Kilning invokes oxidative changes in malt proteins

Beneath glycation, oxidation reactions may take place at cereal proteins during production of malt. The extent of oxidative chemical changes at malt proteins has not yet been studied. In the present short communication, malt protein was characterized by the determination of free thiol groups and degree of methionine oxidation as well as the sites that are reactive to covalent modification by 2,4-dinitrophenylhydrazine (DNPH, “protein carbonylation”). Protein carbonylation in pale malts was around 1.5 nmol/mg protein and increased with increasing malt colour. Investigations on the protein pellet isolated for determination of carbonylation revealed that solubility and colour may disturb the quantification of carbonyl sites in roasted malts. Free thiols decreased with increasing malt colour already in pale malts (EBC < 10). The formation of methionine sulfoxide (MetSO) was intensified with increasing malt colour. An amount of 7–20% of methionine was converted to MetSO in pale and dark malt, whereas nearly 60% of methionine was oxidized to MetSO in roasted malts. The formation of methionine sulfone was negligible. This study shows that malt proteins suffer from oxidation during kilning, and future studies will have to show whether this supports the pro- or antioxidant activity of malt.


Introduction
Malting is the initial step in beer production. During this step, germinated cereal grains are kilned to obtain malt that can be stored for up to several years and to produce flavour and coloured compounds. Malting permits chemical reactions involving several cereal constituents, with glycation being the most important and wanted reaction. Beneath glycation, different other post-translational reactions can take place such as protein oxidation [1]. These non-enzymatic reactions do not stop after malting, but may take place even during production and storage of beer and contribute to flavour deterioration [2].
Protein oxidation is a multi-stage reaction [3,4]. First, radicals are formed within the protein either at backbone or side-chain loci by the action of reactive oxygen species (ROS) or reactive nitrogen species (RNS). The radicals may then be stabilized by reaction with molecular oxygen to peroxides and further to hydroperoxides or be transferred within the protein. Radicals, hydroperoxides and peroxides are normally short-lived and may not easily be detected in foods, but some stable radicals were detected in model systems and dark roasted malt [5,6]. A prooxidative effect of highmolecular weight melanoidins in barley malt was described [7]. The next step in protein oxidation is the formation of Contribution for the Special Issue: The chemistry behind malt and beer production-from raw material to product quality. hydroxylated, carbonylated and carboxylated compounds during degradation of the peroxides and hydroperoxides. Covalent attachment is also observed during methionine oxidation by formation of methionine sulfoxide (MetSO) and methionine sulfone (MetSO 2 , Fig. 1A). Oxidation of methionine side-chains is involved in the development of the unpleasant "light-struck" off-flavour, because this reaction may lead to the formation of sulfhydryl radicals. These may react with photooxidation products of isohumulones and form 3-methyl-2-butene-1-thiol as the actual odorant. [8]. Beneath methionine, especially carbonylated species have gained much attention in the literature because of their easy quantitation by UV spectrometry. The major compounds behind "protein carbonylation" are glutamic and aminoadipic acid semialdehydes (Fig. 1B) [9,10]. Further important oxidized amino acids are 3,4-dihydroxyphenylalanine, o-and m-tyrosine, which originate from tyrosine and phenylalanine, and the tryptophan oxidation products N-formyl kynurenine, kynurenine, and oxindolylalanine [11]. Protein oxidation in food has often been assessed by unspecific sum methods such as fluorescence spectroscopy or measurement of protein carbonylation. The significance of this value, however, is more and more put into question, and data on individual structures are urgently needed to precisely assess the oxidation mechanisms [12][13][14].
Moreover, protein oxidation has been described during plant development. Reactive oxygen species occur during early seed germination and were discussed as to act as signalling molecules during seed development and plant ageing [15,16]. In barley seeds, ROS, mainly the superoxide radical anion produced by NADPH oxidase, have a regulative function during germination in gibberelline synthesis [17]. ROS participate in the regulation of seed dormancy in barley seed development [18]. As malt production comprises germination, and germination and early plant development is linked to higher protein oxidation in seedlings [19], protein oxidation in malts is very likely.
Earlier analyses of glycation compounds in different brewing malts revealed that the formation of individual glycation compounds cannot explain the disappearance of amino acids such as lysine and arginine completely [20]. Malt was also taken as a model food in the present study due to its potential susceptibility to oxidation as described above. In the present study, first insight into the relevance of protein oxidation in malt should be gained to evaluate its contribution to the overall oxidation events in malt.

