Health Beneficial Phytochemicals in Dioscorea caucasica Lipsky Leaves and Tubers and Their Inhibitory Effects on Physiologically Important Enzymes

Dioscorea caucasica Lipsky is a tertiary relict endemic plant naturally growing in the western part of the trans-Caucasus regions; it has adapted and successfully grows in the temperate region of the Baltic countries. Information about its phytochemical composition and bioactivities is rather scarce. This study reports the results of the identification of 41 compounds in D. caucasica leaf and tuber hydroethanolic extracts using UPLC-QTOF/MS. Organic acids were found in both extracts; hydroxycinnamates and flavonoids were the main phytochemicals in the leaves, while steroidal glycosides, fatty acids (mainly hydroxylated) and carbohydrates were found in the tubers. Leaf extracts inhibited enzymes in a dose-dependent manner and were remarkably stronger inhibitors of physiologically important enzymes, namely α-amylase (48.6% at 480 µg/mL), α-glucosidase (IC50 = 41.99 and 47.95 µg/mL with and without 0.1 M Na2CO3), acetylcholinesterase (45.85% at 100 µg/mL) and angiotensin-converting enzyme (IC50 = 829.7 µg/mL), most likely due to the presence of some quantified polyphenolic antioxidants. The mode of inhibition of α-glucosidase and acetylcholinesterase was assessed via kinetic studies based on Lineweaver–Burk inhibition plots. Leaf and tuber extracts acted as mixed-type and competitive inhibitors of α-glucosidase, respectively; the leaf extract demonstrated an uncompetitive inhibition mode of acetylcholinesterase. It is expected that this new knowledge of D. caucasica will serve for its valorization in developing new health beneficial ingredients for functional foods and nutraceuticals.


Introduction
The interest in health beneficial natural substances has been steadily increasing during the last few decades. This tendency has been fostered by the fast developments in the area of functional foods and nutraceuticals, which have contributed in shifting healthcare concepts towards the increasing role of preventive medicine. In addition, such a generic characteristic of nutritional and healthcare products as 'naturalness' has become very popular among consumers [1]. Due to the vast number of plant species, which are the main sources of bioactive phytochemicals, this tendency has stable and sustainable support, particularly considering the presence of many under-investigated plants.
Plant phytochemicals exert various bioactivities, which may provide numerous health benefits, particularly in preventing and/or delaying the development of age-dependent degenerative and chronic diseases [2]. Among the mechanisms of their bioactivities, antioxidant and antimicrobial properties, molecular signaling, inhibition of enzymes and other effects have been reported for various plant preparations and purified natural compounds [3][4][5][6][7]. Many studies have reported the inhibition of physiologically important mexicana [26], while piscidic acid was reported in D. nipponica tuber [27]. However, this acid was not reported previously in D. caucasica tubers. The MS data of 2-deoxy-2,3-dehydro- 3 min is also in agreement with the previously reported data [28,29]. These compounds were found only in the leaf extract. Table 1. Identification data of the constituents detected in D. caucasica leaf and tuber extracts by UPLC/Q-TOF-MS in negative ion mode.   [37]; they are included in Table 1 as unseparated sugars.
Long-chain fatty acids and their hydroxylated derivatives are common in various plant materials, including yam tubers of Dioscorea species [38,43,44].  [39]). The latter was assigned to a carnitine derivative, hydroxy dodecanoylcarnitine. Conjugates of fatty acids with amino acids occur widely in food from animal sources, but there is limited availability in plants. The [41] and some other plants [40,42]. The content of neochlorogenic acid, chlorogenic acid, rutin and isoquercitrin in the D. caucasica leaf extract was 5.89, 116.6, 4.02 and 187.2 µg/mL, respectively. Thus, quercetin glycoside and isoquercitrin were the most abundant phenolic constituents recovered from the plant leaves. In general, these data agree with the UPLC-QTOF/MS analysis, which demonstrated the highest m/z intensities for the major quantified compounds. As may be judged from the peak area in the chromatographic profile (Figure 1), some other constituents may also be present in the extract at high concentrations. Based on the peak m/z intensities in the UPLC-QTOF/MS chromatograms, these compounds might be quinic acid, which forms esters with phenolic acids, and quercitrin, which eluted after isoquercitrin (Figure 1), i.e., similarly as in the UPLC-QTOF/MS chromatogram. The content of biologically active phenolic compounds in the leaves was remarkably higher than in the tubers. For instance, the content of neochlorogenic and chlorogenic acids in the tuber extract was only 0.687 and 1.869 µg/mL, respectively. Quantification of other tuber phytochemicals was beyond the scope of this study; however, based on the peak m/z intensities in the UPLC-QTOF/MS chromatograms, the major compounds in the tuber extract might be some sugars, piscidic acid, linolenic acid and two steroidal glycosides with t R 3.7 and 6.6 min ( Table 1).   [45]), from baker's yeast (Saccharomyces cerevisiae) Type I, belongs to the Glycoside Hydrolase 13 (GH_13) family [46], whose specificity is directed mainly towards the exohydrolysis of (1→4)-α-glucosidic linkages [47]. Under the conditions used (T = 37 °C; pH 6.8), α-glucosidase catalyzes the cleavage of the p-nitrophenyl-α-D-glucopyranoside (pNPG) substrate, resulting in the formation of α-D-glucose and p-nitrophenol (pNp). The amount of yellow color pNp liberated during the reaction was monitored by absorbance measurements at λ = 405 nm [48] All in vitro experiments were performed at extract concentrations of 3.125, 6.25, 12.5, 25, 50 and 100 µg/mL. The extracts inhibited α-glucosidase activity in a dose-dependent manner; in the assays without adding 0.1 M Na2CO3 (Figure 2(A1) and Table 2), enzyme activity as compared to control (without inhibitor) was reduced from 11.94 ± 1.22 % to 78.49 ± 2.39% by increasing the concentration from 3.125 to 100 µg/mL. In case of the addition of 0.1 M Na2CO3, enzyme inhibition for the maximal and minimal applied concentrations was 59.30 ± 1.1% and 16.93 ± 1.43% (Figure 2(A2) and Table 2). It may be observed that the differences in enzymatic activity with/without using 0.1 M Na2CO3 were not remarkable; however, they were statistically significant, except for the 25 µg/mL concentra-

