Inhibition effects of Maillard reaction products derived from L-cysteine and glucose on enzymatic browning catalyzed by mushroom tyrosinase and characterization of active compounds by partial least squares regression analysis

Abstract

Inhibition of tyrosinase activity by Maillard reaction products derived from cysteine and glucose (Cys-MRPs) was studied. Pre-incubation of mushroom tyrosinase with Cys-MRPs decreased enzyme activity with increasing reaction time. We show that Cys-MRPs irreversibly block the active site of mushroom tyrosinase and that the competitive inhibitors dithiothreitol and kojic acid protect the enzyme from Cys-MRPs inactivation. Correlation of tyrosinase inhibition ability, volatile compounds, non-volatile compounds (HMF, DDMP and maltol), and Maillard reaction conditions of Cys-MRPs was analyzed by partial least squares regression (PLSR). 3-Ethyl-2-formylthiophene, α-dimethylformylthiophene, 2,6-dimethylpyrazine, ethylpyrazine, 2-ethyl-6-methylpyrazine, 2-methyl-3-(2-thienyldithio) thiophene, and furfural showed a significant and positive contribution to inhibition ability, while 2-propionylfuran and α-dimethyl-2-formylfuran showed a significant but negative correlation with inhibition ability. Of the three non-volatile compounds analyzed, only 2,3-dihydro-3,5-dihydroxy-6-methyl-4(H)-pyran-4-one (DDMP) showed a significant and positive correlation with inhibition ability, while HMF and maltol showed a weak negative correlation. The reaction temperature and time showed a significant and positive correlation with inhibition rate, whereas the ratio of sugar to amino acid showed a negative effect within the experimental range.

---

1. Introduction

Almost half of the world's fruit and vegetable crops are lost because of postharvest deteriorative reactions.¹ Enzymatic browning is essentially an oxidation process in which polyphenoloxidase (PPO) catalyzes phenolic substrates to the corresponding o-quinones, which are unstable and automatically converted to melanin. Although melanin itself is harmless to human health, the quinones produced during the process show strong oxidation activity and can interact with amines, amino acids, peptides, and proteins, as well as antioxidant compounds, thereby, resulting in loss of nutritional quality.² Tyrosinase (PPO) catalyzes either the o-hydroxylation and subsequent oxidation of monophenolic compounds, or the oxidation of o-diphenolic compounds. The catalytic center of mushroom tyrosinase, a type-3 copper protein, contains two copper ions, with each coordinated by a highly conserved three histidine residue.³

The control of enzymatic browning has always been a challenge for the food industry. Sulfites are used extensively as effective inhibitors of enzymatic browning. However, use of sulfites has been restricted by the Food and Drug Administration (FDA) because of potential associated hazards.⁴ Therefore, several studies have been carried out to find non-sulfite, natural anti-browning agents. Ascorbic acid and its derivatives, kojic acid and 4-hexyl resorcinol (4HR) have been used commercially in the food industry with limited success.^5–7 As use of the term “tyrosinase inhibitor” varies among different researchers, Chang et al. classified them into six types: reducing agents (such as ascorbic acid), o-quinone scavenger (such as cysteine), alternative enzyme substrate, nonspecific enzyme inactivators (such as acids or bases), specific tyrosinase inactivators (suicide substrate), and specific tyrosinase inhibitor, of which the last two are considered to be “true inhibitors”.⁸ Kojic acid, tropolone, and dithiothreitol (DTT) have been used as positive control tyrosinase inhibitors by different researchers.

The Maillard reaction has an important role in improving the appearance and taste of foods. It is initiated by reaction between amino and carbonyl compounds, followed by a cascade of reactions, which produce a wide range of products. It has been demonstrated that Maillard reaction products (MRPs) can inhibit enzymatic browning.^9–11 However, the inhibitory efficiency is affected significantly by type of amino acids and processing parameters (such as temperature, pH, and time).^12–14 Cysteine (Cys) is the only amino acid that can inhibit enzymatic browning by itself. Previous studies showed that MRPs derived from cysteine and reducing sugar (Cys-MRPs) possessed the most potent inhibitory efficiency on enzymatic browning compared with the other amino acids. This is in accordance with the findings of Maillard et al., who compared the inhibition efficiency of MRPs derived from glycine, arginine, proline, and lysine and found that Cys-MRPs was the most effective inhibitor of mushroom tyrosinase and apple PPO.¹⁵ Maillard reaction processing parameters significantly affected the properties of the final product mixture. Cheriot et al. used surface response methodology to study the effect of Maillard reaction conditions on the browning inhibition capacity of final products on eggplant PPO, and reported that lower pH (pH 2), high temperature (115 °C), and proper reaction time (200 min) are favorable for high inhibitory activity.¹⁶

The initial phase of the Maillard reaction leads to formation of Amadori compounds, which undergo further reactions following two main decomposition pathways: 1,2-enolization (pathway A) and 2,3-enolization (pathway B). 5-(Hydroxymethyl)-2-furaldehyde (HMF) is formed from hexoses via 3-deoxyhexosones (pathway A), whereas 1-deoxy-2,3-hexosuloses generate 2,3-dihydro-3,5-dihydroxy-6-methyl-4(H)-pyran-4-one (DDMP) as a typical marker of pathway B. Further dehydration of DDMP will generate maltol. As the early stage of the Maillard reaction influences the course of the whole reaction, understanding these phenomena is crucial to study of the tyrosinase inhibition mechanism of MRPs.¹⁷ Measurement of HMF, DDMP, and maltol content can help to investigate which pathway is more favorable for formation of an efficient tyrosinase inhibitor.

