds chnged s the concentrtions chnged. The ttribute cooked bens ws lso influenced by vrious compounds. 3MH suppressed cooked bens when present t higher concentrtions (Figures 3.15 nd 3.16). On the other hnd, when IBMP nd methionl were present in the sme medium, there ws n enhncing effect tht cn be seen in Figure Grpefruit Methionl (µg/l) MH (ng/l) 7 > 6 < 56 < 36 < 16 < -4 Figure 3.12 Response surfce plot of the grpefruit ttribute s influenced by 3MH nd methionl. 87

102 Guv ripe Methionl (µg/l) MH (ng/l) 7 > 8 < 76 < 56 < 36 < 16 < -4 Figure 3.13 Response surfce plot of the guv ripe ttribute s influenced by methionl nd 3MH. Green pepper MH (ng/l) IBMP (ng/l) 4 45 > 5 < 44 < 34 < 24 < 14 < 4 < -6 Figure 3.14 Response surfce plot of the green pepper ttribute s influenced by 3MH nd IBMP. 88

103 5 4 3 Cooked bens MH (ng/l) IBMP (ng/l) 4 45 > 2 < 16 < 6 < -4 Figure 3.15 Response surfce plot of the cooked bens ttribute s influenced by 3MH nd IBMP Cooked bens MH (ng/l) Methionl (µg/l) 16 > 4 < 34 < 24 < 14 < 4 < -6 Figure 3.16 Response surfce plot of the cooked bens ttribute s influenced by 3MH nd methionl. 89

104 5 4 Cooked bens IBMP (ng/l) Methionl (µg/l) 16 > 4 < 32 < 22 < 12 < 2 Figure 3.17 Response surfce plot of the cooked bens ttribute s influenced by IBMP nd methionl. The contribution of 3MH to green chrcteristics hs been reported before (King et l., 211). Some studies found no correltion between IBMP nd IPMP levels in wine nd the vegetl ttribute nd it ws suggested tht compounds other thn the methoxypyrzines cn be responsible for the perception of this rom (Ferreir et l., 22; Preston et l., 28). The present study confirms the contribution of 3MH to these chrcteristics. As mentioned before, the enhncing effect of IBMP nd methionl in the perception of cooked bens nd cooked potto could hve importnt implictions during wine ging. As IBMP is not oxidtion sensitive (Mris, 1998), the odours tht occur due to this compound remin in the wine for long period of time. With the formtion of methionl the enhncing effect could occur, cusing odours such s cooked bens nd cooked potto to be perceived t stronger intensity level cusing the wine to smell over ged even t lower methionl concentrtions. This should be kept in mind when wines with higher methoxypyrzine content re stored for lter consumption. Certin compounds could suppress individul ttributes in unique wy. By using sttisticl networks, these individul influences cn be observed while this is not possible when nlysing biplots s ll the ttributes tht rise from one compound remin grouped. For instnce, IBMP cn enhnce 3MH relted compounds such s guv green nd tomto lef stlk, while suppressing 9

105 guv ripe. The concentrtions of the compounds involved should lso be kept in mind when nlysing the dt s the interctions observed re dependent on the concentrtions. The nlyticl mesurement of rom compounds to predict the sensory composition of wine should be used s n pproximte tool s this study clerly showed the complex interctions between vrious compounds influencing the perception of the overll rom. The exct mechnisms of the interctions remin unknown. The interctive effects could occur orthonslly or chemiclly in the medium. Further investigtions re needed to elucidte this phenomenon. 3.4 CONCLUSION The im of this study ws to improve the understnding of interctive effects between rom compounds. In the present study interctive effects between typicl Suvignon blnc rom compounds nd compounds tht occur due to oxidtion were investigted. The perception of flvours in wine is complicted by the vst number nd combintions of rom compounds present. The mtrix used to test the interctions between compounds should not be underestimted s vrious compounds (voltile or non voltile) present, even t below threshold concentrtions, could potentilly influence the perception of the compound tested. The chllenge thus remins to investigte the sensory interction between these compounds in rel wine. This study highlighted the importnce of protecting young, fruity Suvignon blnc wine from forming typicl ging or oxidtion relted compounds s these compounds (especilly methionl) significntly suppressed the fruity flvours. Further studies should include n investigtion of other oxidtion nd ging compounds tht could ffect the rom composition of the wine, whether these interctions lso occur in rel wine nd whether consumer studies yield similr results. 3.5 ABBREVIATIONS USED 3 mercptohexn 1 ol (3MH); 3 mercptohexn 1 ol cette (3MHA); 4 mercpto 4 methylpentn 2 one (4MMP); 3 isobutyl 2 methoxypyrzine (IBMP); 3 isopropyl 2 methoxypyrzine (IPMP); 3 secbutyl 2 methoxypyrzine (SBMP); PCA, principle component nlysis. 91

106 3.6 REFERENCES Alberts, P., Stnder, M.A., Pul, S.O. & De Villiers, A., 29. Survey of 3 lkyl 2 methoxypyrzine content of South Africn Suvignon blnc wines using novel LC APCI MS/MS method. Journl of Agriculturl nd Food Chemistry. 57, Aronson, J. & Ebeler, S.E., 24. Effect of polyphenol compounds on the hedspce voltility of flvors. Americn Journl of Enology nd Viticulture. 55, Augustyn, O.P.H., Rpp, A. & Vn Wyk, C.J., Some voltile rom components of Vitis Vinifer L. cv. Suvignon blnc. South Africn Journl of Enology nd Viticulture. 3, Benkwitz, F., Nicolu, L., Lund, C., Beresford, M., Wohlers, M. & Kilmrtin, P.A., 212. Evlution of key odornts in Suvignon blnc wines using three different methodologies. Journl of Agriculturl nd Food Chemistry. 6, Blnchrd, L., Drriet, P. & Dubourdieu, D., 24. Rectivity of 3 mercptohexnol in red wine: Impct of oxygen, phenolic frctions, nd sulfur dioxide. Americn Journl of Enology nd Viticulture. 55, Buttery, R.G., Seifert, R.M., Gudgni, D.G. & Ling, L.C., Chrcteriztion of some voltile constituents of bell peppers. Journl of Agriculturl nd Food Chemistry. 17, Cmpo, E., Ferreir, V., Escudero, A. & Ccho, J., 25. Prediction of wine sensory properties relted to grpe vriety from dynmic hedspce gs chromtogrphy olfctory dt. Journl of Agriculturl nd Food Chemistry. 53, Coetzee, C. & Du Toit, W.J., 212. A comprehensive review on Suvignon blnc rom with focus on certin positive voltile thiols. Food Reserch Interntionl. 45, Coetzee, C., Lisjk, K., Nicolu, L., Kilmrtin, P.A. & Du Toit, W.J., 213. Oxygen nd sulfur dioxide dditions to Suvignon blnc must: effect on must nd wine composition. Flvour nd Frgrnce Journl. 28, Culleré, L., Ccho, J. & Ferreir, V., 27. An ssessment of the role plyed by some oxidtion relted ldehydes in wine rom. Journl of Agriculturl nd Food Chemistry. 55, Dll'Ast, C., Cirlini, M., Morini, E. & Glvern, G., 211. Brnd dependnt voltile fingerprinting of Itlin wines from Vlpolicell. Journl of Chromtogrphy A. 1218, Dubourdieu, D., Toming, T., Msneuf, I., Peyrot des Gchons, C. & Murt, M.L., 26. The role of yest in grpe flvour development during fermenttion: The exmple of Suvignon blnc. Americn Journl of Enology nd Viticulture. 57, Ellmn, G.L., Tissue sulfhydryl groups. Archives of Biochemistry nd Biophysics. 82, Esbensen, K.H., 22. Multivrite Dt Anlysis. CAMO Process AS, Osl, Norwy. 92

107 Escudero, A., Asencio, E., Ccho, J. & Ferreir, V., 22. Sensory nd chemicl chnges of young white wines stored under oxygen. An ssessment of the role plyed by ldehydes nd some other importnt odornts. Food Chemistry. 77, Escudero, A., Cmpo, E., Frin, L., Ccho, J. & Ferreir, V., 27. Anlyticl chrcteriztion of the rom of five premium red wines. Insights into the role of odor fmilies nd the concept of fruitiness of wines. Journl of Agriculturl nd Food Chemistry. 55, Escudero, A., Hernndez Orte, P., Ccho, J. & Ferreir, V., 2. Clues bout the role of methionl s chrcter impct odornt of some oxidized wines. Journl of Agriculturl nd Food Chemistry. 48, Ferreir, V., Ortin, N., Escudero, A., Lopez, R. & Ccho, J., 22. Chemicl chrcteriztion of the rom of Grenche rosé wines: Arom extrct dilution nlysis, quntittive determintion, nd sensory reconstitution studies. Journl of Agriculturl nd Food Chemistry. 5, Ferreir, V., Pet'k, A. & Ccho, J., 26. Intensity nd persistence profiles of flvor compounds in synthetic solutions. Simple model for explining the intensity nd persistence of their ftersmell. Journl of Agriculturl nd Food Chemistry. 54, Fischer, U., 27. Flvours nd Frgrnces: Chemistry, Bioprocessing nd Sustinbility. Springer Berlin Heidelberg, Berlin Heidelberg. Frncis, I.L. & Newton, J.L., 25. Determining wine rom from compositionl dt. Austrlin Journl of Grpe nd Wine Reserch. 11, Frivik, S.K. & Ebeler, S.E., 23. Influence of sulfur dioxide on the formtion of ldehydes in white wine. Americn Journl of Enology nd Viticulture. 54, Guth, H., Quntittion nd sensory studies of chrcter impct odornts of different white wine vrieties. Journl of Agriculturl nd Food Chemistry. 45, Hein, K., Ebeler, S.E. & Heymnn, H., 29. Perception of fruity nd vegetl roms in red wine. Journl of Sensory Studies. 24, Herbst, M., Kilmrtin, P.A. & Nicolu, L., 28. Arom stbility in Suvignon blnc wines. The Austrlin nd New Zelnd Grpegrower nd Winemker. 6, Jckowetz, J.N. & De Orduñ, M., 213. Survey of SO 2 binding crbonyls in 237 red nd white tble wines. Food Control. 32, King, E.S., Osidcz, P., Curtin, C., Bstin, S.E.P. & Frncis, I.L., 211. Assessing desirble levels of sensory properties in Suvignon Blnc wines consumer preferences nd contribution of key rom compounds. Austrlin Journl of Grpe nd Wine Reserch. 17,

108 Lcey, M.J., Allen, M.S., Hrris, R.L.N. & Brown, W.V., Methoxypyrzines in Suvignon blnc grpes nd wines. Americn Journl of Enology nd Viticulture. 42, Lee, S.J., 23. Finding key odornts in foods: Gs chromtogrphy olfctometry (GC/O). Food Science nd Biotechnology. 12, Lund, C.M., Nicolu, L., Grdner, R.C. & Kilmrtin, P.A., 29. Effect of polyphenols on the perception of key rom compounds from Suvignon blnc wines. Austrlin Journl of Grpe nd Wine Reserch. 15, Lund, C.M., Thompson, M.K., Benkwitz, F., Wohler, M.W., Triggs, C.M., Grdner, R., Heymnn, H. & Nicolu, L., 29b. New Zelnd Suvignon blnc distinct flvor chrcteristics: Sensory, chemicl, nd consumer spects. Americn Journl of Enology nd Viticulture. 6, Mg, J.A., Sensory nd stbility properties of dded methoxypyrzines to model nd uthentic wines. Chrlmbous, G. (ed). Proc. Flvors nd Off flvors, 6th Interntionl Flvor Conference, 5 7 July 1989, Rethymnon, Grete, Greece. pp Mris, J., Effect of grpe temperture, oxidtion nd skin contct on Suvignon blnc juice nd wine composition nd wine qulity. South Africn Journl of Enology nd Viticulture. 19, Mris, J. & Swrt, E., Sensory impct of 2 methoxy 3 isobutylpyrzine nd 4 mercpto 4 methylpentn 2 one dded to neutrl Suvignon blnc wine. South Africn Journl of Enology nd Viticulture. 2, Mteo Vivrcho, L., Zpt, J., Ccho, J. & Ferreir, V., 21. Anlysis, occurrence, nd potentil sensory significnce of five polyfunctionl mercptns in white wines. Journl of Agriculturl nd Food Chemistry. 58, Nikolntonki, M., Chichuc, I., Teissedre., P.L. & Drriet, P., 21. Rectivity of voltile thiols with polyphenols in wine model medium: Impct of oxygen, iron, nd sulfur dioxide. Anlytic Chimic Act. 66, Prr, W.V., Green, J.A., White, K.G. & Sherlock, R.R., 27. The distinctive flvour of New Zelnd Suvignon blnc: Sensory chrcteristion by wine professionls. Food Qulity nd Preference. 18, Pineu, B., Brbe, J. C., Vn Leewin, C. & Duboudieu, D., 27. Which impct for β dmscenone on red wine rom? Journl of Agriculturl nd Food Chemistry. 55, Preston, L.D., Block, D.E., Heymnn, H., Soles, G., Noble, A.C. & Ebeler, S.E., 28. Defining vegetl roms in Cbernet Suvignon using sensory nd chemicl evlutions. Americn Journl of Enology nd Viticulture. 59,

109 Ribéreu Gyon, P., Boidron, J.N. & Terrier, A., Arom of Musct grpe vrieties. Journl of Agriculturl nd Food Chemistry. 23, Shnnon, P.M., A., Ozier, O., Blig, N.S., Wng, J.T., Rmge, D., Amin, N., Schwikowski, B. & Ideker, T., 23. Cytoscpe: softwre environment for integrted models of biomoleculr interction networks. Genome Reserch. 13, Silv Ferreir, A.C., Guedes de Pinho, P., Rodrigues, P. & Hogg, T., 22. Kinetics of oxidtive degrdtion of white wines nd how they re ffected by selected technologicl prmeters. Journl of Agriculturl nd Food Chemistry. 5, Silv Ferreir, A.C., Hogg, T. & Guedes de Pinho, P., 23. Identifiction of key odornts relted to the typicl rom of oxidtion spoiled white wines. Journl of Agriculturl nd Food Chemistry. 51, Silv Ferreir, A.C., Oliveir, C., Hogg, T. & Guedes de Pinho, P., 23b. Reltionship between potentiometric mesurements, sensoril nlysis, nd some substnces responsible for rom degrdtion of white wines. Journl of Agriculturl nd Food Chemistry. 51, Subileu, M., Schneider, R., Slmon, J.M. & Degryse, E., 28. New insights on 3 mercptohexnol (3MH) biogenesis in Suvignon blnc wines: Cys 3MH nd (E) hexen 2 l re not the mjor precursor. Journl of Agriculturl nd Food Chemistry. 56, Swiegers, J.H., Frncis, I.L., Herderich, M.J. & Pretorius, I.S., 26. Meeting consumer expecttions through mngement in vineyrd nd winery: the choice of yest for fermenttion offers gret potentil to djust the rom of Suvignon blnc wine. Austrlin nd New Zelnd Wine Industry Journl. 21, Swiegers, J.H., Kievit, R.L., Siebert, T., Lttey, K.A., Brmley, B.R., Frncis, I.L., King, E.S. & Pretorius, I.S., 29. The influence of yest on rom of Suvignon blnc wine. Food Microbiology. 26, Tomic, O., Lucino, G., Nilsen, A., Hyldig, G., Lorensen, K. & Næs, T., 21. Anlysing sensory pnel performnce in proficiency tests using the PnelCheck softwre. Europen Food Reserch nd Technology. 23, Toming, T., Recherches sur l rôme vriétl des vins de Vitis vinifer L. cv. Suvignon blnc et s genèse à prtir de précurseurs inodores du risin. Thesis, Université Victor Seglen, Bordeux. Toming, T., Drriet, P. & Dubourdieu, D., Identifiction of 3 mercptohexyl cette in Suvignon wine, powerful romtic compound exhibiting box tree odor. Vitis. 35,

110 Toming, T., Furrer, A., Henry, R. & Dubourdieu, D., Identifiction of new voltile thiols in the rom of Vitis vinifer L. vr. Suvignon blnc wines. Flvour nd Frgrnce Journl. 13, Toming, T., Msneuf, I. & Dubourdieu, D., 24. Powerful romtic voltile thiols in wines mde from severl Vitis vinifer L. Cv. Suvignon blnc. Proc. ACS Symposium Series, pp Vlentin, D., Personl communiction. Vn der Merwe, C.A. & Vn Wyk, C.J., The contribution of some fermenttion products to the odor of dry white wines. Americn Journl of Enology nd Viticulture. 32, Vn Strten, S., De Beuveser, J.C. & Visscher, C.A., Voltile compounds in food, supplement 6. Centrl Institute for Nutrition nd Food Reserch, Zeist, The Netherlnds. Vn Wyngrd, E., 213. Voltiles plying n importnt role in South Africn Suvignon blnc wines. Thesis, University of Stellenbosch, Stellenbosch. 96

111 Chpter 4 Reserch results A chemicl nd sensory study on the evolution of romtic nd non romtic compounds during the progressive oxidtive storge of Suvignon blnc wine 97

112 Chpter 4: Reserch results A chemicl nd sensory study on the evolution of romtic nd nonromtic compounds during the progressive oxidtive storge of Suvignon blnc wine 4.1 INTRODUCTION MATERIALS AND METHODS Oxidtion of the wine Smpling Voltile thiol nlysis Methoxypyrzine nlysis Ester, cid nd lcohol nlysis Monoterpene nlysis Sulphur dioxide nd scorbic cid nlysis Glutthione nlysis Phenolic nlysis Crbonyl compounds nd cetl nlysis Spectrophotometric mesurement Sensory romtic descriptive nlysis Sensory colour nlysis Dt nlysis RESULTS AND DISCUSSION Oxygen concentrtion Chemicl nlyses Voltile Thiols Methoxypyrzines Esters, cids nd lcohols Monoterpenes Sulphur dioxide nd Glutthione Polyphenols Aldehydes Sotolon

113 Acetls Colour Combined chemicl content Sensory evlution Aromtic descriptive nlysis Sensory nlysis ccording to wine colour Chemicl nd sensory dt combined CONCLUSION ABBREVIATIONS USED REFERENCES 17 99

114 4.1 INTRODUCTION The shelf life of especilly white wine is of gret concern for the wine industry. Oxidtion my ply n importnt role in white wine s composition nd its sensory chrcteristics. The shelf life of young white wines nd the rte of oxidtion in generl re influenced by fctors such s the composition of the wine (ntioxidnt content, ph etc.), storge time, storge temperture, type of closure (oxygen ingress nd dsorption), colour of the bottle nd the storge position (Silv Ferreir et l., 22b; Skouroumounis et l., 25; Lopes et l., 26; Ppdopoulou & Roussis, 28; Mury et l., 21; Dimkou et l., 213; He et l., 213). During oxidtion, decrese in certin rom compounds occurs followed by chnge in wine colour (Singleton & Krmling, 1976; Singleton et l., 1979; Escudero et l., 2b). The chnge in rom will typiclly result in the wine initilly losing some of the vrietl chrcter (Blnchrd et l., 24; Nikolntonki et l., 21) fter which n increse in unwnted oxidtion rom compounds such s ldehydes nd sotolon will occur (Wilderndt & Singleton, 1974; Escudero et l., 2; Silv Ferreir et l., 23c; Loscos et l., 21; Escudero et l., 22; Silv Ferreir et l., 23c; Lvigne et l., 28). Oxygen modultes the oxidtion rections, however other chemicl rections not involving oxygen could lso occur during wine ging, thus certin degree of chnge in wine composition will occur even in n environment devoid of oxygen (Uglino et l., 29). Suvignon blnc wines cn hve lrge diversity of flvours such s fruity ( grpefruit, pssion fruit, gooseberry, citrus, tropicl ) nd green chrcters ( green pepper, grssy, sprgus ) (Mris, 1994; Toming et l., 1998; Lund et l., 29; Coetzee & Du Toit, 212). These descriptors hve been ttributed to key chemicl rom nd flvour compounds occurring in the wine. The voltile thiols (4 mercpto 4 methylpentn 2 one (4MMP), 3 mercptohexn 1 ol (3MH) nd 3 mercptohexn 1 ol cette (3MHA)) cn contribute to tropicl/fruity style wines, while the methoxypyrzines (3 isobutyl 2 methoxypyrzine (IBMP), 3 isopropyl 2 methoxypyrzine (IPMP), 3 sec butyl 2 methoxypyrzine (SBMP)) cn give the wine green pepper, grssy nd sprgus odour (Lcey et l., 1991; Toming et l., 1998; Swiegers et l., 26; Coetzee & Du Toit, 212). Both of these odour groups re often sought fter for the production of qulity Suvignon blnc wines (Lund et l., 29; King et l., 211). The voltile thiols re known to be oxidtion sensitive nd cn lso decrese due to cid hydrolysis during ging (Blnchrd et l., 24; Toming et l., 24; Herbst et l., 28). The methoxypyrzines re not sensitive to oxidtion nd studies hve shown methoxypyrzine concentrtion to remin the sme during oxidtive hndling of juice nd wine (Mris, 1998; Coetzee et l., 213). These two groups of compounds re considered 1

115 to be chrcter impct compounds for Suvignon blnc wines, however they re not solely responsible for the overll rom. Vrious other chemicl groups such s esters, higher lcohols, ftty cids nd terpenes cn lso contribute significntly to the romtic composition (Swiegers et l., 26). The origin nd rection of the vrious rom groups hve been discussed in Chpter 2 nd comprehensive review on Suvignon blnc rom (with focus on voltile thiols) nd the rection of vrious rom groups to oxidtion hs been recently published (Coetzee & Du Toit, 212). During nd fter the disppernce of the positive rom, the formtion of unwnted rom compounds such s cetldehyde, methionl, phenylcetldehyde nd sotolon cn occur during wine oxidtion nd cn be detrimentl the wine rom, cusing flvours described s green pple, potto, honey nd curry to dominte the wine rom bouquet (Frivik & Ebeler, 23; Silv Ferreir et l., 23c). The presence of these oxidtion compounds could lso interct with remining plesnt rom compounds leding to the suppression of certin positive ttributes s ws seen in the cse of methionl in Chpter 3. The formtion of these oxidtion relted roms should be voided to preserve the fresh nd fruity chrcter of the wine. In some cses, certin mount of oxygen cn significntly improve the wine rom nd qulity by preventing the formtion of reductive off odours in the bottle. Arom ttributes ssocited with these off odours re rotten egg, cbbge nd grlic nd the compounds, H 2 S nd methyl mercptn hve been indicted s being primrily responsible for post bottling reduction (Lopes et l., 29; O'Brien et l., 29b; Uglino et l., 211). Studies investigting vrious types of bottle closures (with different oxygen trnsmission rtes) concluded tht closures tht re less oxygen permeble llow better preservtion of fresh nd fruity roms, especilly the preservtion of voltile thiols (Lopes et l., 29). However, these types of closures could promote the production of reductive off odours (Godden et l., 21; Skouroumounis et l., 25; Kwitkowski et l., 27; Lopes et l., 29; O'Brien et l., 29b). The effect of oxygen nd ging on vrious spects of dry white wines hve been investigted (Silv Ferreir et l., 23c; Brjkovich et l., 25; Herbst et l., 28; Lopes et l., 29; Dimkou et l., 211; Herbst Johnstone et l., 211; Uglino et l., 211; Mkhotkin et l., 212; Frcsetti et l., 213) nd the evolution of wine sensory qulity is thought to rech pek fter period in the bottle, however the time period necessry nd the mount of oxygen required to rech this pek remin unknown. This identifiction is complex, s it depends on how the wine ws intended to evolve, when it is going to be consumed nd how sensitive the wine is towrds oxygen. Understnding wine s oxygen needs enbles the winemker to control wine evolution in the bottle by choosing the optiml closure in ccordnce with the desired wine style nd shelf life. 11

