Introduction

Fermentation-derived aroma compounds are one of the defining features of wine quality. Primary aroma and flavour compounds such as acetic acid, ethanol and glycerol are largely dependent on the quality and amount of carbon in the starter must and on fermentation conditions. Nonetheless, the jury is still out on the foundation of some distinctive flavours and aromas usually associated with particular varieties and locales (Jackson, 2014). Scientists and winemakers have long suspected that yeast nutrition, along with other factors, plays a crucial role in determining the chemical profile of wine aroma. Nitrogen concentration in the grape must has been identified as one nutritional factor crucial in aroma biosynthesis during winemaking.

The important nitrogen in grape must, referred to as yeast assimilable nitrogen (YAN), determines yeast growth, development and subsequently aroma production. As a result, addition of diammonium phosphate (DAP) as a nitrogen supplement is now standard practice in most cellars. Yeast can meet its nitrogen needs by breaking down proteins, and by using amino acids and peptides present in the must. Amino acids are a significant source of YAN, but are not all used at the same rate. Depending on how readily yeast metabolise the amino acids, they can be classified as preferred, intermediate, non-preferred and non-utilised as shown in Table 1. Furthermore, their use by yeast will result in the production of various aroma compounds, and many amino acids are direct precursors of specific aromatic features. This report is a summary of a study that investigated the influence of individual amino acids and amino acid classes on the aroma profiles of synthetic grape must (SGM) fermented by two industrial Saccharomyces cerevisiae strains.

The goal was to determine the effect of individual amino acids as the sole source of YAN on the growth and aroma profile, and consequently determine the predictability of aroma using various nitrogen concentrations and sources under standard wine fermentation conditions. This is essential for the further optimisation of yeast nutrition, because amino acids form the largest group of YAN compounds (Bell & Henschke, 2005) and their metabolism is directly related to many flavour compounds, and thus the ultimate quality of wine (Styger et al., 2011). As such, it is worthwhile to assess the amino acid and aroma relationship to improve forecasting of wine aroma by using grape must characteristics. Knowledge of such relationships will enable winemakers to make more informed decisions on how to proceed with the fermentation process.

Growth kinetics

Two Saccharomyces cerevisiae strains were used in this study to ferment synthetic grape must. The effect of individual amino acids shown in Table 1 on growth was assessed for the two yeast strains using yeast nitrogen base media (YNB) containing 10% (w/v) glucose at a pH of 3.8. Only one of the amino acids was added to give a YAN concentration of 150 mg L-1 and YNB with only ammonium used as the control. Optical density (representing biomass production), duration of lag phase were used as indicators of growth. The slope of the exponential growth phase was used to calculate the growth rate.

Fermentation kinetics

The defined grape must was used to determine the aroma production pattern of the two yeasts when presented with single amino acids as nitrogen sources. Just like with the growth experiment, amino acids were added to the media to contribute a YAN of 150 mg L-1, but the media were fermented to dryness, i.e. until sugar levels were less than 4g L-1. At the end of fermentation, final biomass was quantified to determine the effects of different amino acids on biomass production. Productivity showed notable variations between different amino acids. For both strains, the preferred amino acids, glutamine, glutamate, alanine aspartate, arginine and serine exhibited high growth rates which were however not directly correlated to biomass accumulation. For instance, phenylalanine had a final biomass of 2.87 g L-1, but had a comparatively low growth rate of 0.15 as shown in Figure 1. Nevertheless, the preferred amino acids were efficiently utilised in comparison to the other classes and had higher growth rates. The preferred amino acids also led to the shortest time to complete fermentation.

winearoma_thumbnail
FIGURE 1. A comparison of mean biomass accumulation (A) and growth rate (B) of Strain A and Strain B cultured in media with only one amino acid or ammonium as the sole source of nitrogen. Sample text FIGURE 2. Relationship between BCAAs and related volatile metabolite produced by VIN13 during fermentation. TABLE 1. Amino acid classification based on the efficiency of utilisation by wine yeast (Ljungdahl & Daignan-Fornier, 2012). The classes are important in predicting the wine aroma when single amino acids are used as sole sources of nitrogen. FIGURE 3. Effect of amino acids classes (YAN = 200 mg/L) on the accumulation of major volatile compounds and the predictability of BCAAs-linked aroma compounds.

The amino acid consumption pattern is broadly consistent for both strains, especially for the preferred amino acids, with the exception of arginine. There was no significant difference in the final biomass accumulation between the strains as is shown in Figure 1 except for fermentation driven by arginine, leucine, isoleucine, tyrosine and threonine which had strain B producing more cells at the end of fermentation. From this it can deduced that strain B is a better fermenter than strain A given that they had similar media to begin with.

These data highlight the fact that total YAN only has limited potential to predict the fermentation performance of yeast. Indeed, two musts with the same YAN, but very different amino acid composition, and in particular a different ratio of preferred vs. non-preferred amino acids, will lead to different fermentation dynamics.