In general, there are two mechanisms for inhibition of enzymatic browning by adding chemical reagents, as shown in Fig. 1: directly acting on PPO as an enzyme inhibitor, or reacting with the initial product o-quinone forming adduct products or reducing quinone back to phenolic substrate to block or delay the formation of dark pigment.¹⁸ In the former situation, polarographic methods can be used to determine PPO activity by measuring dissolved oxygen consumption (i.e. substrate utilization), whereas for the latter case, spectrophotometric and polarographic methods can be used together to verify possible interaction between o-quinone and inhibitors. For example, in the case of cysteine, a decline curve of dissolved oxygen content with time can be monitored, where a flat curve of absorbance in the visible region against time is seen when enough cysteine is provided. Therefore, more information can be obtained from a combination of polarographic and spectrophotometric assays.

Fig. 1 Schematic representation of diphenolase activity on catechol, together with possible mechanisms of inhibition of enzymatic browning.

In the literature, confusing results for enzymatic browning inhibition have been obtained using spectrophotometric measurement because of the non-linear relationship between absorbance and o-quinone formation.¹⁹ As MRPs are a complex combination of various compounds, it cannot be ruled out that there are different inhibitors exhibiting different inhibitory mechanisms as demonstrated by Wu et al.,²⁰ who found that multiple components may contribute to the tyrosinase inhibitory effects of MRPs using activity-guided chromatographic analysis.

Although tyrosinase inhibition has always been a hotspot of research in the food, cosmetic, and medical industries, only a small number of reports have been published on use of MRPs as tyrosinase inhibitors. In the present study, Cys-MRPs, the most potent tyrosinase inhibitor among MRPs, was chosen for study of inhibitory characteristics on mushroom tyrosinase to gain more knowledge on its inhibition mechanism. The correlation between tyrosinase inhibition ability and volatile, non-volatile compounds, and Maillard reaction conditions was established using partial least squares regression analysis to characterize the active compounds of Cys-MRPs for efficiently inhibiting tyrosinase.

2. Materials and methods

2.1. Materials

Mushroom tyrosinase was obtained from Worthington (Lakewood, NJ, USA), pyrocatechol (CA), sodium sulfite (Na₂SO₃), L-ascorbic acid, L-cysteine, L-glutathione (GSH), dithiothreitol (DTT), and glucose were purchased from Shanghai Chemical Reagent (Shanghai, China). All chemicals were of analytical grade.

2.2. Preparation of Maillard reaction products (MRPs)

Glucose (0.25 M) and L-cysteine (0.25–1.25 M) were dissolved in water, and the solution was transferred into 25 mL screw-sealed Pyrex vials and pH adjusted to different pH values (1–13) using concentrated NaOH and H₃PO₄. The Pyrex vials were then heated in an oil bath at different temperatures (80–140 °C) for varying times (0–270 min). Afterwards, the vials were removed and immediately cooled in ice water. The Maillard reaction conditions for MRPs used for gas chromatography-mass spectrometry (GC-MS) and high performance liquid chromatography (HPLC) analysis, are shown in Table 1. Aliquots of the final products (Cys-MRPs) were used immediately to determine their effects on tyrosinase activity. The remainder were stored at −20 °C until further use as initial studies suggested that this storage condition did not significantly alter the inhibition potency of MRPs for up to 1 week.

Table 1 Maillard reaction conditions of 12 samples

ID	Ratio of glucose to cysteine	Temperature (°C)	Time (h)
RC1	2	110	2
RC2	1	110	4
RC3	0.5	110	6
RC4	2	120	2
RC5	1	120	4
RC6	0.5	120	6
RC7	2	130	2
RC8	1	130	4
RC9	0.5	130	6
RC10	2	140	2
RC11	1	140	4
RC12	0.5	140	6

---

2.3. Enzyme determination

2.3.1. Spectrophotometric measurements. Oxidation of phenolic substrate was assessed by monitoring the increase in absorbance at 410 nm against time resulting from oxidation of pyrocatechol (CA) in the presence of oxygen⁹ using a UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan) connected to a computer. Typically the assayed mixture contained: CA (1 mM), assayed inhibiting compounds (Cys or Cys-MRPs, 0.2 mM), and tyrosinase (15 units per mL) dissolved in 0.1 M phosphate buffer (pH 6.8). During measurement, CA and/or assayed inhibitor was first added to quartz cuvettes, and then tyrosinase was added to initiate the reaction. The maximal slope from the linear part of the progress curve was designated as enzyme activity. All assays were carried out at room temperature.

2.3.2. Polarographic measurements. The polarographic analysis was carried out according to the method of Janovitz et al.,²¹ with some modifications. The reaction rate was calculated from the initial slope of the progress curve giving oxygen consumption versus time using a Clark-type oxygen microelectrode (Strathkelvin Instruments Ltd, North Lanarkshire, UK). Oxygen consumption data was recorded using 782 Oxygen System v 3.0 software (Strathkelvin Instruments). The reaction was conducted in a RC350 respiration cell (Strathkelvin Instruments) maintained at 25 °C, with a magnetic stirrer. The total volume of 3.0 mL testing system contained: 3 mM CA and 50 units per mL tyrosinase dissolved in air-saturated phosphate buffer (0.1 M, pH 6.8). Mushroom tyrosinase was injected into the reaction cell to initiate the reaction. Activity was expressed as nmoles of oxygen consumed per second under the assay conditions. For inhibition study, inhibitors were added to the reaction system before the enzyme.

2.4. Pretreatment of tyrosinase with different inhibitors

Tyrosinase was pre-incubated with different inhibitors, as specified below. After incubation, inhibitors were removed by diafiltration, using centrifugal filters (Amicon Ultra, 0.5 mL, 10 kDa molecular mass cut-off (Millipore)). Samples were centrifuged at 14 000g, 4 °C for 15 min, and the filtrate was discarded. The retentate was re-suspended in 500 μL 200 mM PB pH 6.8, and filters were centrifuged again. A total of four cycles of centrifugation and re-suspension was performed. After the final centrifugation, the retentate was re-suspended in the initial sample volume, using 200 mM PB, pH 6.8.