Enzyme Inhibitory Properties of D. caucasica Extracts
α-Glucosidase (otherwise known as α-1,4-glucosidase; IUMB enzyme nomenclature number EC 3.2.1.20 [45]), from baker's yeast (Saccharomyces cerevisiae) Type I, belongs to the Glycoside Hydrolase 13 (GH_13) family [46], whose specificity is directed mainly towards the exohydrolysis of (1→4)-α-glucosidic linkages [47]. Under the conditions used (T = 37 • C; pH 6.8), α-glucosidase catalyzes the cleavage of the p-nitrophenyl-α-D-glucopyranoside (pNPG) substrate, resulting in the formation of α-D-glucose and pnitrophenol (pNp). The amount of yellow color pNp liberated during the reaction was monitored by absorbance measurements at λ = 405 nm [48] All in vitro experiments were performed at extract concentrations of 3.125, 6.25, 12.5, 25, 50 and 100 µg/mL. The extracts inhibited α-glucosidase activity in a dose-dependent manner; in the assays without adding 0.1 M Na 2 CO 3 ( Figure 2A1 and Table 2), enzyme activity as compared to control (without inhibitor) was reduced from 11.94 ± 1.22% to 78.49 ± 2.39% by increasing the concentration from 3.125 to 100 µg/mL. In case of the addition of 0.1 M Na 2 CO 3 , enzyme inhibition for the maximal and minimal applied concentrations was 59.30 ± 1.1% and 16.93 ± 1.43% ( Figure 2A2 and Table 2). It may be observed that the differences in enzymatic activity with/without using 0.1 M Na 2 CO 3 were not remarkable; however, they were statistically significant, except for the 25 µg/mL concentration (Table 2). Obviously, these differences may be related to the concentration of biologically active and other compounds present in the extracts. The IC50 values of D. caucasica leaf extracts with/without 0.1 M Na2CO3 at the tested concentrations were 41.99 µg/mL and 47.95 µg/mL, respectively. For comparison, acarbose at 200 µg/mL concentration reduced α-glucosidase activity by 71.17%. Most likely, the high inhibitory activity of D. caucasica leaf extracts may be attributed to the chlorogenic acids, quercitrin and isoquercitrin, although the contribution of other non-identified compounds cannot be discounted. In contrast, the tuber extracts of D. caucasica displayed rather weak inhibitory activity against α-glucosidase; for instance, at 500 µg/mL, enzyme activity was reduced by 40.79 ± 2.71% ( Figure 3).   The IC 50 values of D. caucasica leaf extracts with/without 0.1 M Na 2 CO 3 at the tested concentrations were 41.99 µg/mL and 47.95 µg/mL, respectively. For comparison, acarbose at 200 µg/mL concentration reduced α-glucosidase activity by 71.17%. Most likely, the high inhibitory activity of D. caucasica leaf extracts may be attributed to the chlorogenic acids, quercitrin and isoquercitrin, although the contribution of other non-identified compounds cannot be discounted. In contrast, the tuber extracts of D. caucasica displayed rather weak inhibitory activity against α-glucosidase; for instance, at 500 µg/mL, enzyme activity was reduced by 40.79 ± 2.71% (Figure 3). the high inhibitory activity of D. caucasica leaf extracts may be attributed to the chlorogenic acids, quercitrin and isoquercitrin, although the contribution of other non-identified compounds cannot be discounted. In contrast, the tuber extracts of D. caucasica displayed rather weak inhibitory activity against α-glucosidase; for instance, at 500 µg/mL, enzyme activity was reduced by 40.79 ± 2.71% ( Figure 3).
D. caucasica leaf extracts displayed a mixed-type non-competitive mode of inhibition of α-glucosidase ( Figure 2A2,B2). By virtue of the Lineweaver-Burk plot in the enzymatic reaction system I (Table 3), the V max(app) decreased from 0.157 ± 0.03 to 0.134 ± 0.004 mg/L·min after increasing the inhibitor concentration from 15 to 25 µg/mL; K m(app) values were 29.76 ± 1.54 and 33.61 ± 0.75 mg/L, respectively. In contrast, in the enzymatic reaction system II (Table 3), the increase in extract concentration resulted in an increase in the V max(app) value, from 0.402 ± 0.041 to 0.449 ± 0.026 mg/L·min. The K m(app) value in this case was more than two-fold higher (86.91 mg/L) at 25 µg/mL than at 15 µg/mL (39.984 mg/L). It was established that an increased K m and unchanged V max show competitive inhibition, while the decreased V max and increased/decreased K m values (Table 3) indicate the mixedtype inhibition model. Despite significant differences in kinetic constant (V max(app), K m(app) ) values among enzymatic reaction systems, noticeable is the hallmark of non-competitive inhibition. The double reciprocal plot showed a group of straight lines with different slopes, which intersect at the third quadrant ( Figure 2B2)/second quadrant ( Figure 2A2), suggesting that the extracts act as mixed-type inhibitors [50].   Table 3). The decreased rate of catalysis confirms the interaction of the inhibitor with α-glucosidase, which reduces product formation. The kinetic reaction model was also elaborated for D. caucasica tuber extract. At the concentrations of 200 and 500 µg/mL, it competitively inhibited α-glucosidase: the K m(app) values increased from 69.34 ± 0.31 to 83.33 mg/L, whereas V max(app) 0.99 ± 0.22 and 1.034 ± 0.38 mg/L·min were not significantly different (Table 3). Further, the Lineweaver-Burk plot provides information about 1/V max and −1/K m for the α-glucosidase kinetics. As may be observed (Figure 3), the increased K m value and insignificant changes in the V max value in the case of increasing inhibitor concentrations suggests the mechanisms of a competitive mode of inhibition [50,51]