116 In commercil cellr, oxygen cn come into contct with wine during different winemking processes, such s trnsfer of wine, filtrtion nd bottling (Cstellri et l., 24; Du Toit, 26). In the pst number of studies investigted the effect of dding lrge mounts of oxygen to white wines nd/or storge t elevted tempertures (Ferreir et l., 1997; Escudero et l., 22; Silv Ferreir et l., 23c; Loscos et l., 21; Cejudo Bstnte et l., 213). Although this type of pproch enbles more rpid rte of oxidtion nd yielded vluble dt, not much is known bout repetitive oxidtion of Suvignon blnc wine over prolonged period of time t oxygen levels nd tempertures which would normlly occur in commercil wine cellr. Furthermore, the collection of lrge vriety of dt from romtic to non romtic chemicl compounds s well s sensory ssessment including both orthonsl nd visul investigtions, re first for white wine ging reserch. This is the first study tht we know of tht monitored Suvignon blnc wine evolution in controlled environment fter which both chemicl nd sensory nlysis were performed. Successive mild oxygen dditions were pplied nd oxygen concentrtions were specificlly mesured while the wines were kept t mild temperture over long period of time. Frequent smpling lso llowed the investigtion of the evolution of the romtic nd non romtic compounds during the time of the study nd the inclusion of sensory dt t ech of these stges provided comprehensive ssessment of wine oxidtive ging in terms of chemicl composition nd sensory effects. 4.2 MATERIALS AND METHODS Oxidtion of the wine A Suvignon blnc wine ws oxidised t vrious stges over time to compre chemicl content nd sensory impct. Wine ws collected from commercil cellr (De Grendel Wine Estte, Tygerberg, South Afric) directly fter the 211 hrvest. The wine ws trnsported to the experimentl cellr of the University of Stellenbosch fter which sulphur dioxide (SO 2 ) nd scorbic cid content were mesured (section 4.2.7). The wine ws then divided using nitrogen gs (Afrox South Afric) into 33 x 5 litre glss bottles. Figure 4.1 shows the lyout of the study done in triplicte. After distribution of the wine into the bottles, the first tretment (Control ) ws smpled nd frozen t 2 C for nlyses. This tretment would serve s the beginning of the tril (no time lpse). All of the Ox tretments then received oxygen by using micro oxygen sprger (Figure 4.2) connected to 12

117 cylinder contining high purity (99.5%) medicl oxygen (Afrox South Afric). No oxygen ws dded to the Control smples. All tretments were stored in the drk t 15 C. The next oxygention took plce when the dissolved oxygen concentrtions reched level below 1 mg/l. Overll the study proceeded for 7 months. Figure 4.1 Experimentl lyout During sprging, the oxygen concentrtion ws constntly monitored by using oxygen sensor spots (Pst3; PreSens, Regensburg, Germny) fitted inside ech bottle llowing non invsive mesurement (by not opening the vessel) over time (Figure 4.2). Presens uses the oxoluminescense principle to provide relible mesurements of dissolved oxygen in the wine through the technicl cpbility tied up in the oxygen sensor spot. A light emitting diode (LED) provides blue light excittion pulse to the oxygen sensor spot, which returns fluorescent red light signl indictive of the oxygen concentrtion. Sprging stopped s soon s the required dissolved oxygen concentrtion ws reched. The bottles were then hermeticlly seled using tight seling plstic screw cp. As soon s ll the oxygen in the Ox smples ws consumed, the next phse (T1 Control nd T1 Ox) ws smpled nd frozen t 2 C. The remining Ox tretments (T2 Ox to T5 Ox) were oxygented gin to the required level of dissolved oxygen nd the process ws repeted until ll tretments were smpled nd frozen for nlyses. 13

118 Figure 4.2 Oxygen dosing process Smpling Smpling ws done t the end of ech time frme (T1, T2, T3, T4 nd T5) nd ll smples were frozen until the complete set of 33 smples could be nlysed t the sme time. To the smples tken for the nlyses of voltile thiols, phenols nd glutthione (GSH), 1 mg/l sulphur dioxide ws dded. The reminder of the wine (± 4 litres) were trnsferred into 5 L plstic continers, seled under CO 2 gs nd frozen t 2 C for sensory nlysis Voltile thiol nlysis Anlysis ws done to quntify three voltile thiols, 3MH, 3MHA nd 4MMP. Mesurement of voltile thiols were done using smple preprtion nd quntifiction s first described by Toming et l. (1998b). The method ws djusted ccording to Suklje et l. (212). Some modifictions such s the ddition of deuterted internl stndrds were implemented for the quntifiction of 3MH [3 mercpto(1 2 H 2 )hexnol] nd 3MHA [3 mercpto(1 2 H 2 )hexyl cette] (Hebditch et l., 27). 4 Methoxy 2 methyl 2 mercptobutne (4MMB) ws used s internl stndrd for 4MMP. 14

119 To 5 ml wine, 5 ml of 1mM p hydroxymercurybenzote (p HMB) solution nd then.5 ml of 2 nm butylted hydroxynisole (BHA) solution ws dded. After stirring the smple, deuterted lbelled isotopes were dded. The wine smple ws percolted on n nion exchnge column using Dowex resin (Sigm Aldrich) fter which the thiols were eluted with 5 ml of 5 mm cysteine solution nd extrcted into dichloromethne prior to concentrtion. Orgnic phses were evported under reduced pressure (25 mbr) to pproximtely.5 ml nd trnsferred into 1.5 ml drk vils. The Soxhlet flsk ws rinsed with.5 ml dichloromethne nd then plced in n ultrsonic bth for one minute. Smples were then further concentrted under reduced pressure (1 mbr) to pproximtely 3 µl. The concentrted smples were injected onto gs chromtogrph (GC) (Agilent Technologies 789A) coupled to mss spectrometric detector (Agilent Technologies 5975C upgrded with Triple Axis Detector, Agilent, Snt Clr, CA, USA). The thiols were seprted on HP Innowx column (6 m x.25 mm x.25 μm) with helium s crrier gs t flow rte of.6 ml/min. Injector temperture ws set t 24 C. The initil oven temperture (5 C for 5. min) ws rmped to 115 C t rte of 3 C/min, then incresed to 15 t 4 C/min (held for 3. min), fter which it ws incresed to 25 C t rte 3 C/min nd finlly incresed to 25 C t 1 C/min (held for 19.6 min) before dropping to 5 C t 4 C/min (held for 3. min). The ion source temperture ws 23 C, the uxiliry temperture t 25 C nd the qudrupole temperture djusted to 15 C. For qulittive determintion retention time nd mss spectrum in selective ion monitoring mode (SIM) were used Methoxypyrzine nlysis IBMP nd IPMP nlyses were outsourced to n ccredited lbortory (VinLAB Pty Ltd, Stellenbosch, South Afric). The method used ws djusted ccording to vrious published methods (Allen & Lcey, 1998; Kotserides et l., 1999; Roujou de Boubée et l., 2; Hrtmnn et l., 22). Fifty millilitres of wine contining isopropylethoxypyrzine s internl stndrd were extrcted using C18 solid phse extrction (SPE) crtridges (ISOLUTE, Interntionl Sorbent Technology). The extrct ws concentrted under gentle flow of nitrogen to 5 μl, of which 1 μl ws injected into the instrument. Gs chromtogrphy (Vrin 39 GC) coupled to mss spectrometer (MS) (Vrin, Sturn 21T) ws used. The instrument contined COMBIPAL xt uto smpler with split/splitless 1177 injection port. Seprtion ws done using CPWAX52 column (3 m x.25 mm x.25 µm) from Agilent J&W. The crrier gs consisted of helium t flow rte of 1.3 ml/min. The injection port ws heted to 22 C nd the MS trnsfer line to 245 C. Oven temperture strted t 5 C for 15

120 5. min which ws then incresed t 5 C/min to 11 C fter which it ws rmped to 245 C t 25 C/min. Ech compound ws quntified by comprison with clibrtion curve constructed using pure stndrds. The reltive pek re nd internl stndrd pek re were then compred to the clibrtion curve in order to quntify ech compound. The limit of quntifiction (LOQ) nd detection limit (LOD) were 2.5 nd 1 ng/l respectively Ester, cid nd lcohol nlysis Chemicls, stndrds, nd model wine mtrix hve been described (Louw, 27; Louw et l., 29). Five millilitres of wine with internl stndrd, 4 methyl 2 pentnol (1 μl of.5 mg/l solution in 12 % v/v ethnol, 2.5 g/l trtric cid, deionised Milli Q Wter (Millipore Filter Corp., Bedford, MA, USA)), ph djusted to 3.5 using.1 M NOH), ws extrcted with 1 ml of diethyl ether by sonicting the ether/wine mixture for 5 minutes. The wine/ether mixture ws then centrifuged t 36 g for 3 min. The ether lyer ws removed nd dried on nhydrous N 2 SO 4. The extrct ws then trnsferred to vil contining n insert nd cpped. Ech extrct ws injected into the instrument in duplicte. A gs chromtogrphy flme ioniztion detector (GC FID) ws used for the nlysis. Vlidtion of the method, in terms of selectivity, linerity, LOD, LOQ, recovery, robustness, nd repetbility, hs been described (Louw, 27; Mlherbe, 211). A J&W DB FFAP cpillry GC column (Agilent, Little Flls, Wilmington, DE) with dimensions 6 m.32 mm.5 µm nd Hewlett Pckrd 689 Plus GC (Agilent, Little Flls, Wilmington, DE) equipped with split/splitless injector nd FID detector were used. Three µl of the diethyl ether extrct ws injected t 2 C. The initil oven temperture ws 33 C for 8. min, fter which the temperture ws incresed by 21 C/min to 13 C, t which it ws held for 1.3 min. The temperture ws then rmped t 21 C/min to 17 C held for 1. min then up to 24 C held for 2.5 min. Post run occurs t 24 C for 5. min. The split rtio ws 15:1, nd the split flow rte ws 49.5 ml/min. The column flow rte ws 3.3 ml/min using hydrogen s crrier gs. The detector temperture ws 25 C nd with column flow of 6. ml/min, clened the column of high boiling contminnts. Quntifiction ws performed by compring the rtio of the pek re nd internl stndrd pek re with clibrtion grphs constructed using pure stndrds (Louw, 27; Mlherbe, 211). 16

121 4.2.6 Monoterpene nlysis Solid phse extrction ws performed in Visiprep SPE vcuum mnifold 2 port model from Supelco, in which there 2 positions re vilble for performing the SPE simultneously. Crtridges (Strt SDB L, Phenomenex, Torrnce, CA, USA) were conditioned by rinsing with 4 ml dichloromethne, 4 ml methnol nd finlly 4 ml of model wine solution (12% ethnol wter mixture). Internl stndrd, 5 μl of 2,6 dimethyl 6 hepten 2 ol (25 mg/l in ethnol) ws dded to 5 ml of wine nd mixed briefly. The wine ws then rinsed through the crtridge by vcuum suction (.5 kp). Clen up ws obtined by flushing the crtridge with 4 ml Milli Q Wter. The crtridge ws then dried under vcuum ( 1 kp) for 15 minutes. Finlly, terpenoids were eluted from the solid phse using 2 ml dichloromethne. The dichloromethne elute ws dried on sodium sulphte crystls nd injected into the GC FID. Ech extrct ws injected into the GC FID in duplicte. Seprtion nd quntifiction were performed on Hewlett Pckrd (Plo Alto, CA) 589 Series II gs chromtogrph equipped with 6 m.32 mm.5 µm fused DB FFAP cpillry column (J&W Scientific, Folsom, CA), nd flme ioniztion detector. Seprtion conditions were s follows: injector temperture 2 C; GC column temperture 4 C (12 min) t 12 C/min to 19 C, followed by temperture rmp of 15 C/min to finl temperture of 25 C held for 2 min; crrier gs helium t 4 kp. Ech compound ws quntified by comprison with clibrtion curve constructed using pure stndrds. The reltive pek re nd internl stndrd pek re were then compred to the clibrtion curve in order to quntify ech compound (Louw, 27) Sulphur dioxide nd scorbic cid nlysis Free nd totl SO 2 were nlysed potentiometriclly (Ripper method) using the Metrohm titrtion unit (Metrohm Ltd., Switzerlnd) (Amerine & Ough, 1974). Ascorbic cid nlysis ws done by n ccredited commercil lbortory (Vinlb, Stellenbosch) using n utomted enzymtic procedure Glutthione nlysis A 2 µl wine smple ws trnsferred into 2 ml eppendorf tube (Eppendorf, Hmburg, Germny) fter which it ws diluted 5 times by dding 8 µl HPLC grde Milli Q Wter supplemented with 1 mg/l SO 2 nd freshly prepred 5 mg/l scorbic cid solution (Sigm Aldrich, St. Louis, MO, 17

122 USA) to protect the smple ginst oxidtion. The diluted smple ws then centrifuged (Centrifuge 5415 D, Eppendorf, Hmburg, Germny) t 12 8 rpm for 5. minutes t 2 C fter which 95 µl were trnsferred to n mber coloured vil which ws sprged with CO 2 gs before cpping. Ultr performnce liquid chromtogrphy tndem mss spectrometric (UPLC MS/MS) nlysis ws performed on Wters Acquity UPLC (Milford, MA) connected to Wters Xevo triple qudrupole MS using electronspry ioniztion in the positive mode. Seprtion ws chieved on Wters Acquity BEH Phenyl column (1 x 2.1 mm x 1.7 μm), using.4% trifluorocetic cid (Solvent A) to cetonitrile (Solvent B) grdient. The injection volume ws 3 μl. The MS settings were optimized for best sensitivity, Cone voltge of 18 V ws used for reduced GSH nd 2 V for oxidised GSH (GSSG). Dt ws cquired in multiple rection monitoring mode (MRM). A MRM trnsition of 38.1>179.1 t collision energy of 17 V ws used for GSH. A MRM trnsition of > t collision energy of 2 V ws used for GSSG (Kritzinger et l., 213) Phenolic nlysis Reverse phse high performnce liquid chromtogrphy (RP HPLC), dpted from the method of Peng et l. (22) ws performed on Agilent 12 series HPLC system equipped with diode rry detector (Agilent Technologies, Plo Alto, CA, USA). Dt processing ws done with Chemsttion softwre (Agilent, Wldbronn, Germny). Seprtions were crried out on polystyrene/divinylbenzene reversed phse column (PLRP S, 1Ǻ, mm, 3 µm) from Polymer Lbortories (Ltd) (Shropshire, UK) protected with gurd crtridge (PLRP S, mm) (Polymer Lbortories (Ltd), Shropshire, UK) with the sme pcking mteril. The following mobile phses were used: solvent A, contining 1.5% o phosphoric cid (Merck) nd solvent B consisting of cetonitrile (Sigm). A flow rte of 1 ml/min ws used nd column temperture of 35 C. Phenols were quntified using externl stndrds: (+) ctechin hydrte (Fluk), ( ) epictechin (Sigm) nd cffeic cid (Sigm). Monomeric flvnols were quntified t 28 nm s mg/l (+) ctechin units with quntifiction limit of 1.5 mg/l, nd ( ) epictechin s ( ) epictechin with quntifiction limit of 1.5 mg/l. Trns cftric cid nd grpe rection product (GRP) hve mximl bsorbnce t 316 nm nd were quntified s mg/l cffeic cid. Wine smples were filtered through.45 µm filter fter which the smple ws plced in 1.5 ml drk glss vil nd protected with nitrogen gs from oxidtion. The LOQ ws determined s the smllest re tht could be ccurtely integrted (<3% stndrd devition), or expressed s signl to noise rtio of t lest 1 (Peng et l., 22). 18

123 4.2.1 Crbonyl compounds nd cetl nlysis The extrction procedure ws done ccording to previous publiction by Silv Ferreir et l. (23). In short, 5 ml of wine smple were spiked with 5 µl of 3 octnol in hydrolcoholic solution (447 mg/l) s the internl stndrd. Anhydrous sodium sulphte (5 g) ws dded to increse the ionic strength of the smple fter which the wine ws extrcted twice with 5 ml of dichloromethne. The two orgnic phses were then blended nd dried with nhydrous sodium sulphte. A totl of 2 ml of the orgnic extrct were concentrted to.4 ml under constnt nitrogen strem. Extrcts were nlysed using Vrin 45 GC, equipped with mss spectrl detector, Vrin 24 MS, nd the Vrin MS worksttion softwre version Seprtion ws done using Stbilwx DA column (6 m x.25 mm x.25 µm) fused silic (Restek, Bellefonte, PA). The injection port ws heted to 22 C. The injection volume ws 2 µl in splitless mode. Helium ws used s crrier gs (Gsin, Portugl) t constnt flow rte of 1. ml/min. Oven temperture strted t 4 C for 1. min which ws then incresed t 2 C/min to 22 C nd held for 2 min. Mss spectr were cquired in the electron impct mode (ioniztion energy, 7 ev; source temperture, 18 C). Ion trp tempertures were set t 23, 45 nd 17 C respectively. Mss rnge ws m/z 33 35, with scn rte of 6 scns per second in full scn mode. The emission current ws 5 µa, nd the electron multiplier ws set in reltive mode to the utotune procedure. Mximum ioniztion time ws 25 µs, with n ioniztion storge level of m/z 35. Compound identifiction ws chieved by compring retention times nd mss spectr obtined from smple contining pure, uthentic stndrds. Kovts indices were clculted nd lso used s vlidtion when compred to mss spectr, s reported in the Ntionl Institute of Stndrds nd Technology (NIST) 5 MS librry dtbse. Quntifiction ws done on the bsis of stndrd clibrtion curves. No stndrd curves were constructed for benzldehyde, cis nd trns dioxne, cis nd trnsdioxolne, methionol, furfurl nd 5 hydroxymethylfurfurl. For these compounds, the pek re to internl stndrd rtio ws used to compre the vrious tretments. Totl cetldehyde ws nlysed using n Aren 2XT enzyme robot (Thermo Electron Oy, Finlnd). The kit used for the nlysis is the Boehringer Mnnheim / R Biophrm Succinic cid (Roche ctlogue number mnufctured by R Biophrm AG, Drmstdt). Acetldehyde is quntittively oxidised to cetic cid in the presence of ldehyde dehydrogense (Al DH) nd nicotinmide denine dinucleotide (NAD). This enzyme buffer is t ph 9 which releses ll bound cetldehyde from the sulphite. The mount of NADH formed by this rection is stoichiometric to the mount of cetldehyde nd NADH is determined by mens of its light bsorbnce t 34 nm. 19

124 The clibrtion rnge for the nlysis rnged from 216 mg/l nd n r 2 vlue of.998 ws obtined. Anlysis nd utomtion were done ccording to the mnufcturer s instructions Spectrophotometric mesurement Two millilitres of wine smple ws plced in 1 mm opticl pth length cuvette. The colour of the wines were mesured t 42 nd 44 nm ginst the blnk (Milli Q Wter ) using Thermo Spectronic GENESYSTM 1 Ultrviolet (UV) spectrophotometer (Anlytic Jen Specord 5 UV/VIS Spectrophotometer; Jen, Germny). The bsorbnce t 42 nm is commonly used in the wine industry to mesure brown colour in white wines (Singleton & Krmling, 1976; Zoecklein et l., 1995; Ilnd et l., 24) while 44 nm corresponds to the mximum bsorbnce of coloured pigments in model wine contining (+) ctechin (Brdshw et l., 21) Sensory romtic descriptive nlysis The wines were subjected to descriptive nlysis for rom profiling. Twelve judges prticipted in the evlution. Four trining sessions of two hours ech were conducted to obtin consensus between the judges. At ech trining session the pnel ws presented with ll eleven wines. The first section of the trining ws imed t generting suitble ttributes for ll the wines. The wines were then scored on 1 mm unstructured line scle ccording to the ttribute intensity. Scling the ttribute t indicted the bsence of the descriptor while scling t 1 indicted to very high intensity. Reference stndrds were mde in order for the pnel to fmilirize themselves with the nturl scent. After trining, the wines were subjected to testing (no discussion) where the results were collected on pper bllots. Ech repliction ws tested in duplicte. This would men tht the smple set ws subjected to 6 sessions of testing (ech biologicl repetition tested twice). The evlution ws done in booths with rtificil dylight lighting nd temperture control t 2±2 C. Blck ISO wine tsting glsses were used to void ny potentil bis on the pnel due to visul cues, with the risk tht if crried out under nturl lighting drker wines my hve been rted s higher in ttributes relted to oxidtion. Smple glsses were mrked with rndom three digit code (unique for ech judge) nd glsses were covered with plstic lid prior to sensory ssessment to prevent the rom of contminting the lbortory environment. The order of the smples ws rndom nd blnced cross the ssessors. Pnellists evluted the smples orthonslly only nd regulr breks were tken between smples to prevent ftigue. 11

125 Sensory colour nlysis Sorting ws done to clssify the wines ccording to colour. Ech biologicl repetition ws sorted by 17 judges which resulted in 51 nswers (3 biologicl repetitions). Judges consisted of winemkers, finl yer nd post grdute viticulture nd oenology students nd wine reserchers. Judges were chosen ccording to their knowledge of wine nd were sked to rrnge the 11 wines from lest oxidised to most oxidised. The sorting sheet lso stipulted the wines be rrnged ccording to colour only (no sniffing). The sorting ws done in booths with stndrd rtificil dylight lighting nd 4 ml of the wine smple were presented in cler ISO wine tsting glsses. Smple glsses were mrked with rndom three digit code (unique for ech judge) nd glsses were covered with plstic lid Dt nlysis For sttisticl nlysis of the chemicl dt, one wy nlysis of vrince (ANOVA) ws conducted to investigte differences between tretments. For the sensory dt, ssessor performnce ws evluted using PnelCheck (Version V1.4., Nofim, Tromsø, Norwy) ccording to the workflow s described by Tomic et l. (21). ANOVA, correltion nlysis, prtil lest squres regression (PLS) nd cluster nlysis (CA) were performed using Sttistic softwre (Sttistic 1, Sttsoft Inc., Tuls, USA), while principle component nlysis (PCA) ws done using UBbipl pckge (Gower et l., 211). The PCA biplots were creted using the correltion mtrix of the men dt nd in the PLS, the chemicl dt re presented s x vribles (predictor vribles) nd sensory descriptive dt s y vribles. Post hoc Fisher s lest significnt difference (LSD) tests were used to test for significnce nd p vlue threshold of.5 (p<.5) ws used for the determintion of sttisticl significnce. 111

126 4.3 RESULTS AND DISCUSSION Oxygen concentrtion Oxygen ws dded to the Ox tretments t five different intervls, while the Control did not receive ny oxygen. Tble 4.1 shows the totl mount of oxygen consumed by the tretments s well s the time lpse between ech oxygention. Tble 4.1 Totl mount of oxygen consumed (mg/l) nd time between smpling of ech stge. Oxygen concentrtion vlues re mens of triplicte nlysis with stndrd devition; different letters indicte significnt differences t p<.5. Totl consumed oxygen concentrtion (mg/l) Tretment nd dys ged Control Ox T dys T1 64 dys 6.59 e ±.33 T2 148 dys d ±.6 T3 19 dys c ±1.4 T4 24 dys b ±1.4 T5 218 dys ±.41 Totl dissolved oxygen concentrtions dministered during this study vried from n verge of 6.59 mg/l up to mg/l. After lcoholic fermenttion, oxygen cn dissolve t vrious concentrtions due to winemking prctices such s rcking, filtrtion, centrifugtion, wine trnsfers, cold stbiliztion nd closure types (Cstellri et l., 24), while working conditions such s the temperture of the wine, the level of the wine in the tnk nd the protection with inert gs cn lso hve significnt influence (Lisjk, 27). 112

