Comparative analysis of taste components of three seasoning bases prepared via stir- frying, enzymatic hydrolysis, and thermal reaction

Mung bean sprouts have gained the interests of consumers owing to their umami taste and high nutritional value. In this work, high- performance liquid chromatography and liquid chromatography quadrupole time- of- flight mass spectrometry were employed to investigate the taste components of three seasonings prepared via stir frying, enzymatic hydrolysis, and thermal reaction. The results indicated that enzymatic hydrolysis released taste compounds from raw materials more effectively than high- temperature stir frying. The thermal reaction improved the fraction of umami components in the enzymatic hydrolysate, which resulted in the highest equivalent umami concentration and improved the taste contributions of glutamic acid, inosine 5'- monophosphate, and guanosine 5'- monophosphate disodium salt hydrate. Total of 26 peptides were identified, including Ala- Met, Ala- Asp, Glu- Asp, Glu- Ala- Glu, Ala- Pro- Ser, Ala- Glu, Ser- Ala- Ser, Ser- Asp- Ala, His- Ile, Asp- Val, Ala- Asp, Glu- Ala- Ala-Ala, Gly- Ala- Glu- Asp- Gly- Gly, Ala- Glu- Ser, Glu- Ser- Asp- Val- Ala, Ser- Ser- Ser- His- Phe, Gly- Asp- Cys- Ser- Asp- Asp, Ala- Ala- Lys, Ala- Ser- Tyr, Ser- Ala- Met- Gly, Glu- Ser- Asp-Val- Ala, Thr- Ser- Ser- Ala- Ile- Ser, Ser- Gly- His- Glu- Asp- Glu, Ile- His- Glu- Ala, Ser- Arg-Ser, and Ser- Ala- His- Pro- Gly- Thr. The sweet and umami amino acids were the main residues of the N- terminus positions of peptides. Double and triple continuous umami amino acid sequences could also be important for the umami taste of peptides.


| INTRODUC TI ON
Mung bean sprouts is one of the most popular sprouts because of their high content of proteins, polypeptides, polysaccharides, and polyphenols. Mung bean sprouts is widely consumed as fresh salad vegetables or common side dishes and also used as health food and medication (Tang et al., 2014). During the germination of mung beans, proteins are converted into free amino acids, which are more easily absorbed by the human body via the action of various enzymes (Nonogaki et al., 2010). Moreover, these free amino acids, which are precursors in Maillard reactions are important for flavor formation in food. Combining the demands of modern consumers for "nutrition, delicacy, and convenience" and the important role of mung bean sprouts for health care, the efficient utilization of mung bean sprouts and industrialization of mung bean sprout products are foreseeable.
Meat flavoring, which uses animal and plant proteins as precursors, is an important raw material for the industrialization of traditional Chinese food (Lieske & Konrad, 1994). Multileveltargeted enzymatic hydrolysis of proteins during the preparation of meat flavoring is an effective way for improving the utilization ratio of animal and plant proteins. The glutamic acid (Glu) and Glu-rich oligopeptides from the hydrolysis of vegetable proteins present typical umami taste (Sonklin et al., 2011). Sonklin et al. indicated that Glu and aspartic acid (Asp) were the main components of mung bean protein isolate (MPI) (Sonklin et al., 2011).
Enzymatic bromelain mung bean meal protein hydrolysate (eb-MPH) produced from MPI presents bouillon, salty, umami, and sweet tastes (Sonklin et al., 2011). Moreover, the amino acid profile of eb-MPH was similar to that of meat, and thus, eb-MPH could be used as a condiment for direct food enhancement or precursor for thermal process flavoring. Mung bean sprouts possess more secondary metabolites than mung beans (Tang et al., 2014).
In-depth analysis of taste components, particularly umami taste components of mung bean sprout dishes and their enzymatic hydrolysis products is the key to developing meat flavoring using mung bean sprouts as a plant protein source. However, only few studies have been reported on the taste components of mung bean sprout products.
In this study, three mung bean sprout products were prepared and their taste components were analyzed. The taste components of the three samples were investigated using high-performance liquid chromatography (HPLC) and liquid chromatography quadrupole time-of-flight mass spectrometry (LC-Q-TOF/MS). This work aimed to: (1) quantify the taste compounds from three mung bean sprout product samples, (2) isolate the umami peptides from three mung bean sprout product samples, (3) elucidate the umami and umami-enhancing effect of each fraction separated from the three mung bean sprout product samples, and (4) investigate the relationship between the umami components and umami taste properties of the three mung bean sprout product samples.