α-Amylase
As the tuber extract possessed remarkably weaker α-glucosidase inhibitory effects, further studies were focused on the leaf extract. It is evident that D. caucasica leaf extracts inhibited the α-amylase enzyme ( Figure 4); at 480, 400, 240, 200, 160 and 80 µg/mL concentrations, it reduced α-amylase activity by 48.6 ± 2.2%, 43.4 ± 1.7%, 42.6 ± 2.3%, 42.2 ± 1.3%, 19.0 ± 0.9% and 12.8 ± 0.9%, respectively. The inhibitory effect of the plant was compared with the standard drug, acarbose, which reduced enzyme activity by approximately 93.8% at the concentration of 50 µg/mL. The inhibition increased with extract concentration; however, there was no linear dependence. At the concentration of ≥200 µg/mL, the changes in percentage inhibition were not significant. Most likely, it depends on the substrate composition and the mode of inhibition, which will be demonstrated further in the studies of reaction kinetics with other two tested enzymes. From these data, we can conclude that D. caucasica leaf extract was a moderate inhibitor (40-50% inhibition at ≥200 µg/mL concentration) of α-amylase, with a remarkably lower effect than acarbose.
1.3%, 19.0 ± 0.9% and 12.8 ± 0.9%, respectively. The inhibitory effect of the plant was compared with the standard drug, acarbose, which reduced enzyme activity by approximately 93.8% at the concentration of 50 µg/mL. The inhibition increased with extract concentration; however, there was no linear dependence. At the concentration of ≥200 µg/mL, the changes in percentage inhibition were not significant. Most likely, it depends on the substrate composition and the mode of inhibition, which will be demonstrated further in the studies of reaction kinetics with other two tested enzymes. From these data, we can conclude that D. caucasica leaf extract was a moderate inhibitor (40-50% inhibition at ≥200 µg/mL concentration) of α-amylase, with a remarkably lower effect than acarbose.

Acetylcholinesterase (AChE) Inhibition Assay
AChE (IUMB enzyme nomenclature number EC 3.1.1.7), also known as acetylcholine hydrolase, choline esterase I, cholinesterase, acetylthiocholinesterase and acetyl-βmethylcholinesterase, belongs to the cholinesterase (ChEs) family, which is most widely known for hydrolyzing the neurotransmitter acetylcholine (ACh) to choline (Ch) and acetic acid in the synaptic cleft. The principle of this assay method is based on the production of tiocholine (SCh), which is formed from acetylthiocholine during enzymatic hydrolysis with Ellman's reagent (DTNB) and produces yellow-colored chromophore 5-thio-2-nitrobenzoate. The rate of appearance of the yellow derivative is measured spectrophotometrically at 412 nm [52].
The samples for the assay were prepared from the lyophilized powder and tested for their ability to inhibit AChE catalysis in the concentration range of 25-100 µg/mL. The results are listed in Table 4. It may be observed that the inhibition noticeably increased after increasing the concentration of the leaf extract from 25 µg/mL (inhibition = 27.55%) to 40 µg/mL (inhibition = 40.58%), while a further increase in the concentration to 80 µg/mL resulted in a significant increase in AChE inhibition that was, however, less remarkable. However, at 80 and 100 µg/mL extract concentrations, there were no significant (p > 0.05) differences in decreased enzyme activity. The inhibitory effect of D. caucasica solution on AChE was compared with the synthetic inhibitor Donepezil HCl. As expected, the positive control was a more potent AChE inhibitor than the extracts of D. caucasica: at a 16 µg/mL concentration, it reduced enzyme activity by 70.62 ± 2% and V max to 0.0593 µM/min/mg protein.

