Vitis 33, 31-35 (1994) Model wine soltions: ffect of slphr dioxide on color and composition dring ageing by NN PCNLL, J. BKKR and P. BRDL nstitte of Food Research, Reading Laboratory, Reading, Berkshire, United Kingdom S m m a r y : The effect of slphr dioxide and acetaldehyde on the interaction between malvidin 3-glcoside, the main anthocyanin in red wine made from Vitis vinifera grapes, and ( + )-catechin was examined. n all model wines the molar osses of malvidin 3-glcoside were significantly greater than the osses of catechin. The concentration of malvidin 3-glcoside decreased fastest in the presence of acetaldehyde while the most rapid lass of catechin occrred in the model containing malvidin 3-glcoside and catechin only. n the presence of slphr dioxide, these osses still occrred, bt mch more slowly, indicating that condensation reactions may take place even in the presence of slphr dioxide. Polymerisation was most prominent in the model containing malvidin 3-glcoside, catechin and acetaldehyde. Concrrent with the osses in anthocyanin and catechin, qalitative and qantitative changes in visible color were also observed. Changes in color monitared by measring he angle and chroma are also reported. K e y w o r d s : model wines, anthocyanins, acetaldehyde, slphr dioxide, catechin, HPLC, color measrements, ageing. lntrodction One of its most important qality indicators of a red wine is its color. Dring matration, red wine color changes from bright red to a reddish-brown tint, de to the anthocyanins extracted from the grape skins dring fermentation forming polymeric pigments, by mechanisms smmarised and discssed previosly (BKKR et al. 1993). Slphr dioxide is widely sed in wine making as an antioxidant and an inhibitor of ndesirable microbial growth; if added prior to fermentation, it can increase the extraction of color from grape skins when making red wines, and stabilise the color when added at bottling (BRROUGHS 1974). However, this is achieved at the expense of reversible bleaching of the monomerk anthocyanin color of red wine de to the formation of anthocyanin-4-bislphite, a stable colorless compond (JRD 1964; TMBRLK and BRDL 1967). Frthermore, slphr dioxide binds strongly to acetaldehyde (BRROUGHS and SPRKS 1973), hence the presence of slphr dioxide wold be expected to inflence the color changes dring matration. The effect of slphr dioxide on wine polymerisation reactions has not been stdied hitherto. The aim of this work is to extend the stdy of polymerisation in model red wine soltions and examine the effect of slphr dioxide on these reactions. We stdied interactions between malvidin 3-glcoside, the main anthocyanin in red wine made from Vitis vinifera grapes, and (+)-catechin; the roje of acetaldehyde and slphr dioxide was also investigated. Highperformance liqid chromatography (HPLC) was sed to measre the anthocyanin and catechin concentrations. Color changes were monitared sing spectrophotometry and tristimls measrements. Materials and methods S a m p 1 e p r e p a r a t i o n : nthocyanins were extracted from the skins of varios Portgese grape varieties, with methanol and formic acid (97:3, v/v), and malvidin 3-glcoside was separated and prified by preparative HPLC (BKKR et al. 1993). Reagents were analytical grade: ( + )-catechin (Sigma), acetaldehyde (ldrich) and sodim metabislphite (Merck). Solvents were HPLC grade. M o d e 1 s o 1 t i o n s : Reactants were dissolved in a sterilised model wine medim of potassim bitartrate (Merck) bffer (.2 M, ph = 3.7), containing ethanol (1 % v/v). The reactants were sed in amonts calclated to give the following final concentrations in reaction vials (3 ml sample ina 7 ml vial): malvidin 3-glcoside.25 mm (.12 mg/ml), ( + )-catechin.4 mm (.12 mg/ml), acetaldehyde.9 mm (.4 mg/ml) and slphr dioxide.9 mm (.5 mg/ml). Five different experimental mixtres were prepared, filtered (.45 flm sterile membrane filters) and dispensed into 5 sets of vials as follows: () malvidin 3-glcoside (control), (C) malvidin 3-glcoside and ( + )-catechin, () malvidin 3-glcoside and (+)-catechin and acetaldehyde, (CS) malvidin 3-glcoside and ( + )-catechin and slphr dioxide, (S) malvidin 3-glcoside and ( + )-catechin and acetaldehyde and slphr dioxide. ll samples were allowed to react in the dark at ambient temperatre, and were analysed in dplicate from separate vials over a period of 95 d. C o 1 o r m e a s r e m e n t s : Philips PU874 spectrophotometer and 2 mm path Jength glass cells were sed for CLB 76 color measrements (BKKR et al. Correspondence to: Dr. JoKJ BKKR, nstitte of Food Research, Reading Laboratory, arley Gate, Whiteknights Road, Reading, Berkshire RG6 2F, United Kingdom.
