Extraction and characterization of low molecular weight bioactive carbohydrates from mung bean (Vigna radiata)

Accepted Manuscript Extraction and characterization of low molecular weight bioactive carbohydrates from mung bean (Vigna radiata) Cipriano Carrero-Carralero, Drashti Mansukhani, Ana-Isabel Ruiz-Matute, Isabel Martínez-Castro, Lourdes Ramos, María Luz Sanz PII: DOI: Reference:

S0308-8146(18)30937-3 https://doi.org/10.1016/j.foodchem.2018.05.114 FOCH 22947

To appear in:

Food Chemistry

Received Date: Revised Date: Accepted Date:

12 January 2018 21 May 2018 25 May 2018

Please cite this article as: Carrero-Carralero, C., Mansukhani, D., Ruiz-Matute, A-I., Martínez-Castro, I., Ramos, L., Sanz, M.L., Extraction and characterization of low molecular weight bioactive carbohydrates from mung bean (Vigna radiata), Food Chemistry (2018), doi: https://doi.org/10.1016/j.foodchem.2018.05.114

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Extraction and characterization of low molecular weight bioactive carbohydrates from mung bean (Vigna radiata) Cipriano Carrero-Carralero1, Drashti Mansukhani1, Ana-Isabel Ruiz-Matute1, Isabel Martínez-Castro1, Lourdes Ramos1, María Luz Sanz1* 1

Instituto de Química Orgánica General (CSIC), Juan de la Cierva 3. 28006 Madrid (Spain)

*Corresponding author: [email protected] Tel (+34) 91 562 2900 Fax (+34) 91 564 4853

Running title: MAE and GC-MS analysis of bioactive carbohydrates from mung bean

1

Key words Mung bean (Vigna radiata), gas chromatography-mass spectrometry (GC-MS), cyclitols, methyl-scyllo-inositol, glycosyl-methyl-inositols, glycosyl-inositols, microwave assisted extraction, solid liquid extraction.

2

Abstract Due to the great interest in obtaining natural bioactive carbohydrates to be used as functional ingredients, a selective microwave assisted extraction (MAE) method was optimized to ensure the exhaustive extraction of inositols andgalactooligosaccharides (-GOS) from mung bean. Thereafter, a comprehensive characterization of these compounds was carried out by gas chromatography coupled to mass spectrometry (GC-MS). Apart from free inositols and -GOS, several glycosyl-methyl-scyllo-inositols and glycosyl-inositols were detected for the first time in this legume. Under optimized MAE conditions (0.5 g dry sample, 2 cycles of 3 min, 50 ºC, 10 mL 50:50 ethanol:water, v:v), bioactive carbohydrates yields were similar to those found using solid–liquid extraction (SLE), but with shorter analytical times. Concentrations of bioactive carbohydrates in MAE extracts from samples of different geographical origins ranged between 74.1 and 104.2 mg g -1 dry sample. MAE was proved a good alternative to SLE to obtain extracts enriched in bioactive carbohydrates.

3

1. Introduction Mung bean (Vigna radiata L. Wilczek) is an important annual legume native to Asia and widely consumed in different countries (Tang, Dong, Ren, Li, & He, 2014). Several properties, such as antitumor (Soucek, Skvor, Pouckova, Matousek, & Slavík, 2006) and antioxidant activities (Randhir & Shetty, 2007), have been attributed to mung bean. The consumption of this legume has also been associated with a low increase in blood glucose in humans, which is why this legume is considered useful for diabetic patients (Tang et al., 2014). Mung bean is an excellent source of carbohydrates (Randhir & Shetty, 2007; Tang et al., 2014). Its carboyhdrate content is higher than in other legumes, predominantly starch (42%). Moreover, many bioactive carbohydrates, such as -galactooligosaccharides (-GOS) (Åman, 1979), which are well known prebiotics, and cyclitols such as scyllo- and methyl-scyllo-inositol (Ueno, Hasegawa, & Tsuchiya, 1973), myo-inositol and galactinol (Åman, 1979; Schweizer & Horman, 1981) have also been described in mung bean. Several bioactive properties have been attributed to these cyclitols such as antihyperglycemic (Bates, Jones, & Bailey, 2000), antioxidant (Agarie et al., 2009) and hepatoprotective effects (Dhanasekaran, Ignacimuthu, & Agastian, 2009). Both scyllo- and methyl-scyllo-inositol can also be considered as potential therapeutic agents against Alzheimer’s disease (Tanaka et al., 2015). Solid liquid extraction (SLE) with polar solvents has been commonly applied to extract bioactive carbohydrates, including inositols, from mung bean (Ueno et al., 1973). However, the use of advanced extraction techniques such as microwave assisted extraction (MAE) is currently gaining

4

great attention as an alternative to these well accepted, but tedious and time-consuming SLE-based processes (Al-Suod, Ligor, Rațiu, Rafińska, Górecki, & Buszewski, 2017). MAE has been demonstrated to be useful for the extraction of high molecular weight carbohydrates (pectins, inulin, xylans, etc.) from different natural matrices such as coffee grounds (Passos & Coimbra, 2013) and artichocke bracts (Ruiz-Aceituno, García-Sarrió, Alonso-Rodriguez, Ramos, & Sanz, 2016). However, to the best of our knowledge, to date only one application of this technique for the extraction of low molecular weight carbohydrates (LMWC) (Ruiz-Aceituno et al., 2016) can be found in the literature. In that work, the MAE process was optimized to simultaneously extract three free inositols (chiro-, scyllo- and myo-inositol) and inulin from artichoke bracts. In general, MAE provides high yields of bioactive carbohydrates in short times and using small solvent volumes (Ruiz-Aceituno et al., 2016). However, in these treatments, non-bioactive carbohydrates, such as glucose, fructose or sucrose, which can interfere in the bioactivity of obtained extracts, were coextracted, requiring an additional fractionation step. Regarding carbohydrate analysis, gas chromatography coupled to mass spectrometry (GC-MS) is usually the technique of choice, due to its resolution power, sensitivity, robustness and identification capabilities (Beveridge, Ford, & Richards, 1977; Schweizer & Horman, 1981; RuizAceituno et al., 2016). However, identifying carbohydrates is not straightforward due to their similar structures and the presence of isomeric compounds such as cyclitols. A recent work published by our group (Ruiz-Aceituno, Carrero-Carralero, Ruiz-Matute, Ramos, Sanz, & MartínezCastro, 2017) provided new insight on GC (retention indexes, IT, values) and MS data and their relationship with the chemical structure of different cyclitol glycosides. Criteria established in that study were useful to identify unknown compounds.

