Impact of using yeast derivatives on the development of atypical aging in wines

Yeast derivatives (YDs) are commercial products obtained from yeast colonies grown in bioreactors then inactivated and dried. They are employed in winemaking to supply nitrogen (N), enhance yeast growth and remove unwanted compounds. Since glucose is the primary carbon and energy source for yeast growth, agricultural byproducts are commonly used in their production. Atypical aging (ATA) is a wine aroma defect arising from the oxidation of indole-3-acetic acid (IAA) to 2-aminoacetophenone (AAP). It entails the appearance of unpleasant odours as well as the rapid loss of varietal aromas. As IAA is one of the main plant hormones and YDs are often manufactured using material of plant origin, it can be assumed that AAP precursors are contained in those commercial formulations. If this is the case, the use of YDs could potentially increase the risk of obtaining ATA-tainted wines. To explore this hypothesis, 28 YDs were screened for the presence of AAP-related compounds, as well as their amino acidic content. Since several ATA precursors were detected, fermentation experiments were carried out in which different amounts of YD were added to different grape musts. Interestingly, depending on the nature of the commercial formulation, AAP development was observed to be either enhanced or diminished. More specifically, as well as concentration of AAP and precursors, the presence of cell walls within the YDs seemed to have an influence on the development of this defect. Furthermore, as diammonium phosphate (DAP) is also used to provide grape must with N, the effect of supplementing must with DAP was investigated. The addition of increasing doses of DAP did not significantly affect the accumulation of AAP. These results are very promising for both winemakers and YD producers, and they highlight the need for further research on the use of YDs in oenology.


INTRODUCTION
Yeast derivatives (YDs) are obtained from colonies of Saccharomyces and/or non-Saccharomyces spp, which are grown on sugar-rich media and thermally inactivated before autolytic, plasmolytic and hydrolytic techniques and subsequent drying are applied in order to produce commercial formulations (Münch & Schieberle, 1998;Nagodawithana, 1992).YDs represent a heterogeneous family of products and, depending on how they are produced, are grouped into different classes: inactivated yeast (thermally or pH degraded and dried), yeast autolysate (from biomass autolysis), yeast cell wall (insoluble part without the cytoplasmic portion) and yeast protein extracts (soluble fraction, mainly cytoplasm) (Ángeles Pozo-Bayón et al., 2009;Rigou et al., 2021).While this classification reflects the variety of products that can be obtained from the bacterial biomass, each manufacturer has its own formulation, which can include one or more classes along with other compounds, such as vitamins and minerals.Furthermore, the compositional parameters of each product are highly influenced by the yeast strain used, the conditions in which the culture is grown and the manufacturing process itself (Guilloux-Benatier & Chassagne, 2003;Klis et al., 2006;Rigou et al., 2021).
YDs are generally employed to provide assimilable nitrogen, stimulate yeast growth and prevent stuck fermentations (Ángeles Pozo-Bayón et al., 2009).Moreover, they can provide protection against chromatic and sensorial faults (Andújar-Ortiz et al., 2014) and enhance colloidal stability due to their mannoproteins content (Morata et al., 2019;Pérez-Magariño et al., 2015).Additionally, their ability to bind unwanted molecules and their capacity for delivering volatile compounds to the wine have recently been demonstrated (Comuzzo et al., 2006).Industrially, YDs are produced in bioreactors, where yeast colonies are grown in large amounts and transformed into the final products.As glucose is the main carbon and energy source required for the growth of yeasts, the substrates utilised for biomass production typically include agricultural and food waste by-products.More specifically, starch, molasses, wood, fruits and vegetable waste are among the most common raw materials employed in their manufacture (Bekatorou et al., 2006;Jay, 1992).
Atypical aging (ATA) is a sensory defect that occurs in young white wines and is described as a loss of varietal aroma and the advent of unpleasant notes reminiscent of naphthalene, dirty cloth and acacia.Although other compounds might be involved in ATA formation, research has identified 2-aminoacetophenone (AAP) as the main chemical and sensory descriptor (Schneider, 2014).AAP arises from the oxidation of indole-3-acetic acid (IAA), a natural hormone present in plants.As IAA is responsible for regulating several biological mechanisms, its availability is controlled via methylation and/or binding with amino acids and sugar moieties.During vinification, the bound IAA contained in must is freed up by yeast, which uses the amino acids for its metabolism (Simat et al., 2004).Subsequently, free IAA is released into the wine and upon sulphoration a cascade of chemical reactions can lead to the formation of AAP (Hoenicke et al., 2002b).Additionally, yeast has recently been demonstrated to be capable of the de novo synthesis of IAA and AAP (Álvarez-Fernández et al., 2020).
Considering that YDs can be manufactured using byproducts of plant origins as raw material, it is plausible that they contain both free and bound IAA.Moreover, the production of YDs may involve thermal treatments; thus, the IAA present could be converted into AAP (Christoph et al., 1999).If this is the case, using YDs to enhance the quality of wine would also increase the likelihood of obtaining ATA-tainted products.Hence, the adoption of YDs in winemaking might represent a double-edged sword.To test this hypothesis, 28 commercially available YDs were specifically screened for the presence of AAP and some ATA precursors.The samples were also profiled according to amino acid (AA) content and compared for their yeast available nitrogen (YAN) contribution.In order to understand the effect of using YDs on ATA formation, fermentations were carried out using varying amounts of YDs and measuring the AAP content in the finished wines.Additionally, the effects of supplementing different grape musts with YD and the use of diammonium phosphate (DAP) prior to fermentation were also evaluated in relation to AAP development.

