Ferentation Kinetics of Different Sugars by Apple Wine Yeast Saccharoyces cerevisiae D. Wang 1, Y. u 1,2, J. Hu 1 and G. Zhao 1 ABSTRACT J. Inst. Brew. 11(4), 34 346, 24 A non-linear kinetic odel to predict the consuption of different sugars (glucose, fructose and sucrose) as a substrate, during an apple wine yeast ferentation with Saccharoyces cerevisiae strain CCTCC M2122 is proposed. This odel was used to predict sugar utilization by this yeast beginning at various initial sugar concentrations. After observation of the experiental data, a odel based on the logistic equation of yeast growth, growthassociated production of ethanol with a lag tie, and consuption of sugars for bioass foration and aintenance, was developed. After experiental odel fitting, kinetic paraeters in the odel were estiated. The experiental verification of the odel was perfored using flask-scale ferentations, and the odel obtained predicted the ferentation perforance effectively, using different sugars as the substrate set at various initial sugar concentrations. Based on estiated kinetic paraeters and the characteristics of sugar utilization, the yeast exained appeared to be glucophilic. The effects of different sugars with various initial concentrations on the ferentation perforance by this yeast were investigated, and soe applications of kinetic paraeters are discussed. Key words: Apple wine, ferentation kinetics, sugar, yeast. INTRODUCTION Apple wine is a ferented beverage ade fro fresh or concentrated apple juice. It has had a long tradition in Europe and has taken an iportant place in the global fruit wine industry 11. It has becoe the second largest fruit wine industry with an increasing deand in China. Today the use of selected pure cultures of yeast for ferentation of apple wine as starters, and the technological advances in other parts of the ferented beverage industry have influenced the apple wine aking process, however the available inforation is not sufficient yet to perit a full understanding and control of the process 5,11. Apple juice contains any sugars, including fructose, glucose, sucrose as well as other carbohydrates, in varying concentrations. Unlike ash, the sugar in the highest aount in apple juice is fructose, up to 7% of the total 1 Lab. of Brewing Microbiology and Applied Enzyology, School of Biotechnology, The Key Laboratory of Industrial Biotechnology, Ministry of Education, Southern Yangtze University, 17 Huihe Rd., Wuxi, P.R. China, 21436. 2 Corresponding author. E-ail: yxu@sytu.edu.cn Publication no. G-24-134-255 24 The Institute & Guild of Brewing ferentable sugar of 1~15 g/l, plus glucose and sucrose 5,11. Like wine yeast, the priary function of apple wine yeast (Saccharoyces cerevisiae is the ajor industrial strain) is to catalyze the rapid, efficient and coplete conversion of sugars to alcohol without the developent of ferentation off-flavors. However, slow and incoplete alcoholic ferentations of juice (i.e., sluggish or stuck ferentations) are a chronic proble for the fruit wine industry 3. This can lead to unscheduled loss of tank capacity due to extended processing ties and the potential for icrobial instability and off-taste of the final product due to residual sugars. In these cases, fructose is the ajor factor, causing a high residual sugar concentration. In spite of the iportance of fructose ferentation for apple wine production, few studies have addressed this subject. Furtherore, other ferentable sugars fro cane, beet or hydrolyzed corn syrup are coonly used as adjuncts in apple wine production. The changes in sugar content will affect the ferentation process. Hence the need for a practical investigation into the echanis of sugar uptake and utilization and the kinetic behavior of ferentation by yeast used in this process is significant. Over the past 2 years, nuerous papers concerning sugar uptake and utilization by yeast have been published 1. It is stated that the rate of alcohol production by yeast is liited priarily by the rate of sugar uptake, especially the uptake of fructose 16,17. In general, while both glucose and fructose are utilized siultaneously, glucose