127 4.3.2 Chemicl nlyses Chemicl nlyses were conducted on romtic s well s non romtic compounds. Tble 4.2 shows the concentrtion of ll compounds s well s the colour bsorbnce vlues t ech smpling stge. Tble 4.3 shows vrious romtic compounds, the perception thresholds (s reported in literture) s well s the medium in which the thresholds were tested long with the sensory ttributes ssocited with ech compound. Usully more thn one rom detection threshold hs been published nd where possible, the vlue determined in synthetic wine medium ws used for consistency. As mentioned erlier, the voltile thiols nd the methoxypyrzines re considered to be rom impct compounds for Suvignon blnc wines specificlly. These compounds re often sought fter in commercil Suvignon blnc wines nd winemkers will ttempt to produce wines with high concentrtions of these compounds, while trying to preserve them s best possible (Lund et l., 29). Other rom groups such s the esters re lso importnt in the romtic bouquet of the wine, while cids nd lcohols re not lwys considered s being mjor contributors, however their role s n romtic buffer hs been reported (Ferreir et l., 28). Monoterpenes re usully ssocited with the Musct fmily of grpes. However, these compounds could be present in significnt mounts in Suvignon blnc wines nd could thus influence the rom (Benkwitz et l., 212). The ldehydes especilly cn influence the overll romtic bouquet during oxidtion, while other compound groups such s the cetls might not lter the composition (Silv Ferreir et l., 22; Silv Ferreir et l., 23c). 113

128 Tble 4.2 Concentrtion nd bsorbnce vlues of ll chemicl compounds mesured. Vlues re mens of triplicte nlysis; different letters in row indicte significnt differences t p<.5. Compound unit T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox Oxygen Totl mount of O 2 consumed mg/l 6.59 e d c b Voltile Thiols 4 Mercpto 4 methylpentn 2 one (4MMP) ng/l b 15.4 cde c 9.61 f 12.2 fd b cd 1.22 fe 9.71 f 9.49 f 3 Mercptohexylcette (3MHA) ng/l b bc c c c bc d de f fe 3 Mercptohexnol (3MH) ng/l c c b d e f f g Esters Isomyl cette mg/l b 5.6 c 5.8 cd 4.78 ed 4.67 efd 6.27 b 4.93 ed 4.44 ef 4.4 ef 4.13 f Hexyl cette mg/l.9.87 b.85 c.83 cd.83 ed.82 efd.87 b.83 ed.81 ef.81 ef.8 f 2 Phenylethyl cette mg/l b 2.26 bc 2.17 dc 2.9 dec 2.2 de 2.43 b 2.11 dec 1.99 de 1.99 de 1.92 c Ethyl cette mg/l b b 82.7 b 73.4 b b b b b b b Ethyl butyrte mg/l c.55 b.53 c.51 c.54 c.53 c.5 c.52 c.51 cb.49 c Ethyl lctte mg/l 5. c 5.43 c 8.5 b b c 7.18 b 8.35 b Ethyl hexnote mg/l Ethyl octnote mg/l 1.36 d 1.87 b 1.62 bc 1.61 bd 1.56 dc 1.36 dc 1.86 b 1.48 dc 1.86 b b Ethyl decnote mg/l 2.12 de db 2.45 dc 2.33 dc 1.56 e 3.4 b 2.27 de 3.31 b bc Diethyl succinte mg/l.45 g.67 f 1.3 e 1.28 d 1.3 d 1.44 c.72 f 1.9 e 1.51 c 1.77 b 1.94 Acids Acetic cid mg/l c b b c b c cb 42.2 c c b Propionic cid mg/l Butyric cid mg/l Isobutyric cid mg/l b.76 b.76 b.7 b.74 b.74 b.71 b.71 b.76 b.74 b Isovleric cid mg/l b.6 b.61 b.58 b.6 b.59 b.56 b.57 b.6 b.57 b Hexnoic cid mg/l Octnoic cid mg/l c 7.82 c 7.94 b 7.84 c 7.85 c 7.27 c 7.24 c 7.5 cb 6.98 c 7.46 c Decnoic cid mg/l ce 2.54 cb 2.54 cb 2.53 cbd 2.66 b 2.18 ce 2.28 ceb 2.2 e 2.7 e 2.14 ed Alcohols Methnol mg/l b b b 9.69 b b b b b b b Propnol mg/l b b b 5.21 b 56.4 b b b b b b Butnol mg/l b b 1.38 b 1.56 b 1.55 b 1.47 b 1.49 b 1.57 b 1.56 b Isobutnol mg/l b b b 15.9 b b b b 17.3 b b 17.5 b Isomyl lcohol mg/l Hexnol mg/l.91 cb.91 c.97 c.99 b.96 c.99 c.91 cb.91 cb.95 c Phenylethnol mg/l Ethoxy 1 propnol mg/l b b b 1.78 b 1.95 b 1.95 b Methionol pek/is rtio

129 Tble 4.2 Continued. Compound unit T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox Monoterpenes Linlool mg/l 1.46 e 11.5 cd b 13.7 b cd ed cd 12.8 cb ed 1.88 ed Gerniol mg/l b bc c bc Frnesol mg/l b b b 2.8 b Antioxidnts Free SO 2 mg/l b 32. b bc dc d e 1. f 7. g 6.67 gh 5. h Totl SO 2 mg/l b 9.67 b b b b 82. c 6.67 d d de e Reduced glutthione mg/l b 7.78 c 6.12 d 5.55 d 5.42 d 5.39 d 1.52 e.56 e.62 e.57 e Oxidised glutthione mg/l.9 f.27 dce.16 fe.15 fe.23 de.15 fe.21 df.33 dc.38 bc.48 b.7 Grpe Rection Procuct Cftric cid units 2.4 c 3.45 b 3.94 b 3.86 b b 3.82 b 4.13 b b Phenols Gllic cid mg/l.29 b c (+) Ctechin mg/l 2.62 c c 3.3 b 3.25 b 3.12 c 2.83 cb 2.86 cb 2.83 cb 2.64 c Cffeic cid mg/l 6. b b 6.39 b 6.64 b 6.64 b 6.35 b 5.93 b 6.38 b 6.34 b 6.46 b trns Cftric cid Cffeic cid equivlents dc d d 12.4 b b 12.1 bc d db 12.8 bc dc p Coumric cid mg/l 3.96 b b 4.17 b b 3.78 b 4.7 b 4.3 b 4.11 b p Coutric cid Coumric cid equivlents.75 e.85 bc.84 bc.88 b ed.77 e.8 ec.83 bcd.76 e Polymeric phenols Ctechin equivlents 2.6 b cd cf fc fe cb cf cde fd f Crbonyl compounds Acetldehyde mg/l d d d 4.43 d 4.95 d 4.6 d d d c b Methionl ug/l d 1.31 c 2.64 b Phenylcetldehyde ug/l nd nd nd nd nd nd nd nd nd nd nd Benzldehyde pek/is rtio d dc dc bc 17.2 bd bd 2.46 b dc 6.78 e Furfurl pek/is rtio 1.91 e 18.6 d d c c c d b Hydroxymethylfurfurl pek/is rtio 4.32 d 6.91 bd 6.7 dc 6.82 bd 9.29 bc 8.8 bc 7.41b d 1.55 b Sotolon ug/l Acetls Cis dioxne pek/is rtio c c c c c c c c 26.4 c b 94.1 Cis dioxolne pek/is rtio 6.82 c 2.82 c 2.2 c 2. c 1.86 c 1.78 c 7.3 c 5.34 c 28.8 c b 182. Trns dioxne pek/is rtio 3.36 cd 3.7 d 3.8 cd 3.18 cd 3.4 d 2.97 d 3.84 cd 3.28 cd 5.97 c 9.75 b Trns dioxolne pek/is rtio 4.91 c 2.12 c 1.4 c 1.34 c 1.14 c 1.7 c 4.25 c 3.1 c c b Spectrophotometric mesurements 42 nm bsorbnce units.53 g.55 gf.59 ef.58 ef.59 ef.61 e.62 e.77 d.85 c.93 b nm bsorbnce units.39 g.4 gf.43 fe.42 efg.42 fe.44 de.47 d.59 c.64 b.7.73 IS = Internl stndrd 115

130 Tble 4.3 List of the romtic compounds found in the wine, rom perception thresholds nd ttributes used to describe the vrious odours. Compound Perception threshold Threshold determined in Descriptors Reference Voltile Thiols 4 Mercpto 4 methylpentn 2 one (4MMP).8 ng/l 12% ethnol, ph 3.5 Pssion fruit, broom, blck current Drriet et l., Mercptohexylcette (3MHA) 4.2 ng/l 12% ethnol, ph 3.5 Pssion fruit, grpefruit, gooseberry, guv Toming et l., Mercptohexnol (3MH) 6 ng/l 12% ethnol, ph 3.5 Pssion fruit, grpefruit, gooseberry, guv Toming et l., 1998 Methoxypyrzines 3 Isobutyl 2 methoxypyrzine (IBMP) 2 ng/l wter Green pepper Buttery et l., Isopropyl 2 methoxypyrzine (IPMP) 2 ng/l wter Asprgus Buttery et l., 1969; Allen et l., 1998 Esters Isomyl cette.5 mg/l 12.5% ethnol, ph3.2 Bnn, fruity, sweet Benkwitz et l., 212b Hexyl cette.4 mg/l 12.5% ethnol, ph3.2 Apple, cherry, per, flower Benkwitz et l., 212b 2 Phenylethyl cette.25 mg/l 1% ethnol Rose, honey, tobcco, flower Guth, 1997 Ethyl cette 12.3 mg/l 1% ethnol, ph 3.2 Pinepple, fruity, vrnish, solvent Escudero et l., 24 Ethyl butyrte.2 mg/l 1% ethnol Acidic, fruity, pple Guth, 1997 Ethyl lctte 146 mg/l 14% ethnol, ph3.5 Lctic, buttery, fruity Moyno et l., 22 Ethyl hexnote.14 mg/l 11% ethnol, ph3.4 Green pple peel, fruit, bnn, brndy Ferreir et l., 2 Ethyl octnote.5 mg/l 11% ethnol, ph3.4 Sweet, ripe bnn, per, sopy Ferreir et l., 2 Ethyl decnote.2 mg/l 11% ethnol, ph3.4 Fruity, florl, grpe, sopy, brndy Ferreir et l., 2 Diethyl succinte 1.2 mg/l 1% ethnol, ph 3.5 Fruity, melon Peindo et l., 24 Acids Acetic cid 3 mg/l 1% ethnol, ph 3.2 Vinegr Escudero et l., 24 Propionic cid 8.1 mg/l 11% ethnol, ph3.4 Rncid, pungent, soy Ferreir et l., 2 Butyric cid.173 mg/l 11% ethnol, ph3.4 Rncid, cheese, swet Ferreir et l., 2 Isobutyric cid 2.3 mg/l 11% ethnol, ph3.4 Acidic Ferreir et l., 2 Isovleric cid.33 mg/l 11% ethnol, ph3.4 Blue cheese Ferreir et l., 2 Hexnoic cid.42 mg/l 11% ethnol, ph3.4 Swet, cheesy, ftty Ferreir et l., 2 Octnoic cid.5 mg/l 11% ethnol, ph3.4 Swety, rncid, hrsh, ftty Ferreir et l., 2 Decnoic cid 1. mg/l 11% ethnol, ph3.4 Rncid, ftty Ferreir et l., 2 116

131 Tble 4.3 Continued. Compound Perception threshold Threshold determined in Descriptors Reference Alcohols Propnol 36 mg/l 1% ethnol, ph 3.5 Alcohol, ripe fruit Peindo et l., 24 Butnol 15 mg/l 1% ethnol, ph 3.5 Fusel odour, medicinl Peindo et l., 24 Isobutnol 4 mg/l 1% ethnol Fusel, lcohol, nil polish Guth, 1997 Isomyl lcohol 3 mg/l 1% ethnol Whiskey, mlt, burnt Guth, 1997 Hexnol 8 mg/l 1% ethnol Grssy, green, resin, flower, woody Guth, 1997 Phenylethnol 14 mg/l 1% ethnol, ph3.5 Honey, spice, rose, lilc Peindo et l., 24 Methionol 1 mg/l 1% ethnol, ph3.5 Culiflower, cooked cbbge, sweet, potto Peindo et l., 24 Monoterpenes Linlool 25.2 µg/l 11% ethnol, ph3.4 Fruity, citrus, florl, lvender Ferreir et l., 2 Gerniol 3 µg/l 1% ethnol Rose, gernium Guth, 1997 Frnesol 1 mg/l 1% ethnol, ph3.5 Florl, oily Peindo et l., 24 Crbonyl compounds Acetldehyde.5 mg/l 1% ethnol Sherry, nutty, bruised pple Guth, 1997 Methionl.5 µg/l 11% ethnol,ph 3.4 Bked pottoes Escudero et l., 2b Phenylcetldehyde 1 µg/l 1% ethnol, ph 3.2 Honey Culleré et l., 27 Benzldehyde 2 mg/l 1% ethnol, ph3.5 Liquor, checmicl, bitter cherry Escudero et l., 22; Peindo et l., 24 Furfurl 15 mg/l beer Cooked vegetbles, woody, pper, green fruits Câmr et l., 24 5 Hydroxymethylfurfurl 1 mg/l beer Aldehyde, chemicl, woody Câmr et l., 24 Sotolon 2 µg/l 12% ethnol, ph 3.5 Spice, curry, nutty Pons et l., 24 Acetls Cis dioxne, Cis dioxolne, Trnsdioxne, Trns dioxolne 1 mg/l (mixture of 4 isomers) Estimted by sniffing from GC O sweet, old port like Silv Ferreir et l.,

132 Voltile Thiols The voltile thiols contribute to lrge extent to the fruity rom in Suvignon blnc wines, often described s tropicl, pssion fruit, guv nd grpefruit. These compounds form n importnt prt of the overll rom bouquet of Suvignon blnc wine nd the retention of these compounds is importnt during ging, especilly due to the thiols being oxidtion sensitive (Blnchrd et l., 24; Nikolntonki et l., 21). The decrese of these compounds during oxygen exposure nd ging cn be due to the direct rection with hydrogen peroxide in the presence of metl ion (Jocelyn, 1972); the thiols cn rect with polyphenolic compounds vi nucleophilic, cidctlysed substitution rections (Ribéreu Gyon, 1998, 24b) nd they cn degrde t fst rte due to the rection with phenolic oxidtion products such s the very rective o quinones (Nikolntonki et l., 21). Any of these rections will result in loss of thiol derived vrietl chrcter. Recent observtions reported the o quinone trpping s the min mechnism ccounting for 3MH loss in wine under oxidtive conditions, while other rections seem to contribute mrginlly (Kreitmn et l., 213). In this study, the vlues obtined for 3MHA flls within the rnge of concentrtions found in Suvignon blnc wines originting from round the world, however 3MH concentrtions re lower when considering wines from New Zelnd, but were on pr with concentrtions found in generl in South Africn Suvignon blnc wines (Tble 4.2) (Lund et l., 29; Vn Wyngrd, 213). After bout 7 months of storge t 15 C in the drk, 3% decrese in 3MHA concentrtion ws seen for the Control smples while the Ox smples concentrtion decresed by bout 75% (Figure 4.3). A decrese of up to 46% hs lso been observed in New Zelnd Suvignon blnc wines tht hve been stored for 7 months t 15 C (Herbst Johnstone et l., 211). Another study done in French rosé wines reported 6% loss in 3MHA within 3 months of storge (Murt, 25). The loss in 3MHA cn be due to oxidtion of the voltile thiol with previously described mechnisms (probbly the rection tking plce in the Ox smples) nd/or due to hydrolysis rection. 3MHA, being n cette ester, cn hydrolyse to form 3MH nd cetic cid (Figure 4.4). The oxidtion of 3MH could lso ccelerte this rection by removl of the hydrolysis product. This hydrolysis rection is probbly responsible for the decrese in 3MHA concentrtion especilly in the Control smples. 118

133 12 3MHA concentrtion (ng/l) A b bc c c c bc d de f fe 2 1 Tretment Oxygen (mg/l) T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 T5 Control Tretment T1 Ox T1 Ox 6.59 e T2 Ox T2 Ox d T3 Ox T3 Ox c T4 Ox T4 Ox b T5 Ox T5 Ox B b 3MH concentrtion (ng/l) c c d e f f g Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 Ox T4 Ox b T5 Ox T5 Ox MMP concentrtion (ng/l) C b cde c f fd b cd fe f f 1 5 Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Figure 4.3 3MHA (A), 3MH (B), 4MMP (C) concentrtion (ng/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<

134 Figure 4.4 Hydrolysis of 3 mercptohexn 1 ol cette to 3 mercptohexn 1 ol nd cetic cid (Herbst Johnstone et l., 211). To the contrry, 3MH concentrtions remined stble (Tble 4.2, Figure 4.3) from T Control to T1 Control (up to 64 dys) then strted to increse with the mximum of ng/l t T3 Control fter which it decresed to level of 79. ng/l t T5 Control which, interestingly, ws still significntly higher thn the initil concentrtion of ng/l t T Control. This increse is likely due to the hydrolysis of 3MHA to 3MH nd similr results hve been reported in previous studies (Herbst et l., 28; Herbst Johnstone et l., 211). Herbst Johnstone et l. (211) found slight decrese (5%) in 3MH concentrtions during the first 3 months of storge fter which the concentrtion incresed up to 7 months of storge. Another study by the sme uthor reported n increse in 3MH levels of up to 63% during the first 3 months of storge fter which 3MH levels decresed (Herbst et l., 28). Other thn the hydrolysis of 3MHA, the possibility of the relese of 3MH from remining precursors or ny 3MH disulphide present in the wine during ging exists, however, the thiol precursors pper to be stble t wine ph, so tht the free form should not be ble to form vi this mechnism (Uglino et l., 211). The role of known precursors is still to be clrified nd the possibility of other not yet identified precursors which could lso contribute to the increse in 3MH concentrtions during ging needs to be elucidted (Cpone et l., 21; Srrzin et l., 21; Coetzee & Du Toit, 212). These mechnisms should be investigted especilly considering tht the 3MHA quntittive loss does not mke up for the gin in 3MH concentrtion on 1:1 rtio. In the Ox smples, 54% decrese in 3MH concentrtion ws observed from T Control to T5 Ox. The rectivity of 3MH nd the oxidtion thereof hs been reported (Blnchrd et l., 24; Nikolntonki et l., 21) nd is probbly the cuse of this decrese. Given perception threshold of 4.2 ng/l (Tble 4.3), 3MHA my ply n importnt role in especilly young Suvignon blnc wines. The perception threshold of 3MH is lso reltively low (6 ng/l; Tble 4.3) nd the sensory impct my be considerble in young wines. The degrdtion of 3MHA nd the subsequent production of 3MH by hydrolysis, is expected to lower the intensity of some tropicl roms such s pssion fruit nd swety roms (Coetzee et l., 212). The vlues obtined for 4MMP (Tble 4.2, Figure 4.3) fll within the rnge of concentrtions found in South Africn nd French Suvignon blnc wines (Toming et l., 1998b; Vn Wyngrd, 213). In 12

135 the Control smples, 4MMP content decresed by 74% from T to T5. The ddition of oxygen in the Ox smples did not seem to ccelerte this decrese (s is the cse with 3MH nd 3MHA), but rther decresed to similr degree when compred to the Control smples. 4MMP hs been shown not to be s sensitive to oxidtion when compred to the other thiols (Nikolntonki et l., 21; Coetzee et l., 213). Rection rtes between the o quinone nd the voltile thiols hve been reported to increse s steric hindrnce declines (Nikolntonki et l., 21; Nikolntonki & Wterhouse, 212). 4MMP is tertiry thiol, therefore it is not s oxidtion sensitive when compred to 3MH, which is secondry thiol. The decrese observed in this study could be due to nturl degrdtion of the molecule over time. A decrese of 74% due to nturl degrdtion is significnt nd should be kept in mind when storing wine. Overll the loss in voltile thiols could chnge the romtic composition of the wine substntilly Methoxypyrzines Methoxypyrzines hve been shown not to be sensitive to oxidtion s even hyperoxidtion (using H 2 O 2 ) of n IBMP spiked Chenin blnc wine did not hve ny effect on the IBMP concentrtion (Mris, 1998). Oxidtion of Suvignon blnc juice lso hd no effect on the methoxypyrzine content in the corresponding wines (Coetzee et l., 213). Methoxypyrzine nlysis ws thus done on select few smples to verify whether ny differences between extreme smples existed. T Control hd n IBMP concentrtion of 12.9 ng/l, while T5 Control nd T5 Ox contined 13.9 nd 13. ng/l respectively. This would suggest tht there ws no tretment effect on the concentrtion of this compound. Smll differences observed would not hve n effect on the sensoril perception nd the differences seen here could lso be considered s smll vritions in mesurements t these low concentrtions. No IPMP ws detected in these smples. The continuous presence of IBMP would indicte the presence of the green rom chrcters in ll the wine smples. Sensory profiling will be discussed in section 4.3.3, however it is interesting to note tht correltion between pyrzine concentrtion nd the green ttributes in wines does not lwys exist, suggesting the contribution of other rom compounds to these ttributes (Ferreir et l., 22; Preston et l., 28). 121

136 Esters, cids nd lcohols Three cette esters, isomyl cette, hexyl cette nd 2 phenylethyl cette (Tble 4.2, Figure 4.5) were significntly degrded during the 7 month storge. Aging time thus seems to be the min driver of degrdtion in most cses s the Control tretment did not differ significntly from its Ox counterprt. The process for the loss of cette esters during ging is gin expected to be hydrolysis of the ester to cetic cid nd n lcohol, which occurs redily t wine ph. The decline in cette esters during storge hs been reported in literture previously (Mris & Pool, 198; Rmey & Ough, 198; Ferreir et l., 1997; Moio et l., 24; Lmbropoulos & Roussis, 27; Mkhotkin & Kilmrtin, 212; Ptrinkou & Roussis, 213). The preservtion of these esters is importnt s they contribute to the plesnt rom of white wines often described s bnn, fruity, per, pple, rose nd flower (Frncis & Newton, 25; Swiegers et l., 25; Benkwitz et l., 212b). Ethyl lctte nd diethyl succinte concentrtions incresed during the study (Tble 4.2, Figure 4.6). This occurred in both the Control nd Ox smples. Diethyl succinte concentrtions incresed to higher concentrtion in the Ox smples when compred to the Control counterprts especilly in the lst three smpling stges. These two compounds re usully ssocited with mlolctic fermenttion (Louw et l., 21) nd contribute odours such s fruity, florl nd brndy. Their occurrence could lso rise due to the trnsformtion of lctic nd succinic cids to form ethyl lctte nd diethyl succinte during fermenttion nd mturtion (De Villiers et l., 23). The increse of these compounds during ging hve been reported in literture previously (Rpp, 1988; Ferreir et l., 1997; Pérez Coello et l., 23; Hernnz et l., 29) nd interestingly these compounds hve been identified s being the min discriminnt vribles in 9 white wines with vrious ges (Pérez Coello et l., 23). No significnt chnge in concentrtion ws observed for ethyl hexnote, while ethyl octnote nd ethyl decnote concentrtions exhibited inconsistency during the storge time (Tble 4.2, Figure 4.7). There seems to be slight decrese in ethyl octnote nd ethyl decnote when considering the Control smples only (T1 Control to T5 Control). The disppernce of the ethyl esters of some higher lcohols during storge hs been reported before (Mris & Pool, 198; Rmey & Ough, 198; Ferreir et l., 1997; Moio et l., 24; Uglino et l., 28; Ptrinkou & Roussis, 213), while the stble concentrtion of ethyl hexnote during ging of Riesling nd Cbernet frnc wines hs lso been reported elsewhere (Blke et l., 21). In the Ox smples, n increse in these ethyl esters ws seen when compring the Control smples with the Ox counterprts, especilly t smpling stges T4 nd T5. A chnge in the hydrolysis esterifiction equilibri could ccount for this 122