2.4.1. Pretreatment of tyrosinase with Cys-MRPs/DTT. Tyrosinase (3000 U mL⁻¹) was incubated with Cys-MRPs (1 mM) or DTT (1 mM) in 200 mM PB, pH 6.8 (1 h, 25 °C). Subsequently, Cys-MRPs or DTT was removed by diafiltration, followed by addition of CuCl₂ (1 mM). After incubation (1 h, 25 °C), CuCl₂ was removed by diafiltration.

2.4.2. Pretreatment of tyrosinase with Cys-MRPs in the presence of competitive tyrosinase inhibitors. Tyrosinase (3000 U mL⁻¹) was incubated with Cys-MRP, tropolone, and kojic acid (each 1 mM), alone or in combinations. After incubation for 1 h at 25 °C, the inhibitors were removed by diafiltration.

2.5. GC-MS spectrometry

The volatile compounds were isolated according to the procedure reported by Song et al.,²² with slight modifications. All analyses were performed on Finnigan Trace GC-MS (Finnigan, USA). Separation was achieved with a DB-WAX capillary column (30 m × 0.25 mm × 0.25 μm, J&W Scientific, Folsom, CA, USA). The SPME fiber (75 μm, carboxen/poly-dimethylsiloxane) was desorbed at 250 °C for 10 min with a splitless mode. The column temperature program was initially at 40 °C for 3 min, 40–60 °C at 5 °C min⁻¹, 80–160 °C at 10 °C min⁻¹, 160–250 °C at 4 °C min⁻¹, and 230 °C for 7 min. Helium was used as carrier gas with a constant flow rate of 1.8 mL min⁻¹. Mass spectra were obtained by electron ionization at 70 eV and a scan range from 20 to 380 m/z. Kovats indices (KIs) were calculated using an n-alkane series (C5–C26) under the same chromatographic conditions as the samples. Compound identification was based on comparison of KIs with those of the standards or those reported in the literature.

2.6. HPLC analysis

4-Hydroxy-5-methyl-3(2H)-furanone (HMF), DDMP, and maltol were determined by a Waters E2695 HPLC system equipped with a diode array detector. Separation was carried out on a Zorbax SB-C18 column (5 μm, 250 mm 4.6 mm i.d., Agilent Inc.), with a linear methanol/water gradient of 5–85% methanol over 30 min. The column was washed between runs with 100% methanol. Reaction mixtures were filtered through 0.45 μm membrane, and the filtrations (20 μL) were injected to the column without further pretreatment. Chromatograms were monitored at 254 nm. HMF, DDMP, and maltol were identified by direct comparison of retention time with that of the authentic compound under identical conditions and by spiking the samples with the standard. The concentrations of three reference compounds were calculated from the calibration curve determined under the same HPLC conditions.

2.7. Statistical analysis

Analysis of variance (ANOVA) was used to determine significant differences (P < 0.05) among population means. All statistical analyses and calculations of means and standard deviations were performed using Statistical Analytical System software (SAS Inst. Inc., Cary, NC, USA). The correlations between inhibition ability of tyrosinase and volatile compounds, non-volatile compounds (HMF, DDMP, and maltol), and Maillard reaction conditions were analyzed by PLSR. The detailed PLSR analysis method was described by Lin et al.²³ using the Unscrambler version 9.7 (CAMO ASA, Oslo, Norway).

3. Results and discussion

3.1. Characterization of enzymatic browning inhibition activity of Cys-MRPs

Cysteine has been extensively studied as an inhibitor of enzymatic browning. It is commonly believed that cysteine interacts with the o-quinone, the initial product of enzymatic browning, to form colorless adduct compounds which are stable and no longer participate in polymerization reactions to form melanin.²⁴ As depicted in Fig. 2A, a typical enzymatic browning progress curve was observed in the presence of cysteine, with a lag period appearing when cysteine concentration was increased.¹¹ Richard et al. suggested that the inhibition effect of cysteine is affected by the molar ratio of cysteine to phenolic substrate: when the molar ratio of cysteine to phenolic substrate is greater than or equal to 1, cysteine could, in theory, permanently inhibit enzymatic browning. On the contrary, the o-quinone, formed in excess, can interact with colorless o-quinone-cysteine adduct compounds, leading to regeneration of phenolic substrate with concomitant formation of a deep color.^24,25 However, the present study showed that when Cys concentration reaches 0.42 mM at a constant substrate concentration of 0.67 mM, enzymatic browning can be completely inhibited, suggesting that the critical point of molar ratio of cysteine to phenolic substrate is smaller than 1. These results are in agreement with some previous research, which suggested that cysteine had no, or only very slight, direct effect on activity of PPO from various sources.^21,26–28 When excess dose of substrate was added to the system, the initial browning rate decreased after the lag period with increased addition of cysteine, suggesting that the adduct product formed by cysteine with o-quinones might possess a certain degree of inhibition activity on enzymatic browning. These results are in agreement with earlier findings of Richard et al., who reported that the adduct product of cysteine and 4-methylcatechol was a competitive inhibitor for apple PPO, as assessed by polarography experiments.²⁵ However, Kuijpers found that the adduct products from cysteine and chlorogenic acid exhibited no inhibitory effect on mushroom tyrosinase.²⁸

Fig. 2 Spectrophotometric determination of CA oxidation at 410 nm in the presence of varying concentrations of cysteine (A) or Cys-MRPs (B and C). (D) Dose–response curve for the anti-browning effect of cysteine and Cys-MRPs on mushroom tyrosinase. MR condition for (B) Cys 0.25 M + glucose 0.5 M, initial pH 4.0, 125 °C incubated for 90 min; MR condition for (C) Cys 0.25 M + glucose 0.75 M, initial pH 7.0, 140 °C incubated for 150 min. The 3 mL reaction mixture contained 0.1 M phosphate buffer (pH 8.0), 0.67 mM CA, 50 units per mL tyrosinase, and the indicated concentrations of cysteine, at 25 °C.