Analysis of AAP and ATA precursors
The AAP and ATA precursors were quantified using a method developed by Nardin et al. (2022).In brief, an ultra-high performance liquid chromatographer (UHPLC; Thermo Ultimate R3000, Thermo Scientific; Waltham, MA, USA) coupled to a high-resolution mass spectrometer (HRMS; Q-Exactive; Thermo Scientific; Waltham, MA, USA) equipped with a heated electrospray ionisation chamber (HESI-II) was operated in positive ionisation mode.Details regarding the standards are given in Supplementary Table S1.
AAP and ATA precursors present in both the YDs and the wines at the end of the alcoholic fermentation (AF) were quantified.For each of the YDs, 1 g of dry powder was put in a falcon tube and dissolved by adding 50 ml of deionised water (Arium Pro Lab Water System; Sartorius AG, Germany).Then, after placing the tube in a mechanical shaker (Multi Reax; Heidolph, Germany) for 30 min, it was centrifuged (5000 rpm, 5 min, 4 °C) and the content was filtered directly into HPLC vials using 0.45 μm PTFE filters.For each YD, the extraction process was performed in triplicate.The samples of the wines were centrifuged and then filtered following the same procedure used for the YDs.

Amino acids quantification in YD samples
The method developed by Delaiti et al. (2024) was used to measure the AAs content of the YDs.Briefly, the method entailed the use of a UHPLC-HRMS system with an ESI-II interface.The column was a Raptor Polar X (Bellefonte, US) and the mobile phase consisted of water (Arium® 98  YD1 and YD2: yeast derivatives used in the fermentation experiments. Pro Lab Water System; Sartorius AG, Goettingen, Germany) with 0.5 % formic acid (Honeywell; Charlotte, NC, USA) and acetonitrile (Honeywell; Charlotte, NC, USA).Samples were prepared following the same procedure reported above.Details regarding the standards used are given in Supplementary Table S2.
The alcoholic fermentation (AF) was carried out at a constant temperature of 18 -19 °C and monitored by checking the sugar levels (° Brix readings) every three days.At the end of the AF, 70 mg/L of SO 2 were added to each tank and 50 ml of each sample were stored at -20 °C until analysis.
3.2.Experimental set-up for the study of the impact of DAP supplementation and the matrix effect on AAP production.
Ten grape musts obtained from different grape varieties were selected and fermented in triplicate.30 L of each matrix were divided into three batches, inoculated with 200 mg/L of Vin 13 (Oenobrands; Montferrier-sur-Lez, France), and supplemented with 100 (n = 10), 250 (n = 10) and 400 (n = 10) mg/L of DAP.Fermentation was carried out at a constant temperature of 18 -19 °C and monitored by checking the sugar levels (° Brix readings).At the end of the AF, 70 mg/L of SO 2 were added to each batch and 50 ml of each sample were stored at -20 °C until analysis.
To study the matrix effect, another experiment was set up.Ten grape musts from different grape varieties were selected and fermented in triplicate.30 L of each matrix were inoculated with 200 mg/L of Vin 13 (Oenobrands; Montferrier-sur-Lez, France), divided into three batches and then supplemented with 400 mg/L of YD1.Fermentation was carried out at a constant temperature of 18 -19 °C and monitored by checking the sugar levels (° Brix readings).At the end of the AF, 70 mg/L of SO 2 were added to each batch and 50 mL of each sample were stored at -20 °C until analysis.