137 observtion (Rmey & Ough, 198), especilly if chnge in cetic cid concentrtion occurred over the storge period (Esclon et l., 22). An increse in ethyl octnote during 12 month storge period hs lso been reported previously (Mris & Pool, 198; Blke et l., 21). 123

138 A Isomyl Acette (mg/l) b c cd ed efd b ed ef ef f Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Hexyl Acette (mg/l) B b c cd ed efd b ed ef ef f Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 Ox T4 Ox b T5 Ox T5 Ox C 2-Phenylethyl cette (mg/l) b bc dc dec de b dec de de e Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Figure 4.5 Isomyl cette (A), hexyl cette (B) nd 2 phenylethyl cette (C) concentrtion (mg/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<

139 A Ethyl Lctte (mg/l) c c b b c b b Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox B b Diethyl Succinte (mg/l) f e d d c f e c.6 g.4.2 Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Figure 4.6 Ethyl lctte (A) nd diethyl succinte (B) concentrtion (mg/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<

140 1.7 Ethyl Hexnote (mg/l) A Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox B Ethyl Octnote (mg/l) d b bc bd dc dc b dc b b Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e d T2 T2 Ox Ox c T3 T3 Ox Ox T4 T4 Ox Ox b T5 T5 Ox Ox C Ethyl Decnote (mg/l) de db dc de e b de b bc Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Figure 4.7 Ethyl hexnote (A), ethyl octnote (B) nd ethyl decnote (C) concentrtion (mg/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<

141 Even though there were no significnt differences in hexnoic cid concentrtions, slight but not significnt decrese in concentrtion ws observed for octnoic cid, while there ws significnt decrese in decnoic cid when compring the Control smples to their Ox counterprts especilly during the lte oxidtive stges (Tble 4.2, Figure 4.8). The sme tendency ws observed in other studies (Mris & Pool, 198; Ferreir et l., 1997; Câmr et l., 26; Blke et l., 21; Lee et l., 211) nd reported the results s surprising due to the expecttion of these compounds to increse s result of the hydrolysis of the corresponding ethyl ester. As mentioned erlier, the ethyl ester concentrtion did not decrese, to the contrry, in some cses the concentrtion incresed nd would thus not result in the formtion of the corresponding cid. The possibility of ftty cid utoxidtion to yield ldehydes hs been contemplted (Nykänen, 1986) nd cn possibly explin the slight decrese in cid concentrtion observed. Other short chin ftty cids, such s cetic cid, propionic cid, butyric cid, isobutyric cid nd isovleric cid, were not significntly ffected by the tretment (Tble 4.2). Ftty cids hve been reported not to oxidise esily since the oxidtion of these compounds under lbortory conditions required the use of oxidnts stronger thn H 2 O 2 (Ferreir et l., 1997). With the exception of hexnol (Figure 4.9), the lcohols did not seem to be ffected by oxidtion (Tble 4.2). This hs lso been observed by other studies (Ferreir et l., 1997), which mens tht the oxidtion of the lcohols to the corresponding ldehydes (Wilderndt & Singleton, 1974) could only ccount for smll frction s the concentrtions did not decrese significntly. Hexnol contributes to the grssy nd green odours of wine nd cn be formed due to oxidtion of linoleic nd linolenic cids nd its formtion during ging hs been reported previously (Mris & Pool, 198; Oliveir et l., 26), however in the present study the hexnol concentrtions did not exceed the perception threshold (Tble 4.3) nd probbly did not ply n importnt role in the sensory perception of the wine. Methionol concentrtions hs been reported to decrese during ging (Ferreir et l., 1997), however in this study, there ws no significnt difference in methionol concentrtions (Tble 4.2). 127

142 6.2 Hexnoic Acid (mg/l) A Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 T1 Ox Ox tretment/time 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox B c b c c Octnoic Acid (mg/l) c c c cb c c Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox C 3. b Decnoic Acid (mg/l) ce cb cb cbd ce ceb e e ed Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Figure 4.8 Hexnoic cid (A), octnoic cid (B) nd decnoic cid (C) concentrtion (mg/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<

143 Hexnol (mg/l) cb c c b c c cb cb c Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Figure 4.9 Hexnol concentrtion (mg/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p< Monoterpenes Monoterpene concentrtions re known to chnge considerbly during wine ging (Rpp & Güntert, 1986) nd the loss in rom my occur due to oxidtion or trnsformtion of the terpene compounds (Versini et l., 1981; Ppdopoulou & Roussis, 21). These compounds re sensitive to cidic conditions nd storge time (Mris, 1983) nd remrkble loss of monoterpene lcohols occurs due to the formtion of terpene oxides tht hve sensory thresholds bout 1 times higher thn the precursors (Ppdopoulou & Roussis, 21). In this study, linlool concentrtion initilly incresed during the ging period fter which it strted to decrese (Tble 4.2, Figure 4.1). The decrese in linlool concentrtion during wine oxidtion nd ging hs been reported previously (Ppdopoulou & Roussis, 21, 28; Loscos et l., 21; Cejudo Bstnte et l., 213) nd my be due to oxidtion or the progressive replcement of linlool by α terpineol (Rodopulo et l., 197; Mris, 1983; Rpp & Güntert, 1986), however the mount of α terpineol detected in these smples ws under the quntifiction limit. A substntil mount of bound terpenols still remin in the wine fter fermenttion nd ging could induce the hydrolysis of these bound terpenols to produce the free, romtic terpenes (Gunt et l., 1986). The extent of the hydrolysis will depend on the specific type nd nture of the glycone (Gunt et l., 1986). This could explin the initil increse in linlool concentrtion t the beginning of the ging period. There were no significnt differences between the Control smples nd their Ox counterprts for linlool 129

144 nd gerniol which could indicte the minor role the ddition of oxygen plyed. Gerniol nd frnesol concentrtion remined the sme for the first three smpling stges (Tble 4.2, Figure 4.1) fter which it strted to decrese probbly due to degrdtion of the terpene lcohol. 13

145 A b b Linlool (mg/l) e cd cd ed cd cb ed ed Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox B 16 Gerniol (mg/l) b bc c bc Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 Ox T5 Ox C Frnesol (mg/l) b b b b Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Figure 4.1 Linlool (A), gerniol, (B), frnesol (C) concentrtion (mg/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<

146 Sulphur dioxide nd Glutthione No scorbic cid ws detected in the wines. The evolution of free nd totl SO 2 is shown in Tble 4.2 nd Figure At the beginning of the study, the wine contined dequte levels (ccording to commercil stndrds) of both free nd totl SO 2. In the Control smples, the free SO 2 concentrtion initilly decresed from T Control (43.33 mg/l) to T1 Control (31.67 mg/l) (64 dys). Therefter the concentrtion remined stble with slight decrese towrds the end of the ging period (to mg/l in T5 Control). From T Control to T5 Control, the totl SO 2 concentrtion decresed from 1.33 mg/l to mg/l. This resulted in 35% nd 8% decrese in free nd totl SO 2 concentrtions respectively. This decrese in SO 2 concentrtion grees with previous studies (Brjkovich et l., 25; Lopes et l., 29; Herbst Johnstone et l., 211) nd even though cre ws tken during bottling to chieve minimum oxygen exposure (in the Control smples), some oxygen could hve dissolved during trnsfer nd smll mount of oxygen could hve remined in the hedspce fter bottling. The most importnt loss in SO 2 hs been reported to occur within the first month of bottling, fter which the SO 2 undergoes slower rte of decrese or remins stble, which is consistent with the results from this study (Brjkovich et l., 25; Lopes et l., 26; Kwitkowski et l., 27; Dimkou et l., 211; Herbst Johnstone et l., 211). With the ddition of oxygen, significnt decrese in SO 2 concentrtion ws observed. Free SO 2 concentrtions decresed drmticlly to rech 1. mg/l t T2 Ox fter which it decresed further to rech 5 mg/l in T5 Ox resulting in n 88% loss. Godden et l. (21) reported the considerbly higher oxidised chrcter in Semillon wines with free SO 2 content below 1 mg/l. Totl SO 2 decresed with 49% reching minimum of mg/l in the T5 Ox smple. With the SO 2 reching these low concentrtions, very little ntioxidnt protection remins in the wine. The potentiometric method used to nlyse the free nd totl SO 2 content could perhps led to slight over estimtion of the SO 2 content in the smple due to the bility of phenols nd sugrs to be oxidised by the iodide. The smller quntities mesured in the lter Ox stges could indicte tht in relity very little to no SO 2 ws present in these smples leving the wine with effectively no SO 2 ntioxidnt protection. 132

147 A Free SO 2 (mg/l) b b bc dc d e f g gh h Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox B b b b b b 9 c Totl SO 2 (mg/l) 8 7 d d 6 de e 5 4 Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Figure 4.11 Free (A) nd Totl (B) SO 2 concentrtion (mg/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<.5. Another ntioxidnt tht helps protect wine from oxidtion is GSH (Kritzinger et l., 213). GSH is very oxidtion sensitive nd diminishes rpidly during storge of wines (Lvigne et l., 27). In this study, GSH concentrtion rnged from.56 mg/l to mg/l (Tble 4.2, Figure 4.12). The initil concentrtion grees with previous study tht found the verge GSH concentrtion in 28 young Suvignon blnc wines to be round 12.5 mg/l (Jnes et l., 21). 133

148 In the Control smples, GSH concentrtions decresed with 68% while it lmost completely disppered in the Ox smples, decresing with 97% nd correlted well with free SO 2 concentrtions (r=.995). As mentioned before, the decrese in GSH content from T Control to T1 Control could be due to oxygen exposure during the bottling phse. However, unlike the free SO 2 content which remined stble therefter, the GSH content continued to decline to the end of the study (T5 Control). GSH content in the Ox smples decresed more drmticlly, with T1 Ox still contining some residul GSH (5.39 mg/l), while further oxidtion led to levels of 1.52 mg/l nd lower. Together with the decrese in reduced GSH, there ws n increse in oxidised GSH nd GRP (Tble 4.2, Figure 4.12). The increse in GRP ws slight nd not consistent. This phenomenon hs lso been reported in bottled Suvignon blnc wine stored for two yers showing firly similr vlues compred to concentrtions found prior to bottling (Herbst et l., 28). The increse in the oxidtion products does not ccount for the totl mount of reduced GSH lost. This hs lso been seen in other studies (Frcsetti et l., 213) nd could indicte to the rection of GSH with other substrtes such s hydroxycinnmic cids or other compounds not mesured. Sonni et l. (211) reported the formtion of methyl glutthionyl methine (t) ctechin complex in model wine due to the rection of GSH with (+) ctechin. Furthermore, GRP hs been reported to undergo hydrolysis in model nd rel wine nd could explin the low percentge of ccounted oxidtion products (Cejudo Bstnte et l., 21). As thiol, GSH is ble to reduce o quinones nd H 2 O 2. Therefore, it cn be hypothesised tht GSH competes in these rections with voltile thiols, thus preventing the loss of the unique rom brought by these compounds. 134

149 A 14 Reduced GSH (mg/l) b c d d d d 4 2 e e e e Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox A 14 Reduced GSH (mg/l) b c d d d d 4 2 e e e e Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c b T4 T4 Ox Ox T5 T5 Ox Ox C GRP (cftric cid units) c b b b b b b b 1 Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Figure 4.12 Reduced glutthione (mg/l) (A), oxidised glutthione (mg/l) (B) nd grpe rection product (cftric cid units) (C) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<

150 Polyphenols Phenol oxidtion cn hve detrimentl effects on the rom of wine due to the formtion of rective o quinones tht cn then in turn rect with other wine constituents such s the romtic voltile thiols. Seven phenolic compounds were identified in the wines of this study (Tble 4.2). Gllic cid ws present in very low concentrtions (<1 mg/l) when compred to concentrtions found in white wines in generl, which ws found to verge round 1 mg/l (Wterhouse & Teissedre, 1997). The concentrtion of this compound hs been reported to remin reltively stble during ging (Wterhouse, 22), which hs lso been observed in the current study. Only from T Control to T1 Control ws significnt increse in gllic cid concentrtion observed (from.29 mg/l to.52 mg/l). This could be due to the hydrolysis of gllte esters during the first periods of ging (Wterhouse, 22). (+) Ctechin, trns cftric cid, p coutric cid, cffeic cid nd p coumric cid concentrtions were lso quntified in the wines (Tble 4.2, Figures 4.13 nd 4.14). (+) Ctechin concentrtions decresed in the Ox smples when compred to the Control counterprts (Figure 4.13), however in some cses the decrese ws not significnt. The flvn 3 ols prticipte in oxidtion nd polymeriztion rections nd hve been linked to the susceptibility of white wines to browning (Simpson, 1982). ( ) Epictechin ws not detected in ny of the smples. The tretment hd little effect on cffeic cid or p coumric cid concentrtions, while decrese ws observed in trns cftric cid nd p coutric cid concentrtions (compring the Control to Ox counterprts); however these decreses (especilly trns cftric) cid were not lwys significnt. The loss of flvn 3 ols during ging nd the more stble responses of the hydroxycinnmtes hve been reported previously (Recmles et l., 26; Herbst et l., 28; Hernnz et l., 29; Herbst Johnstone et l., 211). This is likely due to the greter susceptibility of the flvn 3 ols to prticipte in slow oxidtive or other degrdtive processes during bottle storge. In contrst, n increse in the free hydroxycinnmic cids, cffeic cid nd p coumric cid, ws reported by Herbst et l. (28) scribing it to the hydrolysis of their corresponding trtrte esters. In the current study, coutric cid concentrtions incresed from T Control to T5 Control (.75 to.95 p coumric cid equivlents) nd slight increse ws lso observed from T1 Ox to T4 Ox (.77 to.83 p coumric cid equivlents). 136

151 4.2 (+)-Ctechin (mg/l) c c b b c cb cb cb c Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 T5 Control Tretment T1 Ox Ox 6.59 e T2 Ox Ox d T3 Ox Ox c T4 Ox Ox b T5 Ox Ox Figure 4.13 Flvn 3 ol: (+) Ctechin concentrtion (mg/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<.5. Similr observtions were found for other white wine vrieties during 12 months of storge (Recmles et l., 26; Hernnz et l., 29), however the decrese ws still evident compring the Control smples with Ox smples. Other studies lso reported minor chnges in phenolic composition fter 6 dy storge period of oxygented wines (Frcssetti et l., 213). Sulphur dioxide rects with the o quinone, reducing it bck to the ctechol. The concentrtion of the specific phenol would thus not pper to decrese s it is oxidised. It ws therefore proposed tht oxidtion involves n equilibrium, which is driven forwrd by removl of the o quinone (Dnilewicz, 212). In the cse of (+) ctechin, 96% of the o quinone ws reduced bck to the phenol in n oxidised model wine medium contining SO 2 (Dnilewicz, 21), however this hs not been tested in rel wine situtions nd needs further investigtion. No flvonols (e.g. quercetin derivtives) were detected in these wines even though they hve been detected in hevier pressed Suvignon blnc juice frctions (Ptel et l., 21). The polymeric phenol content of the wines decresed during the study (Figure 4.15), however the difference between the Control smples nd the Ox counterprts were mostly not significnt. Frcssetti et l. (213) lso found decreses in oxygen levels in white wines to not lwys correlte well with decreses in phenolic compounds. 137

152 A b b 6.9 b b b b b b Cffeic cid (mg/l) b b Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox trns-cftric cid (Cffeic cid equivlents) dc B d d b b bc d db bc dc 1. Tretment Oxygen (mg/l) T T Control Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 T5 Control Tretment T1 T1 Ox Ox 6.59 e d T2 T2 Ox Ox c T3 T3 Ox Ox b T4 T4 Ox Ox T5 T5 Ox Ox p-coumric cid (mg/l) C b b b b b b b b Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox

153 1.5 p-coutric cid (p-coumric cid equivlents) e D bc bc b ed e ec bcd e.65 Tretment T T Control T1 T1 Control T2 T2 Control T3 Control T4 Control T5 Control T1 T1 Ox Ox T2 T2 Ox Ox T3 T3 Ox Ox T4 T4 Ox Ox T5 T5 Ox Ox Oxygen (mg/l) Tretment 6.59 e d c b Figure 4.14 Hydroxycinnmic cids: Cffeic cid (mg/l) (A), trns cftric cid (cffeic cid equivlents) (B), p coumric cid (mg/l) (C) nd p coutric cid (p coumric cid equivlents) (D). Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p< Polymeric phenols ((+)-Ctechin equivlents) b cd cf fd fe cb cf cde fd f 13 Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 T5 Control Tretment T1 Ox Ox 6.59 e T2 Ox Ox d T3 Ox Ox c T4 Ox Ox b T5 Ox Ox Figure 4.15 Polymeric phenols concentrtion ((+) ctechin equivlents) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<

154 Aldehydes In wine tht ws not exposed to oxygen, the cetldehyde present is considered to originte from lcoholic fermenttion by yests (Mrglith, 1981). Acetldehyde levels produced by yest cn rnge from mg/l for Scchromyces cerevisie yest strins (Liu & Pilone, 2). In the presence of free SO 2, cetldehyde will immeditely bind to form hydroxysulphonte, contributing to the bound cetldehyde present in wine (Liu & Pilone, 2). In this study, the cetldehyde present in the Control smples probbly originted from lcoholic fermenttion nd is most likely in the bound form considering the free SO 2 content verging t mg/l for ll the Control smples. The verge concentrtion of totl cetldehyde (41.28 mg/l) of the wine t T flls within rnge of cetldehyde concentrtion found in dry white wines (Tble 4.2, Figure 4.16) (Jckowetz & De Orduñ, 213). This bound form of cetldehyde is thought to be odourless (Peynud, 1984; Somers, 1998), however the contribution of the odourless product to the rom of wine, either directly or by wy of interctions, hs not been elucidted A Totl cetldehyde (mg/l) d d d d d d d d c b Tretment T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 T1 Ox Ox T2 T2 Ox Ox T3 T3 Ox Ox T4 T4 Ox Ox T5 T5 Ox Ox Oxygen (mg/l) Tretment 6.59 e d c b Figure 4.16 Totl cetldehyde concentrtion (mg/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<.5. 14

155 During the oxidtive ging, the totl cetldehyde concentrtion incresed especilly during the lst three smpling stges of the Ox tretment nd reched mximum of mg/l t T5 Ox. This increse is probbly due to the oxidtion of ethnol (Wilderndt & Singleton, 1974; Ribéreu Gyon, 1998, 24b). The increse in totl cetldehyde will initilly be due to the increse in bound cetldehyde s the free SO 2 is consumed. Once ll free SO 2 is consumed, n increse in free cetldehyde will occur. Free cetldehyde concentrtion ws not determined s there seems to be confusion in literture bout the mesurement of this compound especilly in terms of the mesurement of either totl or free cetldehyde. The different forms of cetldehyde complicte the nlysis. Totl cetldehyde cn be determined chemiclly (iodimetry) or enzymticlly (ldehyde dehydrogense). However, the chemicl method hs proven to give results 1 2% higher thn the enzymtic method (Ough & Amerine, 1988; Liu & Pilone, 2). Furthermore, the enzymtic method is considered s one of the best methods for this nlysis s it hs been reported to be more ccurte nd specific (Liu & Pilone, 2). Free cetldehyde cn reportedly be mesured by gs chromtogrphy, however the results obtined by these mesurements seem to be overestimted t times s most of the wines mesured contin significnt mount of free cetldehyde in conjunction with free SO 2. These contrdictory results hve to be investigted further. Nevertheless, it hs been reported tht bound cetldehyde cn ccount for 68.8% of the totl SO 2 (Liu & Pilone, 2; Bkker et l., 1993). The bound cetldehyde could thus be clculted theoreticlly. The vlue obtined by this clcultion should be confirmed using more dvnced nlyticl techniques s the percentge bound cetldehyde cn be dependnt on vrious other fctors such s the concentrtion of phenolics, presence of other ntioxidnts s well s competition for the rection with sulphur dioxide. The percntge could lso chnge over time s the composition of the wine chnges. Nevertheless, the theoreticl vlue of the bound cetldehyde ws clculted using the specified percentge in order to roughly estimte the free cetldehyde content by subtrcting the bound cetldehyde from the totl cetldehyde mesurement. The stepwise clcultion of molr conversion followed by the.688 rtio clcultion nd conversion bck to mg/l cn be seen in Tble 4.4 together with free SO 2 concentrtion for interprettion. 141

156 Tble 4.4 Clcultion of free cetldehyde from verge totl SO 2, ssuming 68.8% of the totl SO 2 in ll the smples re bound to cetldehyde. Compound unit T Control T1 Control T2 Control T3 Control Free SO 2 mg/l b 32. b bc dc d e 1. f 7. g 6.67 gh 5. h Totl SO 2 mg/l b 9.67 b b b b 82. c 6.67 d d de e Totl SO 2 mol SO 2 : Acet.688 Bound Acetldehyde mol Bound Acetldehyde mg/l Totl Acetldehyde mg/l d d d 4.43 d 4.95 d 4.6 d d d c b Bound Acetldehyde mg/l T4 Control T5 Control T1 O x T2 O x T3 O x T4 O x T5 O x Free cetetldehyde = Totl Bound mg/l In the Control smples, the bound cetldehyde concentrtion ccounts for 1% of the totl cetldehyde mesured using the enzymtic method. This observtion coincides with the fct tht sufficient free SO 2 ws vilble in the Control smples to bind ny formed cetldehyde t this stge. This would indicte no free cetldehyde present in the wine, which is confirmed by the theoreticl vlue clculted for these smples. During the oxidtive storge, there ws n increse in the theoreticl free cetldehyde concentrtion. Figure 4.17 shows the theoreticl free cetldehyde concentrtion together with the totl cetldehyde s well s the free nd totl SO 2 content. T1 Ox hd very low concentrtions of free cetldehyde, while concentrtions incresed t lrger intervls from T2 Ox onwrds. The low free cetldehyde observed in T1 Ox is probbly due to the presence of sufficient free SO 2 in the smple, thereby binding excessive cetldehyde. The rest of the Ox smples experienced further decrese in free SO 2 with the continued increse in free cetldehyde. As mentioned previously the low SO 2 content observed from T2 Ox onwrds could be due to the overestimtion of this compound concentrtion when using the Ripper method s the concentrtion is expected to be zero t this stge. The increse in free cetldehyde should hve importnt sensoril effects on the wine rom. The odour perception threshold of cetldehyde in synthetic wine solution hs been reported s.5 mg/l (Guth, 1997). At low levels this compound cn contribute plesnt fruity roms while t higher concentrtions it is described s green pple, overripe bruised pple, grssy, pungent, nutty nd sherry (Tble 4.3) (Henschke & Jirnek, 1993; Miyke & Shibmoto, 1993; Frivik & Ebeler, 23). The sensory results of these wines will be discussed in section

157 12 1 Concentrtion (mg/l) Tretment Oxygen (mg/l) T Control 1 T1 Ox 6.59 e 2 3 T2 Ox T3 Ox Tretment d c 4 T4 Ox b 5 T5 Ox Free SO 2 Totl SO 2 Totl Acetldehyde Theoreticl Free Acetldehyde Figure 4.17 Concentrtion chnges of totl SO 2, free SO 2, totl cetldehyde nd theoreticl free cetldehyde (ll in mg/l) due to vrious tretments. Methionl nd phenylcetldehyde re relted to the typicl rom of oxidtive spoiled white wine (Silv Ferreir et l., 22b). These two compounds cn be formed due to the oxidtion of the respective lcohol (methionol nd phenylethnol) (Mrchnd et l., 2; Jrut et l., 25) or vi the rection of dicrbonyl (such s the o quinone formed from the oxidtion of phenols) with the respective mino cids (methionine nd phenyllnine) (Silv Ferreir et l., 22b; Rizzi, 26). In this study, no methionl ws detected in the Control smples, however the methionl content incresed significntly from T1 Ox (.58 µg/l) to T5 Ox (4.1 µg/l) (Tble 4.2, Figure 4.18). As mentioned previously, this increse is probbly due to the Strecker degrdtion rection, especilly seeing tht methionol concentrtions remined the sme throughout the tretment (Tble 4.2). Sufficient methionine concentrtions were present in the T Control smples (>4 mg/l; results not shown), mking this pthwy fesible. Surprisingly, no phenylcetldehyde ws detected in ny of the smples, even in the presence of phenyllnine concentrtions exceeding 13. mg/l in 143