| Materials and chemicals
Xiaotangshan mung bean sprouts were purchased from Beijing Buffer salts I and II for the analysis of organic acids and nucleotides, respectively, were prepared according to Kong's method (Kong et al., 2017).

| Preparation of three seasoning bases
Stir-fried pork with mung bean sprouts (SFPM) was prepared as follows. m pork : m mung bean sprouts = 3375:2000. Pork (337.50 g) was cut into 8 cm long, 0.6 cm wide, and 0.6 cm thick strips. Fresh Welsh onion and fresh ginger were shredded. After cooking oil (56.00 g) was added to an iron pan and was heated at 160°C, the pork strips were stir-fried for 15 s (the meat could discolor during stir-frying).
Subsequently, the mung bean sprouts (200.00 g), shredded fresh Welsh onion (15.00 g), shredded fresh ginger (10.00 g), and salt (9.00 g) were added to the pan and all ingredients were stir-fried for 4.5 min. The dish was cooled to room temperature, mixed with ultrapure water (1:1 (w/v)), and slurried using a juicer. After four repeated experiments, the prepared SFPM sample was 3,812.54 g, which was used for further analysis.
Enzymatic hydrolysate (EH) was prepared as follows. The ratio of pork and mung bean sprouts was the same as SFPM sample.
Ground pork (33.75 g), mung bean sprouts (20.00 g), fresh Welsh onion (1.50 g), fresh ginger (1.00 g), salt (0.90 g), and deionized water (33.75 g) were mixed and hydrolyzed using 0.57 g of an enzyme complex (m neutral protease : m cellulase : m pectinase = 12:4:3, activity 10 5 U/g) at 50°C for 1 hr. After hydrolysis, the system was heated at 90°C for 10 min to inactivate the enzyme, and then it was cooled to 25°C. After six repeated experiments, the prepared EH sample was 877.48 g, which was used for further analysis.

| Free amino acid analysis
The free amino acids in the samples were analyzed according to the method reported by (Wang et al., 2020;Pu et al., 2021a) with some sample preparation modifications. The supernatants of the samples were collected and refrigerated at 4°C for 12 hr. After the oil and fat were removed, the supernatants were centrifuged at the relative centrifugal force of 10,000g for 15 min at 4°C. The supernatants (1 mL) were diluted with 0.1 mol/L HCl to prepare samples with a total free amino acid concentration of approximately 1-2 mg/ mL. After they were filtered through a 0.22 µm nylon filter membrane (Cleman, Beijing, China), 500 µL of sample and 50 µL of internal standard solution were added to a 2 mL injection vial using a pipette (Eppendorf, Hamburg, Germany) and were subsequently mixed using a MX-S (DragonLab, Beijing, China) vortex mixer for HPLC analysis. The quantitative analysis of the free amino acids was performed using the internal standard method. All standard curves (five data points, n = 3) were linear and their R 2 values were higher than 0.999. All samples were analyzed in triplicate.

| Organic acid analysis
The organic acids in the samples were analyzed according to a previously described method with some sample preparation modifications (Kong et al., 2017). The supernatants of the samples were collected and refrigerated at 4°C for 12 hr. After the oil and fat were removed, the supernatants were centrifuged at 10,000g for 15 min at 4°C. The clear supernatant was filtered through a 0.22 µm syringe filter (Cleman, Beijing, China) twice and was diluted before organic acid analysis. The organic acids were quantified using external calibration curves. All standard curves (five data points, n = 3) were linear and their R 2 values were higher than 0.999. All samples were analyzed in triplicate.