Acetylcholinesterase Inhibition Kinetics
The results obtained indicate that the AChE inhibition mechanism of D. caucasica leaf extracts may be rather complex; therefore, enzyme inhibition kinetic studies were performed by monitoring enzyme activity at varying concentrations of acethylthiocholine iodide in the range of 45-160 µmol/L. The Lineweaver-Burk inhibition plots of 1/V versus 1/[ATChI] gave the following equations: without inhibitor: y = 569.49 + 2.3524, r 2 = 0.9874; with 50 µg/mL of leaf extract: y = 483.75 + 6.7466, r 2 = 0.9755. D. caucasica leaf extracts displayed an uncompetitive mode of inhibition with respect to AChE ( Figure 5). By virtue of the Lineweaver-Burk plot in the enzymatic reaction system with pure enzyme, the V max, K m values were 0.425 ± 0.11 µM/mg protein/min and 242.08 ± 15.76 µmol/L, respectively; meanwhile, in the case of using extracts at the concentration of 50 µg/mL, the V max(app), K m(app) values were remarkably lower, 0.148 ± 0.015 µM/min/mg protein and 71.56 ± 6.55 µmol/L, respectively. This profile of inhibition indicates that the inhibitor only binds to the enzyme-substrate complex [51].

Angiotensin-Converting Enzyme (ACE)
D. caucasica leaf extracts also inhibited ACE in a dose dependent manner ( Figure 6): at the concentrations of 250, 500, 1000 and 1250 µg/mL, the enzyme inhibition was 35.46 ± 0.76%, 38.45 ± 0.52%, 51.12 ± 0.81% and 58.34 ± 1.01%. The inhibitory effect of plant extracts was compared with the synthetic antihypertension drug Captopril; at 100 µg/mL, the inhibition level was 74.3 ± 0.57%, while the effective concentration IC 50 (ACE inhibition 50%) for the botanical extract was 829.7 µg/mL. Nevertheless, D. caucasica leaf extract may be considered a promising ingredient for the soft control of low/moderate hypertension. D. caucasica leaf extracts also inhibited ACE in a dose dependent manner ( Figure 6): at the concentrations of 250, 500, 1000 and 1250 µg/mL, the enzyme inhibition was 35.46 ± 0.76%, 38.45 ± 0.52%, 51.12 ± 0.81% and 58.34 ± 1.01%. The inhibitory effect of plant extracts was compared with the synthetic antihypertension drug Captopril; at 100 µg/mL, the inhibition level was 74.3 ± 0.57%, while the effective concentration IC50 (ACE inhibition 50%) for the botanical extract was 829.7 µg/mL. Nevertheless, D. caucasica leaf extract may be considered a promising ingredient for the soft control of low/moderate hypertension.

Discussion
The search for and evaluation of biologically active nutrients in under-investigated plant materials is an important task, which may serve as a good platform in developing new effective ingredients for nutraceuticals and functional foods. In general, studies on the enzyme inhibitory activities of Dioscorea preparations are rather scarce. For instance, the extracts of D. alata and D. bulbifera tubers were tested as α-glucosidase and α-amylase