Recent studies showed that the enzymatic browning inhibition ability of MRPs is greatly influenced by Maillard reaction parameters such as temperature and pH.^11,29 Our preliminary response surface methodology experiments revealed that the optimal Maillard reaction conditions for generating the maximal inhibition ability were 140 °C, pH 7.0 for 150 min reaction time. The browning inhibition characteristics of Cys-MRPs prepared under the optimal conditions were compared with those of Cys-MRPs prepared under normal conditions (Fig. 2B and C). Cys-MRPs prepared under different Maillard reaction conditions showed different inhibition patterns. The inhibition pattern of Cys-MRPs in Fig. 2B is similar to that of cysteine with a lag period, whereas the Cys-MRPs as shown in Fig. 2C, had no lag period. A possible explanation is that Cys-MRPs prepared at 125 °C, pH 4.0, for 90 min (Fig. 2B) still contain a certain amount of unreacted cysteine, whereas all cysteine has reacted in Cys-MRPs prepared at 140 °C, pH 7.0 for 150 min (Fig. 2C). It is well known that Cys-MRPs can directly inhibit PPO activity.¹¹ The Cys-MRPs as depicted in Fig. 2B exhibited the same inhibition characteristic as Cys (Fig. 2A), and also gave a similar dose–response curve (Fig. 2D). While for the Cys-MRPs as depicted in Fig. 2C, the inhibitory capacity was greatly improved, and 0.052 mM Cys-MRPs could completely inhibit the enzymatic browning, indicating that the newly formed compounds have potent browning inhibition ability. As depicted in Fig. 2D, anti-browning capacity and mechanism of Cys-MRPs is greatly influenced by Maillard reaction conditions.

3.2. Time-dependent manner, reversibility of inhibition of Cys-MRPs on tyrosinase, and effects of supplementation of copper ion

To investigate whether Cys-MRPs inhibit tyrosinase in a time-dependent manner, tyrosinase was pre-incubated with Cys-MRPs for different times. The initial reaction rate was determined by measuring the oxygen consumption rate. The reaction rate rapidly decreased with increasing pre-incubation time. After 2 min pre-incubation, the activity decreased by 50%, and further to 20% after 15 min (Fig. 3). These results are in accordance with the findings of Roux et al., who reported that pre-incubation of apple PPO with Cys-MRPs or heated Cys decreased enzyme activity over time.³⁰ Billaud et al. found that pre-incubation of apple PPO with cysteine (5–10 mM) for 5–180 min at 4 °C did result in inactivation of the enzyme.¹⁹

Fig. 3 Effect of Cys-MRP prepared from cysteine/glucose (0.25 M/0.75 M) aqueous model system heated at 125 °C for 90 min on tyrosinase inactivation during a pre-incubation time varying from 0 to 15 min. In the assay medium, tyrosinase was pre-incubated with Cys-MRP (100 mL) at 25 °C. After pre-incubation for the appropriate time, enzyme activity was measured. Activity was measured by polarography at 25 °C and at pH 6.8, using CA (0.67 mM) as the substrate and 50 units per mL PPO. Bars represent the standard deviation from triplicate determinations.

To investigate whether the observed time-dependent mode of inhibition of tyrosinase by Cys-MRPs is reversible, tyrosinase was first incubated with Cys-MRPs and then separated by diafiltration. Any remaining activity of tyrosinase was determined and the diafiltered, CuCl₂-treated tyrosinase was taken as control. Results showed that tyrosinase incubated with Cys-MRPs followed by diafiltration exhibited decreased activity with declining oxygen consumption rate (Fig. 4), indicating an irreversible inhibitory effect. These results are in accordance with Cheriot et al., who demonstrated that inhibition of eggplant PPO by Cys-MRPs prepared by reacting cysteine and glucose (0.25 M/1 M) at 115 °C initial pH 2 for 200 min was almost irreversible, suggesting that the MRPs might irreversibly denature the major part of the enzyme or chelate the copper present in the enzyme to form a complex, which cannot be totally dissociated by dialysis (or gel filtration).¹⁶

Fig. 4 Oxygen consumption progressing curve of CA (0.67 mM) catalyzed by differently pretreated tyrosinase samples (35 U mL⁻¹), as indicated in the legend. The control tyrosinase was incubated without inhibitor, and underwent the same diafiltration and CuCl₂-treatment procedures as the other samples.

To study whether the inactivation of tyrosinase by Cys-MRPs involves action on the copper ions at the active center of tyrosinase, reactivation was attempted by supplementing pretreated tyrosinase with CuCl₂. DTT was used as a control, as it is known to reversibly inhibit tyrosinase.³¹ Comparing the activities of the pretreated tyrosinase samples (Fig. 4), it was found that copper supplementation could only partially reactivate the tyrosinase, which was inactivated by Cys-MRPs, but it could largely reactivate DTT-inactivated tyrosinase. Based on these results, it could be concluded that Cys-MRPs irreversibly inactivated tyrosinase, and that inactivation might be partially attributed to chelation of copper ions. Maillard et al. investigated the copper chelating capacity of several MRPs derived from different amino acids and the relationship with PPO inhibition, and reported that the sulfhydryl groups of cysteine and glutathione are important to the copper chelating properties, making them outstanding among other amino acids.¹⁵

3.3. Inhibition effect of Cys-MRP in the presence of reversible competitive inhibitor

Kojic acid and tropolone are two known reversible competitive inhibitors.^6,32 As competitive inhibitors can compete with substrate for the binding site of enzyme active center, they can be separated by diafiltration, and used to study the possible active site of Cys-MRPs inhibition on tyrosinase. Tyrosinase was pretreated with Cys-MRPs and/or tropolone or kojic acid competitive inhibitors. If the inhibition effect of Cys-MRPs was impaired or eliminated in the presence of tropolone or kojic acid, this could highlight that Cys-MRPs need access to the active center of enzyme to exert an inhibition effect.

