Food Chemistry 121 (2010) Contents lists available at ScienceDirect. Food Chemistry. journal homepage:

Transcription

1 Food Chemistry () 9 9 Contents lists ville t ScienceDirect Food Chemistry journl homepge: An ssessment of the effects of wine voltiles on the perception of tste nd stringency in wine Mrí-Pilr Sáenz-Nvjs,, Ev Cmpo, Purificción Fernández-Zurno, Dominique Vlentin c, Vicente Ferreir, * Itituto de Ciencis de l Vid y el Vino (UR-CSIC-GR), Deprtment of Chemistry, University of L Rioj, Mdre de Dios, E- Logroño, L Rioj, Spin Lortory for Arom Anlysis nd Enology, Deprtment of Anlyticl Chemistry, Fculty of Sciences, University of Zrgoz, E-9 Zrgoz, Spin c UMR CSG CNRS INRA UB, CESG, Rue Hugues Picrdet, Dijon, Frnce rticle info strct Article history: Received June 9 Received in revised form Jnury Accepted Jnury Keywords: Voltile frction Non-voltile frction Tste Recotituted wines Seory nlysis The ojective of this work is mesuring the effect of different voltile extrct compositio on the perception of tste, stringency, glol inteity nd persistence of wine. Six Spnish wines, two from Chrdonny nd four from Temprnillo grpes, ll of them showing different chemicl nd seory chrcteristics, were selected. Wines were seprted into voltile nd non-voltile frctio y solid phse extrction nd lyophilistion nd further liquid extrction, respectively. Eighteen recotituted wines were prepred, comining different voltile extrcts nd different non-voltile mtrices nd djusting ethnol content to % (v/v), nd were further descried y specificlly trined seory pnel. Tste ttriutes (sweetness, cidity, itterness), stringency, rom inteity, glol inteity nd persistence were ssessed in oth, originl nd recotituted wines y using numericl ctegory scle. The seory properties of the originl wines were retined y their corresponding recotituted smples. The seory ssessment of the recotituted wines showed tht the ddition of voltile fruity extrcts from white wines rought out decrese in stringency nd itterness nd n increse in sweet perception in ll cses. While glol inteity nd persistence of white wine mtrices were lso incresed, they did not chnge in red wine mtrices, which suggests tht the voltile frction plys only secondry role in these ttriutes of red wines. Similrly, the effects of replcing the voltile frction of red wine y voltile extrcts from other red wines were smll nd incoistent, which confirms tht tste nd stringency re primrily driven y non-voltile molecules in these wines. Ó Elsevier Ltd. All rights reserved.. Introduction The overll flvour experience perceived during coumption of food is elicited y the simultneous stimultion of severl sees. It hs een widely ccepted tht interctio cn, nd do, occur within stimuli (Nole, 99) (rom, tste, ppernce, or mouth-feel). The presence of rom tste interctio hs een lrgely studied nd evidenced y the scientific literture. These interctio my result from physicochemicl interctio (structure nd inding effects) in the product itself, interctio t the receptor level or cognitive interctio (Smll & Prescott, ). Since competition t the receptor site is highly unlikely ecuse different receptors re involved mong seory modlities, perceptul interctio re more conceivle. It hs een demotrted tht the oritofrontl cortex is the structure most likely involved in these perceptul interctio. Steveon, Bokes, nd Prescott (998) studied the ssocitive lerning etween odour nd tste in experiments * Corresponding uthor. E-mil ddress: vferre@unizr.es (V. Ferreir). including conducting period. They were le to demotrte the implicit nture of this lerned synesthesi. In other words, the sweet tste ws demotrted to e processed long with the retrol perception of the odour to produce unitry setion in the prticipnt. Mny studies hve shown tht odours cn suppress, enhnce or hve no effect on tstes (Cporle, Policstro, & Monteleone, ; Le, Dmevin, Vccher, Morgenegg, & Mrtin, ). These interctio hve een demotrted to occur in synthetic solutio (Welge-Lüssen, Drgo, Wolfeerger, & Hummel, ) nd in rel smples, such s olive oil (Cporle et l., ), itter coco nd milk everges (Le et l., ) or diry desserts (Lethuut et l., ). Moreover, interctio etween rom nd other seory modlities, such s touch, hve recently een descried. Kor, Ltrille, Souchon, nd Mrtin () crried out study on texture flvour interctio in yogurts, reveling tht olfctory perception enhnced product-perceived stringency. When the sujects perceived the flvour coisting of notes such s green pple, they my hve ssocited this lst perception with the stringency of unripe pple nd given higher score to the 8-8/$ - see front mtter Ó Elsevier Ltd. All rights reserved. doi:./j.foodchem...