| Nucleotide analysis
The nucleotides in the samples were analyzed according to a previously described method (Kong et al., 2017). Sample pretreatment was performed using the procedure described in Section 2.4. The nucleotides were quantified using external calibration curves. All standard curves (five data points, n = 3) were linear and their R 2 values were higher than 0.999. All measurements were performed in triplicate.

| Separation and purification of umami peptides
Samples were centrifuged as described in Section 2.4. The supernatants of the samples were ultrafiltrated (25°C, 0.2 MPa) using 5k, 3k, and 1kDa molecular weight (MW) membranes (Millipore, Bedford, MA, USA). Four fractions with different MWs were obtained from each sample, and the most intense umami fraction evaluated by sensory evaluation was then lyophilized.
The freeze-dried powder was redissolved in pure water to a concentration of 200 mg/mL. Afterward, 2 mL of solution was filtered through a 0.22 µm syringe filter twice and was loaded onto a Sephadex G-15 (1.6 cm × 100 cm, Qingpuhuxi Instrument Factory, Shanghai, China) gel filtration chromatography (GFC) column at 25°C and a flow rate of 1.0 mL/min with pure water as the eluent. The ultraviolet (UV) absorbance of the effluent was monitored at 220 nm with a sensitivity of 1.0, using an HD-21-2 (Qingpuhuxi Instrument Factory, Shanghai, China) UV detector because the absorbance of the peptide bond yielded the highest sensitivity at this wavelength (Strong et al., 2005). The experiment was repeated 21 times, and the fractions were collected, pooled, and lyophilized for sensory evaluation and subsequent separation. The most intense umami fractions were redissolved to a concentration of 10 mg/mL, and were subsequently separated using an LC3000 HPLC instrument equipped with a COSMOSIL 5C18-MS-II column (10 mm × 250 mm) at 25°C to obtain several sub-fractions that contained umami peptides. The subfractions were analyzed at 214 nm. The mobile phase consisted of ACN/pure water (10:90, (V/V)) and its flow rate was 1.0 mL/min. The injection volume was 1 mL. Each fraction was collected and freezedried for subsequent identification.
The dried HPLC fractions were reconstituted to 1.0 mg/mL using pure water. After they were filtered through 0.22 µm syringe filters, 1 µL of each fraction was injected into an Agilent 6530 Accurate-Mass (Karlsruhe, Germany) Q-TOF LC/MS system equipped with a ZORBAX SB-C18 (2.1 mm × 150 mm, 3.5 μm; Agilent, Beijing, China) C18 column. Gradient elution that consisted of mobile phases A (H 2 O with 0.1% formic acid) and B (ACN) was used as follows: 5%-40% B from 0 to 15 min, 40%-5% B from 15 to 25 min, and 5% B from 25 to 26 min. The flow rate was 0.3 mL/min. Spectra were recorded in positive ion mode in a mass/charge (m/z) range of 50-1,000 Da.
The LC-Q-TOF/MS instrument was run in positive ion and reflectron modes. All peptides eluted from the reverse-phase column were analyzed online via MS, and the peptides of interest were selected for MS/MS sequencing. umami-enhancing tastes of the samples were scored using a 10-point scale (Wang et al., 2020;Zhuang et al., 2016). The standard umami solution consisted of an aqueous solution of salt (0.35% (w/V)) and MSG (0.35% (w/V)) (Su et al., 2012), which was ascribed a score of 5.
The lyophilized fraction (1.0 g) was dissolved in 1 L of standard umami solution before tasting to evaluate the umami-enhancing effect.
To obtain taste thresholds, taste dilution analysis was used to determine the taste dilution (TD) factors of each fraction. The lyophilized ultrafiltered and GFC fractions were dissolved in pure water to concentrations of 10 mg/mL, and then, serial dilutions (1:2, 1:4, 1:8, 1:16, 1:32, etc.) were prepared using pure water as solvent. The maximum number of dilutions that still allowed the detection of the umami taste was recorded as the TD value.
All sensory evaluations were performed in a sensory panel room at 25°C and 55% humidity. The panelists were asked to take one sip of the sample, hold it in their mouth for 10 s, and spit it out. To avoid fatigue and the carryover effect, the panelists were asked to rinse their mouths with 50-60 ml of drinking water and take 15 s breaks between samples (Dang et al., 2015). Eating, drinking, or smoking was not allowed for 1 hr before the sensory evaluation testing. All score cards were collected at the end of each session, and the average descriptor values from all 10 judges were used for multivariate statistical analysis.