158 the T Control smple (results not shown). The formtion of Strecker ldehydes by the rection of mino cids with o quinones t 1 C hs been described s essentilly zero by (Nikolntonki et l., 212) nd could be n indiction of the inhibition of the formtion of especilly phenylcetldehyde t lower tempertures s is the cse in this study. Methionl cn contribute boiled vegetble nd rotten potto rom notes nd could significntly influence the rom of wine considering the low perception threshold of.5 µg/l (Tble 4.3). Further sensory contributions will be discussed in section Benzldehyde concentrtion mostly did not differ between the Control tretments (Tble 4.2, Figure 4.18), however the concentrtion initilly incresed during the oxidtive storge (T1 Ox to T2 Ox), fter which it significntly decresed. The increse in benzldehyde content in white wines hs been reported previously (Ferreir et l., 1997), however the decrese therefter hs not been observed. This phenomenon is not esy to explin. The formtion of benzldehyde hs been ttributed to phenyllnine oxidtion (Loyux et l., 1981), while other uthors reported its origin from mygdline (Nykänen & Suomlinen, 1982). Therefter, benzldehyde cn rect with H 2 S in the process forming methyl mercptn which is responsible for the smoky/empyreumtic rom in wine. However, this mechnism hs only been suggested nd not conclusively demonstrted (Toming, 23; Toming et l., 23). 144

159 A 4. Methionl (µg/l) c b c.5. Tretment T1 T1 Ox Ox T2 T2 Ox Ox T3 Ox T3 Ox Tretment T4 Ox T4 Ox T5 Ox T5 Ox Oxygen (mg/l) 6.59 e d c b B Benzldehyde (pek/is rtio) d dc dc bc bd bd b dc e 5 Tretment T Control T Control T1 Control T1 Control T2 Control T2 Control T3 Control T3 Control T4 Control T4 Control T5 Control T5 Control Tretment T1 Ox T1 Ox T2 Ox T2 Ox T3 Ox T3 Ox T4 Ox T4 Ox T5 Ox T5 Ox Oxygen (mg/l) 6.59 e d c b Figure 4.18 Methionl (A) concentrtion (µg/l) nd benzldehyde (B) (pek/is rtio) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<.5. Furfurl nd other furnic ldehydes usully hve their origin in heting ok during brrel tosting (Ribéreu Gyon et l., 26). However, their occurrence hs lso been suggested to originte from the degrdtion of crbohydrtes during wine ging (Câmr et l., 24) nd good correltion between time ged nd the concentrtion of these furnic compounds were found in Port nd Mdeir wines (Silv Ferreir et l., 23; Câmr et l., 24). Furfurl hs lso been identified s being one of the compounds linked with the tendency of wine to brown (Ferreir et l., 1997). 145

160 Furfurl correlted with the cooked vegetbles descriptor nd could lso contribute to the woody note of ged wines (Escudero et l., 22), while hydroxymethylfurfurl (5 HMF) gives odours such s ldehyde nd crmel (Meilgrd, 1975; Câmr et l., 24). In the current study, furfurl nd 5 HMF concentrtions incresed during the storge of the Control smples, however this increse ws even more prominent in the Ox smples, especilly for the lst three smpling stges (Tble 4.2, Figure 4.19) A 7 Furfurl (pek/is rtio) e d d c c c d b 1-1 Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Hydroxymethylfurfurl (pek/is rtio) d B bd dc bd bc bc bd b -2 Tretment Oxygen (mg/l) T Control T1 Control T2 Control T3 Control T4 Control T5 Control Tretment T1 T1 Ox Ox 6.59 e d T2 T2 Ox Ox c T3 T3 Ox Ox b T4 T4 Ox Ox T5 T5 Ox Ox Figure 4.19 Furfurl (A) nd hydroxymethylfurfurl (B) (pek/is rtio) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<

161 Sotolon Sotolon is very powerful odournt tht smells of curry nd myrrh t high concentrtions with rosting, mple syrup, burnt sugr nd crmel t lower concentrtions (Escudero et l., 2; Ghidossi et l., 212). Young Suvignon blnc wines were found to contin below.5 µg/l sotolon while Suvignon blnc wines stored under vrious closures during 24 month period contined sotolon content rnging from µg/l (Lopes et l., 29). The rom detection threshold for this compound hs been reported to be 2 µg/l in model wine (Tble 4.3) (Pons et l., 21) nd 8 µg/l in dry white wines (Lvigne et l., 28). In this study, no sotolon ws detected in ny of the Control smples. Sotolon ws lredy detected in the T1 Ox smple nd incresed up to T5 Ox (lthough this increse ws not significnt) (Tble 4.2, Figure 4.2). The occurrence of sotolon could be due to the condenstion rection between α ketobutyric cid nd cetldehyde followed by lctonistion (Tkhshi et l., 1976; Phm et l., 1995; Cutzch et l., 1998), especilly considering the production of cetldehyde in the Ox smples Sotolon (µg/l) Tretment Oxygen (mg/l) T1 Ox T1 Ox 6.59 e T2 Ox T2 Ox d T3 T3 Ox Tretment c T4 Ox T4 Ox b T5 Ox T5 Ox Figure 4.2 Sotolon concentrtion (µg/l) of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<.5. Other pthwys seem less likely due to the fct tht the wine ws fermented dry, leving miniml sugr content for the Millrd rection between cysteine nd vrious sugrs (Hofmnn & Schieberle, 147

162 1995, 1997) to occur, while the fct tht no scorbic cid ws detected in the smples ruled out the oxidtive degrdtion of scorbic cid s pthwy for sotolon formtion. The mximum level quntified in the smples ws in T5 Ox, reching.41 µg/l, which is below the detection threshold in both model wine s well s dry white wine. It is thus unlikely tht sotolon contributed to the romtic composition of the wine, however sensory interctions between vrious rom compounds cn occur nd the role of sotolon in these interctions (especilly in dry white wines) needs to be investigted. The formtion of this compound in white wines seems to be somewht dependnt on temperture nd concentrtion of round 8 µg/l were obtined during forced ged experiment conducted t 6 C (Silv Ferreir et l., 23b). In study where South Africn white wines were ssessed for sotolon levels, only those exposed to prolonged high tempertures hd sotolon levels higher thn the sensory threshold reported white wine (Lvigne et l., 28; Gbrielli et l., 213). It would seem s if the conditions of this study, especilly the temperture, were not high enough to produce substntil mounts of sotolon Acetls Dioxnes nd dioxolnes re formed by the condenstion rection between glycerol nd cetldehyde in n cid medium. During oxidtion there is significnt formtion of cetldehyde through the oxidtion of ethnol nd lrge mounts of the dioxnes nd dioxolnes re expected to be produced seeing tht sufficient quntity of regents re vilble. These compounds serve s oxidtion mrkers in ging Port wines nd could possibly lso serve s oxidtion mrkers for white wines with low SO 2 content. The odour is described s sweet nd old port like (Silv Ferreir et l., 22). In this study, dioxne formtion minly occurred during the lst three smpling stges of the Ox smples (Tble 4.2, Figure 4.21), incresing in significnt mounts. This corresponds to the increse in cetldehyde concentrtion which strted to increse in significnt mounts from the T2 Ox stge nd correltion coefficients of.9457 to.9519 were clculted between the different cetls nd cetldehyde. 148

163 12 1 AA Cis-dioxne (pek/is rtio) c c c c c c c c c b Tretment T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 T1 Ox Ox T2 T2 Ox Ox T3 T3 Ox Ox T4 T4 Ox Ox T5 T5 Ox Ox Oxygen (mg/l) Tretment 6.59 e d c b B Cis-dioxolne (pek/is rtio) c b c c c c c c c c Tretment T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 T1 Ox Ox T2 T2 Ox Ox T3 T3 Ox Ox T4 T4 Ox Ox T5 T5 Ox Ox Oxygen (mg/l) Tretment 6.59 e d c b C Trns-dioxne (pek/is rtio) cd cd d cd d d cd cd c b Tretment T Control T1 Control T2 Control T3 T3 Control T4 T4 Control Control T5 T5 Control Control T1 T1 Ox Ox T2 Ox T2 Ox T3 Ox T3 Ox T4 Ox T4 Ox T5 Ox T5 Ox Oxygen (mg/l) Tretment 6.59 e d c b

164 12 1 DD D Trns-dioxolne (pek/is rtio) c c c c c c c c c b Tretment Oxygen (mg/l) T T Control T1 Control T2 Control T3 Control T4 T4 Control T5 T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 Ox T3 Ox c T4 Ox T4 Ox b T5 Ox T5 Ox Figure 4.21 Cis dioxne (A), cis dioxolne (B), trns dioxne (C) nd trns dioxolne (D) pek/is rtios of the respective tretments. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p< Colour The opticl density t 42 nm (A 42 ) is commonly used in the wine industry to mesure yellow (or brown) colour development due to oxidtion of white wines (Singleton & Krmling, 1976; Zoecklein et l., 1995; Ilnd, 24). In this study, the mesurement t 44 nm ws lso included s this wvelength corresponded to the mximum bsorbnce of coloured pigments in model wine contining (+) ctechin (Brdshw et l., 21). The bsorbnce vlues t both wvelengths cn be seen in Tble 4.2. A very high correltion (r=.9973) ws found between the bsorbnces, thus only the results from A 42 will be discussed s mrker of oxidtion. In the Control smples, n increse in A 42 cn be seen from T Control which ws.53 bsorbnce units (AU) to T5 Control which ws mesured s.61 AU (Figure 4.22). A stedy increse ws lso recorded for the oxidised smples ending with n bsorbnce of.98 AU for T5 Ox, corresponding to 46% increse. High risk of dvnced oxidtion colour formtion in wines contining free SO 2 content less thn 1 mg/l hs been reported (Godden et l., 21). In the present study, significnt increse in yellow/brown colour ws mesured lredy t T2 Ox which corresponded with 1 mg/l free SO 2. 15

165 .11.1 b Absorbnce units t 42 nm g gf ef ef ef e e d c.5.4 Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 T5 Control Tretment T1 T1 Ox Ox 6.59 e T2 T2 Ox Ox d T3 T3 Ox Ox c T4 T4 Ox Ox b T5 T5 Ox Ox Figure 4.22 Absorbnce mesurements t 42 nm. Vlues re mens of triplicte nlysis; different letters indicte significnt differences t p<.5. Browning in white wines cn be due to three mechnisms. The first is the polymeriztion rections occurring during oxidtion of the polyphenols tht could led to the formtion of more intensely coloured products thus leding to the browning of the wine. The second mechnism is the oxidtion of trtric cid forming glyoxylic cid, which medites the condenstion of phenolic molecules which cn led to vrying degrees of polymeristion nd thus contribute to the development of yellow/brown colour in the wine. The production of cetldehyde cn lso enhnce oxidtive colourtion by inducing condenstion of phenolic compounds (Es Sfi et l., 1999; Lopez Toledno et l., 24; Mongs et l., 25). A sensory pnel lso evluted the wines ccording to colour. These results re discussed in section Combined chemicl content Figure 4.23 shows the PCA biplot constructed of ll the chemicl dt obtined throughout the study. The first two principle components ccount for 58% of the vrition with the most of the vrition being explined by PC1 (37%). Locted on the left of PC1 group of positively correlted compounds cn be seen which include the voltile thiols, SO 2, reduced GSH nd certin cette esters. Correlting negtively with this group re compounds locted to the right of PC1. This group 151

166 includes oxidised GSH, furfurl, 5 HMF, A 42, the cetls, cetldehyde, some ethyl esters, methionl nd sotolon. The ltter group re ll mostly unplesnt smelling oxidtion relted compounds. In the middle of PC1 group of compounds including vrious cids nd lcohols cn be seen nd is locted slightly towrd the left side of the PC. For this group of compounds, there seems to be greter vrition in the mesurements s replictions (dt points) re locted in wider rnge. Some points re locted to the high end of the compound concentrtion, while others re locted t the lower end. This vrition in repets ws not seen in the other two groups (locted on the left nd right of PC1). This could indicte the middle group not significntly contributing to the fresh nd fruity or oxidtive chrcteristics of the wines. PC 2(21%) 3MH 3MHA Totl SO2 Free SO2 Reduced GSH 4MMP Hexyl Acette Isomyl Acette 2-Phenylethyl cette Decnoic Acid Ethyl Butyrte Octnoic Acid Isovleric Acid Ethyl Acette Isobutnol Hexnoic Acid Butnol Isobutyric Acid Propnol Isomyl Alcohol Butyric Acid Acetic Acid Propionic Acid 2-Phenylethnol Ethyl Hexnote Hexnol PC 1(37%) Ethyl Lctte Sotolon Trns-Dioxolne Cis-Dioxolne Acetldehyde Methionl Cis-dioxne Trns-dioxne Oxidised GSH Furfurl 5-Hydroxymethylfurfurl AU t 44 nm AU t 42 nm Oxygen consumed Diethyl succinte T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox Control Ox Figure 4.23 PCA biplot of chemicl dt of vrious tretments. T Control is seprted from the rest of the tretments nd is ssocited with the highest concentrtion of fresh nd fruity rom compounds (locted on the left of PC1). The rest of the Control smples were locted towrds the middle of PC1, indicting decrese in the fruity rom 152

167 compounds nd ntioxidnts s time pssed (T T5). According to the chemicl dt, T1 Ox contined higher concentrtions of the fresh nd fruity compounds when compred to T4 Control nd T5 Control nd seems to correspond well with T2 nd T3 Control even though it received oxygen t the beginning of the tril. The rest of the tretments tht received oxygen (T2 T5 Ox) experienced decrese in concentrtion of fruity rom compounds nd ntioxidnts s time pssed with corresponding increse in oxidtion relted compounds. Most of the phenols, some ethyl esters nd the terpenes were not included in the PCA biplot s they did not contribute to the vrition between the tretments Sensory evlution Aromtic descriptive nlysis Pnel performnce ws nlysed using PnelCheck. Tucker plots of the sensory dt obtined from the three tests showed ll descriptors to be significnt (P<.1) (Figure 4.24). Figure 4.24 Tucker plot showing significnce of ll the ttributes generted by the pnel. 153

168 PCA ws pplied to the dt obtined from the descriptive nlysis performed on the wines (Figure 4.25). The first two principle components ccounted for 96.8% of the vrition with most of the vrition being explined by PC1 (9.5%). Guv Pssion Fruit Lemon Fresh Green Potto Bg T Control Green Apple Pinepple Grpefruit T1 Ox T5 Ox Sherry T4 Ox PC2(6.3) T4 Control T2 Control T3 T1 Control Control T5 Control T2 Ox T3 Ox Risins Cooked Green Bnn Honey Dried fruit Yellow Apple PC1(9.5) Syrup scores lodings Figure 4.25 PCA biplot of dt obtined from descriptive nlysis of wines oxidised to vrious degrees over time. Descriptors cn be divided into 3 min groups. The first group showed positive correltion between the ttributes grpefruit, pinepple, guv, pssion fruit, lemon nd fresh green. These ttributes re probbly result of vrious romtic compounds such s sulphur contining compounds like 3MH, 3MHA nd 4MMP (grpefruit, guv nd pssion fruit), esters (pinepple, lemon) nd methoxypyrzines (fresh green) nd re often used to profile young, fresh nd fruity Suvignon blnc wines (Lund et l., 29). The second group of ttributes (t the bottom of the plot) re cooked green, bnn, honey, dried fruit, yellow pple nd syrup. These ttributes correlte positively to ech other nd negtively to the first group of descriptors (risins hs weker correltion with the 154

169 rest of the ttributes). These ttributes re often used to describe wines grdully forming the typicl ged nd oxidised chrcter (especilly honey, dried fruit, pple, syrup nd risins). The third group of ttributes includes potto, green pple nd sherry. The ltter group of ttributes re mostly relted to dvnced wine oxidtion nd re often used when profiling Port or sherry wines where oxidtion roms re signture trit of tht style of wine (Mrglith, 1981; Henschke & Jirnek, 1993; Miyke & Shibmoto, 1993; Frivik & Ebeler, 23). The Control smples re grouped on the left of PC1. T Control is seprted from the rest of the group due to high intensities of the group 1 descriptors. T5 Control is lso seprted from T1 T4 Control smples due to higher intensities of group 2 descriptors. There re no cler seprtion between T1, T2, T3 nd T4 Control smples indicting little sensory difference between them. T Control hs high intensity of fruity roms while T1 T4 hs less of the group 1 ttributes nd developed more towrds the group 2 ttributes. T5 Control hs even less of the group 1 ttributes when compred to the rest of the Control group. There ws no development towrds the group 3 ttributes for the Control wines. The development over time is clerly defined for the oxidised wines on PC1. T1 Ox is ssocited with group 1 ttributes while T2 nd T3 Ox evolved towrds the group 2 ttributes, consequently losing the fresh chrcters of group 1 nd obtining ged chrcters such s honey, syrup nd risins. Only T3 Ox seems to strt showing some group 3 ttributes. T4 nd T5 Ox hve very high intensity of the group 3 ttributes nd none of the other descriptor groups. Interestingly, even though T1 Ox received significnt mount of oxygen, it ws perceived to be even more fresh nd fruity when compred to T1 T5 Control. This result is interesting s T1 Control nd T1 Ox both underwent the sme time lpse before smpling, however T1 Ox mintined more fresh nd fruity rom compred to T1 Control despite the fct tht it consumed oxygen t the beginning of the study. Considering the chemicl content of T1 Ox (Tble 4.2, Figure 4.23), it would hve been expected for T1 Ox to be positioned more towrds the group 2 ttributes which is not the cse. To the contrry, not only does it not ssocite strongly with T2 nd T3 Control (s suggested by the chemicl dt in Figure 4.23), but it is ctully perceived to be more fresh nd fruity correlting better with T Control. Another interesting observtion is tht T1 Ox contined significntly lower concentrtions of free nd totl SO 2 when compred to ll the Control smples (bout 1 mg/l less free SO 2 ; Tble 4.2) however, this is not reflected in the PCA biplot constructed from the chemicl dt (Figure 4.23) s overll concentrtions of other compounds drove the dt points of T1 Ox to the left of PC1. This could men tht the SO 2 rected with oxidtive species nd in this wy preserved the positive nd oxidtion sensitive rom compounds. The dissolved oxygen concentrtion in the T1 Ox smple ws 155

170 thus low enough not to deteriorte the rom profile but still high enough to ffect the ntioxidnt content (both SO 2 nd GSH). The presence of other romtic compounds, such s negtive sulphurcontining compounds (e.g. methyl mercptn) which my msk positive fruity relted ttributes could lso hve significnt impct on the overll romtic composition of the wine nd needs to be investigted further. These compounds could potentilly msk the fresh nd fruity ttributes. Figure 4.26 shows the men intensity scores for ech individul ttribute, from which the differences in tretments cn be seen in more detil. Attributes which reched the highest intensities re pinepple (mximum of 35.6 units for T Control), green pple (mximum of 49.1 units for T5 Ox) nd sherry (mximum of 48.9 units for T5 Ox). The overll rom of the Control smples seems to be dominted by ttributes such s guv, pinepple, grpefruit nd cooked green when considering ttribute intensities. 156

171 25 1 Fresh Green b b b b b c Lemon cb cb cb b cb cb cd 5 cd cd d 2 d d d 35 T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox 25 T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox Grpefruit b b b b b c Guv bc bc b b dc b d 1 5 d e e 5 e e e 25 T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox 5 T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox Pssion Fruit b b b b c c Pinepple cb c cb cb d b d e 5 d d d 1 f f T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox 12 3 Bnn bc b bc b bc c c Cooked Green d cb b cd cd cd e 2 5 T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox e 157

172 3 25 Yellow Apple e cb ce ce ce ed cd b ce cb Dried Fruit f bcd ed ec efd bcd ef bc b ec T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox 16 T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox 6 14 Honey ed bc ed ed e ed ec b bcd ed Syrup e ce cd ed ed ce ed c b b T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox Risins 15 1 b Potto Bg 6 5 c c c c c bc c 4 2 d dc bd d bd b d dc bc T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox b 5 b 4 4 Green Apple 3 2 c Sherry 3 2 c 1 de e de de e e e d 1 e e e e e e e d T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox T Control T1 Control T2 Control T3 Control T4 Control T5 Control T1 Ox T2 Ox T3 Ox T4 Ox T5 Ox Figure 4.26 Men vlues for ech sensory ttribute. Tests were done in triplicte; different letters indicte significnt differences t p<

173 Attributes tht scored very low in generl included lemon, bnn, honey nd potto bg. For these ttributes smples were never scored more thn 1 intensity units. When compred to the Ox smples, Control smples were scored higher in ttributes fresh green, lemon, pssion fruit, guv, pinepple nd grpefruit. However, in most cses this is not true for T1 Ox, which is mostly not significntly lower thn the Control smples. T1 Ox ws ctully rted significntly higher thn T1 T5 Control for the ttributes fresh green, pssion fruit nd grpefruit. As seen in the PCA biplot (Figure 4.25), there ws not much difference between the ttributes for T1 T5 Control. T5 Control did in some cses show significnt difference nd would explin the seprtion observed in the PCA biplot. The only ttribute which seems to show some significnt tendency or pttern in the Control smples is cooked green. Although this is not cler from the PCA biplot, the cooked green ttribute incresed in intensity from T Control to T5 Control with mostly significntly lower intensities in the Ox smples. The compound responsible for the perception of this ttribute is not exctly known, however the development of reductive odours which re often described to imprt cooked vegetbles, onions nd cbbge roms re cused by sulphur contining compounds such s H 2 S, mercptns, disulphides nd dimethyl sulphides which occur due to low oxygen content in wine (Ruhut, 1993; Brjkovich et l., 25). These compounds hve low perception thresholds nd re known to negtively influence young white wines, imprting unplesnt odours (Ruhut, 1993; Godden et l., 21). As the Control smples did not receive ny oxygen t the beginning of the study (mintining low oxygen content), it is very likely tht these compounds developed in the wine. Interestingly, even though T1 Ox seems to be better ssocited with the Control smples thn the Ox smples, it scored significntly lower in the cooked green ttribute when compred to T2 T5 Control. The prevention of mercptn formtion by the ddition of oxygen t the beginning of the study could explin this observtion, however such specultion certinly needs to be tested by pproprite experiments designed to investigte this hypothesis. Fresh green lso vried between the tretments. The persistence of IBMP concentrtions in the smples would indicte tht this ttribute would remin t higher intensities throughout the tretment, however this is not the cse. As mentioned previously, the green ttributes in wines hve sometimes shown wek correltion to the methoxypyrzine concentrtions, which could be the cse in this study. The suppressive effect of other compounds (such s the reductive odours or ldehydes in the oxidtive tretments) on the fresh green ttribute should lso not be underestimted. Bnn, honey, yellow pple nd dried fruit were used more often for the first oxidtion stges, while syrup, risins, potto bg, green pple nd sherry were used more often for the lst oxidtion stges. Attributes such s green pple nd sherry becme quite potent nd 159