Pretreatment of tyrosinase with tropolone or kojic acid followed by diafiltration did not inactivate the enzyme, indicating that these inhibitors are indeed reversible (Fig. 5). When tyrosinase was incubated with equimolar Cys-MRPs and tropolone together, efficiency of inhibition of the tyrosinase was lower than that of Cys-MRPs alone, and a similar trend was observed when tropolone was replaced by kojic acid but to a lesser extent. All these results suggest that the inhibitory effect of Cys-MRPs on tyrosinase is partially attributed to action on the active center of the enzyme or, more accurately, on the copper ions of the active center. The difference between tropolone and kojic acid probably results from higher affinity of tropolone to the enzyme. Chen et al. reported that the inhibitor constant K_i of kojic acid ranged from 0.02 to 0.71 mM depending on the enzyme and substrate, in particular the K_i value for mushroom tyrosinase and substrate DL-dopa is 0.02 mM.⁷ Espín et al. reported that the K_i value of tropolone for mushroom tyrosinase and substrate 4-tert-butylcatechol is 8–11.5 μM.³³

Fig. 5 Oxygen consumption progressing curve of incubations of CA (0.67 mM) catalyzed by differently pretreated and subsequently diafiltered tyrosinase (35 U mL⁻¹), as indicated in the legend. The competitive tyrosinase inhibitors tropolone and kojic acid were used in the pretreatments.

3.4. Correlation between volatile compounds formed and their browning inhibition ability

To study the relationships between browning inhibition rate and GC-MS profiles, 12 MRPs prepared under different temperature (110–140 °C), times (2–6 h), and ratios of sugar to amino acid (0.5–2) were analyzed by GC-MS. More than 100 compounds were identified on GC-MS profiles, and 59 compounds were detected in at least three reaction temperatures – these were used as variables in the subsequent PLSR analysis.

The volatiles formed from the Maillard reaction between ribose and cysteine, were dominated by sulfur-containing heterocyclics with thiophene and its derivatives.³⁴ As shown in Table 2, 59 volatile compounds were generally classified as thiols, thiazoles, thiophenes, disulfides, pyrazines, furans, and other sulfur-containing compounds.

Table 2 Approximate quantities^a of volatile compounds detected in MRPs as affected by Maillard reaction conditions^b in ng g⁻¹