2 M.-P. Sáenz-Nvjs et l. / Food Chemistry () 9 9 stringent inteity of the product in question. This oservtion ws ttriuted to cognitive ssocition etween the two perceptio (stringency nd rom). Mny studies hve delt with seory tste rom interctio of molecules present in wines, e.g. sucrose intercting with vnillin (Welge-Lüssen et l., ) or the prune rom (Prescott, Johtone, & Frncis, ), itterness interctio with coconut rom (rich in c-lctones) (Le et l., ) or cut grss odornt (cis-- hexen--ol) (Cporle et l., ). Furthermore, studies on the interctio etween protei, polyscchrides or polyphenols nd selected rom sustnces isolted from red wines hve een crried out (Dufour & Byonove, 999; Dufour & Byonove, 999), reveling the existence of complexes driven minly y hydrophoic forces. Therefore, even if mny studies on tste rom interctio hve delt with molecules present in wines, to our knowledge no one hs focused on the rom tste nd rom stringency interctio in rel wine smples. In this context, the im of this study is to otin preliminry mesurement of the effect of the voltile composition of wine on some in-mouth seory ttriutes, such s tste, stringency, inteity nd persistence. In prticulr, the work will try to evlute whether replcing the voltile composition of given wine y other voltile extrcts, e.g. tken from different wine, hs ny mesurle effect on those in-mouth seory properties of the recotituted wine nd, in tht cse, to ssess the type, mgnitude nd wine to wine coistency of such effects.. Mterils nd methods.. Chemicls nd regents The chemicl stndrds were supplied y Aldrich (Gillinghm, UK), Fluk (Buchs, Switzerlnd), Sigm (St. Louis, Mo), Lncster (Strsourg, Frnce), Polyscience (Niles, IL), Chem Science (West Chester, PA), Interntionl Express Service (Alluch, Frnce) nd Firmenich (Genev, Switzerlnd), s indicted in Tle. Dichloromethne, methnol, nd ethnol, LiChrosolv qulity, were from Merck (Drmstdt, Germny). Pure wter ws otined from Milli-Q purifiction system (Millipore, Bedford, MA, USA). Polypropylene crtridges ( ml), prepcked with LiChrolut EN resi, were lso otined from Merck (Drmstdt, Germny), wheres mmonium sulphte nd NHCO were supplied y Pnrec (Brcelon, Spin)... Wines A set of six commercil Spnish wines, with mrked technologicl, seory nd romtic compositionl differences, ws selected. The wines were yer-old monovrietl Chrdonny wine fermented in stinless steel vts (W), yer-old monovrietl Chrdonny wine fermented in ok rrel (W), yer-old monovrietl Temprnillo red wine (W), high qulity yerold (8 months in ok rrel) 9% Temprnillo-% Cernet Suvignon red wine (W), yer-old (8 months in ok rrel) monovrietl Temprnillo red wine with mrked stringency (W) nd yer-old ( months in ok rrels) monovrietl Temprnillo red wine with mrked woody rom (W). W ws selected s the model for white wine, W s the model for protein-rich white wine, W s the model for neutrl red, W s the model for highly structured polyphenol-rich red wine, W s the model for very stringent wine, nd W ws exclusively selected ecuse of its typicl woody rom. Conventionl oenologicl prmeters (ethnol concentrtion, ph, reducing sugrs, titrtle nd voltile cidities) were determined in ccordnce with officil OIV prctices (O.I.V., ). L-mlic nd lctic cids were determined y enzymtic methods in ccordnce with officil AOAC nlysis methods (AOAC,, chp. ). Totl polyphenol index (TPI) ws estimted s sornce t 8 nm (Riéreu-Gyon, 9)... Recotituted wine preprtion... Generl The voltile extrcts of the six wines, nmed A, A, A, A, A nd A, respectively, nd the non-voltile extrcts of the five of them coidered more relevnt for the study selected s models of very different wine non-voltile mtrices, nmed M, M, M, M nd M, were seprtely otined s detiled elow (Sectio.. nd..).... Voltile extrcts preprtion SPE crtridges (in ml reservoirs) filled with mg LiChrolut EN resi were put in the extrction unit (VAC ELUT Sttion from Vrin) nd conditioned y pssing (slowly) ml of ethnol nd ml of hydrolcoholic solution (% ethnol (v/v), g l of trtric cid, ph djusted to. with. M NOH). After this, ml of wine were loded. The crtridge ws then ried with ml of the hydrolcoholic solution nd voltile compounds were finlly eluted with ml of ethnol, using positive pressure to void ir contct. The extrct ws spiked with BHA t mg l, nd ws stored in vils with no hedspce, seled nd stored t C prior to smple preprtion.... Non-voltile extrcts preprtion Fifty millilitres of wine were lyophilised in ml round flsks nd, fter this, smples were extrcted with ml of dichloromethne in order to eliminte remining voltile compounds. Afterwrds, dichloromethne ws completely eliminted y forcing strem of pure nitrogen (c. ml min ) to pss through the smple for min. The totl sence of dichloromethne ws ssessed y hedspce solid phse micro extrction (Croxen/ PDMS lm t C min) nd GC with n electron cpture detector (overll system detection limit ng/smple). The extrct ws then dissolved in minerl wter (Evin Ò, Evin-les Bi, Frnce) nd rought up to ml (five times concentrted). After this, smples were plced in vils, with no hedspce, in order to void smple-oxygen contct nd stored t C prior to smple preprtion.... Smple recotitution Recotituted wines were prepred y mixing ml of ethnolic voltile extrct (corresponds to the voltile extrct of ml of wine), ml of non-voltile extrct (corresponds to ml of wine) nd ml of ethnol, nd ringing the mixture to ml with ottled minerl wter (finl ethnol content is % (v/v)). Eighteen smples were prepred y comining different voltile nd non-voltile extrcts from different wines, s shown in Tle. Comintio (rom mtrix) were selected, seeking for those exerting most likely seory impct on in-mouth ttriutes. Efforts were therefore concentrted on the red wine mtrices, prticulrly in the most stringent (M) in order to evlute possile chnges in stringency. Smples were stored t C in ottles with no hedspce nd hermeticlly closed in order to void contct with oxygen prior to seory evlution... Wine chrcteristion... Generl The chrcteristion of the six wines used for the study ws crried out y oth seory nd chemicl nlyses.