| Statistical analysis
Statistical calculations were performed using the SPSS 17.0 (SPSS Inc., Chicago, IL, USA) statistical package for one-way ANOVA (Pu et al., 2021b). Data were expressed as means ± standard deviations of triplicate determinations. The mean values were considered significantly different at p < 0.05.

| Free amino acids
Free amino acids play a crucial role in food taste. As illustrated in Table 1, the amount of total free amino acids in the SFPM sample was the lowest.
The contents of most free amino acids in the SFPM raw material were increased significantly after enzymatic hydrolysis, and the contents of lysine (Lys), Asp, and leucine (Leu) increased the most (876.31, 53.47, and 28.25 times, respectively). The total content of free amino acids of EH was 10.81 times higher than that of SFPM. These results indicated that enzymatic hydrolysis was more effective for releasing the free amino acids from raw materials than high-temperature stir-frying. Moreover, studies have revealed that blanching can reduce the content of free amino acids of mung bean sprouts (Farhangi & Valadon, 1982). This might be another reason for the content of free amino acids of SFPM being lower than that TA B L E 2 Taste thresholds and taste activity values (TAVs) of the taste compounds of the stir-fried pork with mung bean sprouts (SFPM), enzymatic hydrolysate (EH), and thermal reaction flavoring (TRF) samples

| Organic acids
Organic acids are also important taste compounds. As presented in Table 1, the amount of total organic acids in the SFPM sample was the lowest of all analyzed samples. The contents of organic acids in the SFPM raw material increased significantly after enzymatic hydrolysis, and the contents of succinic, acetic, and lactic acids increased the most (28.59, 11.63, and 3.86 times, respectively). In addition to the five organic acids (oxalic, tartaric, succinic, acetic, and lactic acids) detected in the SFPM sample, formic, malic, and citric acids were also detected in the EH sample. The total organic acid content of the EH sample was 7.16 times higher than that of the SFPM sample. This demonstrated that enzymatic hydrolysis was a more effective way to release organic acid from raw materials than high-temperature stir-frying. After the thermal reaction, the contents of oxalic, formic, malic, lactic, and pyroglutamic acids were significantly increased owing to the introduction of the TRF formula. Tartaric acid, an important plant organic acid, might be derived from mung bean sprouts particularly after enzymatic hydrolysis. Tartaric acid was also detected in tilapia frame protein hydrolysate, but could not be detected before hydrolysis (Chuesiang & Sanguandeekul, 2015). Because succinic and lactic acids were reported to present sour and umami taste (Park et al., 2001), they could contribute to the umami taste of the EH and TRF samples.

| 5'-Nucleotide
The amount of total nucleotides in the SFPM sample was the lowest of all analyzed samples (Table 1). After enzymatic hydrolysis, the 5'-CMP and 5'-GMP contents of the sample increased significantly (7.86 and 16.44 times, respectively). The content of total nucleotides

F I G U R E 1 (Continued)
of the EH sample was 5.17 times higher than that of the SFPM sample. This indicated that enzymatic hydrolysis was more effective at releasing 5'-CMP and 5'-GMP from raw materials than high-

| Taste activity value
The taste activity value (TAV) was calculated using the method reported by Schlichtherle-Cerny and Grosch (Schlichtherle-Cerny & Grosch, 1998). The taste thresholds of the free amino acids, organic acids, and 5'-nucleotides in water were retrieved from the literature Kato et al., 1989;Schlichtherle-Cerny & Grosch, 1998;Wang et al., 2020). Compounds with TAV greater than 1 were considered to be food taste-active (Engel et al., 2002). As presented in Table 2, five (Ala, His, Met, Ile, and tartaric acid), two (Gly and tartaric acid) and five (Glu, Arg, 5'-GMP, 5'-IMP, and tartaric acid) taste compounds contributed to the taste of the SFPM, EH, and TRF samples, respectively. Among these taste compounds, Glu, 5'-IMP, and 5'-GMP present umami taste; Gly and Ala are sweet; Met, Ile, Arg, and His are bitter; and tartaric acid is sour (Kato et al., 1989;Liu et al., 2015). The EH sample contained less bitter amino acids than the SFPM sample, which indicated that enzymatic hydrolysis decreased the contribution of the bitter amino acids to the taste of the sample. After the thermal reaction, the taste contributions of Glu, 5'-GMP, and 5'-IMP were increased owing to the addition of soy sauce, HVP, and other ingredients to the sample, which suggested that flavoring was in line with the purpose of developing an umamienhancing TRF.