Discussion
The search for and evaluation of biologically active nutrients in under-investigated plant materials is an important task, which may serve as a good platform in developing new effective ingredients for nutraceuticals and functional foods. In general, studies on the enzyme inhibitory activities of Dioscorea preparations are rather scarce. For instance, the extracts of D. alata and D. bulbifera tubers were tested as α-glucosidase and α-amylase inhibitors [20,53]. Our study substantially expands the knowledge on phytochemicals in D. caucasica leaves and tubers and their effects on physiologically important enzymes, namely α-amylase, α-glucosidase, angiotensin-converting enzyme and acetylcholinesterase. It is evident that the polyphenolic-rich extract of leaves is a remarkably stronger inhibitor of all tested enzymes than that of the tubers. Sterols (campesterol, sitosterol, stigmasterol), triterpene acids (oleanolic and ursolic acid), pentacyclic triterpenoids and their esters (αamyrin, β-amyrin, taraxasterol, taraxerol) were reported in previously published articles on D. caucasica leaf constituents [25], while saponins such as parvifloside, protodeltonin, protodioscin, deltonin, dioscin [54] and diosgenin [54,55] were the dominant compounds in plant tubers. Some studies reported the hypoglycemic and α-glucosidase/α-amylase inhibitory activities of phytosterols [56] present in seed oils and triterpenoid saponins from botanicals [57]; however, it is highly unlikely that liposoluble (hydrophobic) leaf phytochemicals play a significant role as enzyme inhibitors. Therefore, the stronger enzyme inhibition by leaf extracts may be related to the presence of the main polyphenolic antioxidants, 3-O-glycoside flavonoid and hydroxycinnamic acid derivatives (Table 1).
Inhibiting the activity of amylolytic enzymes is important in controlling the postprandial glycemic index, which may help in managing type-2 diabetes mellitus. Chlorogenic and neochlorogenic acids and quercetin derivatives are well-known inhibitors of α-glucosidase and α-amylase [58,59]. For instance, the inhibition of α-glucosidase very strongly correlated with chlorogenic acid present in the young inflorescence tissues of selected Rosaceae plants [60]. However, the inhibitory activity depends on the structural peculiarities of the phenolic compounds. Recently, Zhang et al. [61] reported that the fraction of bound polyphenolics of red quinoa composed mainly of ferulic acid more strongly inhibited α-glucosidase than the fraction of free polyphenolics consisting mainly of hydroxybenzoic acid and its derivatives; the former fraction inhibited α-glucosidase in an uncompetitive mode, while the latter one in a non-competitive mode. This was explained by the peculiarities of the interactions between ferulic acid and enzyme amino acid sites. Flavonols such as quercetin were also reported as potent inhibitors of α-amylase and α-glucosidase in an activity-guided study of chokeberry, pomegranate and red grape extracts [62]. Based on the strong inhibition of glucose release from complex carbohydrates, herbal infusions containing phenolic acids were suggested as natural preparations for preventing type II diabetes [6]. The hypoglycemic effect of di-caffeoylquinic acid derivatives may be also explained by the modulation of α-glucosidase, while chlorogenic acid is a potent inhibitor of glucose 6-phosphate translocase [63].
Quite interesting behavior was observed by evaluating extract-induced α-glucosidase inhibition ( Table 2) and reaction kinetics (Figure 2 and Table 3). It may be noted that the main findings were valid regardless of some modification of the method described in the Sigma-Aldrich protocol [48]. Our data suggest that using 0.1 M Na 2 CO 3 solution may significantly change the manner of extract-mediated α-glucosidase inhibition. Significantly negative differences in the activity shift were observed at concentrations of up to 25 µg/mL, while, at higher concentrations, the inhibition in the reaction without 0.1 M Na 2 CO 3 was remarkably higher. To the best of our knowledge, the peculiarities of botanical extractinduced α-glucosidase inhibition using this approach have not been reported previously. Most likely, the significant differences in inhibition between the reaction systems, which were particularly clearly pronounced in the case of increasing extract concentrations, were due to changes in the ratios of the compounds with different effects on the enzyme.
Lineweaver-Burk, Eadie-Hofstee and Hanes-Woolf plots are widely used for determining important parameters in enzyme kinetics, although they may result in some erroneous data [64]. Chon et al. [65] compared five estimation methods, including Lineweaver-Burk and Eadie-Hofstee plots, for simulation data using additive error and combined error models for the parameters of the Michaelis-Menten equation and concluded that the estimation of V max and K m by nonlinear methods provided the most accurate results, while in case of linear methods, the Eadie-Hofstee plot exhibited some advantages over the Lineweaver-Burk plot, although, for the simulation data incorporating additive error, estimated V max values were similar. In addition, it was noted that the double reciprocal Lineweaver-Burk plot overestimates rate measurements recorded at low substrate concentrations, where the experimental error is liable to be greatest [66], while the Eadie-Hofstee plot is less biased at low [S]; however, it can result in significant errors since both coordinates contain V [64]. Marasović et al. (2017) [67] suggested that the Hanes-Woolf plot is the most accurate of the three; however, its major drawback is that both ordinate and abscissa are dependent on the substrate concentration. Despite the perceived errors of the described methods, they are all presented in the scientific literature and are applied for determining kinetic parameters. We chose the most widely used Lineweaver-Burk method, which enabled us to demonstrate that the D. caucasica leaf and tuber extracts displayed mixed-type and competitive modes of inhibition of α-glucosidase, respectively.
Numerous plant origin bioactive compounds and food-derived peptides may inhibit ACE in a very wide range of IC 50 . Regarding yams, D. opposita autolysate and enzymatic hydrolysates of tuber mucilage were tested as ACE inhibitors [68], while the inhibitory properties of leaf extracts have not been reported. D. causacica leaf extract inhibited ACE in a dose-dependent manner ( Figure 6). Polyphenolic compounds in this case may also play the most important role. For instance, wine lees of Cabernet grape variety, which contained significantly higher amounts of catechin flavanols than other evaluated varieties, effectively inhibited ACE and demonstrated similar potency to the drug Captopril in hypertensive rats [69]. In vivo studies with mice showed that a dietary supplement consisting of selected botanicals, chlorogenic acid and inulin reduced the risk factors of cardiovascular diseases by lowering hepatic angiotensinogen and angiotensin-II levels, among other factors [70]. The hypotensive effects of the phenolic-rich fraction of red wine were also demonstrated by an in vivo study with spontaneously hypertensive rats [5].

