Flavor instability of beer is mainly caused by increasing aldehyde concentrations in freshly bottled beer. Since the preceding study observed the release of aging-relevant aldehydes, this study focused on de novo formation pathways. In the following section, the oxidation of higher alcohols, the Maillard reaction, and the Strecker degradation are discussed because of their contribution to Strecker aldehyde formation during beer aging.
3.1. Oxidation of Higher Alcohols
The first pathway studied is the oxidation of higher alcohols. As already introduced, these compounds can be oxidized in the presence of reactive oxygen species, forming their corresponding aldehydes [
2]. The degradation is an alternative pathway for Strecker aldehyde formation.
Figure S1 shows the concentrations of 2-methylpropanol (2MP precursor), 2-methylbutanol (2MB precursor), and 3-methylbutanol (3MB precursor) during natural and forced beer aging at two barley varieties (B1 and B2), each with three different malt modification levels (P1 < P2 < P3).
There was no consistent trend in the behavior of natural beer aging at all variations. Therefore, the most constant trend did not indicate an oxidation reaction. The oxygen level was below 0.1 mg/L in fresh beer samples at all variations [
1]. Thus, the oxidation reaction was inhibited and no contribution to the aging aldehyde formation could be observed. However, only the P1 levels of 2-methylbutanol and 3-methylbutanol at B2 showed a significant reduction during aging. The degradation could indicate the oxidative formation of 3MB or alternative degradation pathways of 3-methylbutanol favored such as esterification, acetal formation, or elimination reaction. Further, the aging of B2P1 is more influenced by oxygen because it showed the lowest precursor levels (
Figure 2), as well as the lowest SO
2 levels [
1], but oxygen concentrations were low, compared with the other variations in this study. Regarding the effect of malt modification, there was also no consistent trend observable. Surprisingly, the medium malt modification level (P2) showed the highest concentrations for 2-methylpropanol and 2-methylbutanol. The higher alcohol concentration measured in fresh beer is mainly formed during fermentation from their corresponding amino acids via the Ehrlich pathway [
28]. Although the highest amino acid concentrations of valine, leucine, and isoleucine were found in the highest malt modification level (P3) after boiling [
19], a differentiation of the corresponding higher alcohols due to the order in boiled wort (P1 < P1 < P3) was not detected. That confirms that the Ehrlich pathway is affected by various parameters such as yeast strain, pH value, aeration, or additional glucose content, which was not monitored in this study, in boiled wort [
28,
29]. Exemplarily, 3-methylbutanol showed no trend within the different malt modification levels.
Generally, there was no consistent indication for oxidative Strecker aldehyde formation during beer aging for the observed analytes in this study, which confirms the literature study on wine [
11], as already elucidated in the Introduction Section.
3.2. Reactivity of Precursor Compounds in the Maillard Reaction
The possible precursor compounds for flavor-active Strecker aldehydes are primarily the corresponding amino acids and dicarbonyl compounds. The latter can be formed because of direct caramelization, or indirectly via the Amadori product degradation as part of the Maillard reaction from carbohydrates [
5].
Figure 2 shows the free amino acids depending on the malt modification level (P1 < P2 < P3) for two barley varieties (B1 and B2) during natural and forced aging. Valine, methionine, leucine, isoleucine, and phenylalanine were selected because their corresponding Strecker aldehydes are sensory-active beer aging indicators. Furthermore, lysine is shown to be an important reactive and often modified amino acid within the Maillard reaction.
The highest contents of the selected amino acids were determined at the highest malt modification level. The observed effect confirmed the ratio of the amino acids in the final wort [
19]. Although amino acids were metabolized during yeast fermentation [
9], their ratio between different malt modification levels remained constant. It is evident that this ratio differed between the observed amino acids. Thus, the difference in free lysine concentration between the lowest and highest malt modification level in fresh beer was at factor 20, whereas methionine and valine concentrations showed factor 2–5. This may be due to biological differences in protein composition and enzyme activity of barley varieties [
30]. Generally, an increased malt modification level provides higher reaction potential for
de novo Strecker aldehyde formation due to increased amino acid content.
The forced aging of the beers did not lead to significant changes in amino acid concentrations. This is partly due to the low reactivity of amino acids and, further, the relatively short reaction time of 4 days, compared with natural aging. However, natural aging increased Strecker-active amino acids. This could be explained, on the one hand, by the release from imines or Amadori products and, on the other hand, the degradation of peptides and proteins. Thus, it seems likely that the degradation of Strecker aldehydes occurred simultaneously with the release of amino acids. Therefore, an increasing amino acid concentration during aging could additionally increase the de novo formation potential for Strecker aldehydes.
