Metabolite analysis of wheat dough fermentation incorporated with buckwheat

Abstract Dough fermentation represents an important developmental stage in the manufacturing process. In this study, volatile and nonvolatile metabolite analysis were carried out to investigate time‐dependent metabolic changes in the course of wheat dough fermentation incorporated with buckwheat based on gas chromatography–mass spectrometry (GC/MS). A total of 70 nonvolatile metabolites were identified, covering a broad spectrum of polar (e.g., amino acids, sugars, sugar alcohols, and acids) and nonpolar (e.g., fatty acid methyl esters, free fatty acids, and sterols) low molecular weight dough constituents. Meanwhile, sixty‐four volatile metabolites comprising aldehydes, ketones, alcohols, organic acids, aromatic compounds, and furans were identified using solid‐phase micro‐extraction combined with GC–MS. Some differences may exist in the volatile composition between fermented and unfermented dough. Statistical assessment of the nonvolatile data via principal component analysis demonstrated that the metabolic changes during the mixed dough fermentation are reflected by time‐dependent shifts of polar nonvolatile metabolites. And some potential nutritional markers, such as amino acids and sugars, could be developed to optimize and control the industrial dough fermentation incorporated with buckwheat.

. Moreover, buckwheat contains some antioxidants, mainly rutin and quercetin, which was claimed to be effective to strengthen capillary blood vessels and reduce diabetes II (Li & Zhang, 2001;Watanabe, 1998).
Nonetheless, the whole flour buckwheat bread products were found to be bad palatability, digestibility, and baking performances due to buckwheat's content of phytate and tannins, which may decrease the digestibility of buckwheat proteins and confer bitterness to buckwheat products (Li & Zhang, 2001). Therefore, incorporation of buckwheat is considered as a win-win solution for enhancing the overall quality of wheat bread and the palatability of buckwheat bread.
Bread production consists of two main processes: dough fermentation and baking. Particularly, dough fermentation represents an important developmental stage in the manufacturing process (Birch, Petersen, Arneborg, & Hansen,2013 ;Hansen & Schieberle, 2005;Pico, Martínez, Bernal, & Gómez, 2017). This stage is characterized by numerous metabolic processes leading to distinct and time-dependent alterations in metabolite levels. It is also found to be an effective way to enhance the palatability, digestibility and baking performances of buckwheat. Therefore, it is not surprising that metabolomics has proven to be a suitable tool for the investigation of this process. GC-based metabolite profiling techniques have been applied to comparative investigations of many industrial production processes, such as malting, germination of rice (Frank, Meuleye, Miller, Shu, & Engel, 2007;Frank et al., 2007;Gorzolka, Lissel, Kessler, Loch-Ahring, & Niehaus, 2012). Meanwhile, metabolite profiling is also considered to provide valuable data for fermentation-driven metabolic engineering of nutritionally important metabolites in grain food. And metabolites profiling of sourdough fermented wheat flour and rye bread have been performed (Koistinen et al., 2018;Ripari, Cecchi, & Berardi, 2016).
Generally, the incorporation of buckwheat dough in wheat bread has the potential to improve the overall quality of the bread.
However, no volatile and nonvolatile metabolomics-based investigations of wheat dough fermentation incorporated with buckwheat have been conducted. The aim of this study was: from the perspective of food nutrition, to apply metabolic methods to wheat in the dough fermentation process incorporated with buckwheat enabling the analysis of a broad spectrum of low molecular weight metabolites; to identify and to quantify major contributors to the time-dependent metabolic dynamic changes in the mixed dough fermentation; to compare volatile composition differences between fermented and unfermented mixed dough.

| Preparation of dough
Basically, dough was prepared according to AACCI Approved Method 10-10.03 with the following formula: 80.0 g of wheat flour, 20.0 g of buckwheat flour, 1.5% (w/w) sodium chloride, 1.0% (w/w) dry yeast and 52.0% (v/w) water. The gradients were mixed in a 100 g pin bowl mixer for 3 min 50 s. Then, dough was fermented for 2 hr (32°C; humidity of 75%). Samples were taken from 3 different part of dough every 20 min, then put them in liquid nitrogen and freeze-dried for 48 hr. All dried samples were milled and stored at −30°C until analysis.