is utilized faster than fructose by yeast 7,21, and S. cerevisiae appears to be glucophilic, although soe strains have a clear preference for fructose 18. An appropriate evaluation on the discrepancy between the aount of glucose and fructose consued by wine yeast strains during ferentation ight be helpful to solve, at least partially, the probles caused by the slower ferentation of fructose 2. However, few systeatic analyses or quantifications of the preference for glucose or fructose of yeast strains for apple wine have been carried out. More inforation is required to select and evaluate yeasts to iprove ferentation perforance. As a useful tool, the priary objective of a kinetic odel developed for wine ferentation is the prediction of the kinetic behavior of yeast ferentation perforance based on the initial characteristics of the juice. The developent of the corresponding atheatical odels of ferentation kinetics is also iportant in the research of yeast behavior and etabolic regulation 19. An appropriate odel of ferentation, with the technical, econoic and physiological iplications would be a powerful instruent to predict and control proble ferentations, and be 34 JOURNAL OF THE INSTITUTE OF BREWING
helpful to understand the ferentation process 13. Several physical and atheatical odels for wine ferentation have been reviewed, and any factors leading to proble ferentations, such as nitrogen and oxygen liitation, teperature extrees etc., have been well docuented by kinetic ethods 4,6,8,1,13. However, the effects of different sugars, fructose, glucose and sucrose, which are the ain sugars found in apple and grape juice, on the kinetic perforance of yeast ferentation in apple juice are not clear. Therefore the objective of this work was to propose an unstructured odel for apple wine yeast ferentation kinetics using different sugars as a substrate, in which it is assued that the icroorganis can be represented by one state variable, and the consideration of the incorporation of coplicated biocheical knowledge concerning the different regulation echaniss of the cell is not required. Then the ferentation perforance in different initial sugar levels by the apple wine yeast is predicted. Based on kinetic paraeters obtained, utilization of different sugar by this yeast were copared, and the yeast kinetic characteristics discussed. MATERIALS AND METHODS Yeast strain One yeast strain, Saccharoyces cerevisiae strain CCTCC M2122 (China Center for Type Culture Collection) for apple wine ferentation, selected by this laboratory 22, was used for the experients. Mediu In order to investigate the effect of a single sugar on yeast ferentation, a synthetic ediu was prepared to siulate an apple juice concentrate, according to the coposition of an ideal apple juice used for apple wine 5,11. It contained 2~23 g/l single sugar (glucose, fructose or sucrose, respectively), 1.7 g/l yeast extract, 1.7 g/l (NH 4 ) 2 SO 4,.9 g/l quinate, 3. g/l alic acid, 1. g/l lactic acid, and 1 g/l vitain D 2 and B 1. The ph was adjusted to 3.5. Titratable acidity to 8.1 as alic acid was 3.95 g/l. Ferentation conditions Yeast cells were cultured at 25 C in a pasteurized ediu for 24 h, and then a 2% (w/v) inoculu was used to inoculate one litre of ediu in a 2 litre Erleneyer flask, equipped with ferentation air locks. Ferentation was carried out at 15 C and allowed to continue to coplete attenuation. Saples were taken periodically during ferentation for analysis of cell weight, as well as sugar and alcohol concentration. All saples were analyzed in triplicate. The initial sugar concentration was ~1 g/l, and specific details are indicated in the text or figures legends. Analytical techniques Bioass concentration was calculated fro the absorbance easured at 6 n and a calibration was carried out in which the absorbance and dry weight of yeast were correlated 8. All saples were taken after suspending the yeast and washing sufficiently. The yeast was resuspended in.5% EDTA-Na solution to retard flocculation when easuring bioass. Dilutions were ade as necessary to obtain an optical density between.2 and.8. Sugars and