3 M.-P. Sáenz-Nvjs et l. / Food Chemistry () 9 9 Tle Voltile composition of the six studied wines (ll dt re expressed s microgrmmes per litre). Bold numers men compounds ove their odour threshold. Compound W W W W W W Odour description Source Odour threshold Acids -Methylutyric cid. 8. Cheese Aldrich (Ferreir, Lopez, & Ccho, ) -Methylutyric cid Cheese Aldrich (Ferreir et l., ) Isoutyric cid 9 88 Cheese Aldrich (Ferreir et l., ) Butyric cid Cheese Polyscience (Ferreir et l., ) Hexnoic cid 9 9 Cheese Polyscience (Ferreir et l., ) Octnoic cid 9 8 Ftty/unplesnt Fluk (Ferreir et l., ) Decnoic cid Cheese Polyscience (Ferreir et l., ) Alcohols Furfuryl lcohol.8 9 nd.. 9. Hy/mold Fluk (Vn Gemert & Nettenreijer, 9) Methionol Potto/coliflower Aldrich (Ferreir et l., ) Isoutnol,8,,,,, Fusel Merck, (Ferreir et l., ) -Butnol 9 Fusel Aldrich, (Etiévnt, 99) Isomyl lcohol 8,8, 99,, 9,, Fusel Aldrich, (Guth, 99) -Hexnol Grss Sigm 8 (Guth, 99) c--hexenol 8 Grss Aldrich (Guth, 99) Benzyl lcohol 9 9 Sweet, florl Aldrich, (Aznr, Lopez, Ccho, & Ferreir, ) -Phenylethyl lcohol 9,,,,, Roses Fluk, (Ferreir et l., ) Esters Ethyl isoutyrte Fruity/strwerry Fluk (Guth, 99) Isoutyl cette Solvent ChemService (Aznr et l., ) Ethyl -methylutyrte Fruity/green pple Fluk (Guth, 99) Ethyl isovlerte Fruity/nise Fluk (Ferreir et l., ) Ethyl decnote Fruity Polyscience (Ferreir et l., ) Phenylethyl cette Roses ChemService (Guth, 99) Ethyl cinnmte Cinnmte/sweet Aldrich. (Ferreir et l., ) Ethyl utyrte Fruity Aldrich (Guth, 99) Isomyl cette 9 8 Bnn ChemService (Guth, 99) Ethyl hexnote 8 Fruity/nise Polyscience (Guth, 99) Ethyl lctte 8, 9,,, Fruity Aldrich, (Etiévnt, 99) Ethyl octnote Fruity/fresh Polyscience (Guth, 99) Ethyl Fruity Aldrich, (Aznr et l., ) hydroxyutyrte Diethyl succinte 9 9 8, 9 Fruity Fluk, (Etiévnt, 99) Voltile phenols Guicol Phenolic/chemicl Aldrich 9. (Ferreir et l., ) o-cresol nd nd nd Lether/spicy Aldrich (Lopez et l., ) -Ethylguicol nd.8 nd Lether/phenolic Aldrich (Ferreir et l., ) -Propylguicol nd nd nd.9.9. Lether niml Lncster (Aznr et l., ) Eugenol.. nd..9.8 Clove Aldrich (Guth, 99) -Ethylphenol nd nd Bitumen/lether Aldrich (Lopez et l., ) Isoeugenol Spicy Lncster (Aznr et l., ) -Vinylphenol.. nd nd nd nd Almond shell Lncster 8 (Boidron, Chtonnet, & Po, 988) Aldehydes Phenylcetldehyde nd nd nd nd nd. Florl/honey Aldrich (Aznr et l., ) Furfurl Almond Fluk, (Ferreir et l., ) -Hydroxymethylfurfurl nd..... Almond Aldrich, (Vn Gemert & Nettenreijer, 9) Acetldehyde nd nd nd nd nd nd Green pple Aldrich (Guth et l., 99) -Allyl-,- dimethoxyphenol nd..... Spicy smoky Aldrich (Vn Gemert & Nettenreijer, 9) Lctones t-whiskylctone. nd Coconut/woody Aldrich (Etiévnt, 99) c-whiskylctone.. 9 Coconut/woody Aldrich 9 (Ferreir et l., ) d-octlctone Pech Lncster (Vn Gemert & Nettenreijer, 9) c-nonlctone Pech Aldrich (Nkmur, Crowel, Ough, & Totsuk, 988) c-declctone. nd Spicy/woody/ phenolic Fluk 8 (Ferreir et l., ) Norisoprenoids -Dmscenone Bcked pple Firmenich. (Guth, 99) -Ionone..8 nd nd nd nd Tc Sigm. (Etiévnt, 99) -Ionone Violet Sigm.9 (Ferreir et l., ) Terpenols Linlool Florl/musct Aldrich (Guth, 99) -Terpineol Fresh/rosemry Fluk (Ferreir et l., ) -Citronellol Green/lemon Aldrich (Etiévnt, 99) (continued on next pge)

4 M.-P. Sáenz-Nvjs et l. / Food Chemistry () 9 9 Tle (continued) Compound W W W W W W Odour description Source Odour threshold Gerniol.. nd nd nd nd Florl Fluk (Guth, 99) Vnilli Vnillin nd.8 nd... Vnill Aldrich (Guth, 99) Methyl vnillte Vnill Lncster 99 (Lopez et l., ) Ethyl vnillte Vnill/honey Lncster (Lopez et l., ) Acetovnillone Vnill Aldrich (Aznr et l., ) Syringldehyde Vnill Aldrich, (Vn Gemert nd Nettenreijer, 9) Miscellneous Benzoic cid Swet Aldrich (Aznr et l., ) Phenylcetic cid Roses Aldrich (Mg, 9) Ethyl furote Fruity Fluk, (Ferreir et l., ) Dicetyl Butter Aldrich (Guth, 99) Acetoin 8 8 8,,, Ftty/wet Aldrich, (Etiévnt, 99) c-butyrolctone 89 Cheese Aldrich, (Aznr et l., ) Reference from which the vlue hs een tken is given in prenthesis. nd: not detected. Tle Eighteen recotituted wines, non-voltile (MX) nd voltile (AY) extrcts. Smples re referred to in the text s MXAY (formed y ddition of the non-voltile extrcts of wine X nd the voltile extrct of wine Y). M M M M M A x x x x x A x A x x x x x A x x x A x x x A x... Seory nlysis... Pnel trining. In totl, students or stff memers from the University of Burgundy (Frnce) were recruited on the sis of their interest nd their vilility during weeks (one min session per week). They were not pid for their prticiption. Among these pnellists, were selected to crry out the mesuring sessio ( mles nd 8 femles from to 9 yers old). The selection of pnellists ws crried out y clculting the reproduciility index (Ri) proposed y Cmpo, Do, Ferreir, nd Vlentin (8) for romtic ttriutes, where the minimum Ri required to keep judge respoe ws estlished t.. According to this, pnellists were selected nd, with them, three-wy ANOVA for the in-mouth ttriutes, involving smples (S), judge (J) nd replicte (R) s fixed fctors nd ll first order interctio, ws clculted to confirm the pnel performnce. Pnellists ttended eight descriptive seory trining sessio over period of months, during which pnellists worked in sugroups. They were provided with list of terms otined from the literture (Cmpo et l., 8). During trining, different reference stndrds, representtive of rom, tste nd stringency