Malt samples
Malt samples (21 items) were obtained from a commercial manufacturer. The samples comprised malts from barley (17), wheat (3), and spelt (1, Table S1). Malt was ground in a commercial laboratory mill (tube mill, IKA, Staufen, Germany) and only the fraction passing a 0.2-mm sieve was employed for analysis. Malt colour (EBC units) was adopted as indicated on the packages.

Determination of protein carbonylation
Protein-bound carbonyl sites were assessed photometrically by reaction with DNPH after removal of low-molecular reaction products. Based on different literature methods [23][24][25], malt (100 mg) was added with 1800 µL of a solution of 1% Fig. 1 A Structures of methionine and its oxidation products, B modified amino acids with a carbonyl group in their sidechain SDS and 8 M urea in water and shaken for 10 min (Gyrator, Uniequip, Martinsried, Germany). After centrifugation (10 min, 5000 × g, RT), portions of 250 µL of the supernatant were pipetted into reaction vessels. Then, either 250 µL of a solution of 20 mM DNPH in 4 M HCl containing 1% SDS was added or 250 µL 4 M HCl containing 1% SDS. Both incubations were prepared in triplicate. The samples were incubated at room temperature and shaken every 15 min. Then, the mixtures were stored at -18 °C for 15 min and centrifuged (15 min, 9,000 × g, 4 °C). The supernatants were removed and discarded. The pellets were suspended in a mixture of cold ethanol/ethyl acetate (50/50, v/v) and centrifuged as described above. The supernatant was discarded, and the pellet was washed by the same method two more times. Finally, the pellet was taken up in 1 mL of 6 M guanidinium hydrochloride (pH 2.3). The absorbance of the solutions was determined at 280 nm and 370 nm, respectively, using quartz cuvettes. Calibration for the protein content was performed with solutions of bovine serum albumin at concentrations between 0 and 4 mg/mL.
Weights of pellets were determined for the pellets washed with ethanol/ethyl acetate. These were not taken up in guanidinium hydrochloride buffer, but dried at room temperature overnight after extraction. The tubes used for determination of the dry weight were weighed before addition of the sample and after air-drying of the pellet. The protein content of the pellets was estimated by applying a manual ninhydrin method. The pellet was hydrolysed in 2 mL 6 M HCl at 110 °C for 23 h in a sand bath in a drying chamber. After cooling, 2 mL of 6 M NaOH was added. The colour reaction was executed according to the literature [26,27]. A ninhydrin working solution was prepared by mixing 100 mL of a ninhydrin stock solution (8 g ninhydrin dissolved in a mixture of 300 mL ethylene glycol and 100 mL 4 M sodium acetate buffer, pH 5.5) with 2.5 mL of a 10% (w/v) solution of tin(II) chloride in ethylene glycol. Of the hydrolyzate, 100 µL was mixed with 1 mL of the ninhydrin working solution, and the mixture was heated in closed tubes at 104 °C in a drying chamber for 20 min. After cooling, 5 mL of aqueous isopropanol (50/50, v/v) was added. The absorption of the mixtures was measured spectrophotometrically against a reagent blank at 570 nm. Calibration was performed with glycine, and the concentrations of glycine were converted to anhydrous amino acids. The mean molecular weight of 128 Da for an amino acid was used for the calculation, and the condensation of one molecule of water during peptide bond formation was considered.