| Solid-phase micro-extraction sampling
To identify dough volatile compounds, the headspace (HS) SPME technique was employed. For this study, the temperature was maintained at 18°C ± 1°C to avoid formation of flavor compound artifacts. The mixed dough slurry was prepared by blending flour prepared before (3.0 g), 0.5 mol/L sodium chloride solution (2 ml), in a 20 ml flask with a cap and Teflon-faced silicone rubber septa (Supelco, Co.). The flask containing the sodium chloride solution and flour was placed on a magnetic stirring plate (model PC-220) and stirred at 1,100 r min −1 for 20 min. A SPME fiber (DVB/CAR/PDMS) was then exposed to the headspace of the dough slurry for 1 hr in a water bath at 18°C ± 1°C (Dong et al., 2013).

| GC-MS analysis
Agilent 7890A GC-5975C MSD was used to identify the volatile compounds of the mixed dough. The oven temperature program was set at 35 ○ C for 3 min and then to 280 ○ C at 5 ○ C/min, with

| Nonvolatile metabolite analysis
Extraction and fractionation of freeze-dried mixed dough flour were performed in accordance with the procedure developed a previously described procedure developed by Frank et al. (2007), with little minor modifications. The substances were divided into two fractions, namely, polar fractions (amino acids and sugars) and nonpolar fractions (fatty acid methyl ester [FAME], sterols and free fatty acid [FFA]). Samples (300 mg) were added to 5 ml of 80% (v/v) methanol solution for 15 min to extract the amino acids. The mixture was centrifuged for 10 min at 10,000 rpm. Then, 1 ml of the supernatant was poured into a glass vial, and 10 μl of p-chloro-L-phenylalanine in deionized water (0.3 mg/ml) was added as a quantification standard. The methanol solution in the upper layer was dried with nitrogen and used for the subsequent steps. The remaining solid was dissolved into 360 μl of acetonitrile, then 40 μl of MTBSTFA (N-tert-Butyldimethylsilyl-N-methyltrifluoroacetamide) was added and heated at 70°C for 30 min for the silylation of amino acids. As for the sugars, 100 mg of the samples were extracted using the method discussed above, but the supernatant was taken at the volume of 50 μl. Subsequently, 30 μl of phenyl β-D-glucopyranoside (1.6 mg/ml in water) was added to the extract as a quantification standard and then dried with nitrogen. The dry residue was dissolved again in 250 μl of acetonitrile, and 50 μl of TSIM (N-[Trimethylsilyl]imidazole) was added for the silylation of sugars at 70°C for 30 min.
The extraction of the nonpolar fraction was performed by following completely the method of Frank (Frank et al., 2007). The 100 mg sample was extracted with dichloromethane for 15 min, and then the transesterified FAME, sterols, and FFA were separated by transesterification and solid-phase extraction. Finally, silylation was carried out. All of the obtained fractions were analyzed by gas chromatography-mass spectrometry (GC/MS). The GC/MS conditions were in agreement with previously described procedures. And the total ion current chromatograms of different fractions were shown in Figure 1.

| Identification of compounds
Dough constituents both volatile and nonvolatile were identified by comparing retention times and mass spectra with those of reference compounds, or by comparing mass spectra with the entries of the mass spectra library NIST2.2.

| Nonvolatile metabolite analysis in dough fermentation process incorporated with buckwheat
The applied GC metabolite profiling approach allowed the detection of six fractions during the whole dough fermentation process incorporated with buckwheat, including amino acids, sugar and sugar alcohols, acids, fatty acid methyl esters, free fatty acids, and sterols. In total, 70 peaks were identified on the basis of retention times and mass spectrometric data from authentic reference compounds and/or from MS libraries, which was divided into two parts: polar (Table 1) and nonpolar fractions (Table 2). This result reflects the more pronounced influence of the polar constituents on the separation of dough during the fermentation process. In order to assess the major drivers for the time-dependent separation of the fermented dough samples, the PCA loading scores for each single fraction were analyzed (Figure 3b-c). The results show the polar compounds were also found to be major contributors to the time-dependent separation of the fermented mixed dough samples. The later stages of dough fermentation (60-120 min) showed obvious separation from the former ones, with d-mannose, inositol, glyceryl glycoside being the most variable compounds (Figure 3b). As to nonpolar compounds, less contributions were made to the time-dependent separation. They are evenly distributed in the loading plot (Figure 3c).