terms, were presented. Stndrds were either commercilly ville odornts tken from Interntionl Flvour nd Frgrnces (Dijon, Frnce), Sentosphère (Pris, Frnce), Le Nez du Vin (Jen Lenoir, Provence, Frnce) nd Firmenich (Genev, Switzerlnd), or nturl products (fruits, juices, spices, vegetles) prepred t the eginning of ech session. For tste nd stringency, solutio, contining different concentrtio of tle sugr ( g l ) for sweetness, trtric cid (. g l ) for cidity, quinine sulphte ( mg l ) for itterness nd potssium, nd luminium sulphte ( g l ) for stringency stimuli, were presented to the pnel to id with recognition, nd discrimintion etween the different orl setio. The trining period included two phses: generl nd product-specific trining phse. During the generl trining phse (four sessio), pnellists ecme fmilir with rom ttriutes nd with inteity rting of sweetness, cidity, itterness, stringency, romtic nd glol inteity, s well s persistence. During typicl session, pnellists hd to evlute different wines, descriing their odour properties y choosing up to five descriptors in the rom list nd y rting sweetness, cidity, itterness, nd stringency on -point scle ( = sence, = very low nd 9 = very high ), while in-mouth rom inteity, in-mouth glol inteity nd glol persistence were mesured on 9-point scle ( = very low nd 9 = very high ) since, for these lst concepts, the hs no mening. The wines selected for this trining phse presented intee nd esily recognisle tste nd stringency properties nd included red, white nd rosé wines of diverse grpe vrieties nd origi. The session ended with discussion, during which the pnel leder compred the rom descriptors nd the tste inteity scores given y pnellists to descrie ech wine. The specific trining phse coisted of four sessio, during which pnellists ecme fmilir with the type of smples of the study. During this phse, pnellists descried odour properties nd rted the inteities of sweetness, cidity, itterness, stringency, romtic nd glol inteity, s well s glol persistence of ten Spnish commercilly ville wines nd two repetitio (one repetition for session) of five recotituted wines different from those used for the study (formed y the non-voltile extrct of W nd the voltile frctio of W, W, W, W, nd W, respectively).... Wine evlution. Trined pnellists descried wines in duplicte. Ten millilitres wine smples were presented in drk ISO (9)-pproved wineglsses lelled with three-digit rndom codes nd covered y plstic Petri dishes ccording to rndom rrngement. Pnellists were sked to smell ech wine nd to descrie their odour y choosing mximum of five ttriutes from the list of ccording to the cittion frequency method (Cmpo, Bllester, Lnglois, Dcremont, & Vlentin, ; Cmpo et l., 8). Then, they were sked to rte the sweetness, cidity, itterness, stringency, romtic nd glol inteities, s well s the glol persistence of the smples using the ove mentioned structured scles for ech wine. Pnellists pused for min intervls etween smple evlutio to limit dpttion effects. During tht time they were sked to rie their mouths with wter, to hve some plin crckers nd finlly to rie their mouths gin with wter.

5 M.-P. Sáenz-Nvjs et l. / Food Chemistry () 9 9 All wines were served t room temperture nd were evluted in individul ooths. Smples were stored t C. Pnellists were not informed out the nture of the smples eing evluted.... Chemicl quntittive nlysis... Mjor compounds (liquid liquid microextrction nd GC-FID nlysis). Quntittive nlysis of mjor compounds ws crried out, using the method proposed nd vlidted y Orteg, Lopez, Ccho, nd Ferreir (). In ccordnce with this method, ml of wine nd ml of wter were slted with. g of mmonium sulphte nd extrcted with. ml of dichloromethne. The extrct ws then nlyzed y GC with FID detection using the conditio descried elsewhere (Orteg et l., ). Quntittive dt were otined y interpoltion of reltive pek res in the clirtion grphs cotructed y the nlysis of synthetic wines contining known mounts of the nlytes. -Butnol, -methyl--pentnol, -hydroxy--methyl--pentnone nd -octnol were used s internl stndrds.... Minor compounds (SPE nd GC ion trp MS nlysis). This nlysis ws crried out using the method proposed nd vlidted y Lopez, Aznr, Ccho, nd Ferreir (). In ccordnce with the method, ml of wine, contining ll of utyl-hydroxy-nisole (BHA) solution nd ll of surrogte stndrds solution (-octnone, -dmscone, heptnoic cid, nd isopropyl propnote), were pssed through LiChrolut EN crtridge t out ml/min. The sorent ws dried y letting ir pss through (. r, min). Anlytes were recovered y elution with. ml of dichloromethne. An internl stndrd solution ws dded to the eluted smple. The extrct ws then nlyzed y GC with ion trp MS detection under the conditio descried y Lopez et l. ()... Seory chrcteristion of recotituted wines The trined pnel descried the 8 recotituted wines y rting sweetness, cidity, itterness, stringency, romtic nd glol inteities, s well s glol persistence. The pnel trining nd smple evlution were crried out in the sme wy s for wines (see Section...)