Determination of free thiol groups
Based on a literature method [28], ground malt (10-40 mg) was suspended either in 1000 µL of a solution containing 8 M urea, 5 mM DTNB (Ellman's reagent), 3 mM EDTA, 0.2 M Tris-HCl and 1% SDS whose pH had been adjusted to 8.0, or in 1000 µL of a solution containing 8 M urea, 3 mM EDTA, 0.2 M Tris-HCl and 1% SDS whose pH had been adjusted to 8.0. The solution containing DTNB was also subjected to analysis without addition of malt. Each suspension was prepared in triplicate. The solutions were incubated at room temperature for 30 min and shaken regularly. Then, the suspensions were centrifuged, and 100 µL of the supernatants was diluted with 900 µL of the DTNB-free solution. After mixing, 100 µL of the solutions was transferred to a 96-well plate, and the absorbance was read at 412 nm with the microplate reader Infinite pro 200 (Tecan, Switzerland). Calibration was performed with glutathione at final concentrations between 0.005 and 0.5 mM. The reactions in the presence and absence of DTNB were each performed in triplicate. Data for free thiol groups are reported as nmol/ mg malt.

Quantification of methionine, MetSO, MetSO 2 , and pyrraline
Methionine, MetSO, and MetSO 2 were quantitated after anaerobic enzymatic hydrolysis as previously published [22]. The relative oxidation of methionine was calculated as the ratio between MetSO or MetSO 2 and the sum of methionine and its oxidized derivatives after correction for the amount of MetSO formed during enzymatic hydrolysis [22]. In the same enzymatic hydrolysates, pyrraline was quantitated by RP-Phenyl-HPLC with UV-detection as described in the literature [29].