32 NN PCNLL, J. BKKR and P. BRDL 1993). Color measrements of samples containing malvidin 3-glcoside and ( + )-catechin and slphr dioxide were made immediately after the addition of 5 mg/1 acetaldehyde, to ensre that the base color was measred, rather than the color of the anthocyanins bleached by slphr dioxide. T o t a p i g m e n t s : Total pigments and total phenols were determined in triplicate by spectrophotometry sing 1 mm path length UV cells. Sampies were dilted as reqired with.1 M HC and absorbances were measred at 52 and 28 nm respectively. The total pigment absorbance measrement was expressed as concentration (c) of malvidin 3-glcoside (molar mass 529) sing the molar absorptivity vale of 28, at 52 nm (NKTC LKSC and HRZDN 1972). Hence, c(mg/1) = 18.89 absorbance. n t h o c y a n i n a n d ( +) - c a t e c h i n a n a 1- y s i s : Hewlett-Packard 19 M HPLC with an ato injector (25 Jll) and a diode array detector recording at 28 and 52 nm was sed to measre the anthocyanin and ( + ) catechin concentrations sing malvidin 3-glcoside chloride and ( + )-catechin as extemal Standards (BKKR et al. 1993). S p h r d i o x i d e a n a y s i s : Free and total slphr dioxide were determined sing a Dionex 45 ion chromatograph (Dionex Corporation, Snnyvale, C, US). mmediately before analysis samples were dilted as reqired in bffer: for total S 2 a ph 9 bffer was sed (2 mm Na 2 HPO) and for free S 2 a ph 2 bffer was sed (1 mm H 2 SOJ Sampies (5 J.l) were analysed sing a Dionex onpac C-Sl colmn, with 1 mm H 2 S 4 as elent at 1 ml/min flow. Detection was by plsed amperometry, sing a Pt electrode at +.7 V. Qantification was done sing an external Standard prepared from sodim metabislphite in ph 9 bffer. c e t a d e h y d e a n a y s i s: Free acetaldehyde was determined by headspace gas chromatography sing a Perkin imer 87 model, eqipped with a HS-11 headspace sampler. Separation was accomplished on a capillary CP-WX 57CB (5 m x.32 mm i.d.) colmn (Chrompak). The oven was held at 45 oc for 1 min, followed by a 3 C/min increase to 25 C, and held 5 min at this temperatre. Other parameters were: flame ionization detector 3 C, injector temperatre 3 oc, carrier gas (N 2 ) flow rate 1.4 ml/min. The determinationwas calibrated sing an external acetaldehyde standard. Total acetaldehyde was determined by an enzymatic method (Boehringer Mannheim GmbH, food analysis enzymatic kit). S t a t i s t i c a a n a y s i s : Linear regression analysis of variance was done sing SPSS. Reslts and discssion C o o r c h a n g e s : Dring ageing the he angle of all models increased, indicating that they had all developed some browning (Fig. 1 ). The browning reactions started slowly in all models; after abot 2 d of storage there were only small differences between the samples, with the exception of CS; however, there were )arge differences in the extent of browning by the end of the experiment. Model C browned most, while the least browning was observed in this sample when slphr dioxide was also present (CS). The formation of yellow xanthylim pigments, noted by JRD and SoMRS (197) and Jater in the context of long-term wine reactions by TMBRLK and BRDL (1976), is the probable explanation of the browning reactions occrring in mixtre C. t is interesting to note therefore, that this reaction is apparently retarded in the presence of slphr dioxide (CS). Model showed only a very small increase in browning dring the first 5 d, after which time this model had browned Jess than control. This very slow browning in the presence of acetaldehyde can be explained by the sperimposed bleing effect of acetaldehyde - mediated reactions with anthocyanin and catechin. Previosly, BKKR et al. (1993) observed in a model containing.5 %. v/v acetaldehyde, a decrease in he angle of25 o indicating bleing dring the first 25 d of storage, after which a slow increase in he angle was observed indicating the onset