| Equivalent umami concentration
The equivalent umami concentration (EUC, g MSG/100 g sample) is the concentration of MSG equivalent to the umami intensity of mixtures of MSG-like amino acids and 5'-nucleotides (Chen & Zhang, 2007). The EUCs of the samples in this study were calculated using the equation reported by Yamaguchi et al. (1971).
The EUCs of the SFPM, EH, and TRF samples were 5.71, 47.54, and 5,698.91 g MSG/100 g sample, respectively. After enzymatic hydrolysis, the EUCs increased owing to the release of large amounts of Glu, 5'-GMP, and 5'-IMP from the raw material. The EUC of the TRF sample was significantly higher than that of the EH sample owing to the addition of large amounts of Glu, 5'-GMP, and 5'-IMP to the TRF formula. The EUC of the TRF sample (2,929.83 g MSG/100 g sample) was much higher than that reported for most commercial soy sauces (Kong et al., 2018).

| Umami peptides
To investigate the umami and umami-enhancing effect, the samples were partitioned into four peptide fractions using ultrafiltration membranes: U-I (MW < 1 kDa), U-II (MW = 1-3 kDa), U-III (MW = 3-5 kDa), and U-IV (MW > 5 kDa). As depicted in Figure 1a, fraction U-I accounted for the highest mass fractions in the EH and SFPM samples (68.18% and 56.42%, respectively), and only accounted for 28.36% of the TRF sample (the U-II content of the TRF sample was the highest (60.85%)).
Referred to "GB/T 12310-2012 Sensory analysis method-Paired comparison test," the sensory evaluation results of the ultrafiltration components of the three samples were similar. The highest umami and umami-enhancing effect scores of U-I were the highest of all analyzed fractions. Moreover, the scores decreased as MW increased.
However, the umami-enhancing effect of U-III was slightly higher than that of U-II for the TRF sample. This was attributed to some large MW Maillard products, such as glycopeptide cross-linking components, presenting an umami-enhancing effect, mellow taste, and persistent aftertaste.
The U-I fractions of the samples presented the highest umami scores, which were consistent with the results reported by Dang et al. (2015).  For the EH sample, sub-fraction E-2 presented the highest umami fraction, and sub-fraction E-3 presented higher umami-enhancing effect and TD value than subtraction E-2. Therefore, sub-fractions E-2 and E-3 were selected for subsequent separation. Although the mass fraction of sub-fraction F-3 (16.18%) of TRF was not very high, it presented the highest umami level, umami-enhancing effect, and TD value of all TRF sub-fractions. Consequently, sub-fraction F-3 was used for further fractionation using a reverse-phase HPLC system with a preparative C18 column.
The MWs and peptide sequences of the three samples are presented in Table 4. A total of 26 peptides were identified, and some were umami peptides. Among the seven peptides detected in the SFPM sample, Glu-Asp (Kawai et al., 2002), Ala-Glu (Kirimura et al., 1969), Ala-Asp (Kong et al., 2017), and Glu-Ala-Glu (Nishimura & Kato, 1988) have been reported to present umami taste, and Ala-Met (Sforza et al., 2001) has been reported to be bitter. Of the 11 peptides detected in the EH sample, Asp-Val, which was separated from sub-fraction E-3, has been reported to contribute to the bitter taste of Spanish ham (Sforza et al., 2001). Finally, eight peptides were detected in the TRF sample.
Sentandrue et al. reported that most peptides that contained hydrophobic amino acids, such as Phe, Tyr, and Leu presented the same bitter taste as their monomers (Sentandreu et al., 2003). In addition, peptides with hydrophobic groups, such as Phe and Leu, at the Cterminus position, could provide a strong bitter taste to foods. Sour taste peptides are typically associated with umami. The presence of free and dissociated Glu and Asp conferred sour and brothy/umami taste to foods (Kirimura et al., 1969), and sequences, such as Val-Glu, Asp-Ala, Gly-Asp, Val-Asp, and Gly-Glu, might contribute to the sour and umami taste of dry-cured ham (Sentandreu et al., 2003). Although some dipeptides with L-Glu at the N-terminus positions confer a sour