ethanol were deterined according to Ough and Aerine 15. Model siulation Model siulations and data fitting were perfored using the SAS software Syste for Windows version 8.1 (SAS Institute, Cary, NC). Model paraeters are described in the text. Model developent Kinetic odel. Aong the nuerous odels developed, the ajority of the odels are biocheically knowledge based odels 13, which consist of a set of atheatical equations describing the phenoena occurring during wine ferentation. The ain advantage of this type of odel is that they account for biological phenoena. The odel paraeters with soe biological significance can be obtained, but their structures ay be strongly nonlinear, coplex and difficult to verify and validate, and can pose probles in ters of paraeter identification. In the developent of non-linear odeling techniques, there has been an increase in the use of sigoidal shaped growth odels in the predictive icrobiology field to forecast ferentation process. The details of non-linear odeling have been described by Speers et al. 2 In general, the ferentation kinetic odel can be subdivided into a growth odel, a substrate odel, and a product odel. There are three different equations derived to describe the kinetic behavior of the concentration of yeast cells, the sugars (glucose, fructose or sucrose, respectively), and ethanol in this study. Aong the any odels describing the growth kinetics of icroorganiss (e.g. Monod growth odel), the odel structures were chosen. The classical sei-epirical Monod type odels cannot fit processes of ferentation well in any cases, although there are any odified types. Recently the logistic odel, as a sigoidal shaped odel, has been a ost popular one due to its goodness of fit 2 and has been widely used in describing the growth of icroorganis 9,13. Usually, the logistic odel was used to show the self-regression ade by the increase of cell concentration coon in batch-ferentation. In this paper, we develop an equation based on the logistic odel with growth-associated production of ethanol. For cell concentration,, the logistic odel was derived as follows: d = µ 1, (1) dt where µ is the axiu specific growth rate with respect to the ferentation conditions, as the for of the Monod relationship. With the following boundary conditions: t =, =, S = S, P = By integration of equation 1, the kinetic odel can be forulated. The bioass production rate yields the following equation (the logistic equation): VOL. 11, NO. 4, 24 341
e t = µ + e, (2) µ t This equation shows the relationship of bioass and the ferentation tie, which is used to fit the experiental data of bioass concentration. There are two paraeters, µ and, in this equation, and they are estiated fro the experiental data by the atheatical software, SAS Syste. As observed in the present experient (Fig. 1), ethanol concentration increased proportionally to increasing bioass during the ferentation process; especially in the exponential phase of cell growth, and only a very sall proportion of ethanol production was present in the stationary growth phase. A significant relationship was found in our data between the specific ethanol production rate, dp/(dt), and the specific growth rate, µ[d/(dt)], which was expected with growth-associated product foration. Hence ethanol production in this odel is viewed as a growth-associated relation with bioass 12, and inclusion of a non-growth-associated ter in this odel was not justified in this case. However, a delay of ethanol production was found copared with the cell growth, and little ethanol was produced during the yeast lag growth phase. Therefore, a paraeter of the lag tie, t, was introduced to describe the delay of ethanol production to cell growth, and the equation of ethanol production rate was odified as Eq. (3), dp d = Y, (3) p / x dt d( t t) This equation can be integrated using two estiated paraeters fro Eq. (2), µ and, and the odel is described by the Eq. (4). After the experiental data of ethanol production was fitted, two paraeters in this equation, the yield coefficient Y p/x and t, were estiated. Paraeters estiation. The initial values of and S were fixed