Additional precursor compounds that can be formed during the Maillard reaction are Amadori products and the resulting dicarbonyl compounds. Strecker aldehydes can be formed from both substance classes [
7]. Exemplarily, the Amadori products of lysine (ML and FL) and the dicarbonyl compounds 3-DG, 3-DGal, and 3-DM are discussed.
Figure 3 shows the mentioned analytes in their free form, depending on the malt modification level (P1 < P2 < P3), at two barley varieties (B1 and B2) during natural and forced aging.
FL ranged from 3.0 to 11.4 mg/L. ML was determined at 5.5–15.0 mg/L. No significant differences could be seen between the barley varieties with regard to the Amadori product concentration. During forced and natural beer aging, there was also no consistent trend of the Amadori product levels. Note that this study focused on ε-terminal Amadori products of lysine (FL and ML). The α-terminal Amadori products of Strecker-active amino acids might undergo degradation and, therefore, can contribute to Strecker degradation [
7]. Regarding the influence of the malt modification level, the highest level mainly showed the highest concentrations of Amadori products. The effect could be caused by the increased concentration of amino acids (
Figure 2), which was already observed during wort boiling [
19]. This fact indicates that the FL and ML levels could have already been increased in the final wort before fermentation.
As main dicarbonyl compounds, 3-DG (7.9–19.5 mg/L), 3-DGal (1.5–4.7 mg/L), and 3-DM (1.5–9.9 mg/L) were found in concentrations comparable to previous studies [
31,
32]. B2 showed slightly higher contents of dicarbonyl compounds. Regarding the 3-DG and 3-DGal content, a decreasing trend during natural beer aging could be observed, as has been previously reported [
15,
31]. This decrease is caused by the degradation of dicarbonyls, indicating the
de novo formation of Strecker aldehydes. Furthermore, dicarbonyl compounds can undergo conversion to 5-hydroxymethylfurfural (HMF), as previously reported during beer aging [
15]. Nevertheless, the formation of both dicarbonyls cannot be excluded during aging. Exemplarily, Rakete et al. observed the formation of 3-DG and 3-DGal by forced aging model experiments of maltotetraose [
31]. Therefore, the degradation trend of both compounds tends to be underestimated. A higher malt modification level showed a consistent increase in 3-DG concentration during natural and forced aging, whereas 3-DGal content was less affected. However, higher 3-DG content could be caused by a stronger formation during wort boiling at higher malt modification levels [
19,
33]. The resulting higher absolute degradation of 3-DG promotes the
de novo formation of Strecker aldehydes during natural beer aging. Furthermore, this effect can decrease flavor instability, as already shown by artificially increased 3-DG content in pale lager beers [
15]. Dicarbonyl 3-DM showed a slightly increasing tendency during natural beer aging. This could be caused by a new formation from maltose. This phenomenon has already been described in the aging of pilsner beer [
31]. Thus, the participation of 3-DM in Strecker degradation can be confirmed, but the degradation of α-dicarbonyl can occur simultaneously with the dominant formation during beer aging.
In summary, no clear indication for the formation of Strecker aldehydes from 3-DM and Amadori products could be observed. It should be considered that only free dicarbonyls were used for the observation and that only ε-terminal Amadori products were quantified. In contrast, the occurring degradation of 3-DG and 3-DGal indicated a possible de novo formation of Strecker aldehydes.
3.3. Formation of Free Aging Aldehydes during Beer Aging
The elucidated precursor compounds react further during beer aging and form aging aldehydes, causing sensory deterioration. This study focused on the Strecker degradation reaction because this is a typical reaction of the observed precursors amino acids, dicarbonyl compounds, and Amadori products.
Figure 4 shows the concentration of 2MP, 2MB, 3MB, meth, and PA during natural and forced aging in two barley varieties (B1 and B2) and three different malt modification levels (P1 < P2 < P3). Data of all sampling points are provided in the
Supplementary Excel file (Table S1).
The barley variety showed no effect. The formation of the aging aldehydes was comparable for both barley varieties. During aging, aldehydes were formed. As already shown in our preceding study, the rise was caused by the release from bound-state forms [
1]. However, due to the observed decreasing levels of 3-DG and 3-DGal as Strecker aldehyde precursors during aging,
de novo formation was suggested to be occurring simultaneously. In what follows, the aging behavior of the observed Strecker aldehydes is discussed separately.