| Changes of polar nonvolatile metabolite during dough fermentation incorporated with buckwheat
In order to more intuitively see the variation tendency of metabolites at different mixed dough fermentation stage, a heat map was were also considered as the contribution of proteolytical enzyme activities of the grain (Barber, Prieto, & Collar, 1989). After 40 min of fermentation, similar to the expected changes observed for sugars, the unitization of yeast reproductions leads to the observed decrease in mono-saccharides, such as glucose and fructose. Meanwhile, the time for decreased levels of disaccharides like cellobiose and maltose was later than mono-saccharides due to the diauxie of yeast cell.
The levels of acids identified were found to be low and constant during the whole fermentation process. But they were very crucial to the dough flavor because of the esterification with alcohols to form esters, which were major contributor to the dough flavor (Pico, Bernal, & Gómez, 2015).
Generally, the polar amino acids and sugars exhibited the potential to serve as nutritional markers for the wheat dough fermentation incorporated with buckwheat, based on their remarkable change in the fermentation process. Their potential application would meet the bakers' demand to maintain the maximum nutritional level both in quality and quantity.

| Changes of nonvolatile nonpolar metabolite during dough fermentation incorporated with buckwheat
Compared to the relative big changes in polar contents, the non- fermentation. As it was shown in Figure 4, the levels of major fatty acid methyl esters (FAME) including saturated C16, saturated C18, and all the unsaturated C18 were shown decreased during the whole fermentation process. Other FAMEs showed lower levels and no change at same stage of fermentation. The reason for this phenomenon was that both of saturated and unsaturated C16-C18 fatty acid was the major composition of triglycerides both in wheat and buckwheat, and an enzymatic degradation of triglycerides was also a contributor leading to the increasing levels of free fatty acids related, such as stearic acid, oleic acid, linoleic acid, and free hexadecyl fatty acid.
As mentioned above, the levels of free fatty acids increased slightly during the fermentation process, and the levels of other free fatty acids were low and constant at the same fermentation stage.
Regarding the dynamic changes of sterols, dough exhibited lower or constant contents or of the major sterols.

| Changes of volatile metabolite during dough fermentation incorporated with buckwheat
Of the 64 volatile metabolites including 11 aldehydes, 15 alcohols, 8 ketones, 4 furans, 10 aromatic compounds, 16 acids, and esters were finally identified from the mixed dough during the whole fermentation process (Table 3). As it was shown in Table 3, all the aldehydes kept a low and relative constant level during fermentation. They were responsible for the fresh and slightly green notes of dough at a low concentration, and for the butter flavor at higher one (Nor Qhairul Izzreen, Petersen, & Hansen, 2016). Meanwhile, most of volatile alcohols had an increased level at the end of the fermentation compared with the beginning. And some of them exhibited a significantly increased level, such as ethanol, 2-methyl-1-propanol, 3-methyl-1-butanol, and 2-methyl-1-butanol, because they were the by-product of yeast cell reproduction. These volatile

TA B L E 3 (Continued)
alcohols usually contribute a sweet and alcohol odor to the dough (Pico et al., 2017).
In this study, eight volatile ketones were identified. 2,3-pentanedione and 2-heptanone were only detected at the end of the fermentation.
Moreover, similar changing pattern of volatile ketones was observed as aldehydes. All the volatile ketones kept a low and relative constant level during fermentation, except 2,3-butanedione which showed an increased level at the end of the fermentation. This compound contributed a buttery, sulfurous, and pungent odor, which might be involved in the final flavor of the dough (Gassenmeier & Schieberle, 1995).
Acids and esters were most varied in types in dough fermentation. Ethyl acetate, butyrolactone, hexanoic acid ethyl ester, and acetic acid hexyl ester could be detected during the whole fermentation process. Other acids and esters could only be detected at the end of the fermentation, which contributed the major flavor differences between fermented dough and nonfermented one.
Esters usually had a little fruit flavor at lower concentration, even which could be smelled at the end of the fermentation and also be formed by the esterification with alcohols during the fermentation as mentioned above (Birch, Petersen, Arneborg, et al., 2013;Gassenmeier & Schieberle, 1995). Levels of furan and aromatic compounds detected both at the beginning and the end of the fermentation were relative small. And there were no obvious differences between them. Furans usually have a sweet flavor, and aromatic compounds contribute a flower flavor at lower concentration.

| CON CLUS ION
Polar nonvolatile metabolites were the major contributors to the mixed dough fermentation time-driven changes in the metabolic analysis. In addition to the representatives of flavor composition, nutritionally relevant metabolites are covered. They range from the nonpolar free fatty acids to the polar amino acids and sugars. And some potential nutritional markers, such as amino acids and sugars, could be developed in the future, which might finally be correlated to the contents of metabolites responsible for technological properties and the nutritional quality of mixed dough and be applied to optimize and control the dough fermentation. The polar amino acids and sugars exhibited the potential to serve as nutritional markers for the wheat dough fermentation incorporated with buckwheat. Of course, further study should be carried out to quantify each metabolite identified.