Optical densities measurements
To assess the differences in the cell wall content of the YDs, optical density measurements were carried out.Eight solutions containing increasing doses (200, 400, 600 and 800 mg/L) of two distinct YDs (YD1 and YD2) were prepared by accurately weighing the dry powders and diluting them in distilled water.Their optical densities were measured using a spectrophotometer (UV1700; Yoke Instrument, Shanghai, China) set at 660 nm.

RESULTS AND DISCUSSION
1. YD screening for AAP and ATA precursors 28 YDs (Table 1) were screened for the presence of AAP and some ATA precursors.As shown in Figure 1 and Supplementary Table S3, considerable amounts of those compounds were found in most of the samples under examination (Figures 1A and 1C).Furthermore, great compositional variability was observed (Figures 1B and 1C).
TRP and TRH were the most abundant compounds detected in the samples.With a mean value of 13356 mg/kg, TRP concentrations ranged from 1.74 to 41491 mg/kg and in 13 of the 28 samples they were lower than 7000 mg/kg.Regarding TRH, more than half (17 out of 28) of the YD samples contained less than 7 mg/kg, and just one sample contained a conspicuously high concentration (462 mg/kg, sample 15).The IAN values ranged between 0 and 56.26 mg/kg; however, most (20 out of 28) of the YD samples contained less than 10 mg/kg.Meanwhile, regarding ILA, the vast majority of the samples (23 out of 28) showed variable amounts of the compound, but they were all below 2 mg/kg.While TAM concentrations ranged between 0 and 18.89 mg/kg, they were lower than 1 mg/kg in 21 of the 28 YD samples.Only trace amounts of IAM were detected (< 0.37 mg/kg).In contrast to the bound and methylated forms of IAA, which were found in minute concentrations, unbound IAA was detected in 23 samples, with concentrations ranging from 0.07 to 4.48 mg/ kg and a mean value of 1.89 mg/kg.Finally, low amounts of AAP (0.003 -0.05 mg/kg) were also found in all but two of the YD samples; the presence of those compounds can be ascribed to normal yeast metabolism (Álvarez-Fernández et al., 2020;Simat et al., 2004).Furthermore, AAP might result from the YD production process, which can entail a thermal treatment (e.g., in the case of inactivated yeast) leading to the degradation of IAA to AAP (Münch et al., 1997).

Amino acid profiling
The 28 YDs (Table 1) were screened for their amino acid (AAs) and ammonium (NH 4 ) content (Supplementary Table S4).As shown in Figure 2, high variability among the samples was found: this was as expected, as the processing techniques and commercial formulation vary for each company and product.
With a mean value of 26828 mg/kg, NH 4 concentrations ranged from 66 to 178863 mg/kg.However, while only trace amounts were observed in most of the YDs, NH 4 was the most abundant compound in samples 1, 11, 17, 18, 19, 21 and 23, constituting around 65 % of the dry weight (DW).
As regards AAs, on average, the YDs analysed mostly contained glutamic acid (glu), leucine (leu), valine (val) and alanine (ala), with concentrations representing about 10 % of the DW, followed by isoleucine (ile), ornithine (orn) and phenylalanine (phe), constituting 5 % of the DW.The other amino acids were present in varying but small quantities.
The nitrogen (N) fraction available for use by the yeast (known as the yeast available nitrogen (YAN)) was calculated from the results of the AA screening (Verdenal et al., 2021).
The computation was carried out multiplying the ratio of the concentration of each primary amine and ammonium ion over their corresponding atomic mass unit by the atomic weight of N. The trendline in Figure 2B reports the concentrations of YAN calculated for each YD.