Aroma production

Aroma production was determined by using gas chromatography coupled with a flame ionisation detector. The aroma compounds were extracted using ethyl acetate and a total of 32 compounds were analysed. The volatile compound production was more driven by the amino acids used as nitrogen source than by the yeast strain. With the exception of acetic acid and ethyl acetate, aroma compounds were positively correlated to their related amino acid. For instance the two related amino acids, leucine and isoleucine resulted in aroma profiles dominated by isoamyl alcohol (banana), isoamyl acetate (fruity) and isovaleric acid (cheese), compounds whose formation is directly linked to leucine and isoleucine metabolism. A similar trend is observed for phenylalanine and its corresponding volatiles; 2-phenylethanol (floral) and 2-phenylethyl acetate (rose), and valine which increased the synthesis of isobutanol (solvent) and isobutyric acid (rancid). Acetic acid (vinegar) and ethyl acetate (sweet, nail polish remover) appeared positively correlated across for all amino acids and both strains showed no pronounced difference.

Further experiments were conducted with only strain A to investigate the effect of increasing branched chain and aromatic amino acid (BCAA) concentration of the quantity of the resultant aroma metabolites. In these studies SGM contained 100 mg N L-1, 200 mg N L-1 and 300 mg N L-1 YAN provided by individual amino acids. The total YAN was always 300 mg N L-1 and in cases where amino acids were used at lower rates, the YAN was supplemented with ammonium.

Taking 2-phenylethanol, isobutanol, isoamyl acetate (banana) and isoamyl alcohol (banana) to represent volatiles from metabolism of phenylalanine, valine, leucine and isoleucine respectively, Figure 3 shows the strong and positive relationship (R2 is greater than 0.99 for all compounds) between the amino acids and resultant volatiles. Similar relationships were also observed for the other related compound and lead to the development of the theory that if all the BCAAs are added simultaneously to the SGM, all compounds associated with the amino acids included will be positively altered.

To validate the predictability of the linear relationship between BCAAs and aroma compound, another experiment was designed to test the model. Four classes of amino acids (preferred, BCAAs, non-utilised and non-preferred) were used as nitrogen source at a concentration of 200 mg N L-1 and the rate of volatile compound production was compared with the predicted values for BCAAs (Figure 3). Of all the volatiles analysed, only isobutanol and 2-phenylethanol had high predictability. For all other compounds (isobutyric acid, propionoic acid, butanol, 2-phenylethyl acetate, decanoic acid, and valeric acid) there were no similarities between the observed and the predicted concentration.

Thus, predictability of aroma production is possible when individual amino acids are used for the provision of YAN. On the contrary, the predictability is reduced the more complex the nitrogen source. The reason lies with the nature of the metabolic networks involved in aroma production, since many of the amino acids share metabolic enzymes and precursors. In a complex mix, the competition between substrates, the variable preferences of enzymes and the many indirect drivers such as the need to maintain a healthy redox balance result in a more “chaotic” output. However, this study provides baseline data from which to extend further analysis to link amino acid nutrition to yeast aroma production. A better understanding of this relationship is indeed necessary to optimise the use of nitrogen addition to grape must.

Conclusion

Generally, yeast utilisation of amino acid classes is similar across the two strains used in this experiment. The type of nitrogen source played a major role in fermentation kinetic and aroma production, and it would clearly be in the interest of winemakers to be able to monitor levels and types of amino acids in the must before fermentation in order to adjust nitrogen not only with regard to total YAN, but also with regard to potential aromatic impact. The BCAAs are particularly important as they have direct bearing on the production of many major aroma compounds that can either add character or render the wine unpalatable. This study provides a good baseline investigation from which to further explore the effect of more complex nitrogen treatments. Since cultivar and regional differences have been shown to also lead to differences in amino acid present in grape must, this study also contributes to the body of knowledge attempting to explain the basis of varietal and regional wine characteristics.

Acknowledgements

The researchers are grateful to Winetech, NRF and THRIP for the funding provided to conduct this study.

– For more information, contact Florian Bauer at fb2@sun.ac.za.

References

Bell, S.J. & Henschke, P., 2005. Implications of nitrogen nutrition for grapes, fermentation and wine. Australian Journal of Grape and Wine Research 11(3), pp. 242 – 295. Available at: <Go to ISI>://WOS:000233851200001.

Jackson, R.S., 2014. Wine Science 4th ed., Amsterdam: Elsevier Inc.

Ljungdahl, P.O. & Daignan-Fornier, B., 2012. Regulation of amino acid, nucleotide, and phosphate metabolism in Saccharomyces cerevisiae. Genetics, 190(3), pp. 885 – 929.

Styger, G., Jacobson, D. & Bauer, F.F., 2011. Identifying genes that impact on aroma profiles produced by Saccharomyces cerevisiae and the production of higher alcohols. Applied Microbiology & Biotechnology 91(3), pp. 713 – 730.

You may like to read these:

Go Back
Shares