... Dt nlysis... Evlution of wine seory dt On the dt derived from the in-mouth seory nlysis of wines, one-wy nlysis of vrince (ANOVA) (in which wine ws the fctor nd judges (men of oth replictes) were coidered s repetitio) ws performed. On the dt derived from the seory evlution of rom, crried out y the frequency of cittion method, chi-squre (v ) nlysis on the verge cittion frequency (two repetitio) of ech term ws clculted.... Evlution of the recotituted wines seory dt... One-wy ANOVAs. Five one-wy ANOVAs with repeted mesurements were performed on ech of the seven ttriutes verged cross replictes. The first ANOVA ws performed on the recotituted wines formed y the non-voltile extrct of W (MA nd MA), nd the second ANOVA ws computed on the smples formed y the mtrix of W (MA nd MA). The third nd fourth ANOVAs were clculted for two red non-voltile extrcts (M nd M, respectively) to which the voltile extrct of one white wine (A) nd three red wines (A, A, nd A) were dded. Finlly the fifth ANOVA ws performed on the six smples formed y the red non-voltile extrct of W nd the voltile extrcts of the six studied wines (A A). Four one-wy ANOVAs, with repeted mesurements, were clculted on ech of the seven ttriutes verged cross replictes for the four pirs of wine/recotituted smples. The first ANOVA ws clculted on the recotituted wine MA nd the commercil wine W, the second on MA nd W, wheres the third nd fourth ANOVAs were performed on MA nd W, nd MA nd W, respectively. Student Newm Keuls post hoc pirwise compriso (9%) were crried out for significnce effects.... Two-wy ANOVAs. A two-wy ANOVA, with repeted mesurements, ws performed on six ttriutes (sweet, cid, itter, stringent, glol inteity nd persistence) verged cross replictes. The ANOVA model ws computed on the dt derived from the seory nlysis of the recotituted wines formed y the voltile extrct of the white wine W (MA, MA, MA, MA nd MA) nd of the red wine W (MA, MA, MA, MA nd MA). Mtrix colour (white vs red) nd voltile extrct (A vs A) were coidered s fixed fctors nd judges s rndom fctors ccording to the following model: mtrix colour + voltile extrct + mtrix colour voltile extrct. Besides, two-wy ANO- VA, with repeted mesurements, ws clculted with voltile extrct (red vs white) nd white mtrices (M vs M) s fixed fctors nd judges s rndom fctors in order to evlute the effect of the voltile extrct (from red or white wine) on the perceived ttriutes.... Principl component nlysis (PCA). A normlised PCA ws performed on the men rtings over the pnellists for the ttriutes for ech recotituted wine. PC dimeio with n eigenvlue higher thn (Kiser criteri) were retined. All nlyses were crried out with SPAD (version., CISIA-CESRESTA, Montreuil, Frnce) nd SAS (version 9., SAS Ititute Inc. Cry, NY, USA) softwres.. Results nd discussion.. Wine chrcteristion Six quite different wines, in terms of oth rom nd tste properties, were selected for the study. The rom properties of those six originl wines, mesured y recently descried frequency of cittion method (Cmpo et l., ), re given in Fig.. The most discriminnt terms ccording to the v criterion re presented in Supplementry Mteril. These results revel tht the seven rom fmilies, together with the terms lcohol nd vegetle, vried significntly mong the six wines. As shown in Fig., the white wine W ws the fruitiest (FC = ), followed y the second white, W (FC = ), while ll reds showed similr nd rther low frequencies of cittion for this term (from to ). On the otherhnd, reds scored higher thn whites on the urnt-woody term (FC etween nd vs less thn ). Leving side W, which represents quite neutrl nd romticlly poor red wine, the other red smples hd rther specific descriptors, such s vegetl nd vegetles (W), urnt (W), lck nd dry fruits (W nd W) or woody (W). Some of these rom properties cn e explined in terms of the mesured voltile composition, given in Tle. As cn e seen, the two white wines (W nd W) hve the highest concentrtio of utyric, hexnoic nd octnoic cids, s well s of their corresponding ethyl esters, nd lso of isomyl cette. These highest levels in fruity esters, together lso with reltively high levels in -dmscenone cn explin the highest scores in the fruity ttriute in comprison with reds. The white wines lso hve highest levels of terpenols nd of ethyl cinnmte, which contriute to florl nd sweet notes. By contrst, red wines presented highest levels of isoutyric cid, isomyl lcohol, -phenylethyl lcohol, ethyl isovlerte, cetoin nd dicetyl, in ccordnce with previous

6 M.-P. Sáenz-Nvjs et l. / Food Chemistry () 9 9 () Grpefruit EXOTIC FRUITS WINE (W) () BURNT NUTS WINE (W) WOODY YELLOW FRUITS Alcohol Bergmot Per Apple FLORAL WHITE FRUITS CITRUS FRUITS SPICY BLACK FRUITS EXOTIC FRUITS CITRUS FRUITS WHITE FRUITS FLORAL Men cittion frequency Men cittion frecuency (c) UNDERGROWTH ANIMAL WINE (W) (d) Cndied/cooked fruits BLACK FRUITS WINE (W) DRY FRUITS WOODY RED FRUITS ANIMAL BLACK FRUITS Artichoke FLORAL RED FRUITS WHITE FRUITS SPICY SPICY WOODY VEGETABLES VEGETAL Men cittion frequency Men cittion frequency (e) RED FRUITS ANIMAL SPICY Prune VEGETAL WINE (W) (f) ANIMAL VEGETABLES Cherry BURNT VEGETAL Prune WINE (W) UNDERGROWTH DRY FRUITS BLACK FRUITS BURNT RED FRUITS SPICY DRY FRUITS BLACK FRUITS WOODY Men cittion frecuency Men cittion frequency Fig.. Men cittion frequency of the most cited odour ttriutes in the studied wines. Error rs re clculted s s/(n) / ; s, stndrd devition; n, numer of pnellists. reports (Ferreir, Fernndez, & Ccho, 99). Leving side W, whose composition lies hlf-wy etween whites nd reds, red wines re lso richer in some voltile phenols, such s guicol, -ethylguicol, -ethylphenol, -propylguicol, nd eugenol, which is coistent with their highest scores for the urnt-woody term. Leving