Protein carbonylation in malt
During kilning of malt, the Maillard reaction takes place, and individual compounds from this reaction have already been quantified in malt [21,30]. The Maillard reaction also comprises oxidative subprocesses, and oxidized amino acids can result from amino-carbonyl reactions as shown by the example of the formation of aminoadipic acid semialdehyde in the reaction of lysine with methylglyoxal [31]. Oxidation reactions may also occur without parallel glycation by reaction of ROS or RNS with protein-bound amino acids in malt proteins. However, neither sum parameters nor individual oxidized amino acids have been assessed in brewing malt yet.
Protein oxidation is often determined as "protein carbonylation", which, in a narrower sense, is the determination of DNPH-reactive sites on proteins. These can comprise also reaction products of lipid peroxidation-derived carbonyl compounds with proteins. The method is based on the reaction of complete food samples or model proteins with DNPH, whereby protein-bound dinitrophenylhydrazones are formed at protein-bound keto or aldehyde groups. Then, the protein is extracted and the protein pellet purified from low-molecular hydrazones and residual DNPH. Quantification of protein carbonylation is then performed spectrophotometrically with the determination of the protein content by UV spectroscopy in the same sample. In the present study, this analysis was performed on 21 malt samples incorporating as many processing variants as possible. The malt samples had a huge variety of colour (mean EBC 2-1700) and were of barley, wheat, and spelt origin. A literature method was adapted aiming at maximizing solubilization of water-insoluble proteins (prolamins, glutelins) by addition of SDS and urea [26]. Protein carbonylation in pale malts (EBC < 10) was between 0.6 and 3.4 nmol/mg protein (Table 1). Carbonylation rose with EBC to a value of 41 nmol/mg protein until EBC 900 and then declined ( Fig. 2A). This may be due to a decomposition of carbonyl groups during roasting, eventually because these groups are incorporated into the melanoidin skeleton. A small extent of protein carbonylation can already be formed in the seedlings (up to 0.5 nmol/mg protein [32]). In barley seedlings, an increase in protein carbonylation was attributed to increasing levels of ROS during steeping [33]. Higher concentrations of protein-bound carbonyl groups in the malt samples can originate from kilning [34].
The data obtained for protein carbonylation in malt samples have to be considered with care, as they are very much influenced by the solubility of the cereal proteins, which was improved in the present study by addition of SDS and urea during protein extraction. Though the effect of insufficient protein solubilization is accounted for by relating the DNPH adducts to the protein content measured at the same pellet, the comparability between individual malt samples may be disturbed when the solubility of proteins is changed. To address this point, the pellets resulting during determination of carbonylation were also weighed after air-drying to explore differences in the yield of protein during sample work-up. The mean protein concentration in the pellets obtained from 25 mg malt after dissolution in 1 mL guanidinium hydrochloride was 2.1 mg/mL for the whole range of malts (determined from the absorbance at 280 nm), and no increase in the apparent protein concentration was observed with increasing EBC value. Such an apparent increase could have been expected for darker malts due to the general rise Table 1 Concentrations of protein carbonyl sites, free thiol groups, pyrraline, and relative methionine oxidation in malt samples Date are given as ranges and median in parentheses. Individual data were determined in duplicate. n.d., not detectable a Malt samples were classified according to their colour as pale (EBC < 10), dark (10 < EBC < 500) and roasted (EBC > 500) b The relative methionine oxidation is calculated as the ratio between the concentration of MetSO and the sum of methionine and its oxidized derivatives in the specific coefficient of extinction of proteins subjected to browning reactions which may also affect the UV absorbance at 280 nm. The UV absorption of the pellets from dark malts may thus be higher not because of a higher protein content, but because of the higher probability of light absorption due to newly formed chromophores. However, the weights of the pellets isolated from 25 mg malt decreased from 5 to 7 mg for pale malts to 1-2 mg in pellets of malts with EBC values above 280. The apparent protein content of the pellets may be calculated from the protein concentration measured during determination of protein carbonylation and the pellet weight. This apparent protein concentration rose with the EBC value from ca. 25% in a pale malt and exceeded 100% when malt colour exceeded an EBC value of 280. Therefore, a further method was applied to estimate the protein content of selected pellets. Pellets were hydrolysed with HCl and the neutralized pellets brought to reaction with ninhydrin as has already been applied in beer research [28]. With this method, the protein concentration in the pellets was calculated to be between 14.2 and 18.6% in pale and dark malts. In one roasted malt assessed, the protein content had decreased to 2.9% (Table S2).
As the protein content thus must have been overestimated by the determination of the absorption at 280 nm in the pellets of the dark malts due to the very probable absorption of the chromophores at the wavelength used for protein quantitation [34], the protein carbonylation method is very surely not applicable for extremely dark malts, because carbonylation should be underestimated. An apparent increase in the determination of carbonyl sites due to the contribution of malt colour to the colour of the pellet can be excluded, because a blank is subtracted that consists of the protein not reacted with DNPH. This is different for the protein content, which is directly determined at the pellet without the possibility to correct for interferences from sample colour. It would be very important and much interesting also from a chemical point of view to analyse the pellets for their amino acid composition and to gain further insight into the nature of the carbonyl groups formed on the malt proteins during kilning.
Beyond the two amino acids with aldehydic side-chains depicted in Fig. 1B, theoretically all other carbonyl structures in proteins can react. Pyrraline is a glycation compound with a carbonyl group that is able to react with DNPH [35,36] and thus might contribute to protein carbonylation. Pyrraline was quantified in the enzymatic hydrolyzates of the malts as described earlier [21,37]. The pyrraline concentration in seven pale malts (EBC < 10) was between 0.01 and 0.18 nmol/mg malt (median, 0.06 nmol/mg malt), and the carbonyl content was between 0.6 and 3.4 nmol/mg protein (median, 1.5 nmol/mg protein). When pyrraline is related to the protein content of the malts (ca. 10%), these data reveal that pyrraline in glycated proteins may have a substantial impact on the apparent protein carbonylation and underscores that protein carbonylation is not (only) a measure of protein oxidation. The concentration of pyrraline increased with malt colour and exceeded the carbonyl content at EBC values > 100. Whether this is due to an underestimation of protein carbonylation at higher EBC values owing to overestimation of the protein concentration in the pellet for the reasons stated above needs to be explored in the future.