of browning; at the end of the experiment (14 d) the athors reported that the he angle was the same asthat of the starting mixtre. Under the crrent experimental conditions, model did not show a rapid increase in browning ntil after 5 d storage. t this time the concentration of acetaldehyde had decreased to abot mg/1 (Fig. 2), and it is probable that this concentration became a limiting factor, since there was still 5 mg/1 free malvidin 3-glcoside (Fig. 3) remaining in this model. There was only a small difference in browning pattern between samples and S, even thogh eqimolar concentrations of slphr dioxide and acetaldehyde were added in the latter model. Slphr dioxide has a high af- Fig. 1: Changes in he angle (degrees) for all models. (ß): malvidin 3-glcoside (control); C (): malvidin 3-glcoside and ( + ) catechin; (O): malvidin 3-glcoside and (+)-catechin and acetaldehyde; CS (D) : malvidin 3-glcoside and ( + )-catechin and slphr dioxide; S (): malvidin 3-glcoside and ( + )-catechin and acetaldehyde and slphr dioxide. Fig. 2: Changes in total acetaldehyde concentration in models and S., S: see Fig.. Fig. 3: Changes in malvidin 3-glcoside concentration for all samples. The following eqations were calclated sing stepwise linear regression; the standard error is shown in brackets and the mltiple R is given at the end of the eqation. : Y = -.47X (2.76-4) + 1.992 (1.19-2);.995; C: Y = -.52X (2.64-4) + 1.9883 (1.14-2);.966; : Y = -.64X (2.2-4) + 1.9963 (8.62-3);.985; CS: Y = -.19X (1.48-4) + 2.764 (6.46.-3);.933; S: Y = -.43X (2.5-4) + 2.273 (8.85-3);.969., C,, CS, S: see Fig.. Fig. 4: Changes in chroma for models, C and., C, : see Fig.. Fig. 5: Slopes calclated for decrease in anthocyanin and catechin concentrations expressed in mmol/d, with 95 % confidence regions., C,, CS, S: see Fig.. Fig. 6 (right colmn): Decrease in total pigment measrement and anthocyanin concentration determined by HPLC, both expressed in mg/1., C,, CS, S : see Fig. 1.
ffect of S 2 on wine composition dring ageing 33 6 4 2 C S 1 CS. "' "'ii <: > o; 16 model 14 12 total pigments 1 8 6 4 anthocyanins 2 ::J Q -2+------.---~--,------,------.---~ 2 4 6 8 1 4. 3 >- "" CD :!:! 2 1ii <: c >o "" 1. -; 1 : 2.2 2. 1.8 1.6 1.4 o 2 11 g 2 e 4 6 9 8 1 1.2 +--~-.--~-,-~-.--~-,,._, 2 4 6 8, 18 16 14 ~ "" 12 1 4 2 4 CS S C C 6 8 1. Q ::J.,. :!:! >.. 16 +---~------~--- W <~,- -----, 2 4 6 8 1 14 12 1 8 6 plgments model C 4 antho~ 2 16 Q 14. 12 1 -----~ -,--------,---- --.-----.---~- 2 4 6 8 1 8 6 4 2 anthocyanins Model o+-----.------,----.----,-- -~ 2 4 6 8 1 16 Q 14 'C.,..2 ~ 12 1 8 6 4 2 anthocyanins CS o +--------------~---------~- o 2 4 6 8 1 nthocyanins days Catechlns 2,.----------:--------. 1.5 1;' " ö ~ 1. 8. 1ii.5. 16, 14 ;;. 12 1 8 6 4 2 Model S.+-r-~--.--,-~-'--.--,---.--. ~ a ac aca acs acas ac aca acs acas o+-----.--~-.--~-.-----,r-~--. 2 4 6 8 1 days
34 NN PCNLL, J. BKKR and P. BRDL finity for acetaldehyde at this ph (BRROUGHS and SPRKS 1973), and hence shold bind the available acetaldehyde, leaving very little free to react. Ths model S wold have been expected to have a browning pattern similar to model C, if exactly eqimolar concentrations of acetaldehyde and slphr dioxide had been added. However, the average free acetaldehydein S was 7.9 mg/1, whereas in the sample not containing slphr dioxide () all the acetaldehyde is available for reactions with the flavonoid components. The chroma (Fig. 4) of control showed a steady decrease, while model C showed a more gradal decrease dring the first 5 d of storage, followed by an increase in chroma, in agreement with earlier findings (BKKR et al. 1993). n contrast to these earlier observations made in the presence of mch higher concentrations of acetaldehyde, when an increase in chroma was observed dring the first 7 d of storage, indicating an increase in the amont of color expressed by the mixtre, we now note a slow decrease in chroma