Preparation of D. caucasica Extracts
Dioscorea caucasica Lipsky leaves and tubers were collected from the Kaunas Botanical Garden of Vytautas Magnus University in Lithuania (55 • 52 14 N 23 • 54 37 E) [79] in June 2019. The collected materials were air-dried at room temperature and milled in a centrifugal mill, Retsch ZM200 (Haan, Germany), using a sieve with 0.5 mm diameter holes. Twenty grams of powdered sample were extracted with 200 mL 70% (v/v) ethanol in a rotary shaker (200 rpm) during 14 h at room temperature. The extract obtained was filtered through Whatman filter paper no. 1 (Whatman ® International Ltd., Maidstone, UK) to obtain the first extract. This procedure was repeated on the residue using 100 mL 70% ethanol for 1 h to obtain the second filtrate. Both filtrates were combined and the solvent was removed in a rotary vacuum evaporator, the Büchi Rotavapor R-210 (Büchi Labortechnik, Flawil, Switzerland), at 40 • C. The residual water was evaporated by freezedrying. The yield of dry extract powder was 17.86% and 15.65% from leaves and tubers, respectively. Freeze-dried extracts (100 ± 0.1 mg) were dissolved in 10 mL of 70% (v/v) ethanol in a volumetric flask to prepare 10 mg/mL extract stock solutions, which were stored at 4 • C until further use. The extracts were separated using a Waters Acquity UPLC system [80,81] consisting of a binary solvent manager, autosampler, column heater and a photodiode-array detector (PDA) (Waters Corporation, Milford, MA, USA). A Waters Acquity BEH C18 (100 × 2.1 mm; 1.7 µm) [82] column was used for compound separation. Eluent A was 0.4% (v/v) formic acid solution in ultrapure water, and B was acetonitrile. The gradient was formed as follows: initially, the separation was started with 100% A; then, in 9 min, B was increased to 100%, and was held at 100% for 1 min. After this, the column was returned to initial conditions in 1 min and then was allowed to equilibrate for 1 min. The column was equilibrated for 2 min before each run. Flow rate was 0.4 mL/min, injection volume 1 µL and column temperature 40 • C.
The eluted compounds were analyzed on a MAXIS 4G QTOF mass spectrometer (Bruker Daltonik GmbH, Bremen, Germany) equipped with an electrospray ionization (ESI) source. The spectrometer was operated in negative ionization mode, and capillary voltage was maintained at 4000 V. Nitrogen (N 2 ) was used as a nebulizing and drying gas at 2.5 bar pressure and 10 L/min flow rate. Drying gas temperature was maintained at 200 • C. MS were recorded in the range from 80 to 1200 m/z; spectra recording rate was 3 Hz. The compounds profiles were identified by comparing MS results with previously reported data.  (Figure 1). Quantitative determination was performed by using calibration curves, which were produced using reference compounds.  [83], with slight modification. Briefly, 100 ± 0.1 mg of powder was dissolved in 100 mL of cold borate buffer (80 mmol/L, pH 8.3 ± 0.05) and kept at 5 • C temperature overnight. Insoluble matter was centrifuged in a centrifuge, the MPW 260RH (Med. Instruments, Warsaw, Poland), at 6000 rpm for 30 min at 4 • C temperature. After centrifugation, the clear wine-red supernatant was carefully transferred into a clean test tube and used for experiments.

Preparation of Solutions for Enzyme Inhibition Assays
Captopril solution was prepared as described by Donath-Nagy et al. (2011) [84], with slight modifications. Two tablets containing 25 mg of Captopril were ground in a mortar and extracted with approximately 15 mL water in a 25 mL volumetric flask in the ultrasonic bath (Bandelin Sanorex, Berlin, Germany) for 10 min, and then brought to volume with water and filtered through Whatman filter paper. Solution of Captopril (1 mg/mL) was used as a positive control in the ACE inhibition assay.

Donepezil HCl Solution for Acetylcholinesterase (AChE) Inhibitory Activity
Donepezil hydrochloride solution was prepared from two tablets containing 5 mg of Donepezil Accord drug. The tablets were crushed and 100 ± 0.1 mg of fine powder was dissolved in 100 mL of ultrapure water in a 100 mL volumetric flask to obtain a 1 mg/mL Donepezil HCl stock solution. Working solutions of Donepezil HCl were obtained by the appropriate dilution of the stock solution and used as positive controls in AChE inhibition assays.