TA B L E 4
The molecular weight and sequence of peptides detected in stir-fried pork with mung bean sprouts (SFPM), enzymatic hydrolysate (EH), and thermal reaction flavoring (TRF) taste to foods, they could also impart brothy taste in NaCl-containing aqueous solutions at pH 6.0. Thus, sour peptides were considered to be umami peptides (Arai et al., 1973). Ueda et al. reported that Cyscontaining peptides extracted from onions imparted kokumi taste to foods because the sulfhydryl group at the side chain of Cys could cause a slight convergence sense on the tongue, and that can significantly increase the kokumi taste (Ueda et al., 1994).
According to related studies, it is important to make some predictions on the unreported peptides analyzed in this study. These peptides mainly consisted of bitter, sweet, and umami amino acids.
Because the sweet amino acids were the predominant components, these peptides might present sweet and umami taste, or umami-enhancing effect. A total of 26 peptides were identified in the three analyzed samples ( Figure S1). All peptides identified in the SFPM sample were dipeptides or tripeptides. EH consisted mostly of tripeptides and polypeptides with more than three amino acids (72.73%). The peptides identified after the thermal reactions con-  (7), Gly (2), and Thr (1)), 6 contained umami amino acids at the N-terminus positions (Glu (5) and Asp (1)), and 2 contained bitter amino acids at the N-terminus positions (His (1) and Ile (1)). Therefore, the sweet and umami amino acids were the main residues at the N-terminus positions of umami peptides. The Gly-Gly-continuous, Asp-Asp-continuous, Ala-Alacontinuous, and Ser-Ser-continuous amino acid sequences were identified in the polypeptides separated from the EH and TRF samples. The Ala-Ala-Ala-continuous and Ser-Ser-Ser-continuous amino acid sequences were identified in the polypeptides separated from the EH sample. Double and third umami amino acid continuous could be important for the umami taste of peptides.

| CON CLUS ION
The taste components of three seasonings prepared with mung bean sprout via stir-frying, enzymatic hydrolysis, and thermal reaction were investigated by HPLC and LC-Q-TOF/MS. Enzymatic hydrolysis was more effective at releasing the taste compounds from raw materials than high-temperature stir-frying. The thermal reaction improved the proportion of taste components in the EH, which resulted in the EH sample presenting the highest EUC value of all analyzed food samples, and also improved the contribution of Glu, 5'-IMP, and 5'-GMP to the food taste. The sweet and umami amino acids were the main residues at the N-terminus positions of the 26 peptides identified in the analyzed samples. Furthermore, double and triple continuous umami amino acids could be important for the umami taste of peptides. The aforementioned results not only provide insight into the unique taste of three seasonings prepared with mung bean sprout, but also provide guidelines for the development of umami flavorings.

ACK N OWLED G M ENTS
This work was supported by the National Key R&D Program of China (No. 2016YFD0400705). The authors thank Professor Dejian Huang at the National University of Singapore for extending his help to revise the manuscript.

CO N FLI C T O F I NTE R E S T
The authors have declared no conflicts of interest for this article.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data availability statement of "JFPP-08-20-1999.R1, entitled Comparative analysis of taste components of three seasoning bases prepared via stir-frying, enzymatic hydrolysis, and thermal reaction" is as follows: (1) The data that support the findings of this study are available from the corresponding author upon reasonable request.
(2) The data that supports the findings of this study are available in the supplementary material of this article.