dring the entire storage period of. These observations accord well with the fact that no new peaks were observed by HPLC, jst a gradal decline in the concentration of anthocyanins and catechin. cetaldehyde and slphr dioxide c h a n g e s : The total acetaldehyde concentration in models and S decreased dring the first 8 d of the experiment, from 35 mg/1 to less than 5 mg/1 (Fig. 2). The variation between the measring points is de to inter-via! variation. The effect of slphr dioxide on the loss of acetaldehyde cannot be determined from these data. The free acetaldehyde concentration in the S model did not change dring the corse of the experiment, remaining at an average of 7.9 mg/1. The slphr dioxide concentration decreased slbwly in both CS and S, and no obvios effect of acetaldehyde on the rate of loss was observed. nthocyanin and catechin losses: On storage, all soltions exhibited a logarithmic decrease in total anthocyanin concentrations (Fig. 3). The linear loss of anthocyanins in all for model mixtres indicates a first order reaction with respect to this loss, confirming the findings of BRNOWSK and NGL (1983). n models C and the rate of loss was greater than in control, while slphr dioxide had a slowing effect on this process in both cases, althogh in S the rate of loss was only marginally less than in control. The reaction rates k (day 1) are shown in the Table and compared with rates reported previosly (BKKR et al. 1993; ßRNOWSK and NGL 1983). Using a 3.6-fold molar excess of acetaldehyde, the observed increase in reaction rate for the model was less than when the molar excess was 5-fold. Srprisingly, even in the presence of a 3.6-fold molar addition of slphr dioxide (S), the anthocyanin concentration still decreased, althogh at a mch slower rate. The slowing effect of slphr dioxide on the rate of anthocyanin loss is shown even better comparing C and CS reaction rates. The apparent eqilibrim constant for acetaldehyde with slphr dioxide at ph 4 is 1.4 X 1 6(ßURROUGHS and SPRKS 1973), while for malvidin 3-glcoside with slphr dioxide it is 6 x - 5 (BRROUGHS, pers. comm.). Ths in the presence of both acetaldehyde and malvidin 3-glcoside, slphr dioxide wold mostly be bond to acetaldehyde, as can also be determined from the ack of bleaching. lthogh the anthocyanin was to some extent protected against loss by slphr dioxide, its degradation still occrr~d. The osses of catechin in all for models were also logarithmic with time. n order to compare the osses of anthocyanins with those of catechin (not shown), the regression lines were calclated on the concentration expressed in mmol (Fig. 5). The rate of loss of catechin was always significantly slower (p <.5) than for anthocyanins in corresponding models. This is attribtable to the anthocyanin and catechin condensation which is sperimposed on the degradation of anthocyanins alone, as indicated by the loss in model. The extent to which both reactions occr cannot be ascertained, bt the differences in color changes for the models indicate that different reactions occr. dding acetaldehyde increased the slope of anthocyanin loss, bt decreased the slope of catechin loss. The implication appears to be that the condensation involving acetaldehyde, anthocyanin and catechin involves Tab e Reaction rates with respect to the disappearance of total anthocyanins in model wine systems, calclated with respect to malividin 3-glcoside Mixtre k (day-1) k (day-1)t iteratre vale k (day-1) nthocyan in nthocyanins and catechin nthocyanins, catechin and acetaldehyde nthocyanins, catechin and slphr dioxide nthocyanins, catechin, acetaldehyde and slphr dioxide 1.77 X 1-3 12.7 X 1-3 14.73 X 1-3 4.32 X 1-3 9.81 X 1-3 5.66 X 1-3 6.31 X 1-3 8.88 X 1Q-3 6.74 X 1Q-3 254.76 X 1-3 19. X 1-3 t Bakker, Bridle and Picinelli (1993), sing a 4-fold molar excess. Data calclated from Baranowski and Nagel (1983), sing a 5-fold molar excess.