side romtic inteity, for which the pnel ws not le to provide coistent ssessment, the scores otined for the mesured in-mouth properties (tste, stringency, glol inteity nd persistence) re shown in Fig.. As cn e seen, the two white wines, W nd W, re sweet, slightly cid nd not stringent nor itter, while the reds scored very low in sweetness nd reltively high in stringency nd itterness. Wines W nd W re the most stringent nd itter, lthough, in this cse, differences were not significnt. As for glol inteity nd persistence, the most remrkle fct is tht W is the lest intee, nd tht no significnt difference ws found for the ttriute persistence mong the studied wines. Finlly, the conventionl oenologicl prmeters of the studied wines re given in Tle. The ph vlues were, s expected, higher in reds thn in whites, nd rnge from. to., while the highest vlue for the titrtle cidity ws oserved for W (. g l ). It should e noted tht ph is not correlted (r =.; P =.) with the cidity perceived on the studied wines, which is in ccordnce with the results pulished y Etio et l. (8), lthough it should e noted tht white wines hve the highest levels of mlic cid whose tste threshold hs een estimted s.9 mg l (Hufngel & Hofmnn, 8). Reducing sugr contents re low, s expected for dry tle wines, nd the mximum vlue for this prmeter is found in W. All the wines were reltively rich in ethnol, which rnged from.% to.8%. Red wines hd highest levels of TPI (totl polyphenol index), nd the mximum vlue of this prmeter is oserved for W (8.)... Recotituted wines The mjor gol of the smple preprtion procedure ws to chieve completely odourless tstnt frctio nd tsteless odornt frctio. Those ojectives were est chieved y using comintion of lyophilistion nd liquid liquid extrction for smple deromtistion, nd direct solid phse extrction for the extrction of the voltile frction, voiding different strtegies of distilltion, which in most cses induced the formtion of rtifcts or involved quite complex setup during which it is difficult to pre-

7 M.-P. Sáenz-Nvjs et l. / Food Chemistry () 9 9 Men seory rting c c c c w w w w w w c c sweet cid itter stringent glol inteity glol persistence Fig.. Comprison of the men seory rtings of the studied wines. Error rs re clculted s s/(n) / ; s, stndrd devition; n, numer of pnellists. Different letters indicte the existence of significnt difference etween smples (.) (Student Newm Keuls test);, no significnt differences; sme letters indicte tendency in perception (.). Tle Conventionl oenologicl prmeters of the studied wines. ph Voltile cidity Titrtle cidity Reducing sugr Mlic cid Lctic cid Ethnol (v/v) TPI W W W W W W Expressed s g trtric cid per litre. Expressed s g per litre. men scores 8 sweet cid itter stringent glol inteity glol persistence Fig.. Comprison of the men seory rtings of the four pirs of wine smple/recotituted smple. Error rs re clculted s s/(n) / ; s, stndrd devition; n, numer of pnellists. Smples: W (), MA (), W (), MA (), W (), MA (), W () nd MA (8). serve smple integrity. The process worked resonly well, since the odour of the tstnt frction ws residul, s ws the tste of the odornt frctio nd, more importntly, recotituted wines did not differ from norml commercil wine smples from the seoril point of view nd most of the seory properties of the originl wines were lso present in the corresponding recotituted smples. These oservtio re corroorted y the seory scores given in Fig. for four pirs of wine smple/recotituted

8 M.-P. Sáenz-Nvjs et l. / Food Chemistry () 9 9 () men scores MA MA () men scores MA MA sweet cid itter stringent glol inteity glol persitence sweet cid itter stringent glol inteity glol persitence (c) men scores MA MA MA MA (d) men scores MA MA MA MA sweet cid itter stringent glol inteity glol persitence sweet cid itter stringent glol inteity glol persitence (e) men scores MA MA MA MA MA MA cc c sweet cid itter stringent glol inteity glol persitence Fig.. Grph of the men seory rtings of recotituted wines formed y the non-voltile extrct () M, () M, (c) M, (d) M nd (e) M. Different letters indicte the existence of significnt difference etween smples (.) (Student Newm Keuls test);, no significnt differences; sme letters indicte tendency in perception (.). smple. As cn e seen, recotituted smples retin most of the seory properties of the originl wines nd, in fct, only in couple of compriso (out of the possile compriso) were differences significnt. Such differences were oserved for the ttriute cid for the comprison W vs MA, the recotituted smple (MA) eing evluted s the most cid, nd for the ttriute glol inteity for the pir MA-W, in this cse the recotituted smple (MW) eing evluted s less intee... In-mouth seory properties of the recotituted smples The results of the seory nlysis of the recotituted wines re summrised in Fig., which shows the men vlues nd the stndrd error of the ttriutes (except for romtic inteity) ssessed for the 8 recotituted wines. The summries of the different ANOVA sttistics crried out on the dt set re given in the Supplementry Mteril, while Fig. gives the projection of smples nd vriles on the PCA first two dimeio. The PCA plot, ccounting for more thn 89% of the originl vrince, cn e used to gin first ide of the hierrchy nd inteity of the effects cused y different voltile extrcts dded to different non-voltile mtrices. As the PCA plot shows, smples re distriuted primrily, s expected, ccording to the nture of the non-voltile mtrix: the four smples with non-voltile mtrices from white wines re on the left prt of the plne, while the smples mde from red wine