Effects of using YDs on fermentation kinetics
To determine the effects of using a YD on the development of AAP, a fermentation trial was set up.Of the 28 previously analysed samples (Table 1), two were chosen that met the criteria of the presence of large and small quantities of AAP and ATA precursors: sample 10 (YD1) and sample 18 (YD2).
Upon inoculation with yeast and YD supplementation, the fermentation kinetics were monitored.They were found to vary depending on the treatment (Figure 3).Fermentation was slower in the control sample (with no YD present) than in the other treatments.This was expected, as supplementing grape must with YAN is known to improve fermentation performance (Seguinot et al., 2020), and the grape must used for the experiment was N deficient (YAN: 94 mg/L) (Verdenal et al., 2021).Furthermore, this confirms the results of Simat et al. (2004), which showed that nutritional supplements added to must accelerate fermentation.
Although the use of DAP slightly improved fermentation rate, it was not observed to be as effective as using YDs, both YD1 and YD2 being found to dramatically speed up fermentation.Aside from the yeast cell content and cell walls, YDs may also contain other compounds (such as vitamins and minerals) which can affect yeast metabolism and thus the metabolites formed during fermentation.In contrast to YD2, fermentation with YD1 at different concentrations led to a higher pace in the vinification process.Additionally, the fermentations rates increased with increasing amounts of YD1.
When the same dose (400 mg/L) of two different YDs was used, the kinetic rates were significantly different: YD1 addition resulted in a much faster fermentation than YD2 addition.Besides the distinct AA composition and NH 4 content, it was noted that the YAN supplemented with YD1 was higher (36.9 g/kg) than that provided by YD2 (3.06 g/ kg).However, the N contribution of the two YDs was lower than that of DAP.This highlights that despite AAs, NH 4 and YAN play an important role in fermentation (Gobert et al., 2019;Torrea et al., 2011), they are not the only responsible for the nutritional capacity of yeast and other factors might be involved.In this regard, supplementation with fatty acids (Pinu et al., 2019;Redón et al., 2009), as well as the bioavailability of certain metal ions (Birch et al., 2003) both potentially influenced by YD addition (Rigou et al., 2021) -have been shown to have an impact on fermentation performance.

FIGURE 3.
Fermentation kinetics of a non-varietal grape must inoculated with different concentrations of two selected yeast derivatives (YDs) and diammonium phosphate (DAP).

Effects of the use of YDs on AAP formation
Figure 4 is a graphical representation of the mean and standard deviations of the AAP amounts detected in the wines at the end of the AF.Adding the same dose (400 mg/L) of two different YDs before fermentation led to the formation of inconsistent amounts of AAP.Indeed, the addition of YD2 was observed to produce significantly lower concentrations of this compound compared to YD1 (0.103 ± 0.009 vs. 0.173 ± 0.018 μg/L).While this agrees with the higher amounts of AAP and ATA precursors delivered with YD1, it also highlights the significance of the YD composition on the development of this sensorial defect.
Moreover, the formed AAP was found to increase linearly with increasing doses of YD1, and a positive correlation (R 2 = 0.98) was observed.This underlines how the addition of increasing amounts of a YD rich in AAP and ATA precursors is closely correlated with AAP development during AF.Compared to not using YDs, must supplemented with 400 mg/L of YD2 was not found to have any consequences in terms of AAP accumulation (0.103 ± 0.009 vs. 0.111 ± 0.002 μg/L).It was assumed that the AAP and ATA precursors contained in this preparation were not sufficient to produce a consistent effect on the buildup of AAP in the finished wines.The quantity of AAP in the wine made from must supplemented with DAP (200 mg/L) did not differ from that obtained using YD1 at the same concentration (0.131 ± 0.002 and 0.132 ± 0.015 μg/L respectively).Additionally, a double dose of YD2 (400 mg/L) led to an accumulation of significantly lower amounts of AAP compared to the use of DAP (0.103 ± 0.009 vs. 0.131 ± 0.002 μg/L respectively).Finally, Figure 4 also shows the YAN supplemented with each treatment.Interestingly, AAP formation was not found to be dependent on the N status of the grape must prior to fermentation.