ffect of S 2 on wine composition dring ageing 35 relatively more anthocyanin molecles than the condensation occrring in the absence of acetaldehyde. C o m p a r i s o n o f t o t a 1 p i g m e n t m.e a -s r e m e n t w i t h a n t h o c y a n i n c o n c e n t r a- t i o n : The total pigment absorbance measrement expresses the sm of the concentration of monomeric anthocyanins and the color exhibited by polymerised material. Since the molar absorptivity vale of newly forrned polymerised material is not known, the concentration from this absorbance measrement is expressedas malvidin 3-glcoside, assming the samemolar absorptivity vale for both. The difference between the total pigment data and the anthocyanins determined by HPLC both expressed in concentration terms (Fig. 6), gives an indication of the contribtion of the colored polymers to the overall color. Over the corse of the experiment, the largest difference was observed in the model containing anthocyanins, catechin and acetaldehyde, e.g. after abot 5 d, there was doble the color concentration as expressed by the total pigment measrement, i.e. 5 % of the color was de to polymers. n contrast the difference for the model containing anthocyanins, catechin and slphr dioxide remained mch smaller. C o n c s i o n : n the absence of acetaldehyde slphr dioxide redces the rate of anthocyanin loss and slightly redces polymerisation. When acetaldehyde is also present these effects of slphr dioxide are no onger observed. cknowledgements The athors wish to thank N W KLNG and DNCN HDDRLY for statistical analysis, the C for fnding this project nder FLR (project no. 8953) and NTO's Scientific Committee, Spain, for financial spport for Dr. NN PCNLL. References BKKR, J.; PtCJNLL,.; BRDL, P.; 1993: Model wine soltions: Color and eomposition ehanges dring ageing. Vitis 32, 111-118. BRNOWSK,. S.; NGL, C. W.; 1983: Kineties ofmalvidin 3-glcoside condensation in model wine systems. J. Food Sei. 48, 419-421, 429. BRROUGHS, L. F.; 1974: Browning control in frit jiees and wines. Chem. nd. (Sep.) 718-72. --;SPRKS,. H.; 1973: Slphite bindingpower of wines and eiders.. qilibrim constants for the dissociation of earbonyl bislphite eomponds. J. Sei. Food grielt. 24, 187-198. JRD, L. ; 1964: Reactions involved in s1phite bleaehing of anthocyanins. J. Food Sei. 29, 16-19. - -; SOMRS, T. C.; 197: The formation of xanthylim salts from proeyanidins. Phytoehemistry 9, 419-427. NtKTC-LKSC, G. K.; HRZDN, G.; 1972: Qantitativeanalysis of the anthoeyanin content in grape jiees and wine. Lebensm.-Wiss. Techno!. 5, 163-165. TMBRLK, C. F.; BRDL, P.; 1967: FJavylim sajts, ;mthocyanidins and anthocyanins.. Reaetions with slphrdioxide. J. Sei. Foodgrie1t. 18, 479-485. - -; - -; 1976: nteractions between anthoeyanins, phenolie eomponds and acetaldehyde and their significanee in red wines. mer. J. nol. Viticlt. 27, 97-15. Received May 12, 1993