The 2MP concentration showed a strong increase during natural and forced aging. The concentration after forced aging represented levels comparable to the natural aging of 0–3 months. During natural aging, the 2MP concentration increased up to 6 months and was then reduced up to 9 months. A shift to masked forms can be excluded because, after 9 months of natural aging, no bound-state form was observed in the preceding study [
1]. It can be assumed that aldehydes are degraded by alternative reactions such as aldol condensation or further oxidation of the carbonyl group. Furthermore, 2MP surpassed its threshold (86 µg/L) [
34] after 3 months of natural aging at the highest malt modification level (P3) and after 5 months at low (P1) and medium (P2) levels. Sensory attributes showed significant changes after natural aging ranging from 3 to 5 months [
1]. The conformity indicates that 2MP played a key role in the sensory changes during beer aging in the study. It is worth noting that 2MB and 3MB were also formed during natural and forced aging but less strongly than 2MP. The concentration of both compounds after forced aging was comparable to the concentration after 3–6 months of natural aging. As already discussed for the 2MP concentration, the degradation of both aldehydes occurred from 6 to 9 months of natural aging. This effect could also be caused by alternative degradation pathways such as aldol condensation or oxidation reactions. In this study, 2MB and 3MB did not surpass their flavor thresholds (2MB: 45 µg/L; 3MB: 56 µg/L) [
34] but could still contribute to sensory flavor changes by synergistic effects. The limit of detection of meth is 0.8 µg/L. Therefore, the Strecker aldehyde was majorly quantifiable after forced aging in B1 and after 3 months of natural aging in both barley varieties. For meth, forced aging was comparable to the natural aging of 3–6 months. After 3 months of natural aging, a constant increase in meth concentration was observed. The aldehyde surpassed its threshold of 4.2 µg/L [
34] after 9 months of natural aging. This result indicates that meth critically contributes to sensory deterioration in an advanced stage of beer aging (>6 months of natural aging). PA concentration increased constantly after 3 months of natural beer aging. The forced method yielded PA concentrations comparable to 6–9 months of natural beer aging, with an exception of B2P1 and B2P2 (3–6 months of natural beer aging). The aldehyde did not surpass its threshold at 105 µg/L [
34] but could contribute to sensory deterioration by synergistic effects such as 2MB and 3MB.
Generally, an increasing malt modification level caused an enhanced formation of all Strecker aldehydes during natural and forced aging. This effect indicated an accelerated beer aging by both increased release and de novo formation of Strecker aldehydes at higher malt modification levels. Therefore, de novo formation is promoted by a higher content of amino acids and dicarbonyls as Strecker aldehyde precursors due to an increased modification level, as presented in the previous section.
3.4. Formation of AGEs
Glycated amino acids are formed in the late phase of the Maillard reaction and are considered their end products [
12]. Since the formation also occurs in dicarbonyl compounds, it represents a concurrence reaction to the Strecker degradation. In addition to pyrraline, AGEs MG-H1, CEL and CML were investigated. Here, only the free amino acids were determined.
Figure 5 shows the contents of these AGEs depending on the malt modification level (P1 < P2 < P3) in two barley varieties (B1 and B2) during natural and forced aging.
Pyrraline, formed from 3-DG and lysine, was analyzed within 81–188 µg/L. The reaction of MGO with the guanidino group of arginine produces MG-H1, which was quantified at 325–515 µg/L in all samples. In comparison to previous studies, both concentration ranges could be considered low [
15,
35]. CEL and CML are derived from the reaction of MGO and GO with lysine, and thus, similar to MG-H1, are only indirectly involved in the degradation of 3-DG. CEL ranged between 111 and 253 µg/L, and CML was found between 69 and163 µg/L. Here, the contents were in the range described for pale lager beer [
15].
In general, no influence of barley variety on the investigated AGEs could be observed. Higher malt modification tended to result in higher content of AGEs. Natural and forced beer aging revealed no consistent trend on the analytes. This confirmed the results of a study by Nobis et al., in which the concentration of AGEs also remained constant during beer aging. The authors suspected AGE formation during beer aging as a minor reaction [
15]. Furthermore, Rakete et al. observed only a minimal increase in AGEs derived from proline (N-formylproline and N-carboxymethlyproline) during a forced aging experiment [
31]. Thus, the formation of the selected AGEs from the precursor compound 3-DG could be considered as subordinate.
3.5. Influence of the Proteolytic Malt Modification Level on De Novo Formation
The previous sections focused on several individual precursor compounds and aging aldehydes. However, the contribution to
de novo formation should be further regarded for each substance class in a holistic perspective. Hereby, the influence of malt modification level on the aging potential is focused in the following lines.