Free thiols in malt
During oxidation reactions, thiol groups in cysteine may be oxidized to cystine, but also to sulfenic, sulfinic and sulfonic acids. The malt samples in the present study were characterized by the reaction with an excess of the disulfide Fig. 3 Correlations between free thiol groups and A pyrraline (n = 17), and B relative methionine oxidation (n = 18) 5,5ʹ-dithiobis(2-nitrobenzoic acid) (Ellman's reagent), whose reaction with free thiol groups leads to the formation of a UV-active soluble small molecule, namely 5-thio-2-nitrobenzoic acid that is determined photometrically. Thiol groups were at a level between 0.9 and 1.8 nmol/mg in pale malts (EBC < 10, Table 1). This is slightly lower than the values detected in native cereal protein [29], what may be explained by the susceptibility to oxidation of these groups already under the mild conditions applied for production of these pale malts. Thiol groups decreased with increasing malt colour and were no longer detectable in malt with EBC values above 500 (Fig. 2B). Different reasons may be held responsible for the disappearance of thiols such as the formation of disulfide bonds or the oxidation to the above mentioned acids. Moreover, with the increase of pyrraline the content of free thiols decreases (Fig. 3A). Beneath oxidation, chemical conversion of thiol groups and introduction into the melanoidin skeleton may be possible. Literature data indicate the possibility of formation of such adducts, e.g. by thioether formation of pyrraline with cysteine [38][39][40] and the formation of hemithioacetals with methylglyoxal ( Fig. 4) [41].

Oxidation of protein-bound methionine in malt
Methionine is quite susceptible to oxidation also under the conditions applied for protein hydrolysis. Its conversion to MetSO during acid hydrolysis has been reported [42]. Therefore, the comparatively mild protein hydrolysis in the absence of oxygen ("anaerobic enzymatic hydrolysis") was applied as in a previous study [22]. MetSO was quantified in all samples. After correction for artificial oxidation during enzymatic hydrolysis [22], a relative methionine oxidation between 6.9 and 7.9% in seven pale malts was found (Table 1), which may reflect a "background" methionine oxidation which has already been determined in other plant material [43]. Methionine oxidation increased constantly with increasing malt colour (Fig. 2C) with the highest oxidation rates found in the roasted malts. Only very small amounts of MetSO 2 were detected in malts. In pale and dark malts, not more than 0.1% of whole methionine was oxidized to MetSO 2 . Relative methionine oxidation to MetSO 2 increased to up to 0.6% in roasted malts. A comparison between the thiol content and the relative methionine oxidation revealed that thiol groups are more susceptible to oxidation (Fig. 3B). With increasing malt colour in pale malts, the concentration of free thiol groups already begins to decrease while methionine oxidation is still constant. Only after the disappearance of 50% of the free thiol groups, an appreciable change in methionine oxidation begins.
In sum, methionine oxidation can be as high as glycation of lysine to fructosylylsine or maltulosyllysine in malt and beer. In proteins isolated from beer, lysine blockage can rise up to 20% [38]. In malts, lysine blockage is mainly brought about by fructosyllysine and maltulosyllysine, however, not the whole of lost lysine is accounted for by the formation of Fig. 4 Possible reaction products that lead to the disappearance of free thiol groups of protein-bound cysteine. Circles represent the protein backbone these two compounds, and several further glycation products are known to be generated [21]. Oxidation reactions contribute to flavour changes during ageing of beer [2]. The contribution of protein oxidation to oxidative processes at other biomolecules such as lipids or carbohydrates is unclear. However, since oxidative damage at sites on proteins can be transferred within the protein and also to other biomolecules, it would also be possible that oxidized proteins represent a "silent reservoir" for oxidation reactions during processing and storage of beer, e.g. the formation of Strecker aldehydes. Reduction of methionine sulfoxide by yeast methionine sulfoxide reductase(s) may recover native methionine, but this enzymatic reduction reaction is also responsible for the formation of dimethylsulfide (DMS) from dimethylsulfoxide (DMSO). A higher content of MetSO in malt proteins might thus inhibit DMS formation [44]. It is also an open question as to whether the antioxidant capacity, which is important for the aroma stability of beer, increases or decreases with the extent of protein oxidation. Methionine might act as a scavenger that protects critical structures from oxidation. Further studies on the chemistry of oxidized amino acids, especially of MetSO as well as the abundance of individual other amino acid oxidation products such as allysine, glutamic acid semialdehyde, or the tyrosine isomers as well as tryptophan oxidation products need to be performed.

Conflict of interest
The authors declare that they have no conflict of interest.
Compliance with ethics requirements This article does not contain any studies with human or animal subjects.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.