Acarbose Solution for α-Amylase and α-Glucosidase Inhibition Assays
Aqueous solution of acarbose was prepared by dissolving 100 ± 0.1 mg acarbose in a 100 mL volumetric flask to obtain a 1 mg/mL acarbose stock solution. Working solutions of acarbose were obtained by appropriate dilution of the stock solution and used as positive controls in α-amylase/α-glucosidase inhibition assays.

α-Glucosidase
The α-glucosidase inhibition was evaluated by the chromogenic method using pNPG as a substrate [48]. The α-glucosidase activity measurements were conducted at 37 ± 0.2 • C using a 1 mL mixture composed of 0.33 mM pNPG, dissolved in potassium phosphate buffer (pH 6.8) and different concentrations of inhibitors, namely 0, 3.125, 6.25, 12.5, 25, 50 and 100 µg/mL. The assay was initiated by the addition of 5 µL solution of enzyme. α-Glucosidase and product formation (pNPG + α-glucosidase → α-D-glucose + pNp was monitored after 30 min at 405 nm using the Spectronic Genesys 8 spectrophotometer (Thermo Spectronic, Rochester, NY, USA). The data were recorded in two ways: with the stop solution 0.1 M Na 2 CO 3 and without it. The control solution was prepared using the same buffer without inhibitor. The percentage of inhibition of α-glucosidase was calculated similarly to α-amylase.
For kinetic studies, α-glucosidase activity was measured spectrophotometrically by following the absorbance at 405 nm in assay mixture containing various concentrations of substrate: 0.1mM, 0.125 mM, 0.23 mM and 0.33 mM pNPG in 0.1 M K 2 HPO 4 / KH 2 PO 4 phosphate buffer, pH 6.8 ± 0.05, fixed amount of α-glucosidase (5 µL) solution and the absence or presence of the leaf and tuber extracts at 15 µg/mL, 25 µg/mL and 200 and 500 µg/mL concentrations, respectively.

α-Amylase
In vitro α-amylase inhibition was assayed by the method by described in the Sigma-Aldrich protocol [85], with slight modifications. α-Amylase [76] was diluted (1:1000) in 0.02 mM sodium phosphate buffer (pH 6.9) containing 6.7 mM sodium chloride to produce stock solution. The reaction mixture containing 400 µL of α-amylase (0.5 IU/mL) and 400 µL of varying concentrations of extracts (40-480 µg/mL) was incubated in a test tube at 25 • C ± 0.2 • C for 30 min, followed by the addition of 400 µL of 1% (w/v) potato starch solution (0.02 mM sodium phosphate buffer, pH 6.9 ± 0.5), as a substrate, and further incubation for 3 min. The reaction was terminated by the addition of 800 µL of 3.5-dinitrosalicylic acid (DNSA) and heating in a water bath for 10 min at 85 ± 0.2 • C. Afterwards, the mixture was removed from the water bath, cooled under tap water and diluted with 3 mL of distilled water. The absorbance (A) was measured at 540 nm and the inhibition activity was calculated by the following equation: % Inhibition = (A control − A sample )/A control × 100. The blank sample was prepared without enzyme; control samples were those without extracts (100% enzyme activity). All tests were performed in triplicate. Acarbose was used as a positive control.

Acetylcholinesterase (AChE)
The AChE inhibition was estimated in vitro by Ellman's method [52] using ATChI as a substrate. The spectrophotometric method for the estimation of AChE activity is based on the determination of yellow-colored chromophore (5-thio-2-nitrobenzoate; C 7 H 5 NO 4 S; λ = 412 nm) during enzymatic hydrolysis with DTNB reagent (5,5 -dithiobis-(2-nitrobenzoic acid), which reacts with the sulfhydryl groups of the protein. Briefly, AChE activity measurements were carried out at a constant temperature of 20 ± 0.5 • C using a 1 mL mixture composed of 50 mmol/L Tris/HCl (pH 8.0 ± 0.05) buffer, different concentrations (25-100 µg/mL) of inhibitor, 0.05 mmol/L DTNB and AChE at 0.55U. After pre-incubation for 10 min, the reaction was initiated by the addition of 2.5 mM ATChI. The formation of a DTNB-tios complex (C 7 H 5 NO 4 S) was measured at 412 nm by using a GENESYS 50 UV/Vis spectrophotometer (GENESYS Instruments Limited, Cambridge, UK). Donepezil, a standard AChE inhibitor, was used as a positive control and Tris/HCl (pH 8.0 ± 0.05) was used as a negative control. Enzyme-specific activity was calculated using nitrobenzoate (TNB) molar extinction coefficient ε =14150 M −1 cm −1 [86] and expressed as µmol thiocholine formed/min/mg protein. The enzyme percentage inhibition was calculated by comparing the enzymatic activity with/without inhibitor using the following equation: % I = (A control − A sample ) /A control × 100 %, where A contol is AChE activity for the negative control, and A sample is the presence of the plant extract or Donepezil.
For kinetic studies, AChE activity was measured spectrophotometrically by following the absorbance at 412 nm in assay mixture containing various concentrations of substrate, 45, 50, 65, 85 and 160 µM ATChI, in 50 mmol/L Tris/HCl, pH 8.0 ± 0.05, fixed amount of AChE (10 µL) solution and the absence or presence of the leaf extracts at 50 µg/mL concentration.  [83]. Ethanol solutions of plant extract at concentrations of 250, 500, 1000 and 1250 µg/mL were used for the ACE inhibition assay. Three hundred µL of rabbit lung ACE (rabbit lung acetone powder reconstituted in 80 mM sodium borate buffer containing 300 mmol/L sodium chloride, pH 8.3 ± 0.05) and 200 µL of inhibitor solution were added into a test tube and pre-incubated for 3-5 min at room temperature. After pre-incubation, the reaction was started (t = 0 min) by adding 800 µL ACE-specific substrate solution comprising 0.8 mmol/L FAPGG dissolved in a borate buffer, to the final volume of 1300 µL. The mixture was allowed to stand at room temperature for 1 min and absorbance was recorded at 340 nm using the Biochrom Libra S4+ visible spectrophotometer (Biochrom Ltd., Cambridge, UK). Subsequently, the absorbance was monitored after 20 min at 37 ± 0.2 • C. Hydrolysis of FAPGG results in a decrease in absorbance at 340 nm. The rate of decrease in absorbance is directly proportional to ACE activity in the sample. The percentage inhibition of ACE activity by plant extracts was calculated with the following equation: Inhibition, % = [100 − ( ∆ A 340/min inhibitor/ ∆ A 340/min control × 100], where ∆ A is the difference between the initial and final absorbance of the test sample during its incubation.