Effects of DAP supplementation on AAP development
Together with YDs, the addition of diammonium phosphate (DAP) represents one of the most common tools in winemaking when N deficiency poses a threat to fermentation (Bell & Henschke, 2005).To date, research on the impact of the pre-fermentative addition of DAP on ATA development has only been touched upon, and no clear effect has yet been demonstrated.While some authors (Bach, 2005;Schwab et al., 1996) have not found any effects, Rauhut et al. (2003) established that DAP supplementation can actually diminish the formation of AAP.Schneider et al. (2014) hypothesised that this is due to the quenching of oxygen radicals through the action of metabolites generated during the fermentation process; if this is the case, the extent of the effect would be dependent on the dose of DAP added.These results are not conclusive and they emphasise the importance of further investigating the impact of DAP on ATA development.In the first fermentation trial of this study, compared to the control, the use of DAP slightly increased the production of AAP (0.131 ± 0.002 vs. 0.111 ± 0.002 μg/L; Figure 4).However, the increase was very limited (non-significant ANOVA test).The trendline represents the correlation between the increasing amount of YD1 and the AAP detected in the finished wines.
Simone Delaiti et al.Given the uncertainty associated with the use of DAP, another fermentation experiment was set up.As can be seen in Figure 5 (and Supplementary Table S5), increasing doses of DAP (100, 250 and 400 mg/L) added to 10 different grape juices prior to fermentation did not produce any effect on the accumulation of AAP in the finished wines.

Comparison of added versus detected AAP concentrations
As reported above, AAP was detected in most of the analysed YDs (Figure 1, Supplementary Table S3).This indicates that, even though the amounts are limited, supplementing the grape must with such products poses a considerable risk of producing ATA-tainted wines.Additionally, AAP can originate from the oxidation reaction of IAA (Hoenicke et al., 2002a), and yeast is capable of a de novo formation of both compounds (Álvarez-Fernández et al., 2020;Liu et al., 2016).
To investigate the AAP generated solely during fermentation, aside from the AAP contained within the YD itself, the theoretical and actual concentrations were considered.
FIGURE 6. Differences between the AAP supplemented with two yeast derivatives (YD1 and YD2) and actual AAP detected in the wines at the end of alcoholic fermentation, expressed as a percentage.
The theoretical concentrations were calculated based on the values reported in Supplementary Table S3 and computed as follows: [AAP] x dose.The differences between the added and measured AAP values, expressed as a percentage, are shown in Figure 6, where 0 % represents the control (i.e., no YD used).Compared to the supplemented AAP, the use of YD1 at different concentrations (200, 400 and 600 mg/L) was associated with a 16 %, 36 % and 44 % (respectively) higher AAP content in the final wines.Furthermore, a positive trend between the amounts of added YD1 and developed AAP was observed.This is in agreement with the generation of AAP as a result of yeast action during fermentation, and the conversion into AAP of ATA precursors present in the juice (Hoenicke et al., 2002b) and -as in this case -supplemented with YDs.However, the opposite was observed for YD2: compared to the theoretical supplemented amount, there was an 8 % decrease in the AAP content of the wine.This reduction might be due to the formulation of this commercial product.Indeed, according to the specification sheet of the YDs, YD1 mainly constitutes yeast autolysate and YD2 inactivated yeast.The presence of yeast cell walls in YD2 might have contributed to the adsorption of some of the compounds involved in ATA formation (Lubbers et al., 1994;Mazauric & Salmon, 2006).To further explore this hypothesis, an estimation of the number of cells contained in the two previously used YDs was performed.As the most common method used in biotechnology entails the measurement of the optical density (OD 660) of a cell suspension (Hulst & van de Hulst, 1981;Wood & Krieg, 1993), eight solutions containing respectively 200, 400, 600 and 800 mg/L of YD1 (n = 4) and YD2 (n = 4) were prepared and the OD 660 was evaluated.As shown in Figure 7, increasing concentrations of YD2 corresponded to a rise in OD 660 values.Conversely, the OD 660 measured for YD1 did not correspond to the dose.This may indicate that a considerable fraction of YD2 is actually made up of cell walls, thus accounting for the adsorption mechanism involved in the reduction of the AAP content mentioned above.While this might be a promising result regarding finding a solution to counteract the appearance of UTA in wines, further research is needed to better investigate the biochemical mechanisms implicated.
FIGURE 7. Optical densities at 660 nm (OD660) measured for solutions containing increasing doses of yeast derivatives (YDs).