5

Therefore, in this work, an exhaustive characterization of LMWC (including bioactive carbohydrates such as -GOS, cyclitols and cyclitol glycosides) present in mung bean has been carried out by GC-MS for the first time. SLE and MAE have been optimized and compared for the selective extraction of these bioactive LMWC. Finally, MAE has been applied, under optimized conditions, to extract these carbohydrates from mung beans of different geographical origins. 2. Materials and Methods 2.1. Standards and samples Analytical standards of fructose, sorbitol, galactose, glucose, galactinol, myo-inositol, scyllo-inositol, maltose, maltotriose, maltotetraose, manninotriose, nystose, pinitol, raffinose, stachyose, sucrose, and verbascose were obtained from Sigma Chemical Co. (St. Louis, MO, US). Digalactosyl-myo-inositol (DGMI) was extracted from buckwheat (Fagopyrum esculentum) as previously described (Ruiz-Aceituno et al., 2017). Vigna radiata beans from 6 different brands of different geographical origins [Taiwan (MBT1), Argentina (MBA1), Spain (MBS1 and MBS2) and China (MBC1 and MBC2)] were purchased at local markets in Madrid (Spain). 2.2. Extraction procedure

6

Dried Vigna radiata beans were milled until powder with a grinder. The whole sample was sieved through a square mesh of 500 µm before extraction. Water of ultra-pure quality (18.2 MΩcm) was produced in house using a Milli-Q Advantage A10 system from Millipore (Billerica, MA, USA). Ethanol and methanol from analytical grade were acquired from Scharlau (Barcelona, Spain). 2.2.1. Solid liquid extraction (SLE) Dried Vigna radiata sample (MBS1; 0.5 g) was extracted with 10 mL of different solvents [water (100%), methanol (100%), ethanol (100%), and mixtures of methanol:water (50:50, v/v) and ethanol:water (50:50 and 75:25, v/v)] under constant stirring at 25ºC. Using the optimum solvent, the influence of the extraction time was investigated by analyzing extracts treated for 5, 30, 60 and 120 min. The influence of temperature was also evaluated at 25ºC and 50ºC. In all cases, extracts were immediately centrifuged at 4400 ×g for 10 min at 4 ºC and kept in a freezer at -18 ºC until analysis. Unless otherwise specified, all experiments were performed in triplicate. 2.2.2. Microwave assisted extraction (MAE) A MARS 6 (CEM, NC, USA) MAE instrument was used in the study. Microwave power was set at 1250 W in all instances. Mung bean sample (MBS2) was placed in 100 mL Green Chem vessels (CEM) and dissolved in 10 mL of the selected solvent. The effect of three independent factors (temperature, time and amount of sample) on the extraction of bioactive carbohydrates was evaluated following a Box–Behnken experimental design using StatGraphics Centurion XV software (Statistical Graphics Corporation, Rockville, MD, USA). A total of 15

7

experiments were carried out in randomized order. The experimental ranges used for the different factors were: temperature (T) 50 - 120 ºC, time (t) 3 - 30 minutes, and amount of sample (s) 0.01 - 1.00 g.

The quadratic model proposed was: R = 0 + 1T + 2t + 3s +1,1T2 + 2,2t2 + 3,3s2 +1,2Tt + 1,3Ts + 2,3ts + Eq where 0 is the intercept, i are the first-order coefficients, i,i the quadratic coefficients for ith factors, i,j are the coefficients for the interaction of factors i and j and  is the error. Three response (R) variables were individually considered in the optimization of the MAE method: Ri, cyclitol amount (mg.g -1 dry sample); Ra,

-GOS amount (mg g-1 dry sample), and Rc, non-bioactive sugar amount (mg.g-1 dry sample). The experimental conditions that independently maximized Ri and Ra, and minimized Rc were obtained from the fitted models by multiple linear regression (MLR). A desirability function (RD) was also optimized to provide MAE conditions that simultaneously maximize different responses; this function takes values between 0 (completely undesirable value) and 1 (completely desirable or ideal response).

8

Under optimized conditions, the effect of three consecutive extraction cycles was also evaluated. Obtained extracts were immediately cooled down on ice, centrifuged at 4400 ×g for 10 min at 4 ºC and kept in a freezer at -18 ºC until analysis. The optimized method was applied to the treatment of mung bean samples from different origins.

2.3. Analysis of carbohydrates by gas chromatography coupled to mass spectrometry (GC-MS) A two-step derivatization procedure (oximation + silylation) was carried out prior to GC-MS analysis according to Ruiz-Aceituno et al. (2017). Mung bean extracts (2 mL) were mixed with 0.1 mL of a 70% ethanolic solution of phenyl--D-glucoside (1 mg mL-1; internal standard) and dried under vacuum (38-40ºC). Oximation was done using 2.5% of hydroxylamine chloride in pyridine (350 µL) at 75 ºC for 30 min; after this, the silylation of the carbohydrates was carried out using hexamethyldisilazane (350 µL) and trifluoroacetic acid (33 µL) at 45 ºC for 30 min. All reagents were obtained from Sigma Chemical Co. Under these conditions, two peaks corresponding to the syn (E) and anti (Z) forms per reducing sugar were obtained, whereas non-reducing carbohydrates (i.e. cyclitols, cyclitol glycosides, sucrose and oligosaccharides of raffinose family), which do not form oximes, gave a single peak. After derivatization, 1 µL of the derivatized mixture was injected into the GC-MS system. For GC-MS analyses, the method proposed by Ruiz-Aceituno et al. (2017) was followed. A 7890A gas chromatograph coupled to a 5975C quadrupole mass spectrometer detector (Agilent Technologies, Palo Alto, CA, USA) with a HT5 (5% phenyl (equiv.) polycarborane–siloxane)

9

capillary column (25 m × 0.22 mm i.d. × 0.1 µm film thickness; SGE, Ringwood, Australia) was used. Helium at 1 mL min−1 was employed as carried gas. The oven temperature was programmed from 180 ºC (10 min) at 5 ºC min −1 to 200 ºC (15 min), then at 15 ºC min−1 to 270 ºC, at 1 ºC min−1 to 290 ºC, at 15 ºC min−1 to 300 ºC (15 min), and finally at 15 ºC min−1 to 360 ºC (15 min). Injections (1 µL) were carried out in split mode (1:20) at 320 ºC. The MS detector was operated in the electron impact (EI) mode at 70 eV, scanning the 50–650 m/z range. The transfer line was set at 280 ºC and the ionization source at 230 ºC. For data acquisition and treatment, the HP ChemStation software (Agilent Technologies) was used. Carbohydrates were identified by comparison of their corresponding linear retention indices (IT) and mass spectra with those of available standards and/or previously reported in the literature (Ruiz-Aceituno et al., 2017). Tentative identification of new cyclitol glycosides was done on the basis of their IT and MS data. Quantitative data were obtained by the internal standard method. Standard solutions were prepared over the expected concentration range (0.1 - 1 mg.mL-1 for each compound) in mung bean extracts to calculate each corresponding response factor relative to phenyl-β-D-glucoside. Those compounds whose standards were not available were assigned with a response factor of 1. 2.4. Statistical analysis