9 M.-P. Sáenz-Nvjs et l. / Food Chemistry () 9 9 Fig.. PCA on the principl components nd. () Lodings of the in-mouth ttriutes, () projection of the 8 smples. non-voltile mtrices re on the right prt. Next, the PCA shows tht, while red wine mtrix smples re grouped minly ccording to the nture of the mtrix (M smples centre up, M elements centre, nd M elements down nd right), white wine mtrix smples re grouped ccording to the nture of the voltile extrct (A with + vlue in PC, A with negtive vlues of PC); i.e. voltile composition exerts mjor influence on the tste properties of white wines while, in reds, the influence seems to e much lower. Nevertheless, it cn e oserved tht the three red wine mtrix smples contining the voltile extrct A from fruity white wine re in centred re, clerly displced to the left of the other equivlent smples from the sme mtrix, which suggests tht smples contining A re sweeter nd less itter nd stringent thn re the other smples from the sme mtrix. These oservtio re corroorted y Fig.. In the cse of the white wine mtrix M, Fig. shows tht glol inteity nd persistence were significntly perceived s different (P =.89 nd P =., respectively), which me tht replcing A y A (from neutrl red wine), reduces the inteity of oth ttriutes. In the cse of M, replcing A y A rings out significnt decrese of sweetness. It is importnt to note tht there is trend (P =.898 nd P =.9) in the perception of stringency, which increses when A is replced y A in oth white wine mtrices. Such increse could e relted to the decrese in sweetness (significnt only in the cse of M). In generl, repeted mesures AN- OVA, crried out on M nd M (dt not shown) confirms tht in white wine mtrix, replcing A y A rings out significnt decreses of the perceptio of sweetness (P =.), glol inteity

10 8 M.-P. Sáenz-Nvjs et l. / Food Chemistry () 9 9 nd persistence (P =.) nd n increse on the perception of stringency (P =.). The second relevnt oservtion from the PCA ws the displcement of red wine mtrix smples contining A to the left of the plne. As Fig. confirms, such displcement cn e explined y the incresed sweetness of these smples (significnt only in M), nd y decresed levels of itterness (significnt only in M) nd of stringency (significnt in M nd M nd close to it in M). The repeted mesures ANOVA, crried out on these smples, confirms tht the ddition of A to ny red wine mtrix cuses significnt increse of the perception of sweetness nd significnt decreses of itterness nd stringency. It is remrkle tht, in red wine mtrices, there is no significnt effect linked to the presence of A on the inteity nd persistence of the smples, s ws previously oserved with white wine mtrices. This suggests tht persistence nd inteity of in-mouth perception in white wines is highly dependent on romtic composition while, in reds, this does not hppen; however, sweetness, itterness nd stringency re in ll cses relted to the wine voltile extrct composition. The effect of red wine voltile extrcts on the tste nd stringency ttriutes of red wine mtrices is less cler, s oth, PCA, Fig. nd different sttisticl tretments (not shown) indicte. There re some chnges, ut in most cses they re not significnt, nor re they coistent in the different mtrices. The single difference is due to the sweeter chrcter of the smple composed of the most stringent mtrix (M) nd the neutrl wine (A). Remrkly, tht wine ws the richest (mong reds) in the fruity rom components, isomyl cette, ethyl isovlerte nd -dmscenone, s cn e seen in Tle. It should e noted tht we hd expected, perhps nively, tht the voltile extrct of complex nd well-structured wines, such s W, would hve hd some effect on the inteity nd persistence of neutrl nd rther poor wine, such s W. Fig. suggests, however, tht, if there is n effect of the voltile extrct on inteity nd persistence, it is certinly very smll; i.e. inteity nd persistence in red wines re primrily relted to the non-voltile composition. Similrly, it cn e seen tht the voltile extrct of very stringent nd itter wine, such s W, does not elicit more itterness when dded to different wine, nd tht the voltile frctio extrcted from well-structured wines, such s W or W, do not cuse itterness nd stringency to decrese, which implies tht, leving side the strong effect exerted y the voltile extrcts of white wines, stringency nd itterness of red wine re primrily nd minly relted to its non-voltile mtrix composition.. Conclusio This work hs shown tht sweetness of dry wine is closely relted to fruity rom, nd tht, s sweetness most likely ffects the perceptio of stringency nd itterness, these two lst percepts re lso inversely relted to fruity rom. This suggests tht not only re polyphenols respoile for stringency nd itterness perceptio in wines, ut tht such ttriutes re indirectly relted to the stringency-sweetness interction. It cn lso e concluded tht glol inteity nd persistence seem to e closely relted to the voltile composition in white wines ut not in red wines. Nevertheless, the fct tht replcing the voltile extrct of neutrl or n stringent red wine y nother, from well structured wine, hs smll effects on stringency, itterness, inteity nd persistence, suggests tht these percepts re primrily cused y