174 overpowering t the end of the ging period s the fresh nd fruity ttributes (with the exception of yellow pple) were in ll cses scored below 5 intensity units for the lst two smpling periods Sensory nlysis ccording to wine colour In wines, phenolic compounds undergo non enzymtic oxidtion in the presence of oxygen, yielding polymerized polyphenols tht finlly precipitte in the form of brown pigments (Singleton, 1987). In this study the wines were subjected to visul nlysis nd rnked from lest oxidised to most oxidised ccording to their colour only (Figure 4.27). Similr to the PCA, cluster nlysis (CA) is used to clssify objects into groups chrcterized by the vlues of set of vribles (in this cse the observtion of the colour of the wine). CA is therefore n lterntive to PCA for exmining clssifictory structure of dt (Mssrt et l., 1997) nd llows obtining groups bsed on their similrities (Cmiñ et l., 28). Figure 4.28 shows the dendogrm obtined from the smples subjected to the evlution of their colour. The Wrd s method which clcultes the distnces between objects of cluster ws used s mlgmtion criterion, while Eucliden distnce ws used s ssocition criterion. Two min groups seprted the Control smples from the Ox smples. It is thus evident tht there were cler differences between these two groups which the pnellist could esily distinguish. On the x xis the smples re thus rrnged from most oxidised (left) to lest oxidised (right) ccording to their colour. The next grouping seprtes T1 nd T2 Ox from T3, T4 nd T5 Ox. T1 nd T2 Ox were thus clssified s being less oxidised when compred to the other three Ox smples. T1 Control seems to be somewht seprted from T, T2, T3, T4 nd T5 Control. It is difficult to explin this observtion. In order to see individul rnking ccording to the colour of the wine, the dt ws subjected to further nlysis. Figure 4.29 shows the rnking position of ech tretment. 16

175 Figure 4.27 Smple set rrnged from lest oxidised to most oxidised ccording to colour only Linkge Distnce T5 Ox T4 Ox T3 Ox T2 Ox T1 Ox T1 Control T5 Control T4 Control T3 Control T2 Control T Control Tretment Figure 4.28 Dendogrm of wine smples obtined by cluster nlysis. 161

176 12 Men rnking position ccording to severety of oxidtion (colour only) e c d d d d b b Tretment Oxygen (mg/l) T T Control T1 T1 Control T2 T2 Control T3 T3 Control T4 T4 Control T5 T5 Control Tretment T1 Ox Ox 6.59 e T2 Ox Ox d T3 Ox Ox c T4 Ox Ox b T5 T5 Ox Ox Figure 4.29 Averge rnking position of wines subjected to sensory evlution ccording to colour only. Pnellists were sked to rnk the wines from lest oxidised () to most oxidised (11). T Control ws rnked s the lest oxidised smple followed by T2, T3, T4 nd T5 Control which did not differ significntly from ech other. As mentioned previously, T1 Control ws rnked somewht more oxidised compred to the other Control smples. The men rnking of the wines incresed significntly when compring the Control smples to the Ox counterprts. This confirms the bility of the pnel to esily distinguish between these two groups. Of the Ox smples, T1 Ox nd T2 Ox were clssified s being lest oxidised while there were no significnt difference between T3, T4 nd T5 Ox. As seen previously, T1 Ox ws described s more fresh nd fruity thn most of the Control smples when considering the descriptive nlysis pplied to the smples. The visul nlysis did not coincide with the results from the olfctometric nlysis nd the T1 Ox smple ws situted to the oxidised side of the spectrum. Thus, ccording to the visul nlysis, the wine looked oxidised, however nlysing the wines romticlly proved otherwise. This is contrdictory to previous studies which report decrese in certin rom compounds during oxidtion first which is then followed by chnge in wine colour (Singleton & Krmling, 1976; Singleton et l., 1979; Escudero et l., 2b). In the current study, the development of n oxidised colour preceded the disppernce of the plesnt rom nd could indicte the sequence of these occurnces to not tke plce s described previously. The results from this study would suggest chnge of colour fter which decrese in plesnt rom occurs with n increse in oxidtion 162

177 rom fter tht. Differences in the chemicl composition between wines could lso perhps ccount for these discrepncies Chemicl nd sensory dt combined Combining the dt obtined from the chemicl nlysis with the dt from the sensory evlution could revel other findings not observed when looking t the dt sets individully. It is lso of interest to ssess whether there my be predictive bility of the chemicl dt for the degree of fresh nd fruity or oxidtive chrcter of the wine. A prtil lest squre regression (PLS) nlysis ws performed to explore the reltionships between the different concentrtions of chemicl compounds (plced in the X spce) nd the sensory descriptors (plced in the Y spce). The PLS constructed explined 55% of the vrince of X nd 7% of the vrince in Y for dimensions 1 nd 2 (Figure 4.3). There ws strong reltionship between mny of the chemicl mesurements nd ttributes. The ttributes follow similr pttern to the lodings when PCA ws performed using the sensory dt lone (Figure 4.25). Two groups pper on opposite sides of the PLS, showing strong negtive ssocition. On the left, ttributes syrup, risins, sherry, green pple nd potto bg correspond strongly nd on the right, ttributes pinepple, grpefruit, pssion fruit, guv, fresh green nd lemon correlted well. Descriptors dried fruit, yellow pple nd honey hd wek correltion to the oxidtion relted ttributes (on the left). Bnn correlted wekly with the fresh nd fruity ttributes locted to the right of the PLS. Cooked green ssocited strongly with the ltter group. Most of the scores were locted between the two ellipses explining between 5 nd 1% of the vrince. 163

178 DIMENSION 2 (X=18%, Y=9%) Honey Yellow Apple Gllic cid Dried fruit Syrup Furfurl Ethyl Decnote 5-Hydroxymethylfurfurl 44 nm Oxidised GSH 42 nm412 nm GRP Oxygen consumed Ethyl Octnote Risins Methionl Sherry Diethyl succinte Cis-dioxne Sotolon Green Apple Acetldehyde Cis-Dioxolne Potto Bg Trns-dioxne Trns-Dioxolne Ethyl Lctte Hexnol Benzldehyde Bnn Gerniol Linlool Frnesol (+)-Ctechin Methionol Pinepple Polymeric phenols Cftric cid 3MH Grpefruit Cffeic cid Coutric cid Cooked Green 4MMP 3MHA Pssion Fruit Guv p-coumric cid Totl SO2 Hexyl Acette Fresh Green Isomyl Acette Lemon Free SO2 Reduced GSH 2-Phenylethyl cette Ethyl Butyrte Decnoic Acid Isobutnol Methnol Propionic Acid Butnol Octnoic Acid Ethyl Hexnote Propnol Butyric cid Ethyl Acette 3-Ethoxy-1-propnol Isovleric cid Isobutyric Acid Acetic Acid Hexnoic Acid 2-Phenylethnol Isomyl Alcohol DIMENSION 1 (X=37%, Y=61%) Figure 4.3 PLS plot of component 1 nd component 2 contining chemicl informtion in the X spce nd the sensory descriptive dt in the Y spce. Three groupings of chemicl compounds cn be observed. On the left, compounds such s furfurl, 5 HMF, oxidised GSH, methionl, cetls, cetldehyde nd sotolon ssocited strongly with ech other nd the oxidtion relted sensory ttributes. On the right, correlting strongly with the fruity nd fresh ttributes, the compounds 3MH, 3MHA, 4MMP, hexyl cette, isomylcette, 2 phenylethyl cette nd reduced GSH re grouped. The voltile thiols nd few cette esters could be the fctors driving the fresh nd fruity flvours found in the wine. As the concentrtions of these compounds decresed, the ssocited ttributes lso decresed. As seen previously, compounds grouping t the bottom centre of the PLS consisted of vrious cids including hexnoic cid nd octnoic cid s well s vrious lcohols such s isomyl lcohol nd 2 phenylethnol. These compounds did not correlte to either of the sensory groups formed. However, the group in generl ws more closely locted to the fresh nd fruity side (right) of the PLS compred to the oxidtive side (left) nd could thus contribute to the overll fresh, fruity wine rom without contributing to specific rom ttributes identified in this study. 164

179 Interestingly, vrious polyphenolic compounds including trns cftric cid, p coutric cid nd (+) ctechin were locted in the centre of the PLS nd ws not significntly influenced by the tretment. This phenomenon ws lso seen in study done by Frcssetti et l. (213). The monoterpenes (linlool, frnesol nd gerniol) were lso locted inside the inner ellipses contributing to less thn 5% of the vrince. Tble 4.5 provides summry of the correltion coefficients, with consumed oxygen concentrtion, free nd totl SO 2, A 42 nd reduced GSH being correlted with the sensory scores for ll the rom ttributes. For consumed oxygen nd A 42, positive correltions were observed with the ttributes ssocited with ging nd oxidtion, while positive correltions were seen between the fresh nd fruity ttributes nd free nd totl SO 2 s well s reduced GSH. The strongest correltions were seen for A 42, consumed oxygen nd free SO 2 nd would indicte good bility to identify the development of oxidtion roms nd loss in fruity chrcters. Godden et l. (21) lso found good correltions between some of these prmeters nd sensory chrcteristics of white wines. Tble 4.5 Person correltion coefficients mong the sensory vribles with consumed oxygen, free nd totl SO 2, reduced GSH nd A 42. Consumed O 2 Free SO 2 Totl SO 2 Reduced GSH A 42 Fresh Green.7995 ***.8241 ***.8182 ***.8215 ***.8469 *** Lemon.8222 ***.899 ***.8555 ***.8255 ***.8618 *** Pssion Fruit.8349 ***.8211 ***.8286 ***.7922 ***.8743 *** Guv.8997 ***.8835 ***.8963 ***.7922 ***.9275 *** Pinepple.9182 ***.8422 ***.8657 ***.7613 ***.9396 *** Grpefruit.9376 ***.8443 ***.8883 ***.7241 ***.9476 *** Cooked Green.722 ***.5874 ***.78 ***.2722 ns.6738 *** Bnn.2923 ns.147 ns.893 ns.551 ns.245 ns Honey.3398 ns.5299 **.5128 **.3885 *.4193 * Dried Fruit.4362 *.691 ***.5599 ***.6153 ***.589 ** Yellow Apple.444 *.6319 ***.5898 ***.5336 ***.5144 ** Syrup.8 ***.8235 ***.835 ***.7291 ***.8428 *** Risins.956 ***.863 ***.8917 ***.733 ***.931 *** Potto Bg.8739 ***.6978 ***.7219 ***.5922 ***.843 *** Green Apple.9519 ***.7677 ***.8357 ***.6153 ***.931 *** Sherry.964 ***.84 ***.8648 ***.6541 ***.9485 *** ns = not significnt t P=.5; * = significnt t P <.5; ** = significnt t P<.1; *** = significnt t P <.1 165

180 The ddition of low mounts of oxygen to the wine could indeed improve the wine sensory qulity. An improved expression of wine fruity ttributes ws lso observed when Suvignon blnc wine ws exposed to low (< 3mg/L) or moderte (bout 6 mg/l) oxygen exposure in conjunction with decrese in reductive ttributes (Lopes et l., 29). In the current study, the occurrence of cooked green could be n indiction of the development of reductive odours in the Control smples. As seen in the sensory results, T1 Ox ws considered to be more fresh nd fruity when compred to T1 Control. As T1 Ox ssocited better with the Control smples when looking t the chemicl dt, the question remins how does this smple score fresher nd fruitier thn the Control smples s seen in the sensory evlution? The similr chemicl composition would suggest tht the wine be sensorilly more similr to the other Control smples, however the ddition of moderte mounts of oxygen to this wine could inhibit the formtion of reductive off odours. This lso rises the question of the bility of these reductive odours to msk nd suppress fresh nd fruity odours considering the similr chemicl composition otherwise. Follow up studies should investigte the formtion of the reductive odours nd wht specific oxygen concentrtions to dminister to benefit the wine qulity. Interctive effects between the reductive sulphur contining odours nd other fresh ssocited odours such s the voltile thiols nd the methoxypyrzines need to be investigted. The role of exct oxygen mesurements, free nd totl SO 2 s well s A 42 mesurements s predictive tool ws lso investigted. Importntly, the study hs llowed the compositionl mesurements of sulphur dioxide, degree of browning nd oxygen to be directly relted to the sensory properties of the Suvignon blnc wine nd gives resonble bility for predicting the degree of fruitiness or oxidtion during storge of white wines. The mesurement procedures for these prmeters re firly simple providing the necessry equipment is vilble. Furthermore, the strong correltions observed between especilly free SO 2 nd totl SO 2 (r=.9651) nd free SO 2 nd A 42 (r=.9234) indicte tht only the mesurement of free SO 2, reltively simple test tht is commonly performed by mny wineries, needs to be performed in commercil sitution. A certin cut off free SO 2 concentrtion for fruity style white wines hs been contemplted in previous study (Godden et l., 21). The current study confirms the detrimentl sensory effect s soon s wine rech levels below 1 mg/l free SO 2 t which point insufficient SO 2 is vilble to bind newly formed oxidised compounds, llowing the development of unplesnt wine flvour nd rom. This seems to be pplicble to certin wine style: fruity, no ok contct, reltively low ph, hndled reltively nerobiclly nd lso bottled with pproximtely 3 mg/l free SO 2. It would seem tht the periodic mesurement of SO 2 together with consumed oxygen nd A 42 could be used by wine compnies, distributors nd retilers to identify wines tht re pproching criticl level. It 166

181 is lso possible tht by incresing the concentrtion of SO 2 in wines t bottling, winemkers could compenste for the losses over time identified in this study. This could potentilly increse longevity of the bottled wines, however increses in SO 2 concentrtions in wines re highly undesirble s the US FDA estimtes tht s mny s 1% of the generl popultion show n incresed degree of sensitivity to sulphites (Ppzin, 1996). Considering the concentrtion of vrious fruit driven romtic compounds to be higher thn the perception threshold (Tble 4.3) even in the lter oxidtive stges it is evident tht some romtic compounds could hve powerful odourous influences cusing complete msking or suppressing effects leding to significnt decrese in the perception of the fruity ttributes. So, not only is the loss in fruity chrcters likely to be cused by oxidtive dmge to flvour compounds, but lso by interctive effects of especilly ldehydes being formed. On the other hnd, some of the unplesnt oxidtion relted compounds tht developed remined below the perception threshold of tht compound (such s sotolon). The effect of these compounds even when present t very low concentrtions should lso be investigted s enhncing effects could tke plce contributing to the overll oxidtion flvour. In this study, the first dministrtion of oxygen resulted in 6.59 mg/l dissolved oxygen concentrtion which ws ll consumed before the second oxygen dose ws dministered. This concentrtion in white wines terms would be considered to be quite high. Even being t the higher rnge of dissolved oxygen for white wines, the tretment still scored higher in fresh nd fruity ttributes when compred to smples where no oxygen ws dded. It would however be interesting to investigte the effect of smller dosges of oxygen. This would give better ide of t wht exct concentrtion the intensity of the fruity ttributes will rech its pek. The proposed study would probbly result in the wines being very similr in sensoril perceptions, however the chemicl content should gin be monitored to identify the oxygen dose where the best sensory perceptions re mtched with the idel chemicl sitution. After the second oxygen dose (T2 Ox; mg/l consumed oxygen) the fresh nd fruity ttribute intensities dropped significntly, however the wines will probbly only be clssified s n extreme cse of oxidtion between T3 nd T4 Ox (16.99 nd mg/l respectively). This suggests tht even though the wine qulity decresed, the wine ws ble to consume more oxygen thn expected without developing the overpowering oxidtion chrcter. Only in very few studies hs the ctul degree of oxygen exposure been precisely determined or t lest estimted (Godden et l., 21; Brjkovich et l., 25; Lopes et l., 29; Cillé et l., 21; Nygrd et l., 21), however in some of these studies the oxygen mesurements fluctuted cusing substntil vritions between repets. 167

182 Considering the mount of dys ged, the Control smples seem to hve developed the reductive odour within 2 months of ging. There ws no significnt difference between the sensory profiles of the T1 T4 Control smples which would indicte the ging of 64 dys vs. 24 dys hd little effect, however the role of the reductive odours msking nd suppressing bilities could ply n importnt role nd should be investigted before mking ny further conclusions regrding the ging period of the Control smples. Another topic worth investigting would be the effect of further ging on the T1 Ox smple. The ging process ws effectively stopped by freezing the smple fter 64 dys of ging. However, would the sensory perception of the wine remin s desirble for longer ging period nd if so, for wht length of time? In other words, the ging cpcity of this type of wine (where oxygen exposure took plce) should be investigted. 4.4 CONCLUSION The results in this study gree with results reported in other studies. Herbst et l. (211) lso reported the decrese in the voltile thiols during ging s well s the initil increse in 3MH concentrtions, while Silv Ferreir et l. (22) reported the increse in oxidtion compounds in white wines. In the ltter publiction, the results reported gree with the present study in tht methionl ws formed during the ging period, however substntil phenylcetldehyde nd sotolon formtion ws lso reported by Silv Ferreir et l. (22) which ws not observed in the present study. Aging conditions such s higher ging temperture could ccount for this observtion. Results reported by Godden et l. (21) coincide with the present study, especilly considering the SO 2 levels nd formtion of yellow/brown colour. The sensory dt reported in the present study is supported by previous publictions (Godden et l., 21; Lopes et l., 29), however this is the first study where comprehensive overview of both chemicl (romtic nd non romtic) s well s sensory dt hs been obtined during repetitive oxidtion of the wines. This provides insight into the dvntges s well s the disdvntges of vrious levels of oxygen exposure s well s the cpcity of wine to consume oxygen. It is still rther difficult to determine the idel mount of oxygen tht cn be beneficil for wine during bottle ging, especilly considering the wide rnge of vrieties nd wine styles. Review rticles hve ttempted to shed some light on the overll effect of oxygen on the composition of wines during ging (Coetzee & Du Toit, 212; Uglino, 213), however the chemicl mechnisms involved in the formtion of mny key ging compounds remin to be estblished nd further reserch is needed to understnd the vrious mechnisms. The results in this study will dd to the dt bse of wine reserch regrding oxygen nd could serve s n indiction to the mount of 168

183 oxygen tht could be beneficil for especilly fruit driven Suvignon blnc wines. It cn lso give guidnce s to whether enough oxygen exposure hs tken plce to dversely ffect the wine t lter stge of storge. This informtion would be most vluble, not only for wineries but lso for pckging developers. Given the diversity tht occur between individul wines, the gret chllenge tht remins would be the bility to predict the tendency of wine to lose desirble rom compounds nd develop oxidtive odours nd, in generl, to define how much oxygen is needed to chieve the perfect blnce in order to obtin wine t its best potentil nd most blnced in terms of romtic composition. The impct of oxygen on wine is context specific due to fctors such s must nd wine composition, dditives nd timing of oxygen exposure which cn ll influence the outcome. The ddition of oxygen to white wine in commercil wineries is often kept to minimum, which should be re ssessed. Little dt exists regrding n oxygen specifiction level tht white wines require. However, such specifictions would probbly differ from wine to wine. Wine producers should thus conduct their own trils to confirm the influence of oxygen on their wines whenever possible. Oxygen mngement in wine is generting considerble interest especilly regrding topics such s methods to improve the wine qulity, improving the winemker s control nd the formtion of credible solutions on how to mnge the oxygen during winemking (Nygrd, 21). Through integrtion of flvour chemistry nd sensory reserch such s the current study, wine producers re given guides tht enble them to tilor mke wines ccording to the trget mrket. A better understnding of potentil sensoril improvements bsed on oxygen mngement could led to the development nd implementtion of novel winemking prctices. 4.5 ABBREVIATIONS USED 3 mercptohexn 1 ol (3MH); 3 mercptohexn 1 ol cette (3MHA); 4 mercpto 4 methylpentn 2 one (4MMP); 3 isobutyl 2 methoxypyrzine (IBMP); 3 isopropyl 2 methoxypyrzine (IPMP); 3 secbutyl 2 methoxypyrzine (SBMP); light emitting diode (LED); p hydroxymercurybenzote (p HMB); butylted hydroxynisole (BHA); selective ion monitoring mode (SIM); limit of quntifiction (LOQ); limit of detection (LOD); gs chromtogrphy flme ioniztion detector (GC FID); gs chromtogrph (GC); solid phse extrction (SPE); gs chromtogrphy olfctometry (GC O); ultr performnce liquid chromtogrphy tndem mss spectrometric (UPLC MS/MS); mss spectrometer (MS); glutthione (GSH); oxidised glutthione (GSSG); grpe rection product (GRP); multiple rection monitoring mode (MRM); high performnce liquid chromtogrphy (HPLC); reverse phse high performnce liquid chromtogrphy (RP HPLC); ldehyde dehydrogense (Al DH); nicotinmide 169

184 denine dinucleotide (NAD); ultrviolet (UV); nlysis of vrince (ANOVA); prtil lest squre regression (PLS); cluster nlysis (CA); principle component nlysis (PCA). 4.6 REFERENCES Allen, M.S. & Lcey, M.J., Methoxypyrzines of Grpes nd Wines. In Chemistry of Wine Flvor. pp Americn Chemicl Society. Amerine, M.A. & Ough, C.S., Methods for the nlysis of musts nd wine. Wiley Interscience Publiction, New York. Bkker, J., Picinelli, A. & Bridle, P., Model wine solutions: colour nd composition chnges during ging. Vitis. 32, Benkwitz, F., Nicolu, L., Lund, C., Beresford, M., Wohlers, M. & Kilmrtin, P.A., 212. Evlution of key odornts in Suvignon blnc wines using three different methodologies. Journl of Agriculturl nd Food Chemistry. 6, Benkwitz, F., Toming, T., Kilmrtin, P.A., Lund, C., Wohler, M. & Nicolu, L., 212b. Identifying the chemicl composition relted to the distinct rom chrcteristics of New Zelnd Suvignon blnc wines. Americn Journl of Enology nd Viticulture. 63, Blke, A., Kotserides, Y., Brindle, I.D., Inglis, D. & Pickering, G.J., 21. Effect of light nd temperture on 3 lkyl 2 methoxypyrzine concentrtion nd other impct odournts of Riesling nd Cbernet frnc wines during bottle geing. Food Chemistry. 119, Blnchrd, L., Drriet, P. & Dubourdieu, D., 24. Rectivity of 3 mercptohexnol in red wine: Impct of oxygen, phenolic frctions, nd sulfur dioxide. Americn Journl of Enology nd Viticulture. 55, Brdshw, M.P., Scollry, G.R. & Prenzler, P.D., 21. Ascorbic cid induced browning of (+) ctechin in model wine system. Journl of Agriculturl nd Food Chemistry. 49, Brjkovich, M., Tibbits, N., Peron, G., Lund, C.M., Dykes, S.I., Kilmrtin, P.A. & Nicolu, L., 25. Effect of screwcp nd cork closures on SO 2 levels nd roms in Suvignon blnc wine. Journl of Agriculturl nd Food Chemistry. 53, Buttery, R.G., Seifert, R.M., Gudgni, D.G. & Ling, L.C., Chrcteriztion of some voltile constituents of bell peppers. Journl of Agriculturl nd Food Chemistry. 17, Cillé, S., Smson, A., Wirth, J., Dievl, J. B., Vidl, S. & Cheynier, V., 21. Sensory chrcteristics chnges of red Grenche wines submitted to different oxygen exposures pre nd post bottling. Anlytic Chimic Act. 66,