by the experiental conditions. The other paraeters, such as µ,, t, and soe yield coefficients were estiated by the Newton nonlinear regression ethod of SAS 8.1 syste using batch experiental data. In the SAS progra, 5 iterative search techniques, including Newton nonlinear regression ethod, Newton- Gauss, Marguardt, Gradient and Dud technique, were used to iniize the residual su of squares. All the tech- eµ ( t t) e µ t P = Yp / x, + eµ ( t t ) + e µ t (4) For the alcoholic ferentation process, the equation describing the substrate consuption rate takes into account two aspects, the sugar consuption in the foration of bioass and the aintenance of bioass 6,13. The consuption rate of sugar was described as below: ds 1 d = +, (5) dt Y dt x / s Cobined with the Eq. (1), Eq. (3) and estiated paraeters, this equation can be integrated and the sugar consuption equation can be obtained as Eq. (6). The related paraeters were estiated. 1 S = S Y x / s ln µ µ e t + e + e µ t µ t (6) Fig. 1. The ferentation profiles of apple wine yeast using (A) glucose with a deterined initial concentration of 85. g/l, (B) fructose of 91.1 g/l, and (C) sucrose of 96.6 g/l as a substrate, respectively. 342 JOURNAL OF THE INSTITUTE OF BREWING
Table I. Kinetic paraeters estiated fro the experiental data on different sugars. Glucose Fructose Sucrose µ (h 1 ).945 (.23).8 (.19).887 (.17) (g/l) 8.366 (.1164) 7.227 (.111) 7.5215 (.933) Y p/x (g/g) 5.579 (.115) 5.7878 (.182) 5.6637 (.1238) t (h) 15.3 (1.2) 14.59 (1.23) 17.43 (1.21) Y x /s (g/g).22 (.122).1922 (.89).261 (.212) (h 1 ).1144 (.69).715 (.38).789 (.64) N.B. Values in brackets denote the approxial standard errors of the estiated paraeters using the Newton nonlinear regression ethod. niques gave siilar results, however the Newton nonlinear regression ethod tended to give slight better fits as evidenced by exaination of the residual su of squares and F-value. RESULTS AND DISCUSSION The ferentation profiles of apple wine yeast on the different sugars are shown in Fig. 1. There were soe differences in ferentation perforance of this yeast, in cell growth rate, ethanol production rate, and sugar consuption rate, using glucose, fructose or sucrose as sole sugars, respectively. Based on the experiental data of the different sugars, soe kinetic paraeters, including yeast axiu specific growth rate (µ ) and axiu bioass concentration ( ), the yield coefficient (Y p/x ) and lag tie ( t), Y x /s and aintenance coefficient (), were estiated by atheatical software with Eq. (2), Eq. (4) and Eq. (6), respectively. The estiated values of the paraeters are given in Table I. The estiated values of µ agreed with those generally exhibited by yeasts under the present conditions 6,1. By the ethod of ferentation kinetics, the effects of the different sugars are described in detail. Yeast growth kinetics on different sugars Bioass concentrations on different sugars were fitted by Eq. (2) in Fig. 2, using the estiated paraeters. As can be observed fro the figure, the odel fits well to the experiental results, with high significance for odel fitting (all the Pr <.1), and the Correlation Coefficients, r 2, of odel fitting are.991 for glucose,.991 for fructose, and.993 for sucrose as a substrate, respectively. In other words, the odel is able to predict the growth of yeast in these cases. When using glucose as the sole sugar in the ediu, the yeast axiu specific growth rate (µ ) and axiu bioass concentration ( ) were highest copared with those of other sugars,.97 h 1 and 8.15 g/l, respectively. On the contrary, the lowest µ and were obtained when using fructose as the sole carbon source. Judging fro the estiated paraeters, it is suggested that with this yeast strain, glucose is ore suitable for the production of bioass and this yeast appears to be glucophilic. Ethanol production kinetics on different sugars With estiated values of Y p/x and t (given in Table I), Eq. (4) was used to fit ethanol production on the different sugars. A high significance for odel fitting was verified (all the Pr <.1), and r 2 are.99 for glucose,.989 for fructose, and.985 for sucrose as a substrate, respectively. Experiental