10

For statistical analysis, the Univariate analysis of variance (ANOVA) and Duncan test (SPSS software) were used. Differences were considered to be significant at P < 0.05. 3. Results and Discussion 3.1 Identification of LMWC present in aqueous Vigna radiata extracts Figure 1 shows the GC-MS profile of a SLE Milli-Q water extract (2 h extraction at 25ºC under constant stirring) from mung bean seeds (sample MBS1). Up to 45 LMWC were either positively (pure standard was available) or tentatively (based on IT and MS data) identified in this extract (see Table 1 for peak assignment and IT values). Detected LMWC included mono- and disaccharides, sugar acids, linear polyalcohols, cyclitols and oligosaccharides (up to degree of polymerization 6). Regarding the latter, non-reducing bioactive -GOS from the raffinose family (namely raffinose, stachyose, verbascose and ajugose) previously reported in the literature (Åman, 1979; Kotiguda, Peterbauer, & Mulimani, 2006) were confirmed in these extracts by comparison with their corresponding standards. Moreover, reducing oligosaccharides, such as maltotriose and maltotetraose, were also detected. On the contrary, manninotriose, whose presence in mung bean had been previously reported by Åman (1979), was not found in these samples. Regarding cyclitols, three free inositols, namely scyllo-inositol (IT 1979), myo-inositol (IT 2055) and O-methyl-scyllo-inositol (IT 1911) were detected in this extract. The presence of these inositols in this legume has been previously described by Ford (1985). However, to the best of our

11

knowledge, information concerning the presence of cyclitol glycosides in mung bean is very scarce in the literature and, in fact, only galactinol had been previously reported (Åman, 1979; Ford, 1979; Ueno et al., 1973). In this work, 20 compounds with mass spectrometric features characteristic of cyclitol glycosides were identified. Table 2 summarizes relevant mass spectrometric features (i.e., characteristic m/z fragments and ratios of their abundances) of cyclitol glycosides (probably galactosides) found in this mung bean extract. As an example, full mass spectra obtained for some of these compounds are reported in Figure S1 as part of the Supplementary Material. Eleven derivatives of methyl-inositols were tentatively identified on the basis of mass spectrometric data, in particular the presence of the m/z ion at 375 and the relative abundances of m/z ions 133/129 and 260/265, as previously discussed by Ruiz-Aceituno et al. (2017). Considering that only O-methyl-scyllo-inositol has been found as a free methyl-inositol, it is probable that these eleven compounds are derivatives of this cyclitol. According to their retention times, compounds 12, 13, 14 and 16 (IT 2606, 2643, 2658 and 2724, respectively), were assigned as methyl-scyllo-inositol glycosides with substitutions at different positions. Cyclitol glycosides described in legumes are α-galactosides (Peterbauer & Richter, 2001); thereby, these compounds could tentatively be assigned as α-galactosyl-methyl-scyllo-inositols. Nevertheless, due to the symmetry of the methyl-scyllo-inositol ring, only three α-galactosides of this cyclitol are expected. The fourth one could then be a βgalactoside or a glucoside, as described in other plants (e.g. O-β-glucosyl-scyllo-inositol was isolated by Kamano et al. (1971) from O. stenophylla).

12

Meanwhile, compounds 23, 24 and 25 (IT 3524, 3535 and 3641, respectively) could be identified as isomeric digalactosyl-methyl-scyllo-inositol derivatives; compounds 35 and 36 (IT 4108 and 4117) could correspond to trigalactosyl-methyl-scyllo-inositol derivatives; and compounds 41 and 42 (IT 4389 and 4407) could be tetragalactosyl-methyl-scyllo-inositol derivatives. To the best of our knowledge, the presence of methylscyllo-inositol glycosides in this legume has not been previously described. Regarding glycosyl-inositol derivatives, galactinol (peak 19) was positively identified by comparison of its IT value (2832) and mass spectrum with those of the corresponding standard. Compounds 17 and 18, with IT values of 2775 and 2798 respectively, were identified as galactosylinositols, whereas compounds 26, 27, 29 and 31 (with IT values of 3552, 3583, 3642 and 3722) were assigned as digalactosyl-inositols. The final compound (analyte 31) was positively identified as DGMI by comparing its IT value and mass spectrum with those of the corresponding standard. Finally, compounds 38 and 39 (IT 4167 and 4187) were tentatively assigned as trigalactosyl-inositols. In all cases, and as observed in a previous manuscript for other samples (Ruiz-Aceituno et al., 2017), glycosyl-methyl-inositol derivatives were found to elute before glycosylinositol derivatives on the 5% phenyl (equiv.) polycarborane–siloxane stationary phase used for GC separation. Considering that myo- and scylloinositols occur in free form in V. radiata, the glycosyl-inositols mentioned above could be derivatives of both inositols. In this case, the latter should be expected at very low levels, since this cyclitol is a minor compound in this bean. 3.2 Optimization of the SLE of LMWC from mung beans

13

As shown in section 3.1, apart from -GOS and cyclitol derivatives, several LMWC, such as glucose, fructose or maltose, were coextracted. These compounds may (i) cause interference in the instrumental analysis of the bioactive analytes (e.g., maltose coeluted with a glycosyl-methylinositol), (ii) negatively affect the bioactivity of obtained extracts intended as functional foods, and (iii) promote an undesirable increase of the caloric content of these extracts. Therefore, with the purpose of reducing the content of interfering sugars in the obtained extracts and favor a selective extraction of the bioactive carbohydrates (i.e., cyclitols and -GOS), the use of different solvents during SLE under constant stirring for 2 h at 25 ºC was evaluated. Table 1 compares the concentrations (as mg.g-1 dry sample) of the identified LMWC in sample MBS1 after SLE using different solvents, including water, methanol, ethanol, and several hydroalcoholic mixtures. Under the applied extraction conditions, water was the most effective solvent for the extraction of the previously described analytes, while some of these compounds were not extracted using alcohols and hydroalcoholic mixtures. Regarding water extracts, verbascose (54.8 mg.g-1dry sample) and stachyose (8.5 mg.g-1 dry sample) were the most abundant bioactive carbohydrates, followed by methyl-scyllo-inositol (1.2 mg.g-1 dry sample) and ajugose (1.1 mg.g-1 dry sample). Regarding cyclitol glycosides, the derivatives of methyl-inositol were more abundant (0.56 mg.g-1 dry sample) than those of inositol (0.23 mg.g-1 dry sample). However, high concentrations of non-bioactive carbohydrates (52.7 mg.g-1 mung bean) were also coextracted. Moreover, some coelutions of these carbohydrates with methyl-scyllo-inositol (peak 5) and glycosyl-methyl-inositol derivatives (peaks 14, 16, 36 and 38) were observed. The similarity of the mass spectra of these compounds hindered their quantitation in these extracts (Table 1).