KI	Compound	ID	RC1	RC2	RC3	RC4	RC5	RC6	RC7	RC8	RC9	RC10	RC11	RC12
a Approximate quantities in headspace (ng mL^−1 of mixture) given as means of independent experiments; —, below detection limit (0.07 ng mL^−1 of mixture).b Each model system consisted of different reactant ratios of glucose to cysteine and reacted at different times and temperatures. Reliability of the identification proposal was indicated by the following: (A) mass spectrum and Kovats index according to the literature; and (B) mass spectrum compared with NIST98 and Wiley mass spectral databases. KI, Kovats index calculated for the DB-WAX capillary column; ID, identification. RC1–RC12 represents 12 MRPs samples prepared according to Table 1.
Thiol
818	3-Mercapto-2-butanone	TL1	0.86	7.21	5.32	5.36	21.45	7.83	1.78	17.31	7.89	—	6.88	1.94
869	2-Methyl-3-furanthiol	TL2	—	2.78	—	0.76	6.35	4.31	8.14	6.37	4.31	—	16.62	7.3
904	3-Mercapto-2-pentanone	TL3	—	—	—	—	—	—	5.68	8.92	15.1	—	4.74	0.81
909	2-Mercapto-3-pentanone	TL4	0.99	5.14	1.89	9.15	53.15	45.37	163.1	77.46	67.49	1.74	6.32	1.02
913	2-Furanmethanethiol	TL5	—	3.12	—	4.79	6.43	8.55	—	7.32	—	—	22.09	0.85
959	2-Thiophenethiol	TL6	—	0.65	—	6.89	3.15	9.53	—	—	—	—	—	—
980	3-Thiophenethiol	TL7	—	—	—	—	—	1.43	—	—	—	—	25.55	26.99
1066	2-Methyl-3-thiophenethiol	TL8	1.05	2.17	4.82	1.83	3.89	7.45	4.81	15.78	4.56	—	25.37	27.23
Disulfide
542	Bis(2-methyl-3-furyl)disulfide	DS1	—	1.94	—	—	1.6	—	1.89	2.94	6.26	4.37	8.54	0.73
1643	2-Methyl-3-(2-furfuryldithio)furan	DS2	—	0.48	—	0.8	4.18	2.91	8.23	15.6	4.91	6.71	0.48	—
1702	2,3-Dihydro-5-methyl-4-[(2-methyl-3-furyl)dithio]furan	DS3	—	3.2	1.66	3.77	6.32	1.86	2.92	13.57	4.45	2.13	3.2	1.66
1745	2-Methyl-3-[(2-methyl-3-thienyl)dithio]furan	DS4	—	1.76	—	0.57	3.2	—	7.05	5.41	5.73	4.81	9.93	11.23
1874	α-Dithiobisthiophene	DS5	—	0.7	1.37	1.07	3.18	5.4	—	3.17	—	—	3.7	6.37
1921	2-Methyl-3-(2-thienyldithio)thiophene	DS6	—	1.34	2.22	—	—	—	4.38	4.31	9.27	2.88	10.4	12.22
Thiophene
68	2-Thienylmethanol	TP1	—	—	—	—	3.79	—	1.58	8.77	4.74	16.48	22.96	6.6
1095	2-Acetylthiophene	TP2	—	5.47	—	—	5.18	0.72	1.37	12.54	6.28	4.27	3.78	0.99
1128	3-Methyl-2-formylthiophene	TP3	—	—	—	1.53	7.28	2.67	23.75	10.33	17.43	41.53	28.05	10.48
1149	5-Methyl-2-thiophenecarboxaldehyde	TP4	—	2.51	—	1.74	7.91	4.36	—	—	—	1.37	3.13	—
1189	2-Propionylthiophene	TP5	—	—	3.59	—	3.71	7.99	16.95	12.73	5.32	2.04	15.23	7.83
1194	2-Acetyl-3-methylthiophene	TP6	0.78	3.18	—	1.07	9.95	0.83	3.62	8.74	1.48	7.19	9.37	3.51
1208	3-Ethyl-2-formylthiophene	TP7	—	—	—	—	—	—	1.16	5.75	2.73	2.41	8.9	4.82
1249	α-Dimethylformylthiophene	TP8		0.92	—	—	3.26	0.68	1.65	8.71	7.14	3.87	3.66	2.46
1289	3-Acetyl-2,5-dimethylthiophene	TP9	1.17	—	—	3.84	1.03	—	54.83	7.43	2.36	127.4	6.9	1.1
Thiophenone
959	Dihydro-3(2H)-thiophenone	TO1	—	—	—	—	2.73	—	3.61	13.18	4.72	0.97	1.58	1.05
996	Dihydro-2-methyl-3(2H)-thiophenone	TO2	—	4.81	—	—	7.43	0.58	0.63	21.66	6.16	2.52	31.75	5.33
1026	Dihydro-2,(4 or 5)-dimethyl-3(2H)-thiophenone (E or Z)	TO3	—	2.19	0.52	—	1.9	3.17	1.8	12.61	3.55	0.65	6.96	2.72
Fused bicyclic compound
197	2,3-Dihydro-6-methylthieno[2,3-c]furan	FB1	4.71	3.69	7.25	12.54	19.67	15.43	104.53	35.62	3.89	199.78	56.08	1.94
1220	Thieno[2,3-b]thiophene	FB2	10.27	14.51	19.24	2.67	4.73	8.42	4.32	8.56	12.35	2.48	17.36	17.78
1325	α-Dihydrothienothiophene	FB3	2.57	4.31	4.68	16.01	26.89	17.53	57.32	44.67	17.84	271.68	441.52	91.65
1364	2-Methyl-thieno[3,2-b]thiophene	FB4	—	1.75	4.67	—	—	—	—	4.83	—	10.43	21.2	16.32
1381	α-Methyldihydrothienothiophene	FB5	4.26	3.15	1.96	0.75	—	—	—	—	—	11.92	75.56	4.65
1477	α-Dimethyldihydrothienothiophene	FB6	—	—	—	—	—	—	1.95	7.39	8.54	3.39	36.62	2.95
Polysulfur heterocyclic compound
1153	3,5-Dimethyl-1,2,4-trithiolane (E or Z)	PS1	—	3.28	—	—	5.02	2.57	4.02	12.31	6.44	2.15	7.38	5.34
1160	3,5-Dimethyl-1,2,4-trithiolane (E or Z)	PS2	—	1.9	—	—	5.9	2.77	—	7.92	3.67	—	4.73	—
1185	1,2-Dithian-4-one	PS3	—	2.11	—	1.5	8.1	0.67	—	4.29	—	2.16	10.52	1.7
1232	3-Methyl-1,2-dithian-4-one	PS4	—	2.16	—	—	8.96	1.94	—	—	—	1.26	18.6	5.23
1266	3-Methyl-1,2,4-trithiane	PS5	—	—	3.07	—	4.59	9.07	—	7.61	12.33	—	8.9	16.47
1428	3,6-Dimethyl-1,2,4,5-tetrathiacyclohexane (E or Z)	PS6	—	—	—	—	—	—	—	3.17	—	4.61	13.2	2.18
Pyrazine
825	Methylpyrazine	PY1	2.59	6.81	3.74	—	3.38	9.51	—	—	—	0.61	5.75	—
914	2,6-Dimethylpyrazine	PY2	3.46	5.14	6.93	1.07	3.37	4.87	9.41	7.13	10.67	56.18	22.12	34.76
918	Ethylpyrazine	PY3	—	1.72	—	4.13	10.63	15.78	35.17	25.17	29.46	78.72	103.45	89.15
919	2,3-Dimethylpyrazine	PY4	5.17	4.85	17.21	2.14	5.31	1.76	—	5.71	3.28	7.15	11.38	3.79
999	2-Ethyl-6-methylpyrazine	PY5	—	—	—	10.29	15.85	18.42	7.38	17.53	21.78	45.13	78.31	68.35
1003	Trimethylpyrazine	PY6	1.23	5.39	2.17	18.47	27.16	7.35	0.59	9.27	3.16	1.67	5.89	0.93
Thiazole
933	4,5-Dimethylthiazole	TZ1	3.47	7.36	9.75	3.8	6.78	19.6	7.18	1.87	2.99	3.16	5.17	4.57
997	2,4,5-Trimethylthiazole	TZ2	—	6.77	8.59	15.43	52.32	95.7	15.65	39.27	73.1	0.81	1.54	7.18
1022	2-Acetylthiazole	TZ3	—	9.7	9.33	8.93	37.11	29.05	5.9	79.15	49.37	4.95	11.12	3.32
1024	5-Ethyl-2-methylthiazole	TZ4	—		—	2.19	11.86	5.61	4.81	15.82	9.67	—	—	—
1077	5-Ethyl-2,4-dimethylthiazole	TZ5	3.65	4.18	7.44	14.96	18.62	30.17	8.7	7.53	10.84	—	1.94	2.2
1113	2-Acetyl-4-methylthiazole	TZ6	—	—	—	8.49	27.18	12.7	19.08	42.39	25.17	1.25	2.1	1.45
1241	2,5-Diethyl-4-methylthiazole	TZ7	—	3.19	8.29	8.72	9.55	17.43	3.78	6.81	19.36	0.57	2.13	3.14
Miscellaneous
832	Furfural	MI1	1.6	5.79	2.1	12.25	3.19	4.76	27.2	18.51	11.63	51.5	32.7	27.94
850	2-Furanmethanol	MI2	—	—	—	9.28	—	—	24.35	19.56	15.75	3.02	2.29	—
910	2-Acetylfuran	MI3	4.19	0.73	—	10.3	7.18	2.67	—	4.33	—	—	—	—
936	5-Methylfurfural	FB4	1.89	2.94	6.26	17.05	25.41	15.73	14.38	45.31	9.27	2.92	29.57	4.45
952	1-(2-Furyl)-2-propanone	MI5	19.2	8.23	0.52	—	—	—	4.32	5.18	3.76	0.93	2.75	0.52
1004	Benzofuran	MI6	—	2.6	5.81	3.75	7.56	10.5	—	3.71	8.29	—	—	—
1016	2-Propionylfuran	MI7	7.1	16.23	14.96	—	—	—	0.82	3.87	1.4	—	—	—
1086	α-Dimethyl-2-formylfuran	MI8	27.19	39.5	15.43	19.61	8.25	2.1	—	—	—	—	—	—