data and those predicted by the kinetic odel for ethanol production are shown in Fig. 3. There was little difference between the experiental data and the prediction results, and data predictions obtained fro the odel were reasonable and in soe cases very accurate. Hence, the consideration of yeast growth-associated product foration and the introduction of the lag tie paraeter of ethanol production to cell growth, t, were reasonable in the odel developent. According to the yield coefficient of Y p/x, it appeared that fructose and sucrose were ore beneficial to ethanol production given the sae aount of yeast, with the higher values of Y p/x, although they were not as suitable for yeast growth as glucose. Furtherore, the lag ties of ethanol production, t were siilar when fructose and glucose were used as the substrate respectively, but fruc- Fig. 2. Experiental data and kinetic odel predictions for cell growth, using glucose, fructose or sucrose as the substrate, respectively. Fig. 3. Experiental data and kinetic odel predictions for ethanol production, using glucose, fructose or sucrose as the substrate, respectively. VOL. 11, NO. 4, 24 343
tose was consued ore slowly for ethanol production than glucose due to slow growth of yeast on fructose. Sucrose was hydrolyzed extracellularly by the action of S. cerevisiae invertase producing equiolar glucose and fructose initially 1,14, which was then utilized by the yeast. Therefore, the longest lag tie of ethanol production observed on sucrose was understandable in this case. Ethanol in the ediu results fro the foration of ethanol fro sugar etabolis and its excretion into the ediu. The paraeter of lag tie, t, reflects this process and indicates fast or slow ethanol production relative to yeast growth. generally accepted that the process of sugar uptake represents the ajor control echanis for the rate of glycolytic flux under anaerobic conditions 16, and the rate of sugar uptake by yeast is a consequence of the inherent kinetics of the transport process and substrate inhibition (including copetitive inhibition of fructose and glucose Substrate consuption kinetics on different sugars The predicted evolutions of the different sugars by Eq. (6) during the ferentation processes are shown in Fig. 4, together with the experiental data. This figure indicates that the prediction of the odel agreed well with the experiental data in all cases. The odel was verified with high significance (all the Pr <.1), and r 2 of odel fitting were.996 for glucose,.994 for fructose, and.976 for sucrose as a substrate, respectively. Clearly, the consuption of glucose was faster than the other two sugars. This phenoenon can be explained by kinetic paraeters Y x /s and. Glucose was ore beneficial to bioass production with higher Y x /s. Siultaneously, the yeast ore readily etabolized glucose, which was reflected by the higher aintenance coefficient () on glucose. Due to hydrolysis of sucrose to produce equiolar glucose and fructose, sucrose can be regarded as a ixture of glucose and fructose, and the utilization of sucrose was interediate between those of glucose and fructose. While fructose was utilized slowly, the utilization of sucrose was ore siilar to that of fructose, with lower values of Y x /s and. Cobined with the obtained paraeters of µ and, it was further suggested by paraeters Y x /s and that the yeast was glucophilic. Although a few yeast strains have a clear preference for fructose 18, ost glucophilic fruit wine yeasts utilize fructose ore slowly than glucose. It is now Fig. 4. Experiental data and kinetic odel predictions for sugar consuption, using glucose, fructose or sucrose as the substrate, respectively. Fig. 5. Model predictions for substrate consuption with various initial sugar levels, using (A) glucose as a substrate with deterined initial concentrations of 21. g/l, 85. g/l, and 215.4 g/l, respectively, (B) fructose of 21.6 g/l, 91.1 g/l, and 233.5 g/l, respectively, and (C) sucrose of 24.6 g/l, 97.6 g/l, and 243.2 g/l, respectively. 344 JOURNAL OF THE INSTITUTE OF BREWING