14

On the contrary, fewer bioactive carbohydrates and significantly lower concentrations of those extracted were obtained when using both methanol and ethanol; therefore, these solvents were discarded and not further considered in the present study. The evaluated hydroalcoholic mixtures (methanol: water 50:50, v/v; and ethanol: water 50:50 and 75:25, v/v) proved to be more selective than water for the extraction of the bioactive carbohydrates of interest. Similar concentrations of free inositols (2.1-2.3 mg.g-1 dry sample), cyclitol glycosides (0.8-1.0 mg.g-1dry sample) and -GOS (59.2-77.3 mg.g-1 dry sample) as those found in water extracts were obtained using, respectively, ethanol:water 50:50 (v/v) and methanol:water 50:50 (v/v) as extraction solvents, whereas significantly lower concentrations of other non-bioactive carbohydrates (15.615.1 mg.g-1 dry sample) were obtained. Moreover, cyclitol glycosides that coeluted with some carbohydrates in the water extract could successfully be quantified in these samples. Finally, an extra experiment involving ethanol:water 75:25 (v/v) as extractant was carried out with the aim of totally avoiding the extraction of the non-bioactive carbohydrates, while preserving the efficient recovery of the bioactives. As shown in Table 1, although significantly lower concentrations of non-bioactive carbohydrates were obtained with this mixture (5.7 mg.g-1 dry sample), the extraction of the targeted bioactive carbohydrates (free inositols, glycosyl-cyclitol derivatives and -GOS) was also significantly lower (1.3, 0.43 and 24.9 mg.g-1 dry sample, respectively). Considering these results and the fact that ethanol is a greener solvent than methanol, the mixture ethanol:water 50:50 (v/v) was chosen for subsequent studies. In addition, the improved selectivity provided by this solvent made the subsequent removal of coextracted non-bioactive carbohydrates typically required for water extracts unnecessary (Ruiz-Aceituno et al., 2013).

15

Representative compounds of each category (-GOS, free and glycosylated inositols and other non-bioactive LMWC) were selected as target compounds to determine the influence of the extraction time (evaluated in the 5, 30, 60 and 120 min range) and temperature (evaluated at 25 ºC and 50 ºC) in the SLE process. In general, at 25 ºC, concentrations of inositols (both free and glycosylated) and -GOS showed a slight increased up to 1 h of extraction or remained unaltered at all assayed extraction times (Figure 2). On the contrary, non-bioactive LMWC (i.e., glucose and galactose) increased their concentration with time up to 2 h of extraction. Therefore, 1 h was chosen as the optimal extraction time. The effect of temperature (50 ºC) was also evaluated for the extraction of the target compounds, however, no notable differences in the concentration of bioactive and non-bioactive LMWC relative to 25 ºC were observed in this treatment. 3.3 Optimization of the MAE of LMWC from mung beans Considering the results obtained during the SLE treatment of mung bean, 10 mL of ethanol/water 1:1 (v/v) was selected as solvent for MAE of sample MBS2. The influence of three independent variables (T, t and s) on the efficiency of the MAE of α-GOS (Ra), inositols and glycosylinositols (Ri) and non-bioactive sugars (Rc) was evaluated. A total of 15 experiments were carried out in randomized order according to the Box– Behnken experimental design. Concentrations (mg g-1 dry sample) obtained in each experiment for the selected carbohydrates of the three categories previously considered for SLE are shown in Table 3. Ra varied between 5.30 and 21.21 mg.g-1 dry sample, Ri between 1.84 to 4.78 mg g-1 dry sample and Rc varied between 49.83 to 143.81 mg.g-1 dry sample. In general, the lowest concentrations of all analyzed carbohydrates were obtained at 120ºC, probably due to the degradation of these compounds at this high temperature. 16

Response surface methodology was used to calculate the coefficients of the different Ra, Ri and Rc in the model proposed and to estimate the statistical significance of the regression coefficients. In the case of the α-GOS, T and s2 were the most significant coefficients (Ra = 109.032 0.395*T + 164.113*s – 149.194*s2; P < 0.10), whereas T, T2 and s2 were the most significant factors to maximize the concentration of inositols (Ri = 2.465 + 0.030*T + 3.970*s -0.0003*T2 – 3.609*s2; P < 0.10). The optimal conditions were 50 ºC, 3 min, and 0.55 g of dry sample in terms to maximize the extraction of bioactive carbohydrates (including both α-GOS and glycosyl-inositols). Regarding the Rc model, the most significant coefficients (P < 0.05) to minimize the concentration of non-bioactive carbohydrates were T, s, T2 and s2 (Rc = -9.575 + 0.427*T + 37.232*s -0.003*T2 – 28.754*s2); according to this model, the optimal conditions were 120 ºC, 16.5 min and 0.1 g of dry sample. Finally, a multiple response that simultaneously maximized Ra and Ri and minimized Rc was considered. Optimal operating parameters (RD = 0.85) were 50 °C, 3.6 min and 0.5 g of dry sample when using 10 ml of solvent. Under these conditions, three sequential extraction cycles were carried out with the aim of achieving an exhaustive extraction of targeted carbohydrates. Results of this experiment demonstrated that 90 % of all extracted compounds were already recovered from the beans in the first extraction, while 10% were obtained in the second extraction and only trace levels of some of the most abundant compounds were found in the third extraction cycle. Consequently, two sequential extraction cycles were used for subsequent studies. 3.4. Comparison of SLE and MAE methods