Influence of the matrix on AAP development when fermenting with YDs
As ATA formation can occur via the oxidative reaction of the precursors contained in the grape must, the composition of the matrix is a pivotal factor in determining the onset of this defect (Schneider 2014).Indeed, not only do the varietal notes mask this defect (Alpeza et al., 2021), but the vintage year also plays a major role in its onset (Delaiti et al., 2022).
In this study, it has been demonstrated that the addition of YDs prior to fermentation can affect the rate of AAP formation, sometimes either enhancing or suppressing the formation of this compound.In this regard, it has been speculated that the outcome mainly depends on the type of YD used and its composition.These findings contribute highly to the understanding of ATA development, but as they were obtained using a single grape must, a new fermentation experiment involving the use of different grape juices was carried out to validate the robustness of the results.
Ten different grape musts were inoculated with 400 mg/L of YD1 and the AAP content at the end of the AF was measured.As can be observed in Figure 8 (Supplementary Table S6), the previously obtained results were confirmed: overall, the likelihood of obtaining ATA-tainted wines increased with the pre-fermentation addition of a YD that was rich in AAP and ATA precursors and poor in terms of the cell wall component, regardless of must type used (ANOVA test).

CONCLUSIONS
The aim of this study was to assess the use of YDs in winemaking and the possibility of obtaining ATA-tainted wines.A screening of 28 YDs was first performed, and it was demonstrated that both AAP and ATA precursors are contained in those products in variable amounts.In addition, the impact of using YDs during fermentation was explored and it was found that AAP development can be enhanced or reduced depending on the specific YD used.It was hypothesised that, as well as the AAP and ATA precursors contained in the commercial formulation, the presence of other compounds, such as cell walls, might exert an adsorption effect, thus affecting the AAP content in the finished wine.Nonetheless, a YD rich in ATA precursors but poor in terms of the cell wall component was found to lead to a dose-dependent accumulation of AAP.As grape must composition is known to play a crucial role in the development of the ATA-taint, the applicability of these findings was investigated by fermenting several different grape musts supplemented with the same dose of YD.Interestingly, it was found that regardless of matrix composition, the use of YDs consistently affects the accumulation of AAP.Lastly, as DAP is also used to provide grape musts with N, the effect of DAP supplementation was also studied.The addition of increasing amounts of DAP (up to 400 mg/L) was not found to have a significant effect on the accumulation of AAP.
In conclusion, these findings are of great relevance for both YD producers and winemakers.They prove some of the consequences of YD use and highlight the need for further research on the topic.Indeed, a better understanding of the relationship between the raw material used in YD production and the ATA-precursors contained in the commercial products could aid in manufacturing products less likely to develop sensorial defects.Moreover, given their important role in the onset of ATA, the physio-chemical attributes of YDs should also be thoroughly studied.

FIGURE 4 .
FIGURE 4. Statistical distribution (ANOVA test) of the 2-aminoacetophenone (AAP) content and yeast available nitrogen (YAN, orange squares) added to grape musts fermented with and without the addition of diammonium phosphate (DAP) and different concentrations of two yeast derivatives (YDs).

FIGURE 5 .
FIGURE 5. Statistical distribution (ANOVA) of the 2-aminoacetophenone (AAP) content of 10 finished wines obtained from different grape juices fermented with increasing amounts of diammonium phosphate (DAP).

FIGURE 8 .
FIGURE 8. Statistical distribution (ANOVA) of the 2-aminoacetophenone (AAP) content of 10 finished wines obtained from different grape juices fermented with and without the addition of YD1 (400 mg/L).