185 Câmr, J.S., Alves, M.A. & Mrques, J.C., 26. Chnges in voltile composition of Mdeir wines during their oxidtive ging. Anlytic Chimic Act. 563, Câmr, J.S., Mrques, J.C., Alves, M.A. & Silv Ferreir, A.C., Hydroxy 4,5 dimethyl 2(5H) furnone levels in fortified Mdeir wines: Reltionship to sugr content. Journl of Agriculturl nd Food Chemistry. 52, Cmiñ, J.M., Cntrelli, M.A., Lozno, V.A., Boeris, M.S., Irimi, M.E., Gil, R.A. & Mrchevsky, E.J., 28. Chemometric tools for the chrcteristion of honey produced in L Pmp, Argentin, from their elementl content, using inductively coupled plsm opticl emission spectrometry (ICP OES). Journl of Agriculturl Reserch. 47, Cpone, D.L., Sefton, M.A., Hysk, Y. & Jeffery, D.W., 21. Anlysis of precursors to wine odornt 3 mercptohexn 1 ol using HPLC MS/MS: Resolution nd quntittion of distereomers of 3 S cysteinylhexn 1 ol nd 3 S glutthionylhexn 1 ol. Journl of Agriculturl nd Food Chemistry. 58, Cstellri, M., Simonto, B., Tornielli, G.B., Spinelli, P. & Ferrrini, R., 24. Effects of different enologicl tretments on dissolved oxygen in wines. Itlin Journl of Food Science. 16, Cejudo Bstnte, M.J., Hermosín Gutiérrez, I. & Pérez Coello, M.S., 213. Accelerted ging ginst conventionl storge: Effects on the voltile composition of Chrdonny white wines. Food Chemistry. 78, Cejudo Bstnte, M.J., Pérez Coello, M.S. & Hermosín Gutiérrez, I., 21. Identifiction of new derivtives of 2 S glutthionylcftric cid in ged white wines by HPLC DAD ESI MS. Journl of Agriculturl nd Food Chemistry. 58, Coetzee, C. & Du Toit, W.J., 212. A comprehensive review on Suvignon blnc rom with focus on certin positive voltile thiols. Food Reserch Interntionl. 45, Coetzee, C., Lisjk, K., Nicolu, L., Kilmrtin, P.A. & Du Toit, W.J., 213. Oxygen nd sulfur dioxide dditions to Suvignon blnc must: effect on must nd wine composition. Flvour nd Frgrnce Journl. 28, Culleré, L., Ccho, J. & Ferreir, V., 27. An ssessment of the role plyed by some oxidtion relted ldehydes in wine rom. Journl of Agriculturl nd Food Chemistry. 55, Cutzch, I., Chtonnet, P. & Dubourdieu, D., Rôle du sotolon dns l rôme des vins doux nturels, influence des conditions d élevge et de vieillissement. Journl Interntionl des Sciences de l Vigne et du Vin. 32, Dnilewicz, J.C., 21. Further studies on the mechnism of interction of polyphenols, oxygen md sulfite in wine. Americn Journl of Enology nd Viticulture. 61,

186 Dnilewicz, J.C., 212. Review of oxidtive processes in wine nd vlue of reduction potentils in enology. Americn Journl of Enology nd Viticulture. 63, 1 1. Drriet, P., Toming, T., Lvigne, V., Boidron, J. & Dubourdieu, D., Identifiction of powerful romtic compound of Vitis vinifer L. vr. Suvignon wines: 4 Mercpto 4 methylpentn 2 one. Flvour nd Frgrnce Journl. 1, De Villiers, A., Lynen, F., Crouch, A. & Sndr, P., 23. A robust cpillry electrophoresis method for the determintion of orgnic cids is wines. Europen Food Reserch nd Technology. 217, Dimkou, E., Uglino, M., Diévl, J. B., Vidl, S. & Jung, R., 213. Impct of dissolved oxygen t bottling on sulfur dioxide nd sensory properties of Riesling wine. Americn Journl of Enology nd Viticulture. 64, Dimkou, E., Uglino, M., Dievl, J.B., Vidl, S., Agrd, O., Ruhut, D. & Jung, R., 211. Impct of hedspce oxygen nd closure on sulphur dioxide, color, nd hydrogen sulfide levels in Riesling wine. Americn Journl of Enology nd Viticulture. 62, Du Toit, W.J., 26. The effect of oxygen on the composition nd microbiology of red wine. Thesis, University of Stellenbosch, Stellenbosch. Es Sfi, N., Le Guernevé, C., Lbrbe, B., Fulcrnd, H., Cheynier, V. & Moutounet, M., Structure of new xnthylium slt derivtive. Tetrhedron Letters. 4, Esclon, H., Birkmyre, L., Piggott, J.R. & Pterson, A., 22. Effect of mturtion in smll ok csks on the voltility of red wine rom compounds. Anlytic Chimic Act. 458, Escudero, A., Asencio, E., Ccho, J. & Ferreir, V., 22. Sensory nd chemicl chnges of young white wines stored under oxygen. An ssessment of the role plyed by ldehydes nd some other importnt odornts. Food Chemistry. 77, Escudero, A., Ccho, J. & Ferreir, V., 2. Isoltion nd identifiction of odournts generted in wine during its oxidtion: gs chromtogrphy olfctometry study. Europen Food Reserch nd Technology. 211, Escudero, A., Gororz, B., Melus, M.A., Ortin, N., Ccho, J. & Ferreir, V., 24. Chrcteriztion of the rom of wine from Mccbeo. Key role plyed by compounds with low odor ctivity vlues. Journl of Agriculturl nd Food Chemistry. 52, Escudero, A., Hernndez Orte, P., Ccho, J. & Ferreir, V., 2b. Clues bout the role of methionl s chrcter impct odornt of some oxidized wines. Journl of Agriculturl nd Food Chemistry. 48,

187 Ferreir, V., Escudero, A., Cmpo, E. & Ccho, J., 28. Understnding the role plyed by the odornts nd their interctions on wine flvour. Chssgne, S. (ed). Proc. Wine Active Compounds, Mrch 28, Beune, Frnce. Ferreir, V., Escudero, A., Fernnádez, P.E. & Ccho, J., Chnges in the profile of voltile compounds in wines stored under oxygen nd their reltionship with the browning process. Zeitschrift für Lebensmittel Untersuchung und Forschung A. 25, Ferreir, V., Lopéz, R. & Ccho, J.F., 2. Quntittive determintion of the odornts of young red wines from different grpe vrieties. Journl of the Science of Food nd Agriculture. 8, Ferreir, V., Ortin, N., Escudero, A., Lopez, R. & Ccho, J., 22. Chemicl chrcteriztion of the rom of Grenche rosé wines: Arom extrct dilution nlysis, quntittive determintion, nd sensory reconstitution studies. Journl of Agriculturl nd Food Chemistry. 5, Frcssetti, D., Coetzee, C., Vnzo, A., Bllbio, D. & du Toit, W.J., 213. Oxygen consumption in South Africn Suvignon blnc wines: role of glutthione, sulphur dioxide nd certin phenolics. South Africn Journl of Enology nd Viticulture. Frncis, I.L. & Newton, J.L., 25. Determining wine rom from compositionl dt. Austrlin Journl of Grpe nd Wine Reserch. 11, Frivik, S.K. & Ebeler, S.E., 23. Influence of sulfur dioxide on the formtion of ldehydes in white wine. Americn Journl of Enology nd Viticulture. 54, Gbrielli, M., Buic, A., Frcssetti, D., Stnder, M., Tirelli, A. & Du Toit, W.J., 213. Determintion of sotolon content in South Africn white wines by two novel HPLC UV nd UPLC MS methods. Unpublished dt. Ghidossi, R., Poupot, C., Thibon, C., Pons, A., Drriet, L., Riquier, G., De Revel, G. & Mietton Peuchot, M., 212. The influence of pckging on wine conservtion. Food Control. 23, Godden, P., Leigh, F., Field, J., Gishen, M., Coulter, A., Vlente, P., HoJ, P. & Robinson, E., 21. Wine bottle closures: physicl chrcteristics nd effect on composition nd sensory properties of Semillon wine. 1. Performnce up to 2 months post bottling. Austrlin Journl of Grpe nd Wine Reserch. 7, Gower, J.C., Lubbe, S. & Le Roux, N.J., 211. Understnding Biplots. Wiley, Chichester. Gunt, Y.Z., Byonove, C.L., Bumes, R.L. & Cordonnier, R.E., Stbility of free nd bound frctions of some rom components of grpes cv. Musct during the wine processing: preliminry results. Americn Journl of Enology nd Viticulture. 37,

188 Guth, H., Quntittion nd sensory studies of chrcter impct odornts of different white wine vrieties. Journl of Agriculturl nd Food Chemistry. 45, Hrtmnn, P.J., McNir, H.M. & Zoecklein, B.W., 22. Mesurement of 3 lkyl 2 methoxypyrzine by hedspce solid phse microextrction in spiked model wines. Americn Journl of Enology nd Viticulture. 53, He, J., Zhou, Q., Peck, J., Soles, R. & Qin, M.C., 213. The effect of wine closures on voltile sulfur nd other compounds during post bottle geing. Flvour nd Frgrnce Journl. 28, Hebditch, K.R., Nicolu, L. & Brimble, M.A., 27. Synthesis of isotopiclly lbelled thiol voltiles nd cysteine conjugtes for quntifiction of Suvignon blnc wine. Journl of Lbelled Compounds nd Rdiophrmceuticls. 5, Henschke, P.A. & Jirnek, V., Yests: Growth during fermenttion. In Wine Microbiology nd biotechnology. Fleet, G.H. (ed.), pp Hrdwood cdemic publishers, Chur, Switzerlnd. Herbst Johnstone, M., Nicolu, L. & Kilmrtin, P.A., 211. Stbility of vrietl thiols in commercil Suvignon blnc wines. Americn Journl of Enology nd Viticulture. 62, Herbst, M., Kilmrtin, P.A. & Nicolu, L., 28. Arom stbility in Suvignon blnc wines. The Austrlin nd New Zelnd Grpegrower nd Winemker. 6, Hernnz, D., Gllo, V., Recmles, Á.F., Meléndez Mrtínez, M.L., González Miret, M.L. & Heredi, F.J., 29. Effect of storge on the ohenolic content, voltile composition nd clolour of white wines from the vrieties Zlem nd Colombrd. Food Chemistry. 113, Hofmnn, T. & Schieberle, P., Evlution of the key odornts in thermlly treted solution of ribose nd cysteine by rom extrct dilution techniques. Journl of Agriculturl nd Food Chemistry. 43, Hofmnn, T. & Schieberle, P., Identifiction of potent rom compounds in thermlly treted mixtures of glucose/cysteine nd rhmnose/cysteine using rom extrct dilution technique. Journl of Agriculturl nd Food Chemistry. 45, Ilnd, P., 24. Chemicl nlysis of grpes nd wine: Techniques nd concepts. Ptrick Ilnd Wine promotions, Cmpbelltown, SA, Austrli. Ilnd, P., Bruer, N., Edwrds, G., Weeks, S. & Wolkes, E., 24. Chemicl nlysis of grpes nd wine: Techniques nd concepts. Winetitles, Adelide. Jckowetz, J.N. & De Orduñ, M., 213. Survey of SO 2 binding crbonyls in 237 red nd white tble wines. Food Control. 32,

189 Jnes, L., Lisjk, K. & Vnzo, A., 21. Determintion of glutthione content in grpe juice nd wine by highperformnce liquid chromtogrphy with fluorescence detection. Anlytic Chimic Act. 674, Jrut, I., Ccho, J. & Ferreir, V., 25. Concurrent phenomen contributing to the formtion of the rom of wine during ging in ok wood: An nlyticl study. Journl of Agriculturl nd Food Chemistry. 53, Jocelyn, P.C., Biochemistry of the SH group; the occurrence, chemicl properties, metbolism nd biologicl function of thiols nd disulphides. Acdemic Press, London, UK. King, E.S., Osidcz, P., Curtin, C., Bstin, S.E.P. & Frncis, I.L., 211. Assessing desirble levels of sensory properties in Suvignon Blnc wines consumer preferences nd contribution of key rom compounds. Austrlin Journl of Grpe nd Wine Reserch. 17, Kotserides, Y., Beloqui, A., Byonove, C.L., Bumes, R.L. & Bertrnd, A., Effects of selected viticulturl nd enologicl fctors on levels of 2 methoxy 3 isobutylpyrzine in wines. Journl Interntionl des Sciences de l Vigne et du Vin. 33, Kreitmn, G.Y., Lurie, F. & Elis, R., 213. Investigtion of ethyl rdicl quenching by phenolics nd thiols in model wine. Americn Journl of Enology nd Viticulture. 61, Kritzinger, E.C., Stnder, M. & Du Toit, W.J., 213. Assessment of glutthione levels in model solution nd grpe ferments supplemented with glutthione enriched inctive dry yest preprtions using novel UPLC MS/MS method. Food Additives & Contminnts: Prt A. 3, Kwitkowski, M., Skouroumounis, G.K., Lttey, K.A. & Wters, E.J., 27. The impct of closures, including screw cp with three different hedspce volumes, on the composition, colour nd sensory properties of Cbernet Suvignon wine during two yers' storge. Austrlin Journl of Grpe nd Wine Reserch. 13, Lcey, M.J., Allen, M.S., Hrris, R.L.N. & Brown, W.V., Methoxypyrzines in Suvignon blnc grpes nd wines. Americn Journl of Enology nd Viticulture. 42, Lmbropoulos, I. & Roussis, I., 27. Inhibition of decrese of voltile esters nd terpenes during storge of white wine nd model white wine medium by cffeic cid nd gllic cid. Food Reserch Interntionl. 4, Lvigne, V., Pons, A., Drriet, P. & Duboudieu, D., 28. Chnges in sotolon content of dry white wines during brrel nd bottle ging. Journl of Agriculturl nd Food Chemistry. 56, Lvigne, V., Pons, A. & Dubourdieu, D., 27. Assy of glutthione in must nd wines using cpillry electrophoresis nd lser induced fluorescence detection chnges in concentrtion in white 175

190 wines during lcoholic fermenttion nd ging. Journl of Chromtogrphy A. 1139, Lee, D. H., Kng, B. S. & Prk, H. J., 211. Effect of oxygen on voltile nd sensory chrcteristics of Cbernet Suvignon during secondry shelf life. Journl of Agriculturl nd Food Chemistry. 59, Lisjk, K., 27. The role of oxygen in new vinifiction technologies of white nd red wines. Thesis, University of Ljubljn, Ljubljn. Liu, S. Q. & Pilone, G.J., 2. An overview of formtion nd roles of cetldehyde in winemking with emphsis on microbil implictions. Interntionl Journl of Food Science nd Technology. 35, Lopes, P., Sucier, C., Teissedre, P.L. & Glories, Y., 26. Impct of storge position on oxygen ingress through different closures into wine bottles. Journl of Agriculturl nd Food Chemistry. 54, Lopes, P., Silv, M.A., Pons, A., Toming, T., Lvigne, V., Sucier, C., Drriet, P., Teissedre, P.L. & Dubourdieu, D., 29. Impct of oxygen dissolved t bottling nd trnsmitted through closures on the composition nd sensory properties of Suvignon blnc wine during bottle storge. Journl of Agriculturl nd Food Chemistry. 57, Lopez Toledno, A., Villno Vlenci, D., Myen, M., Merid, J. & Medin, M., 24. Interction of yest with the products resulting from the condenstion rection between (+) ctechin nd cetldehyde. Journl of Agriculturl nd Food Chemistry. 52, Loscos, N., Hernndez Orte, P., Ccho, J. & Ferreir, V., 21. Evolution of the rom composition of wines supplemented with grpe flvour precursors from different vrietls during ccelerted wine ging. Food Chemistry. 12, Louw, L., 27. Chemicl chrcteriztion of South Africn young wines. Thesis, University of Stellenbosch. Louw, L., Roux, K., Tredoux, A., Tomic, O., Næs, T., Nieuwoudt, H.H. & Vn Rensburg, P., 29. Chrcteriztion of selected South Africn young cultivr wines using FTMIR Spectroscopy, Gs Chromtogrphy, nd Multivrite Dt Anlysis. Journl of Agriculturl nd Food Chemistry. 57, Louw, L., Tredoux, A.G.J., Vn Rensburg, P., Kidd, M., Næs, T. & Nieuwoudt, H.H., 21. Fermenttion derived rom compounds in vrietl young wines from South Afric. South Africn Journl of Enology nd Viticulture. 31, Loyux, D., Roger, S. & Add, J., The evolution of Chmpgne voltiles during geing. Journl of the Science of Food nd Agriculture. 32,

191 Lund, C.M., Thompson, M.K., Benkwitz, F., Wohler, M.W., Triggs, C.M., Grdner, R., Heymnn, H. & Nicolu, L., 29. New Zelnd Suvignon blnc distinct flvor chrcteristics: Sensory, chemicl, nd consumer spects. Americn Journl of Enology nd Viticulture. 6, Mkhotkin, O., Pineu, B. & Kilmrtin, P.A., 212. Effect of storge temperture on the chemicl composition nd sensory profile of Suvignon blnc wines. Austrlin Journl of Grpe nd Wine Reserch. 18, Mkhotkin, O. & Kilmrtin, P.A., 212. Hydrolysis nd formtion of voltile esters in New Zelnd Suvignon blnc wine. Food Chemistry. 135, Mlherbe, S., 211. Investigtion of the impct of commercil mlolctic fermenttion strter cultures on red wine rom compounds, sensory properties nd consumer preference. Thesis, University of Stellenbosch, Stellenbosch. Mris, J., Terpenes in the rom of grpes nd wines: A review. South Africn Journl of Enology nd Viticulture. 4, Mris, J., Suvignon blnc cultivr rom review. South Africn Journl of Enology nd Viticulture. 15, Mris, J., Effect of grpe temperture, oxidtion nd skin contct on Suvignon blnc juice nd wine composition nd wine qulity. South Africn Journl of Enology nd Viticulture. 19, Mris, J. & Pool, H.J., 198. Effect of storge time nd temperture on the voltile composition nd qulity of dry white tble wines. Vitis. 19, Mrchnd, S., De Revel, G. & Bertrnd, A., 2. Approches to wine rom: Relese of rom compounds from rections between cysteine nd crbonyl compounds in wine. Journl of Agriculturl nd Food Chemistry. 48, Mrglith, P.Z., Flvour Microbiology. Chrles C. Thoms Publishers, Springfield, Illinois. Mssrt, D.L., Vndeginste, B.G.M., Buydens, L.M.C., De Jong, S., Lewi, P.J. & Smeyers Verbeke, J., Hndbook of chemometrics nd qulimetrics. Elsevier, Amsterdm. Mury, C., Clrk, A.C. & Scollry, G.R., 21. Determintion of the impct of bottle colour nd phenolic concentrtion on pigment development in white wine stored under externl consitions. Anlytic Chimic Act. 66, Meilgrd, M.C., Flvour chemistry of beer. Prt II: Flvour nd threshold of 239 rom voltiles. Mster Brewers Assocition of the Americs, Technicl Qurterly. 12, Miyke, T. & Shibmoto, T., Quntittive nlysis of cetldehyde in foods nd beverges. Journl of Agriculturl nd Food Chemistry. 41,

192 Moio, L., Uglino, M., Genovese, A., Gmbuti, A., Pessin, R. & Piombino, P., 24. Effect of ntioxidnt protection of must on voltile compounds nd rom shelf life of Flnghin (Vitis vinifer L.) wine. Journl of Agriculturl nd Food Chemistry. 52, Mongs, M., Brtolomé, B. & Gómez Cordovés, C., 25. Updted knowledge bout the presence of phenolic compounds in wine. Criticl Reviews in Food Science nd Nutrition. 45, Moyno, L., Ze, L., Moreno, J. & Medin, M., 22. Anlyticl study of romtic series in sherry wines subjected to biologicl ging. Journl of Agriculturl nd Food Chemistry. 5, Murt, M.L., 25. Recent findings on rosé roms. Prt 1: Identifying roms studying the romtic potentil of grpes nd juice. Austrlin nd New Zelnd grpegrower nd winemker. 497, Nikolntonki, M., Chichuc, I., Teissedre., P.L. & Drriet, P., 21. Rectivity of voltile thiols with polyphenols in wine model medium: Impct of oxygen, iron, nd sulfur dioxide. Anlytic Chimic Act. 66, Nikolntonki, M., Jourdes, M., Shinod, K., Teissedre, P.L., Quideu, S. & Drriet, P., 212. Identifiction of dducts between n odoriferous voltile thiol nd oxidized grpe phenolic compounds: Kinetic study of dduct formtion under chemicl nd enzymtic oxidtion conditions. Journl of Agriculturl nd Food Chemistry. 6, Nikolntonki, M. & Wterhouse, A.L., 212. A method to quntify quinone rection rtes with wine relevnt nucleophiles: A key to the understnding of oxidtive loss of vrietl thiols. Journl of Agriculturl nd Food Chemistry. 6, Nygrd, M., 21. Oxygen mngement from grpes to glss. Wine Industry Journl. 25, Nygrd, M., Osidcz, P., Roget, W., Frncis, L., Vidl, S. & Agrd, O., 21. The effect of closure choice on consumer preference rting of wines. Austrlin nd New Zelnd grpegrower nd winemker. 563, Nykänen, L., Formtion nd occurrence of flvor compounds in wine nd distilled lcoholic beverges. Americn Journl of Enology nd Viticulture. 37, Nykänen, L. & Suomlinen, H., Arom of beer, wine nd distilled lcoholic beverges. In Hndbook of Arom Reserch. pp 29. Springer, Reidel, Dordrecht. O'Brien, V., Colby, C. & Nygrd, M., 29. Mnging oxygen ingress t bottling. Wine Industry Journl. 24, O'Brien, V., Frncis, L. & Osidcz, P., 29b. Pckging choices ffect consumer enjoyment of wines. Wine Industry Journl. 24,

193 Oliveir, J.M., Fri, M., Sá, F., Brros, F. & Arújo, I.M., 26. C 6 lcohols s vrietl mrkers for ssessment of wine origin. Anlytic Chimic Act. 563, Ough, C.S. & Amerine, M.A., Methods for nlysis of musts nd wines. John Wiley & Sons, New York. Ppdopoulou, D. & Roussis, I.G., 21. Inhibition of the decline of linlool nd α terpineol in Musct wines by glutthione nd N cetyl cysteine. Itlin Journl of Food Science. 13, Ppdopoulou, D. & Roussis, I.G., 28. Inhibition of the decrese of voltile esters nd terpenes during storge of white wine nd model wine medium by glutthione nd N cetylcysteine. Interntionl Journl of Food Science nd Technology. 43, Ppzin, R., Sulfites: sfe for most, dngerous for some. FDA Consumer Mgzine. 3, Ptel, P., Herbst Johnstone, M., Lee, S.A., Grdner, R.C., Wever, R., Nicolu, L. & Kilmrtin, P.A., 21. Influence of juice pressing conditions on polyphenols, ntioxidnts, nd vrietl rom of Suvignon blnc microferments. Journl of Agriculturl nd Food Chemistry. 58, Ptrinkou, M. & Roussis, I.G., 213. Decrese of wine voltile rom esters by oxidtion. South Africn Journl of Enology nd Viticulture. 34, Peindo, R.A., Moreno, J., Bueno, J.E., Moreno, J.A. & Muricio, J.C., 24. Comprtive study of romtic compounds in two young white wines subjected to pre fermenttive cryomcertion. Food Chemistry. 84, Peng, Z., Ilnd, P.G., Oberholster, A., Sefton, M.A. & Wters, E.J., 22. Anlysis of pigmented polymers in red wine by reverse phse HPLC. Austrlin Journl of Grpe nd Wine Reserch. 8, Pérez Coello, M.S., González Viñs, M.A., Grcí Romero, E., Díz Mroto, M.C. & Cbezudo, M.D., 23. Influence of storge temperture on the voltile compounds of young white wine. Food Control. 14, Peynud, E., Knowing nd Mking Wine. John Wiley & Sons. Phm, T.T., Guichrd, E., Pscl, S. & Chrpentier, C., Optiml conditions for the formtion of sotolon from α ketobutyric cid in the french "Vin June". Journl of Agriculturl nd Food Chemistry. 43, Pons, A., Lvigne, V., Lndis, Y., Drriet, P. & Duboudieu, D., 21. Identifiction of sotolon pthwy in dry white wines. Journl of Agriculturl nd Food Chemistry. 58, Preston, L.D., Block, D.E., Heymnn, H., Soles, G., Noble, A.C. & Ebeler, S.E., 28. Defining vegetl roms in Cbernet Suvignon using sensory nd chemicl evlutions. Americn Journl of Enology nd Viticulture. 59,