17

The efficiency of the two optimized extraction methods, i.e. conventional SLE and MAE, concerning the recovery of -GOS, cyclitols and other carbohydrates from mung bean (sample MBS2) was then compared. Results are shown in Table S1 of Supplementary Material. In general, similar concentrations of all carbohydrates were obtained by both methods. Only some significant differences were observed for specific compounds that were better extracted by MAE, like raffinose (1.87 mg.g-1 dry sample by SLE and 5.1 mg.g-1 dry sample by MAE) and a diglycosyl-methyl-inositol (0.13 mg.g-1 dry sample by SLE and 0.45 mg.g-1 dry sample by MAE). On the contrary, the highest concentrations of some non-bioactive carbohydrates were, in general, obtained using SLE, such as glucose+galactose (0.12 vs 0.05 mg.g-1 dry sample by SLE and MAE, respectively). It is worth highlighting that MAE yielded these concentrations using the same amount of mung bean used for SLE (0.5 g) and the same volume of solvent (10 mL), but with the advantage of requiring shorter extraction times (2 cycles of 1 h in SLE vs 2 cycles of 3 min in MAE). Therefore, MAE treatment was selected for the analysis of mung bean samples from different geographical origins. 3.4. Application of the optimized MAE method to the analysis of mung bean of different origins The optimized MAE method was applied to the extraction of bioactive carbohydrates from commercial mung beans from different geographical origins [Taiwan (MBT1), Spain (MBS2), Argentina (MBA1) and China (MBC1 and MBC2)]. As can be seen in Table 4, different concentrations of bioactive carbohydrates were observed in the samples. In general, the mung bean sample from Argentina showed the lowest concentrations of

18

bioactive compounds (-GOS, 68.2 mg.g-1 dry sample; free inositols, 4.0 mg.g-1 dry sample; and glycosyl-inositols, 1.9 mg.g-1 dry sample). Meanwhile, MBS2 showed the highest concentration of -GOS (96.3 mg.g-1 dry sample), MBT1 the highest concentration of free inositols (7.8 mg.g-1 dry sample) and MBC2 the highest concentration of glycosyl-cyclitols (5.2 mg.g-1 dry sample). Verbascose was the most abundant carbohydrate in all samples, with concentrations varying between 50.4 to 74.0 mg.g-1 dry sample, followed by stachyose (concentrations in the range 11.5-15.4 mg.g-1 dry sample). On the other hand, galactinol was the most abundant glycosyl-inositol (0.4-1.1 mg.g-1 dry sample), whereas some diglycosyl-methyl-scyllo-inositols (e.g. peak 24) and triglycosyl-methyl-scyllo-inositols (e.g. peak 36) were also found at relatively high concentrations (0.23-0.8 mg.g-1 dry sample and 0.3-1.3 mg.g-1 dry sample range, respectively). The concentration of the main bioactive carbohydrates found in these extracts are in good agreement with those reported by Yasui, Tateishi and Ohashi (1985) in 80% ethanolic extracts of mung bean from Japan. Regarding non-bioactive carbohydrates, MBT1 showed the highest concentrations. Sucrose was also found at high concentration in all extracts (10.3-17.5 mg.g-1 dry sample), followed by maltose which was particularly abundant in sample MBC2 (0.43 mg.g-1 dry sample). Monosaccharides, maltotriose and maltotetraose were detected at lower levels in all samples. In general, lower concentrations of sucrose and glucose than those found in previous works (Aman, 1979) were detected in the tested samples, probably due to the selective extraction carried out in this study. 4. Conclusions

19

This study reports a detailed characterization of LMWC in mung bean samples from different geographical origins. Several glycosyl-inositols (probably from both myo- and scyllo-inositol) and glycosyl-methyl-inositols not previously described have been identified in the investigated mung bean extracts. Abundance ratios of m/z ions 133/129 and 260/265 and the presence or absence of the m/z 375 ion allowed the characterization of these compounds based on their respective mass spectra. Moreover, the feasibility of using SLE and MAE for the selective extraction of cyclitols (free and glycosylated) and -GOS against other high content non-bioactive LMWC also present in the samples has been evaluated. Ethanol water 50:50 (v/v) was found to be a selective solvent for the extraction of bioactive carbohydrates. After carefully optimizing the several parameters affecting the efficiency of the extraction with both techniques, MAE provided the most satisfactory results. To the best of our knowledge, this is the first time that MAE has been used for the extraction of these bioactive carbohydrates from legumes, in particular from mung bean. MAE has proved to be a good alternative to conventional SLE to obtain extracts enriched in bioactive carbohydrates. The procedure developed in this study could be considered a valuable and greener alternative to more time-consuming SLE methodologies currently in use to enrich food ingredients for industrial purposes.

Acknowledgements

20

This work has been funded by Fundación Ramón Areces (project CIVP17A2843) and by Ministerio de Economía, Industria y Competitividad (MINECO) of Spain (project AGL2016-80475-R). Authors also thank financial support from the Comunidad Autónoma of Madrid and European funding from FEDER program (project S2013/ABI-3028, AVANSECAL). C.C.C. thanks MINECO for a predoctoral contract and D.M. thanks CSIC, Fondo Social Europeo and Iniciativa de Empelo Juvenil for a contract.

Supplementary Material description Figure S1. Full mass spectra of some trimethylsilyl glycosyl-cyclitols of mung bean. Table S1. Comparison of the SLE and MAE efficiency for the extraction of selected compounds of MBS2 under optimal conditions.

21

References Agarie, S., Kawaguchi, A., Kodera, A., Sunagawa, H., Kojima, H., Nose, A., & Nakahara, T. (2009). Potential of the common ice plant, Mesembryanthemum crystallinum as a new high-functional food as evaluated by polyol accumulation. Plant Production Science, 12, 3746. Al-Suod, H., Ligor, M., Rațiu, I.-A., Rafińska, K., Górecki, R., & Buszewski, B. (2017). A window on cyclitols: Characterization and analytics of inositols. Phytochemistry Letters, 20, 507-519. Åman, P. (1979). Carbohydrates in raw and germinated seeds from mung bean and chick pea. Journal of the Science of Food and Agriculture, 30, 869-875. Bates, S. H., Jones, R. B., & Bailey, C. J. (2000). Insulin‐like effect of pinitol. British journal of pharmacology, 130, 1944-1948. Beveridge, R. J., Ford, C., & Richards, G. (1977). Polysaccharides of tropical pasture herbage. VII. Identification of a new pinitol galactoside from seeds of Trifolium subterraneum (subterranean clover) and analysis of several pasture legume seeds for cyclohexitols and their galactosides. Australian journal of Chemistry, 30, 1583-1590. Dhanasekaran, M., Ignacimuthu, S., & Agastian, P. (2009). Potential hepatoprotective activity of ononitol monohydrate isolated from Cassia tora L. on carbon tetrachloride induced hepatotoxicity in wistar rats. Phytomedicine, 16, 891-895.