---

PLSR was performed to process data from GC-MS analysis of MRPs and tyrosinase inhibition rate. The X-matrix was designed as GC-MS data of 12 MRPs derived from various Maillard reaction conditions (time, temperature, and ratio of cysteine to glucose); the Y-matrix was designed as tyrosinase inhibition rate (Fig. 6). The derived PLSR model included two significant PCs and explained 80% of the cross-validated variance. Furthermore, more PCs did not provide any predictive improvement in the Y-matrix obtained. PC1 versus PC2 was explored and is presented in Fig. 6.

Fig. 6 PLSR results: (A) standardized, estimated regression coefficients and significant indications (streaked bars) from PLS1 prediction models for the enzymatic browning inhibition rate from volatile compounds of MRPs. (B) Correlation loadings plot for GC-MS data (X-matrix) and enzymatic browning inhibition rate (Y-matrix). Significant variables are marked with circles. Ellipses show r² = 50% and 100%, respectively.

As shown in Fig. 6B, the volatile compounds of MRPs marked with small circles were found to be significant in browning inhibition ability. The big circles indicate 50% and 100% explained variance, respectively. The measured parameter of browning inhibition rate was located in the first quadrant between the inner and outer ellipses, r² = 50% and 100%. More volatile compounds were located on the right side in PC1, suggesting that the total amount of volatile compounds of Maillard reaction was positively correlated with the enzymatic browning inhibition ability.

Further studies of the relationships were carried out by calculating estimated regression coefficients from the jack-knife uncertainty test. The results from the PLS1 regression analysis described the contribution of volatile compounds of MRPs to the browning inhibition ability. 3-Ethyl-2-formylthiophene, α-dimethylformylthiophene (TP7, 8), 2,6-dimethylpyrazine, ethylpyrazine, 2-ethyl-6-methylpyrazine (PY2, 3, 5), 2-methyl-3-(2-thienyldithio) thiophene (DS6), and furfural (MI1) showed significant and positive contribution to tyrosinase inhibition ability, while 2-propionylfuran and α-dimethyl-2-formylfuran (MI7, 8) showed a significant but negative contribution to tyrosinase inhibition ability.

3.5. Correlation between non-volatile compounds formed and their browning inhibition ability

HMF, maltol, and DDMP were chosen as typical reference compounds to study their relationship with the inhibition ability of final products.¹⁷ While HMF and DDMP are, respectively, typical compounds of two enolization pathways of Maillard, maltol in the glucose-MRPs can be taken as an indicator of Maillard reaction under high intensity.³⁵

As shown in Fig. 7, only DDMP showed significant and positive correlation with the inhibition ability, whereas HMF and maltol showed weak negative correlation. DDMP itself cannot inhibit tyrosinase activity but could be used as an index for evaluating inhibition ability of MRPs derived from cysteine and glucose.

Fig. 7 PLSR results: (A) standardized, estimated regression coefficients and significant indications (streaked bars) from PLS1 prediction models for the enzymatic browning inhibition rate from non-volatile compounds of MRPs and Maillard reaction condition. (B) Correlation loadings plot for non-volatile compounds of MRPs and Maillard reaction conditions (X-matrix) with enzymatic browning inhibition rate (Y-matrix). Significant variables are marked with circles. Ellipses show r² = 50% and 100%, respectively.

The PLSR results of Maillard reaction conditions and inhibition rate, indicated that reaction temperature and time have a significant and positive effect on the inhibition rate of final products within the experimental range. These results are in agreement with previous findings that increasing Maillard reaction intensity contributes to tyrosinase inhibition ability; whereas increasing ratio of sugar to amino acid in a certain range decreases inhibition ability.