Figure 6 shows the molar distribution of the sum of free aldehydes, bound aldehydes (4-VP releasable aldehydes) [
1], and precursor compounds (amino acids, dicarbonyls, and Amadori products) during aging due to different malt modification levels (P1 < P2 < P3) and barley varieties (B1 and B2). For amino acids, the sum of Strecker-relevant amino acids (leucine, isoleucine, valine, methionine, and phenylalanine) was calculated. Dicarbonyls comprised the sum of 3-DG, 3-DGal, and 3-DM concentrations. Amadori products were calculated as the sum of FL and ML content. All three substance classes are precursor compounds contributing to Strecker degradation. The 4-VP releasable aldehydes from the preceding article were included in the evaluation to further elucidate comprehensively the influence of
de novo formation and release for the formation of flavor-active volatiles. All observed compounds contribute to the actual aging potential at several aging stages [
19].
In what follows, all substance classes are considered, starting with precursors. Strecker-active amino acids represented the highest percentage of the overall sum of the calculated compounds. Notably, 73–93% of the sum belonged to this substance class. Furthermore, the concentration of amino acids was 5–16 times higher than that of dicarbonyls, accounting for only 5–20% of the aging potential (sum of all compounds). The ratio of amino acids and α-dicarbonyls indicated that the dicarbonyl concentration was the critical factor for
de novo Strecker degradation from both classes. This fact is further substantiated by the higher reactivity of dicarbonyl compounds [
5]. Dicarbonyls were the only precursor substance class in which degradation was observed for almost all variations. This fact could be shown by the changes in the sum from fresh beer to 9 months naturally aged beer (B1P1: 95 to 75 µM; B1P2: 114 to 103 µM; B1P3: 143 to 120 µM; B2P1: 90 to 94 µM; B2P2: 120 to 113 µM; B2P3: 156 to 134 µM). The degradation indicated Strecker degradation and confirmed previous references that also observed dicarbonyl degradation during beer aging [
15,
36]. The Amadori products accounted for only a minor percentage of 1–8% of the aging potential at all aging stages. Here, it should be considered that only two Amadori products as direct precursors of dicarbonyls were observed. Amadori products formed at the α-amino group of amino acids as direct Strecker aldehyde precursor compounds [
8] might be more important. For example, Wittmann and Eichner (1989) determined fructose-derived Amadori products from valine, leucine, and isoleucine in the sum of 105 µM in fresh beer [
37].
Regarding free and bound-state aldehydes, it was revealed that both forms in sum accounted for only 0.7% of the aging potential. The precursor substance classes were highly elevated, compared with the free aldehydes in this study, whereby the ratio increased toward free aldehydes during aging. In fresh beer, the precursor concentration (sum of amino acids, dicarbonyls, and Amadori products) was 3050–13,790 times higher than free aldehyde concentration. After natural aging of 9 months, the ratio decreased to only 334–787 times. The shift and ratio indicate that de novo formation of aging aldehydes can occur directly starting from freshly bottled beer. The concentration of bound aldehydes increased with higher amino acid concentration, indicating imine formation.
The malt modification level had a strong influence on the entire aging potential. During all stages of natural and forced aging, the sum of free aldehydes, bound-state aldehydes, and precursors was increased by a higher malt modification level. The difference between B1P1 and B1P2 was low because both samples had the lowest difference between the soluble nitrogen content of the used malts [
1]. However, it could be stated that a higher proteolytic malt modification level increases the
de novo formation of free aging aldehydes by an elevated pool of precursors in fresh beer.
Regarding the result of the preceding article [
1] and this study, it could be concluded that the release of the bound-state and
de novo formation of aging aldehydes occurred simultaneously. However, both studies aimed to comprehensively elucidate whether one of both pathways was dominant at a certain aging period. Therefore, a calculation was performed to evaluate the importance of the release of aging aldehydes. The sum of the concentrations of free and bound (4-VP releasable) aging aldehydes [
1] in fresh beer was compared with the concentration of free aldehydes of the following sampling points during natural aging for each barley variety and malt modification level. The aging period, where the calculated sum in fresh beer exceeded the concentration of free aldehydes in the aged beer, was defined as the release-dominated stage. Neglecting the simultaneously occurring
de novo formation, within these months of aging, the free aldehydes were mainly formed by their observed bound-state concentrations in fresh beer.
Table 2 summarizes the release-dominated aging period of each variation. Longer ranges indicate that
de novo formation of aging aldehydes is less important.
A higher malt modification level at B2 tended to favor earlier importance of
de novo formation pathways, whereas no influence was observed for B1 variations. Generally, it could be shown that the 4-VP releasable bound-state aldehyde was predominant toward aging aldehyde formation up to 4 months of natural aging. The result confirmed the findings of the study by Nobis et al., in which Strecker degradation by 3-DG spiking was observable after prolonged storage (>6 months) [
15]. The sample set showed varied release-dominated aging periods. Therefore,
de novo formation became important starting from 1 to 4 months of natural aging in this study.