22

Ford, C. W. (1979). Simultaneous quantitative determination of sucrose, raffinose and stachyose by invertase hydrolysis and gas—liquid chromatography. Journal of the Science of Food and Agriculture, 30, 853-858. Ford, C. W. (1985). Identification of inositols and their mono-O-methyl ethers by gas—liquid chromatography. Journal of Chromatography A, 333, 167-170. Kamano, Y., Tachi, Y., Otake, T., & Komatsu, M. (1971). Isolation and Structure of Two New Glucosides, 1-O-β-D-Glucopyranosylscylloinositol and 1-O-β-D-Glucopyranosyl-proto-quercitol. Chemical and pharmaceutical bulletin, 19, 1113-1117. Kotiguda, G., Peterbauer, T., & Mulimani, V. H. (2006). Isolation and structural analysis of ajugose from Vigna mungo L. Carbohydrate Research, 341, 2156-2160. Passos, C. P., & Coimbra, M. A. (2013). Microwave superheated water extraction of polysaccharides from spent coffee grounds. Carbohydrate polymers, 94, 626-633. Peterbauer, T., & Richter, A. (2001). Biochemistry and physiology of raffinose family oligosaccharides and galactosyl cyclito ls in seeds. Seed Science Research, 11, 185-197. Randhir, R., & Shetty, K. (2007). Mung beans processed by solid-state bioconversion improves phenolic content and functionality relevant for diabetes and ulcer management. Innovative food science & emerging technologies, 8, 197-204.

23

Ruiz-Aceituno, L., Carrero-Carralero, C., Ruiz-Matute, A., Ramos, L., Sanz, M., & Martínez-Castro, I. (2017). Characterization of cyclitol glycosides by gas chromatography coupled to mass spectrometry. Journal of Chromatography A, 1484, 58-64. Ruiz-Aceituno, L., García-Sarrió, M. J., Alonso-Rodriguez, B., Ramos, L., & Sanz, M. L. (2016). Extraction of bioactive carbohydrates from artichoke (Cynara scolymus L.) external bracts using microwave assisted extraction and pressurized liquid extraction. Food Chemistry, 196, 1156-1162. Ruiz-Aceituno, L., Rodríguez-Sánchez, S., Ruiz-Matute, A. I., Ramos, L., Soria, A. C., & Sanz, M. L. (2013). Optimisation of a biotechnological procedure for selective fractionation of bioactive inositols in edible legume extracts. Journal of the Science of Food and Agriculture, 93, 2797-2803. Schweizer, T. F., & Horman, I. (1981). Purification and structure determination of three α-d-galactopyranosylcyclitols from soya bean. Carbohydrate Research, 95, 61-71. Soucek, J., Skvor, J., Pouckova, P., Matousek, J., & Slavík, T. (2006). Mung bean sprout (Phaseolus aureus) nuclease and its biological and antitumor effects. Neoplasma, 53, 402-409. Tanaka, H., Hashimoto, M., Fukuhara, R., Ishikawa, T., Yatabe, Y., Kaneda, K., Yuuki, S., Honda, K., Matsuzaki, S., & Tsuyuguchi, A. (2015). Relationship between dementia severity and behavioural and psychological symptoms in early‐onset Alzheimer's disease. Psychogeriatrics, 15, 242-247.

24

Tang, D., Dong, Y., Ren, H., Li, L., & He, C. (2014). A review of phytochemistry, metabolite changes, and medicinal uses of the common food mung bean and its sprouts (Vigna radiata). Chemistry Central Journal, 8, 4. Ueno, Y., Hasegawa, A., & Tsuchiya, T. (1973). Isolation of O-methyl-scyllo-inositol from mung bean seeds. Carbohydrate Research, 29, 520521. Yasui, T., Tateishi, Y. & Ohashi, H. (1985). Distribution of low molecular weight carbohydrates in the subl~enus ceratotropis of the genus Yigna (Leguminosae). The Botanical Magazine (Tokyo), 98, 75-87.

25

Figure captions Figure 1. GC-MS profile of TMSO carbohydrates extracted by SLE from Vigna radiata (sample MBS1) using water as solvent. See Table 1 for peak assignation. Internal standard (I.S.) Figure 2. Content (mg.g-1 dry sample) of the target analytes in mung bean extracts (MBS1) by SLE (mg g−1 dry sample) at different extraction times at 25 ºC. α-GOS (A), free inositols and glycosyl-inositols (B) and non-bioactive carbohydrates (C).

26

Figure 1. GC-MS profile of TMSO carbohydrates extracted by SLE from Vigna radiata (sample MBS1) using water as solvent. See Table 1 for peak identification. Internal standard, I.S.

27

28

29

30

31

Figure 2. Content (mg.g-1 dry sample) of the target analytes in mung bean extracts (MBS1) by SLE (mg g −1 dry sample) at different extraction times at 25 ºC. α-GOS (A), free inositols and glycosyl-inositols (B) and non-bioactive carbohydrates (C).

(A)

60

Raffinose Stachyose Verbascose Ajugose

40

mg g

-1

20

2

1

0 5

30

60

120

min

32

(B) 3.0 2.5

mg g

-1

2.0

scyllo-Inositol Methyl-scyllo-inositol myo-Inositol Glycosyl-methyl-inositol Digalactosyl-methyl-inositol Digalactosyl-inositol Trigalactosyl-methyl-inositol

1.5 0.5

0.0 5

30

60

120

min

33

(C) 15

Fructose GlucoseGala Sucrose Maltose Maltotriose

mg g

-1

10

Fructose Glucose+Galactose Sucrose Maltose Maltotriose

5 0.2

0.1

0.0 5

30

60

120

min Table 1. Low molecular weight carbohydrate concentration (mg.g -1 dry weight; standard deviations are shown in parenthesis) and retention indices (IT) of mung bean extracts (sample MBS1) obtained using SLE with different solvents. Peak number

Carbohydrate

T

I

Water

Methanol

Ethanol

Methanol: water (50:50, v/v)

Ethanol : water (50:50, v/v)

Ethanol : water (75:25, v/v)

34

1

Fructose

2

Sorbitol

3

Galactose

4

Glucose

5

Methyl-scyllo-inositol

6

Galacturonic acid

7

Glucaric acid

8

scyllo-Inositol

9

myo-Inositol

10

Sucrose

11

Pentosyl-hexose

12

Glycosyl-methyl-scyllo-inositol

13

Glycosyl-methyl-scyllo-inositol

14

Glycosyl-methyl- inositol

15

Maltose

16

Glycosyl-methyl-inositol

17

Galactosyl - inositol

18

Galactosyl - inositol

1826/ 1846 1835 1893/ 1941 1912/ 1941 1911 1923 1933 1979 2055 2517 2549 2606 2643

1.03 b (0.06)* 0.8 b (0.7)

0.007 a (0.002) 0.002 a (0.001)

0.001 a (0.001)