in the juice) 3. Berthels et al. 2 indicated that the cause of the glucose/fructose discrepancy appears to be located in the transport and/or phosphorylation steps of the ferentation pathway, and suggested that deterination of the glucose/fructose discrepancy of candidate wine yeast strains should be a standard procedure in strain evaluation and selection. In the experients reported in this paper, the focus was on the utilization of single sugars, which eliinated the effect of the copetitive inhibition of fructose and glucose. The results suggested that the slow utilization of fructose was due to the nature of this yeast, which was independent as to whether glucose was contained in ediu or not. Due to the high fructose content in apple juice, the evaluation and selection of a fructophilic yeast strain, with a high-level expression of hexokinase I (displaying a faster reaction rate in edia containing fructose) and soe ebers of the HT transporter faily with a higher affinity towards fructose 3, could be significant to the wine industry. The odel developed here and kinetic paraeters are potential tools for strain evaluation and selection in this aspect of yeast ferentation perforance. Until now, this odel has not been used in industrial apple wine ferentation, and further research, such as experients with different sugar ferentations siultaneously, with varying sugar concentrations, need to be perfored. Model predictions for substrate consuptions with various initial sugar levels To verify the prediction capability of this odel, experients of sugar consuption with various initial sugar levels were carried out. Predicted curves of sugar consuption, with the experiental data, and the details of various initial sugar levels are shown in Fig. 5. The predictions for sugar consuption by the odel are illustrated in this figure, except for the higher initial glucose concentration of 215.4 g/l. When glucose with higher initial concentration was utilized, a discrepancy between experiental data and predicted curves ay result due to the inhibitory effect of glucose, which becae ore clear as the glucose concentration increased 3,17. As a result, the utilization rate of glucose declined and did not agree with the odel. In this case, the odel predicted that glucose would be utilized copletely in 225 hours, while it took ~37 hours. However, this inhibition was not observed with high concentrations of fructose or sucrose. According to the prediction of the odel, high concentrations of fructose (233.5 g/l) or sucrose (243.2 g/l) were consued copletely in ~42 hours and 38 hours, respectively. These results suggest that the glucose concentration should not be too high in the ferentation ediu (apple juice) if the desired ferentation efficiency of the sugar is to be reached, since inhibition of glucose can occur. Furtherore, when ferentable sugars are used as adjuncts in soe cases of apple wine production, it is suggested that high levels of glucose could yield this inhibitory effect. Fructose has the caution that ost fruit wine yeasts applied in the apple wine industry are glucophilic. Sucrose ay be a coproise, if high concentrations of adjunct sugar are required. CONCLUSIONS A non-linear kinetic odel for different sugar ferentations by one industrial Saccharoyces cerevisiae apple wine yeast is proposed, based on the logistic equation of yeast growth, growth-associated production of ethanol with a lag tie, and consuption of sugars for bioass foration and aintenance. Soe kinetic paraeters with physiological significance in the odel were estiated by atheatical software. The experiental verification of the odel was perfored using flask-scale ferentations. The results obtained indicated that the odel could predict ferentation perforance using different sugars as the substrate with various initial sugar concentrations. Moreover, soe kinetic paraeters obtained, such as µ,, Y x /s,, could be used as indicators for selection or evaluation of potential glucophilic or fructophilic yeast strains. The predictive capability of this odel has potential as a useful tool for deterining how to address apple wine ferentation issues. ABBREVIATIONS bioass concentration (g L 1 ) axiu bioass concentration (g L 1 ) initial bioass concentration (g L 1 ) s aintenance coefficient (g sugar/g bioass h) t tie (h) P produced ethanol concentration (g L 1 ) S ferentable sugar concentration (g L 1 ) S initial ferentable sugar concentration (g L 1 ) Y p/x yield coefficient of ethanol on bioass (g ethanol/g bioass) Y p/s yield coefficient of ethanol on sugar (g ethanol/g sugar) Y x /s yield coefficient of bioass on sugar (g bioass/g sugar) µ specific growth rate (h 1 ) µ axiu specific growth rate (h 1 ) ACKNOWLEDGEMENTS This work was financially supported by the Ministry of Science and Technology, P.R. China (Project No. 21BA51AF). REFERENCES 1. Barnett, J.A., Sugar utilization by Saccharoyces cerevisiae. In: Yeast Sugar Metabolis, F.K. Zierann and K.D. Entian, Eds., Technoic Publishing: Switzerland, 1997, pp. 35 43. 2. Berthels, N.J., Cordero Otero, R.R., Bauer, F.F., Thevelein, J.M. and Pretorius, I.S., Discrepancy in glucose and fructose utilisation during ferentation by Saccharoyces cerevisiae wine yeast strains. FEMS Yeast Research, 24, 4, 683 689. 3. Bisson, L.F., Stuck and sluggish ferentations. A. J. Enol. Vitic., 1999, 5(1), 17 119. 4. Boulton, R., The prediction of ferentation behavior by a kinetic odel. A. J. Enol. Vitic., 198, 31, 4 45. 5. Cabranes, C., Mangas, J.J. and Blanco, D., Selection and biocheical characterization of Saccharoyces cerevisiae and Kloeckera apiculata strains isolated fro Spanish cider. J. Inst. Brew., 1997, 13, 165 169. VOL. 11, NO. 4, 24 345
6. Caro, I., Pérez, L. and Cantero, D., Developent of a kinetic odel for the alcoholic ferentation of ust. Biotechnol. Bioeng., 1991, 38, 742 748. 7. Cason, D.T., Reid, G.C. and Gatner, E.M.S., On the differing rates of fructose and glucose utilization in Saccharoyces cerevisiae. J. Inst. Brew., 1987, 93, 23 25. 8. Craer, A.C., Vlassides, S. and Block, D.E., Kinetic odel for nitrogen-liited wine ferentations. Biotechnol. Bioeng., 22, 77(1), 49 6. 9. García-Ochoa, F. and Casas, J.A., Unstructured kinetic odel for sophorolipid production by Candida bobicola. Enzye and Microbial Technology, 1999, 25, 613 621. 1. Giovanelli, G., Peri, C. and Parravicini, E., Kinetics of grape juice ferentation under aerobic and anaerobic conditions. A. J. Enol. Vitic., 1996, 47(4), 429-434. 11. Lea, A.G.H., Cideraking. In: Ferented beverage production, A.G.H. Lea and J.R. Piggott, Eds., Blackie Acadeic and Professional Press: London, 1995, pp. 67 96. 12. Luedeking, R. and Piret, E.L., Kinetic study of the lactic acid ferentation: Batch process at controlled ph. J. Biol. Microbiol. Technol. Eng., 1959, 1(4), 393 424. 13. Marín, M.R., Alcoholic ferentation odeling: current state and perspectives. A. J. Enol. Vitic., 1999, 5(2), 166 178. 14. Ostergaard, S., Olsson, L. and Nielsen, J., Metabolic engineering of Saccharoyces cerevisiae. Microbiol. Mol. Biol. Rev., 2, 64, 34 5. 15. Ough, C.S. and Aerine, M.A., Methods for analyses of usts and wines. Wiley: New York, 1988, pp. 365. 16. Pretorius, I.S., du Toit, M. and van Rensburg P., Designer yeasts for the ferentation Industry of the 21st Century. Food Technol. Biotechnol., 23, 41(1), 3 1. 17. Pretorius, I.S., Gene technology in wineaking: new approaches to an ancient art. Agri. Consp. Sci., 21, 66, 1 2. 18. Schütz, M. and Gafner, J., Lower fructose uptake capacity of genetically characterized strains of Saccharoyces bayanus copared to strains of Saccharoyces cerevisiae: A likely cause of reduced alcoholic ferentation activity. A. J. Enol. Vitic., 1995, 46(2), 175 18. 19. Sinclair, C.G. and Kristiansen, B., Ferentation kinetics and odelling. Taylor & Francis: New York, 1987. 2. Speers, R.A., Rogers, P. and Sith, B., Non-linear odeling of industrial brewing ferentation. J. Inst. Brew., 23, 19(3), 229 235. 21. Vidrih, R., Hribar, J., Synthesis of higher alcohols during cider processing. Food Cheistry, 1999, 67, 287 294. 22. Wang, L., u, Y., Zhao, G. and Li, J., Rapid analysis of flavor volatiles in apple wine using headspace solid-phase icroextraction. J. Inst. Brew., 24, 11(1), 57 65. (Manuscript accepted for publication Noveber 24) 346 JOURNAL OF THE INSTITUTE OF BREWING