194 Rmey, D.D. & Ough, C.S., 198. Voltile ester hydrolysis or formtion during storge of model solutions nd wines. Journl of Agriculturl nd Food Chemistry. 28, Rpp, A., Wine rom substnces from gs chromtogrphic nlysis. In Wine nlysis. H.F. Linskens & J.F. Jckson (ed.), pp Springer, Berlin. Rpp, A. & Güntert, M., Chnges in rom substnces during the storge of white wines in bottles. In The Shelf life of foods nd beverges. Chrlmbous, G. (ed.), pp Elsevier, Amsterdm. Ruhut, D., Yests production of sulfur compounds. In Wine microbiology nd biotechnology. Fleet, G.H. (ed.), pp Hrdwood Acdemic, Chur, Switzerlnd. Recmles, Á.F., Sygo, M.L., González Miret, M.L. & Hernnz, D., 26. The effect of time nd storge conditions on the phenolic composition nd colour of white wines. Food Reserch Interntionl. 39, Ribéreu Gyon, P., Glories, Y., Mujen, A. & Dubourdieu, D., 26. Hndbook of Enology. John Wiley & Sons Ltd, Chichester. Ribéreu Gyon, P., Glories, Y., Dubourdieu, D., Mujen, A., 1998, 24b. Trite d'oenologie. Chimie du vin. Stbilistion et tritements. Dunod Pris. Rizzi, G.P., 26. Formtion of Strecker ldehydes from polyphenol derived quinones nd α mino cids in nonenzymtic model system. Journl of Agriculturl nd Food Chemistry. 54, Rodopulo, A.K., Bezzubov, A.A. & Egorov, I.A., 197. Study of romtic substnces in wine. Prikldny Biokhimiy Mikrobiologiy. 6 (214). Roujou de Boubée, D., Vn Leeuwen, C. & Dubourdieu, D., 2. Orgnoleptic impct of 2 methoxy 3 isobutylpyrzine on red bordeux nd loire wines. Effect of environmentl conditions on concentrtions in grpes during ripening. Journl of Agriculturl nd Food Chemistry. 48, Srrzin, E., Sinkhruk, S., Pons, M., Thibon, C., Bennetu, B. & Drriet, P., 21. Elucidtion of the 1,3 sulfnyllcohol oxidtion mechnism: An unusul identifiction of the disulfide of 3 sulfnylhexnol in Suternes botrytized wines. Journl of Agriculturl nd Food Chemistry. 58, Silv Ferreir, A.C., Brbe, J. C. & Bertrnd, A., 22. Heterocyclic cetls from glycerol nd cetldehyde in Port wines: Evolution with ging. Journl of Agriculturl nd Food Chemistry. 5,

195 Silv Ferreir, A.C., Brbe, J. C. & Bertrnd, A., Hydroxy 4,5 dimethyl 2(5H) furnone: A key odornt of the typicl rom of oxidtive ged Port wine. Journl of Agriculturl nd Food Chemistry. 51, Silv Ferreir, A.C., Guedes de Pinho, P., Rodrigues, P. & Hogg, T., 22b. Kinetics of oxidtive degrdtion of white wines nd how they re ffected by selected technologicl prmeters. Journl of Agriculturl nd Food Chemistry. 5, Silv Ferreir, A.C., Hogg, T. & De Pinho, P.G., 23b. Identifiction of key odornts relted to the typicl rom of oxidtion spoiled white wines. Journl of Agriculturl nd Food Chemistry. 51, Simpson, R.F., Fctors ffecting oxidtive browning of white wine. Vitis. 21, Singleton, V.L., Oxygen with phenols nd relted rections in must, wines nd model systems: observtions nd prcticl implictions. Americn Journl of Enology nd Viticulture. 38, Singleton, V.L. & Krmling, T.E., Browning of white wines nd ccelerted test for browning cpcity. Americn Journl of Enology nd Viticulture. 27, Singleton, V.L., Trousdle, E. & Zy, J., Oxidtion of wines I. Young white wines periodiclly exposed to ir. Americn Journl of Enology nd Viticulture. 3, Skouroumounis, G.K., Kwitkowski, M., Frncis, I.L., Okey, H., Cpone, D., Duncn, B., Sefton, M.A. & Wters, E.J., 25. The impct of closure type nd storge conditions on the composition, colour nd flvour properties of Riesling nd wooded Chrdonny wine during five yers' storge. Austrlin Journl of Grpe nd Wine Reserch. 11, Somers, T.C., The Wine Spectrum: n pproch towrds objective definition of wine qulity. Winetitles, Adelide, Austrli. Sonni, F., Clrk, A.C., Prenzler, P.D., Riponi, C. & Scollry, G.R., 211. Antioxidnt ction of glutthione nd the scorbic cid/glutthione pir in model white wine. Journl of Agriculturl nd Food Chemistry. 59, Suklje, K., Bs Cesnik, H., Jnes, L., Kmecl, V., Vnzo, A., Deloire, A.J., Sivilotti, P. & Lisjk, K., 212. The effect of lef re to yield rtio on secondry metbolites in grpes nd wines of Vitis vinifer L. cv. Suvignon blnc. Journl Interntionl des Sciences de l Vigne et du Vin. Swiegers, J.H., Frncis, I.L., Herderich, M.J. & Pretorius, I.S., 26. Meeting consumer expecttions through mngement in vineyrd nd winery: the choice of yest for fermenttion offers gret potentil to djust the rom of Suvignon blnc wine. Austrlin nd New Zelnd Wine Industry Journl. 21,

196 Swiegers, J.H., Willmott, R., Hill Ling, A., Cpone, D.L., Prdon, K.H., Elsey, G.M., Howell, K.S., de Brros Lopes, M.A., Sefton, M.A., Lilly, M. & Pretorius, I.S., 25. Modultion of voltile thiol nd ester roms by modified wine yest. In: Proc. Weurmn flvour reserch symposium, June 25, Roskilde, Denmrk. Tkhshi, T., Tdenum, M. & Sto, S., Hydroxy 4,5 dimethyl 2(5H) furnone, burnt flvouring compound from ged ske. Agriculturl nd Biologicl Chemistry. 4, Toming, T., 23. Contribution of benzenemethnethiol to the smoky roms of certin Vitis vinifer wines. Journl of Agriculturl nd Food Chemistry. 51, Toming, T., Drriet, P. & Dubourdieu, D., Identifiction of 3 mercptohexyl cette in Suvignon wine, powerful romtic compound exhibiting box tree odor. Vitis. 35, Toming, T., Furrer, A., Henry, R. & Dubourdieu, D., Identifiction of new voltile thiols in the rom of Vitis vinifer L. vr. Suvignon blnc wines. Flvour nd Frgrnce Journl. 13, Toming, T., Guimbertu, G. & Dubourdieu, D., 23. Role of certin voltile thiols in the bouquet of ged Chmpgne wines. Journl of Agriculturl nd Food Chemistry. 51, Toming, T., Msneuf, I. & Dubourdieu, D., 24. Powerful romtic voltile thiols in wines mde from severl Vitis vinifer L. cv. Suvignon blnc. In: Proc. ACS Symposium Series, pp Toming, T., Murt, M.L. & Dubourdieu, D., 1998b. Development of method nlyzing the voltile thiols involved in the chrcteristic rom of wines mde from Vitis vinifer L. cv. Suvignon blnc. Journl of Agriculturl nd Food Chemistry. 46, Uglino, M., 213. Oxygen contribution to wine rom evolution during bottle ging. Journl of Agriculturl nd Food Chemistry. 61, Uglino, M., Kwitkowski, M., Trvis, B., Frncis, I.L., Wters, E.J., Herderich, M. & Pretorius, I.S., 29. Post bottling mngement of oxygen to reduce off flvour formtion nd optimize wine style. Austrlin nd New Zelnd Wine Industry Journl. 24, Uglino, M., Kwitkowski, M., Vidl, S., Cpone, D., Siebert, T., Dievl, J., Agrd, O. & Wters, E.J., 211. Evolution of 3 mercptohexnol, hydrogen sulfide, nd methyl mercptn during bottle storge of Suvignon blnc wines. Effect of glutthione, copper, oxygen exposure, nd closure derived oxygen. Journl of Agriculturl nd Food Chemistry. 59, Uglino, M., Siebert, T., Mercurio, M., Cpone, D. & Henscke, P.A., 28. Voltile nd color composition of young nd model ged Shirz wines s ffected by di mmonium phosphte supplementtion before lcoholic fermenttion. Journl of Agriculturl nd Food Chemistry. 56,

197 Vn Wyngrd, E., 213. Voltiles plying n importnt role in South Africn Suvignon blnc wines. Thesis, University of Stellenbosch, Stellenbosch. Versini, G., Inm, S. & Srtori, G., A cpillry column gs chromtogrphic reserch into the terpene constituents of "Riesling Romno" (Rhine Riesling) wine from Trentino Alto Adige: Their distribution within berries, their pssge into the must nd their presence in the wine ccording to different wine mking procedures. Orgnoleptic considertions. Vini d'itli. 23, Wterhouse, A.L., 22. Wine Phenolics. Annls of the New York Acdemy of Sciences. 957, Wterhouse, A.L. & Teissedre, P.L., Levels of phenolics in Cliforni vrietl wine. In Wine: Nutritionl nd Therpeutic Benefits. Wtkins, T. (ed.), pp Americn Chemicl Society, Wshington, DC. Wilderndt, H.L. & Singleton, V.L., The production of ldehydes s result of oxidtion of polyphenolic compounds nd its reltion to wine ging. Americn Journl of Enology nd Viticulture. 25, Zoecklein, B.W., Fugelsng, K.C., Gump, B.H. & Nury, F.S., Wine nlysis nd production. Chpmn & Hll, London. 183

198 Chpter 5 Reserch results Acetldehyde nd its msking/enhncing bilities A sensory study in model wine 184

199 Chpter 5: Reserch results Acetldehyde nd its msking/enhncing bilities A sensory study in model wine 5.1 INTRODUCTION MATERIALS AND METHODS Medium Chemicls nd spiking Experimentl design Sensory nlysis Dt Anlysis RESULTS AND DISCUSSION Pnel evlution Individul compound evlution Multiple compound evlution CONCLUSION ABBREVIATIONS USED REFERENCES

200 5.1 INTRODUCTION Acetldehyde is one of the most importnt crbonyl compounds in wine nd constitutes up to 9 % of the totl mount of ldehydes found in wine (Nykänen, 1986). The possible negtive effect cetldehyde cn hve on the orgnoleptic qulity of wine nd its bility to combine rpidly with sulphur dioxide (SO 2 ) mkes the ccumultion of this compound criticl during winemking (Burroughs & Sprks, 1973). Acetldehyde is formed by the yest during lcoholic fermenttion (Mrglit, 1997) nd cn lso originte from the ctivity of other microbes such s lctic cid bcteri nd cetic cid bcteri (Drysdle & Fleet, 1988; Liu & Pilone, 2). The mount of cetldehyde formed by the yest will depend on mny fctors such s yest strin, temperture, ph, sugr content, oxygen nd SO 2 (Ough & Amerine, 1958; Rnkine & Pocock, 1969; Herriz et l., 1989; Romno et l., 1994; Cbrnes et l., 1998; Grde Cerdán et l., 28). Other winemking processes cn lso hve significnt effect on cetldehyde concentrtions nd moderte to importnt cetldehyde increses usully occur post fermenttion (Jckowetz & De Orduñ, 213) of which the most importnt production is due to the oxidtion of ethnol (Wilderndt & Singleton, 1974; Ribéreu Gyon, 1998, 24b). The oxidtion of ethnol is not direct, but rther vi coupled uto oxidtion of certin phenolic compounds (Wilderndt & Singleton, 1974). Free SO 2 present in the wine will slow this oxidtion by recting with intermedite oxidtion products s well s with cetldehyde, thereby forming n odourless sulphite combintion with cetldehyde (hydroxysulphonte) which is stble in cid medium. It hs been estimted tht only.4% of cetldehyde is in the free form in the presence of 3 mg/l free SO 2 (Blouin, 1965). Free cetldehyde concentrtion will thus increse s soon s ll the free SO 2 hs been converted to the bound form. The free cetldehyde cn hve significnt influence on the rom nd hs been reported to contribute odours such s green pple, overripe bruised pple, grssy, pungent, nutty nd sherry (Mrglith, 1981; Henschke & Jirnek, 1993; Miyke & Shibmoto, 1993; Frivik & Ebeler, 23). The perception threshold of free cetldehyde in 1% ethnol solution hs been reported to be.5 mg/l (Guth, 1997), which cn mke this compound potent contributor to the oxidtion chrcter in wine. In wines, this perception threshold is considerbly higher (1 125 mg/l) (Zoecklein et l., 1995), however the interctive effects between romtic (nd non romtic) compounds could influence this determintion nd might not be pplicble to ll wine types nd styles. A survey done on 127 white wines reported the totl cetldehyde concentrtion (the sum of free nd bound) to rnge between 7 nd 24 mg/l verging t 4 mg/l (Lopes et l., 29; Jckowetz & De Orduñ, 213). Other sources reported verge levels in white wine to be round 8 mg/l (McCloskey & 186

201 Mhney, 1981). Fortified wines hd the highest cetldehyde content (verge 118 mg/l, rnge 12 8 mg/l) (Lchenmeier & Sohnuis, 28). While high levels of cetldehyde re generlly undesirble in tble wines, it is considered to be unique feture of sherry type wines (Sponholz, 1993; Cortes et l., 1998). Suvignon blnc wines present lrge diversity of wine styles nd rom. On the one hnd, the voltile thiols (4 mercpto 4 methylpentn 2 one (4MMP), 3 mercptohexn 1 ol (3MH) nd 3 mercptohexn 1 ol cette (3MHA)) cn led to tropicl/fruity style wines, while the methoxypyrzines (3 isobutyl 2 methoxypyrzine (IBMP), 3 isopropyl 2 methoxypyrzine (IPMP) nd 3 sec butyl 2 methoxypyrzine (SBMP)) cn contribute to the green pepper, grssy nd sprgus odour (Lcey et l., 1991; Toming et l., 1998; Swiegers et l., 26; Coetzee & Du Toit, 212). Both of these odour groups re often sought fter for the production of qulity Suvignon blnc wines (Lund et l., 29b; King et l., 211). The voltile thiols re oxidtion sensitive nd cn lso undergo cid hydrolysis which leds to decrese in these compounds during ging or oxidtion (Toming et l., 24; Herbst et l., 28; Nikolntonki et l., 21). Correspondingly, loss in the fruity odour of the wine will occur. The grpe derived methoxypyrzines re not sensitive to oxidtion nd studies hve shown methoxypyrzine concentrtions to remin stble during oxidtive hndling of juice nd wine (Mris, 1998; Coetzee et l., 213). Even though the concentrtions of especilly the voltile thiols decrese during ging, concentrtions t the end of n oxidtive ging period of eight months still proved to be well bove the odour perception threshold (Chpter 4). The methoxypyrzine concentrtion remined the sme (bove the odour perception threshold) during the ging period. However, the odour of the wine ws dominted by ttributes such s potto bg, green pple nd sherry t the end of the ging period. These sensory ttributes re likely due to the formtion of methionl (potto bg) nd cetldehyde (green pple nd sherry). Interctive effects of rom compounds (suppression or enhncement) hve been reported before (Chpter 3) (Ribéreu Gyon et l., 1975; Ferreir et l., 22; Cmpo et l., 25; Escudero et l., 27; Pineu et l., 27; King et l., 211; Vn Wyngrd, 213) however, the effect of cetldehyde hs not been elucidted. In some studies the contribution of cetldehyde to the oxidtive rom of wines oxidised during short period of time hs been reported s negligible nd ttributed the oxidised rom to the formtion of other compounds (Escudero et l., 22). However, in generl, it is still believed tht cetldehyde is one of the min compounds cusing the typicl oxidtive rom especilly in wines stored for longer period of time (Bro & Quiros Crrsco, 1977; Ribéreu Gyon et l., 1977; Usseglio Tomsset, 1985; Meirlnd & Pernot, 1992). Methionl hs proven to suppress the ttributes tht rise due to the presence of 3MH (Chpter 3) 187

202 nd could possibly enhnce the cooked bens nd cooked potto odour when lone or in combintion with IBMP. The gols of the reserch presented in this chpter re to define the rom ttributes cused by cetldehyde t vrying concentrtions nd to determine the role of cetldehyde in the perception of plesnt roms brought by typicl Suvignon blnc compounds such s 3MH nd IBMP MATERIALS AND METHODS Medium Model wine consisted of distilled wter, 5 g/l trtric cid nd 12% ethnol nd ph djusted to 3.5 (using sodium hydroxide). After prepring the model wine, the composition ws confirmed using WineScn FT 12 instrument (FOSS Anlyticl, Denmrk) Chemicls nd spiking The compounds used in this study were 3MH, IBMP nd cetldehyde. Solutions of mg/l 3MH (Interchim) nd 25 mg/l IBMP (Sigm) were prepred in 99.5% ethnol (Merck Chemicls, South Afric). The exct concentrtion of 3MH ws lso confirmed by Ellmn s regent (Ellmn, 1959). Acetldehyde (Sigm) solutions were mde fresh dily from the undiluted compound. The 3MH nd IBMP solutions were stored in the drk t 8 C, while cetldehyde ws stored t 4 C. These solutions were used to spike the model wine to the desired concentrtions one hour prior to tsting Experimentl design Initilly ech compound ws spiked individully in order to ssess the effect of concentrtion chnges on the romtic perception. The concentrtions used throughout this study were kept constnt nd cn be seen in Tble 5.1. This ws done to compre profiles of both trils (singulr dditions nd multiple dditions). Smples re numbered from 1 to 5 ccording to the level of ech of the compounds. Using the level s the indictor of the concentrtion simplifies the interprettion of the vrious figures in this study. For specific concentrtions of the compounds refer to Tble

203 Tble 5.1 Concentrtions tested Compound Mercptohexn 1 ol (ng/l) Isobutyl 2 methoxypyrzine (ng/l) Acetldehyde (mg/l) A centrl composite design ws used for the sensory nlysis of the smples contining multiple compounds (to ssess interctions). The suitbility of the centrl composite design lies in the fct tht it cn be used to test interctions between different fctors. Idelly fctoril design could be used to test ll combintions, however this design would result in very lrge number of combintions to be ssessed. The centrl composite design only selects certin mount of combintions specificlly chosen by the design. The use of this method led to design with six str points, eight cube points nd one centre point which resulted in 15 smples in totl to be tested (Tble 5.2). Tble 5.2 Compositions (concentrtion levels) of vrious smples in the centrl composite design. Compound Smple 3 Mercptohexn 1 ol 3 Isobutyl 2 methoxypyrzine Acetldehyde Centre Cube Cube Cube Cube Cube Cube Cube Cube Str Str Str Str Str Str

204 5.2.4 Sensory nlysis The sensory pnel consisted of 12 judges. Intensity rting ws done using 1 mm unstructured line scle (rnking intensity from none to intense ). Testing ws done in booths using stndrd ISO wine tsting glsses. The booths hd stndrd rtificil dylight lighting nd temperture control t 2±2 C. Smple glsses were mrked with rndom three digit code (unique for ech judge) nd glsses were covered with plstic lid prior to sensory ssessment to prevent the rom contminting the lbortory environment. The order of the smples ws rndom nd blnced cross the ssessors. Along with the set of spiked smples, blnk glss contining the unspiked model wine only, ws lso provided for comprison. Pnellists evluted the smples orthonslly only nd the dt ws collected on pper bllot. For resons of humn ethics, brief explntion of the ddition of flvour to the smples ws given prior to testing, lthough cre ws tken to exclude ny informtion tht could hve cused bis. Compounds were first nlysed individully. Six trining sessions (one hour ech) were conducted in n ttempt to rech consensus regrding different smples nd terminology used. In this wy, the pnellists were exposed to ech compound twice. At ech trining session, the pnel received ll five levels of compound s well s blnk smple consisting only of model wine. A rnge of reference stndrds ws lso given. During the first discussion session, the pnel generted ttributes nd brief line scling exercise ws done. During the second trining session the descriptors nd the smple s position on the line scle ws finlised. Ech individul compound ws nlysed during forml test session, which contined 3 replictes. The profiling of the smples contining multiple compounds ws done fter the evlution of the singulr compounds. At this stge the judges were lredy fmilir with the compounds. During profiling, the smples were divided into two subsets s ll smples could not be nlysed in one session. The first subset ws str points 1 6 with the second subset contining cube points 1 8. The centre smple lwys ccompnied the sets during trining nd testing. The experimentl lyout nd the combintions of compounds for ech smple nd subset cn be seen in Tble 5.2. Reference stndrds were vilble for the pnellists during the tril. Five trining sessions consisting of 1 hour ech were used per subset (5 hours per subset). During trining, the pnellists were not informed of the composition of ech smple. Smples were scled on 1 mm unstructured line scle using the ttributes generted during the profiling of the singulr compounds. Anlyses were performed s 2 different tests (ech subset seprtely) with 3 replictes ech. With the set of spiked smples (long with the centre smple), blnk glss 19

205 contining model wine only, ws provided for comprison. Str points were tested first, followed by the cube points. The judges took regulr breks between smples to prevent ftigue Dt Anlysis Assessor performnce ws evluted using PnelCheck (Version V1.4., Nofim, Tromsø, Norwy) ccording to the workflow s described by Tomic et l. (21). Sensory descriptive dt for ech ttribute ws nlysed using mixed model nlysis of vrince (ANOVA). Principle component nlysis (PCA) biplots were creted using the correltion mtrix of the men dt. Post hoc Fisher s LSD tests were used to test for significnce of sensoril differences between compound combintions. In custom built perl progrm, t tests were used to test for the significnce of the sensoril differences between compound combintions, nd the significnt sttisticl reltionships were modelled s network using Cytoscpe (Shnnon et l., 23) in which coloured nodes were compound levels tht yielded sttisticlly significnt difference in sensory ttribute s compred to the control levels tht were indicted by n uncoloured node locted t the centre of the network. Response surfce plots were constructed (Sttistic 1, Sttsoft Inc., Tuls, USA) to investigte interctions. A p vlue threshold of.5 ws used for the determintion of sttisticl significnce (p<.5). 5.3 RESULTS AND DISCUSSION Pnel evlution 3 Wy ANOVA ws used to determine significnt ttributes nd Tucker plots were used to ssess the consensus mongst judges (Figure 5.1) nd the grouping indicted tht the pnellists were well trined nd homogeneous in terms of the wine profiling. For most of the ttributes this grouping ws situted between the inner nd outer ellipses. This shows good pnel consensus with 5 1% of the vrince explined by the judges. 191

206 Figure 5.1 Tucker plot of pnellist performnce Individul compound evlution Model wine ws the preferred choice of spiking medium s the complexity of rel wine could interfere with the interction study. Even neutrl or deromtized wine hs sensory properties due to certin romtic compounds tht remin in the wine. For the purpose of this study, the bsence of ny other rom compounds or non romtic compounds (tht could influence the perception of the rom) ws criticl (Lund et l., 29). The concentrtions tested (Tble 5.1) were chosen bsed on concentrtions of the rom compounds found in white wines s reported in literture, s well s preliminry sensory ssessment of the individul rom compounds nd mixtures of the compounds. Pre screenings by experienced wine tsters (ll who completed certificte course in wine evlution) were done before finlising the concentrtions to ensure tht the levels chosen fulfil the requirements for the im of this study. The concentrtions for 3MH nd IBMP in this study were chosen s follows: the first level ws chosen to be below the perception threshold, s this would give n indiction of enhncing or suppressive effects of these compounds even when they re not present in concentrtions high enough to be individully perceived. The second level ws t the perception 192