4. Conclusions

Based on the results of this work, Cys-MRPs can directly inactivate mushroom tyrosinase in a time-dependent irreversible manner. To our knowledge, information concerning the site of action of Cys-MRPs on tyrosinase and the possible interaction between the copper ion of tyrosinase and Cys-MRPs has never been reported. Inactivation of tyrosinase by Cys-MRPs was found to be partially irreversible and can be partially attributed to chelation of copper ions. The results also show that Cys-MRPs might act in the active site of tyrosinase to inhibit enzymatic browning. PLSR analysis between tyrosinase inhibition rate and the GC-MS data of MRPs indicated that some volatile compounds are significantly correlated with the tyrosinase inhibition ability, such as 3-ethyl-2-formylthiophene, α-dimethylformylthiophene, 2,6-dimethylpyrazine, ethylpyrazine, 2-ethyl-6-methylpyrazine, 2-methyl-3-(2-thienyldithio)thiophene, and furfural. Of the three non-volatile reference compounds, only DDMP showed significantly positive correlation with the inhibition ability. Additional studies are needed to elucidate the main mechanisms involved in inhibition of PPO (copper chelating or sequestering properties, action on the protein structure, free radical scavenging properties), this knowledge being essential for better control of enzymatic browning of fruit and vegetables.

Acknowledgements

This study was supported by the National Hi-Technology Research & Development 416 Programme of China (2013AA102204).

References

1. M. V. Martinez and J. R. Whitaker, Trends Food Sci. Technol., 1995, 6, 195–200.
2. Y.-J. Kim and H. Uyama, Cell. Mol. Life Sci., 2005, 62, 1707–1723.
3. M. Friedman, J. Agric. Food Chem., 1996, 44, 631–653.
4. G. M. Sapers, Food Technol., 1993, 47, 75–84.
5. G. M. Sapers and R. L. Miller, J. Food Sci., 1992, 57, 1132–1135.
6. V. Kahn, P. Lindner and V. Zakin, J. Food Biochem., 1995, 18, 253–271.
7. J. S. Chen, C.-i. Wei, R. S. Rolle, W. S. Otwell, M. O. Balaban and M. R. Marshall, J. Agric. Food Chem., 1991, 39, 1396–1401.
8. T.-S. Chang, Int. J. Mol. Sci., 2009, 10, 2440–2475.
9. B. K. Tan and N. D. Harris, Food Chem., 1995, 53, 267–273.
10. M. C. Nicoli, B. E. Elizalde, A. Pitotti and C. R. Lerici, J. Food Biochem., 1991, 15, 169–184.
11. C. Billaud, E. Roux, S. Brun-Merimee, C. Maraschin and J. Nicolas, Food Chem., 2003, 81, 35–50.
12. C. Billaud, C. Maraschin, Y. N. Chow, S. Cheriot, M. N. Peyrat-Maillard and J. Nicolas, Mol. Nutr. Food Res., 2005, 49, 656–662.
13. S. Chériot, C. Billaud, M.-N. Maillard and J. Nicolas, Mol. Nutr. Food Res., 2007, 51, 395–403.
14. J. J. Wu, K. W. Cheng, E. T. S. Li, M. F. Wang and W. C. Ye, Food Chem., 2008, 109, 379–385.
15. M. N. Maillard, C. Billaud, Y. N. Chow, C. Ordonaud and J. Nicolas, LWT–Food Sci. Technol., 2007, 40, 1434–1444.
16. S. Cheriot, C. Billaud and J. Nicolas, J. Agric. Food Chem., 2006, 54, 5120–5126.
17. T. Davidek, N. Clety, S. Devaud, F. Robert and I. Blank, J. Agric. Food Chem., 2003, 51, 7259–7265.
18. M. V. Martinez and J. R. Whitaker, Trends Food Sci. Technol., 1995, 6, 195–200.
19. C. Billaud, E. Roux, S. Brun-Merimee, C. Maraschin and J. Nicolas, Food Chem., 2003, 81, 35–50.
20. J.-J. Wu, K.-W. Cheng, E. T. S. Li, M. Wang and W.-C. Ye, Food Chem., 2008, 109, 379–385.
21. A. H. Janovitz-Klapp, F. C. Richard, P. M. Goupy and J. J. Nicolas, J. Agric. Food Chem., 1990, 38, 926–931.
22. S. Song, L. Yuan, X. Zhang, K. Hayat, H. Chen, F. Liu, Z. Xiao and Y. Niu, Food Chem., 2013, 141, 4278–4288.
23. S. Lin, X. Zhang, S. Song, K. Hayat, K. Eric and H. Majeed, Regul. Toxicol. Pharmacol., 2016, 75, 27–34.
24. F. C. Richard, P. M. Goupy, J. J. Nicolas, J. M. Lacombe and A. A. Pavia, J. Agric. Food Chem., 1991, 39, 841–847.
25. F. C. Richard, P. M. Goupy and J. J. Nicolas, J. Agric. Food Chem., 1992, 40, 2108–2113.
26. E. Valero, R. Varon and F. Garciacarmona, Biochem. J., 1991, 277, 869–874.
27. S. Cheriot, C. Billaud, S. Pochtrager, K. H. Wagner and J. Nicolas, Food Chem., 2009, 114, 132–138.
28. T. F. M. Kuijpers, C. E. Narvaez-Cuenca, J. P. Vincken, A. J. W. Verloop, W. J. H. van Berkel and H. Gruppen, J. Agric. Food Chem., 2012, 60, 3507–3514.
29. B. A. Mogol, A. Yıldırım and V. Gökmen, J. Sci. Food Agric., 2010, 90, 2556–2562.
30. E. Roux, C. Billaud, C. Maraschin, S. Brun-Mérimee and J. Nicolas, Food Chem., 2003, 81, 51–60.
31. Y. D. Park, S. J. Lee, K. H. Park, S. Y. Kim, M. J. Hahn and J. M. Yang, J. Protein Chem., 2003, 22, 613–623.
32. V. Kahn and A. Andrawis, Phytochemistry, 1985, 24, 905–908.
33. J. C. Espín and H. J. Wichers, J. Agric. Food Chem., 1999, 47, 2638–2644.
34. Y. Zhang and C. T. Ho, J. Agric. Food Chem., 1991, 39, 760–763.
35. H. Kim and I. Taub, Food Technol., 1993, 47, 91–99.

---