3.0 c (0.9)

0.14 b (0.08)

1.2 a (0.4) 0.9 a (1.5) 0.97 (0.08) 0.08 b (0.05) 0.51 c (0.07) 27.6 c (9.5) 0.960 b (0.677) 0.048 d (0.003) 0.10 c (0.02)

2658 2697/ 2711

11.4 (5.1)

0.05 a (0.02) 0.009 a (0.002)

0.032 a (0.007) 0.04 a (0.02)

0.017 a (0.008) 0.012 a (0.005)

0.033 a (0.005)

0.09 a (0.02)

0.2 b (0.1)

0.07 a (0.02)

-

-

1.7 a (0.3)

-

-

-

1.6 a (0.3) 0.18 a (0.07)

1.2 a (0.2) 0.04 a (0.01)

-

-

-

-

-

-

a,b

-

-

0.015 a,b (0.003) 2.7 a (1.4)

1.0 a (0.3)

0.57 c (0.09) 13.6 c (2.2)

-

-

-

0.015 b (0.002) 0.024 a (0.003) 0.006 a (0.001)

0.003 a (0.001) 0.005 a,b (0.001)

0.047 d (0.007) 0.10 c (0.01) 0.08 b (0.01)

-

-

-

-

b

a

2724 2775 2798

-

0.013 a,b (0.009) 0.028 a (0.017)

0.006 (0.001) 0.002 a,b (0.000) -

-

0.019 c (0.005) 0.014 b (0.002) 0.020 a,b (0.007)

0.037 (0.007) 0.48 c (0.04) 14.8 c (2.2) 0.011 a (0.006) 0.045 d (0.005) 0.09 c (0.01) 0.06 b (0.03) 0.022 a (0.006) 0.014 b,c (0.004) 0.007 a,b (0.008) 0.018 a,b (0.008)

0.013 a (0.011) 0.16 c (0.02) 5.4 b (1.5) 0.003 a (0.001) 0.029 c (0.001) 0.04 b (0.01) 0.013 a (0.001) 0.007 a (0.003) 0.006 a,b (0.003) 0.004 a,b (0.005) 0.004 a,b (0.003)

35

19

Galactinol

20

Diglycosyl-glycerol

21

Raffinose

22

Maltotriose

23

26

Digalactosyl-methyl-scylloinositol Digalactosyl-methyl-scylloinositol Digalactosyl-methyl-scylloinositol Digalactosyl-inositol

27

Digalactosyl-inositol

24 25

28

Trisaccharide

29

Digalactosyl-inositol

30

Tri-galactosyl-glycerol

31 32

Digalactosyl-myo-inositol (DGMI) Stachyose

33

Reducing tetrasaccharide

34

Tetrasaccharide

35 36

Trigalactosyl-methyl-scylloinositol Trigalactosyl-methyl-scylloinositol

2832 2959 3159 3465/ 3500 3524 3529 3539 3552 3583 3629 3642 3678 3722 3980 4016 4096 4108 4117

0.10 b,c (0.04) 0.06 b (0.01) 0.7 b,c (0.1) 1.8 b (0.4) 0.02 b (0.01) 0.20 d (0.04) 0.008 b (0.005) 0.026 b (0.005) 0.011 c (0.006) 0.08 c (0.03) 0.02 b (0.01) 0.03 b (0.01) 0.05 c (0.02) 8.5 b (0.1) 0.09 (0.13) 0.5 b,c (0.2) 0.07 c (0.03)

0.08 b (0.01) 0.015 a (0.001) 0.61 b (0.02)

0.010 a (0.002) 0.002 a (0.001) 0.11 a (0.02)

0.11 a (0.01) 0.013 c (0.002) 2.0 e (0.3)

-

-

-

2.4 c

-

0.008 a (0.002) 0.057 b (0.008) 0.002 a (0.000) 0.003 a (0.001) 0.001 a (0.001) 0.004 a,b (0.001) 0.013 a,b (0.004) 3.6 a (0.1) 0.23 a,b (0.07) 0.005 a (0.002)

-

a

d

0.001 (0.001)

0.18 (0.02)

-

b

0.004 (0.001) 0.003 a (0.002)

0.024 (0.004) 0.005 b (0.001) 0.037 b (0.006) 0.022 b,c (0.002)

-

-

a,b

-

0.030 b,c (0.006) 14.9 c (1.5) 0.6 c (0.1) 0.006 a (0.009) -

0.19 d (0.01) 0.08 b,c (0.01) 1.5 d (0.1) 0.021 a (0.004) 0.033 b (0.006) 0.100 c (0.007) 0.012 b (0.003) 0.017 b (0.06) 0.010 c (0.010) 0.023 a,b (0.004) 0.025 c (0.004) 0.025 a,b (0.004) 0.027 b,c (0.004) 12.7 c (2.7) -

0.13 c (0.01) 0.015 a (0.006) 1.2 c,d (0.2) 0.016 a (0.001)

-

a,b

0.027 (0.004) 0.05 a (0.01)

0.08 b,c (0.01) 0.007 a,b (0.006) 0.017 a (0.002) 0.021 a,b (0.003) 6.5 b (1.3) -

0.028 a,b (0.001) 0.043 a (0.005)

36

37

Maltotetraose

38

Trigalactosyl-inositol

39

Trigalactosyl-inositol

40

Verbascose

41

Tetragalactosyl-methyl-scylloinositol Tetragalactosyl-methyl-scylloinositol Reducing pentasaccharide

42 43 44 45

Reducing pentasaccharide Ajugose α-GOS Free inositols Cyclitol glycosides Other non bioactive carbohydrates

a-e

4117/ 4154 4167 4187 4305 4389 4407 4445 4476 4770

(0.2) 0.02 a (0.01) 0.016 a (0.005) 54.8 b,c (3.6) 0.028 b (0.005) 0.09 c (0.02) 0.8 (0.6) 0.3 (0.2) 1.1 a (0.2) 65.2 c (4.1) 1.8 b,c (0.5) 0.8 d (0.3) 52.7 d (20.4)

0.02 a (0.02) 0.06a (0.05)

0.12 b (0.02) 0.007 a (0.006)

42.3 b (6.7) 0.10 c (0.03) 0.12 d (0.01)

0.2 b (0.1) 0.015 a (0.003) 0.014 a (0.004) 62.5 c (9.6) 0.036 b (0.008) 0.050 b (0.007)

-

-

-

-

-

-

-

-

-

-

-

-

-

9.5 a (3.3) 0.010 a (0.003) 0.013 a (0.003)

0.32 a (0.06)

-

-

a

0.21 (0.08) 13.9 b (3.6) 0.015 a (0.003) 0.24 b (0.04) 3.1 b (0.5)

a

0.44 a (0.08) 0.022 a (0.003) 1.1 a (0.3)

0.8 (0.1) 59.2 c (8.5) 2.3 c (0.4) 1.0 d (0.2) 15.1 c (2.6)

15.8 a (2.4) -

a

0.6 (0.3) 77.3 c (12.7) 2.1 c (0.3) 0.8 d (0.2) 15.6 c (2.6)

1.3 a (1.7) 24.9 b (5.6) 1.3 b (0.2) 0.43 c (0.07) 5.7 b (1.6)

Different letters indicate significant differences (P < 0.05) among the solvents for carbohydrate concentrations.* Standard deviation in parenthesis.