non-voltile molecules. Acknowledgements The uthors thnk the Itituto de Estudios Riojno, nd Coejerí de Educción, Cultur y Deportes del Goierno de L Rioj (FO- MENTA 8/ project), MEC/FEDER (AGL-9 project) for their finncil support. MPSN thnks the University of L Rioj for her F.P.I. grnt. Appendix A. Supplementry mteril Supplementry dt ssocited with this rticle cn e found, in the online version, t doi:./j.foodchem... References AOAC Officil Methods of Anlysis (). Method 99.: L-mlic cid/totl mlic cid rtio in pple juice (th ed., p. ). Aznr, M., Lopez, R., Ccho, J. F., & Ferreir, V. (). Prediction of ged red rom properties from rom chemicl composition. Prtil lest-squres regression models. Journl of Agriculturl nd Food Chemistry,,. Boidron, J. N., Chtonnet, P., & Po, M. (988). Influence du ois sur certines sustnces odorntes des vi. Conniss. Vigne Vin,, 9. Cmpo, E., Bllester, J., Lnglois, J., Dcremont, C., & Vlentin, D. (). Comprison of conventionl descriptive nlysis nd cittion frequency-sed descriptive method for odor profiling: cse of Burgundy Pinot noir wines. Food Qulity nd Preference,,. Cmpo, E., Do, B. V., Ferreir, V., & Vlentin, D. (8). Arom properties of young Spnish monovrietl white wines: study using sorting tsk, list of terms nd frequency of cittion. Austrlin Journl of Grpe nd Wine Reserch, (),. Cporle, G., Policstro, S., & Monteleone, E. (). Bitterness enhncement induced y cut grss odornt (cis--hexen-l-ol) in model olive oil. Food Qulity nd Preference, (), 9. Dufour, C., & Byonove, C. L. (999). Influence of wine structurlly different polyscchrides on the voltility of rom sustnces in model system. Journl of Agriculturl nd Food Chemistry, (),. Dufour, C., & Byonove, C. L. (999). Interctio etween wine polyphenols nd rom sustnces. An iight t the moleculr level. Journl of Agriculturl nd Food Chemistry, (), 8 8. Etio, I., Elortondo, F. J. P., Alisu, M., Gston, E., Ojed, M., & Schlich, P. (8). Effect of winemking process nd ddition of white grpes on the seory nd physicochemicl chrcteristics of young red wines. Austrlin Journl of Grpe nd Wine Reserch, (),. Etiévnt, P. X. (99). Wine. In H. Mrse (Ed.), Voltile compounds in food (pp. 8 ). Dekker: New York Ferreir, V., Fernndez, P., & Ccho, J. F. (99). A study of fctors ffecting wine voltile composition nd its ppliction in discriminnt nlysis. Food Science nd Technology-Leemittel-Wissechft nd Technologie, 9(), 9. Ferreir, V., Lopez, R., & Ccho, J. F. (). Quntittive determintion of the odornts of young red wines from different grpe vrieties. Journl of the Science of Food nd Agriculture, 8, 9. Guth, H. (99). Quntifiction nd seory studies of chrcter impct odornts of different white wine vrieties. Journl of Agriculturl nd Food Chemistry,,. Hufngel, J. C., & Hofmnn, T. (8). Quntittive recotruction of the non-voltile seometolome of red wine. Journl of Agriculturl nd Food Chemistry,, Kor, E. P., Ltrille, E., Souchon, I., & Mrtin, N. (). Texture-flvor interctio in low ft stirred yogurt: How mechnicl tretment, thickener concentrtion nd rom concentrtion ffect perceived texture nd flvor. Journl of Seory Studies, 8(), 9. Le, D., Dmevin, L., Vccher, C., Morgenegg, C., & Mrtin, N. (). Modultion of perceived tste y olfction in fmilir nd unfmilir everges. Food Qulity nd Preference, ( 8), Lethuut, L., Brossrd, C., Meynier, A., Rousseu, F., Llms, G., Bousseu, B., et l. (). Sweetness nd rom perceptio in diry desserts vrying in sucrose nd rom levels nd in texturl gent. Interntionl Diry Journl, (), 8 9. Lopez, R., Aznr, M., Ccho, J., & Ferreir, V. (). Determintion of minor nd trce voltile compounds in wine y solid-phse extrction nd gs chromtogrphy with mss spectrometric detection. Journl of Chromtogrphy A, 9( ),. Mg, J. A. (9). Tste thresholds vlues for phenolic cids which cn influence flvor properties of certin flours, gri nd oilseeds. Cerel Science Tody, 8,. Nkmur, S., Crowel, E. A., Ough, C. S., & Totsuk, A. (988). Quntittive nlysis of c-nonlctone in wines nd its threshold determintion. Journl of Food Science,,. Nole, A. C. (99). Tste-rom interctio. Trends in Food Science nd Technology, (), 9. O.I.V. (). Recueil des methods interntionls d nlyse des vi et des mouts, Vienne. Orteg, C., Lopez, R., Ccho, J., & Ferreir, V. (). Fst nlysis of importnt wine voltile compounds development nd vlidtion of new method sed on gs

11 M.-P. Sáenz-Nvjs et l. / Food Chemistry () chromtogrphic-flme ionistion detection nlysis of dichloromethne microextrcts. Journl of Chromtogrphy A, 9( ),. Prescott, J., Johtone, V., & Frncis, J. (). Odor-tste interctio: Effects of ttentionl strtegies during exposure. Chemicl Sees, 9(),. Riéreu-Gyon, P. (9). Le dosge des composes phénoliques totux d les vi rouges. Chimie Anlytique,,. Smll, D. M., & Prescott, J. (). Odor/tste integrtion nd the perception of flvour. Experimentl Brin Reserch,,. Steveon, R. J., Bokes, R. A., & Prescott, J. (998). Chnges in odor sweetness resulting from implicit lerning of simultneous odor-sweetness ssocition: An exmple of lerned synesthesi. Lerning nd Motivtion, 9(),. Vn Gemert, L. J., & Nettenreijer, A. H. (9). Compiltion of odour threshold vlues in ir nd wter. TNO report, Ntionl Ititute for wter sypply/centrl ititute for nutrition nd food reserch. The Netherlnds: Zeist. Welge-Lüssen, A., Drgo, J., Wolfeerger, M., & Hummel, T. (). Gusttory stimultion influences the processing of intrl stimuli. Brin Reserch, 8(), 9.