Mathematical Modeling of Enzyme Inhibition Kinetics
The application of kinetic modeling can provide important insights into how interacting components behave in biological systems. The changes in enzyme catalytic activity in a biochemical system are frequently modeled by choosing a mathematical model to evaluate the following fundamental parameters: Michaelis-Menten constant (K m ), maximal velocity (V max ) and kinetic constant (K i ). In this study, kinetic data were estimated by virtue of plotting the data on a double reciprocal graph, called a Lineaweaver-Burk plot [49], to the following equation: 1/V = K m /V max × 1/[S] + 1/V max , where V is the reaction rate, K m is the Michaelis-Menten constant, V max is the maximum reaction rate (mg/L·min), [S] is the pNPG concentration (mg/L). The resultant plot is a straight line, with X-and Y-axis intercepts representing −1/K m and 1/V max , respectively, and the slope is K m /V max . The inhibition type of the inhibitors was determined by analyzing the Lineaweaver-Burk plots, which were competitive, non-competitive or uncompetitive, and the inhibition constant (K i ) values for botanical extracts were estimated by fitting the equation K m(app) = K m (1 + [I]/K i ) for competitive inhibition and equation V max(app) = V max (1 + [I]/K i ) for non-competitive inhibition, where K m(app) and K m are the concentrations of a substrate required to produce 50% of its maximum velocity (V max ) in the presence and absence of an inhibitor, which were determined in parallel; [I] is the compound concentration; K i is the inhibition constant of the reactions. Table 5 summarizes the types of inhibition and their effects on these parameters.

Statistical Data Analysis
Data are expressed as mean ± standard error (SE) and/or standard deviation (SD). Reaction velocities and enzyme kinetics and IC 50 values of the extracts were calculated using Microsoft Excel 2016. Statistical analysis was performed by Student's t-test, pairedsamples t-test and Fisher's least significance difference (LSD) test (p < 0.05).

Conclusions
The first systematic phytochemical study of D. caucasica revealed that the plant leaves are rich in phenolic compounds, mainly caffeic acid esters and quercetin glycosides, while its tubers accumulate steroidal glycosides, as the major secondary metabolites. D. caucasica leaf extracts were remarkably stronger inhibitors of α-glucosidase, α-amylase, acetylcholinesterase and angiotensin-converting enzyme than the tuber extracts. Kinetic studies of enzyme inhibition suggest that the mode of inhibition, depending on the extract origin, may be mixed-type (leaves) and competitive (tubers). Based on the results obtained in this study and previously reported data, it may be sufficiently reasonably assumed that D. caucasica leaf polyphenolics play the most important role in the enzyme inhibition, while the tuber steroidal glycosides might be remarkably weaker enzyme inhibitors or their effects may be hampered by other, antagonistically competitive constituents. In general, the new data on D. caucasica phytochemicals and bioactivities may aid in the valorization of this plant for its wider use in the development of health beneficial ingredients for functional foods and nutraceuticals.
Author Contributions: Conceptualization, A.A. and P.R.V.; methodology, A.A., A.P. and P.R.V.; investigation, A.A. and A.P.; writing-review and editing, A.A. and P.R.V.; resources, P.R.V. and O.R.; supervision, P.R.V. All authors have read and agreed to the published version of the manuscript.