37

Table 2. Mass spectrometric features (i.e., characteristic m/z fragments and ratios of their abundances) of glycosyl-cyclitols found in mung bean extracts.

Peak DP*

Ratios of m/z ions 133/129 260/265

Presence Tentative assignation of m/z 375

12

2

1.27

3.60

Yes

Glycosyl-methyl-scyllo-inositol

13

2

1.49

3.00

Yes

Glycosyl-methyl-scyllo-inositol

14

2

1.20

**

Yes

Glycosyl-methyl-scyllo-inositol

16

2

1.53

4.23

Yes

Glycosyl-methyl-scyllo-inositol

17

2

0.21

0.00

No

Glycosyl-inositol

18

2

0.18

0.00

No

Glycosyl-inositol

19

2

0.28

0.00

No

Galactinol

23

3

1.49

2.75

Yes

Diglycosyl-methyl-scyllo-inositol

24

3

1.05

2.66

Yes

Diglycosyl-methyl-scyllo-inositol

25

3

1.34

**

Yes

Diglycosyl-methyl-scyllo-inositol

26

3

0.35

0.00

No

Diglycosyl-inositol

27

3

0.00

0.00

No

Diglycosyl-inositol

29

3

0.00

0.00

No

Di-glycosyl-inositol

31

3

0.26

0

No

DGMI

35

4

1.00

2.65

Yes

Triglycosyl-methyl-scyllo-inositol

36

4

0.91

3.00

Yes

Triglycosyl-methyl-scyllo-inositol

38

4

0.38

0.00

No

Tri-glycosyl-inositol

39

4

0.34

0.00

No

Triglycosyl-inositol

41

5

1.26

**

Yes

Tetraglycosyl-methyl-scyllo-inositol

42

5

1.16

**

Yes

Tetraglycosyl-methyl-scyllo-inositol

*

Degree of polymerization;

**

m/z 265, not detected.

38

Table 3. Box–Behnken experimental design of the MAE of selected carbohydrates from mung bean (MBS2). Ri, cyclitol amount (mg.g -1 dry sample); Ra, -GOS amount (mg.g -1 dry

sample) and Rc, non-bioactive sugar amount (mg.g-1 dry sample).

Experiment No 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16

t

T

s

Rc

Ri

Ra

3 16.5 30 16.5 3 16.5 30 16.5 3 3 16.5 16.5 30 30 16.5

85 120 120 120 120 50 50 50 50 85 85 85 85 85 85

0.1 1 0.55 0.1 0.55 1 0.55 0.1 0.55 1 0.55 0.55 0.1 1 0.55

5.30 7.95 8.27 6.17 8.42 12.81 15.70 7.43 16.85 14.10 19.14 14.86 8.45 12.66 21.21

2.22 1.84 2.74 2.93 2.29 3.20 4.17 3.59 4.78 2.96 4.49 3.30 3.59 3.76 4.24

69.79 49.83 130.44 83.10 80.27 107.25 143.81 81.78 121.48 123.83 125.81 114.67 99.01 108.23 127.47

Table 4. Concentration (mg.g-1) of LMWC extracted from mung beans of different geographical origins. Peak number

Carbohydrate

MBA1

1

Fructose

0.058 (0.009)* a 0.08 (0.02) a

Sorbitol

0.054 (0.005) a,b 0.07 (0.01) b

3, 4

Galactose & Glucose

0.056 (0.003) a

0.08 (0.01) b

5

Methyl-scyllo-inositol

3.4 (0.3) a

4.6 (0.7) a,b

6

Galacturonic acid

0.6 (0.1) b

0.31 (0.04) a

7

Glucaric acid

0.55 (0.04) a

0.59 (0.06) a

8

scyllo-inositol

0.062 (0.005) a

0.07 (0.01) a

myo-inositol

0.55 (0.04) a

0.65 (0.08) a,b

Sucrose

10.3 (0.3) a

10.7 (1.8) a

Pentosyl-glucoside

0.04 (0.02) a

0.05 (0.04) a

12

Glycosyl-methyl-scyllo-inositol

0.10 (0.02) a

0.17 (0.02) b

13

Glycosyl-methyl-scyllo-inositol

0.22 (0.02) a

0.31 (0.03) a

2

9 10 11

MBC1

MBC2

MBT1

MBS2

0.08 0.14 (0.01) b 0.042 (0.005) a (0.03) a 0.04 (0.02) 0.056 (0.003) a,b 0.031 (0.004) a a,b

0.04 (0.01) a 5.9 (1.0) a,b 0.3 (0.1) a 0.8 (0.1) b 0.10 (0.01) b 0.61 (0.06)

0.06 (0.01) a,b

0.054 (0.005) a

6.9 (1.9) b

4.1 (0.2) a,b

0.3 (0.1) a

0.36 (0.07) a

1.14 (0.07) c

0.40 (0.05) a

0.12 (0.02) b

0.056 (0.007) a

0.7 (0.1) b

0.55 (0.05) a

a,b

13.9 17.5 (2.3) b (1.7) a,b 0.032 (0.007) 0.026 (0.006) a

12.2 (1.4) c 0.021 (0.003) a

a

0.23 0.23 (0.04) d (0.02) c 0.44 0.31 (0.04) a (0.06) b

0.155 (0.003) a,b 0.28 (0.02) a

39

14

Glycosyl-methyl-scyllo-inositol

0.23 (0.04) a,b

0.18 (0.06) a

15

Maltose

0.18 (0.03) a

0.31 (0.08) a

16

Glycosyl-methyl-scyllo-inositol

0.04 (0.01) a,b

0.03 (0.01) a

17

Galactosyl-inositol

0.025 (0.004) a

0.040 (0.007) a,b

Galactosyl-inositol

0.02 (0.01) a,

0.016 (0.002) a

19

Galactinol

0.4 (0.3) a

0.8 (0.1) a,b

20

Diglycosyl-glycerol

0.26 (0.02) a

0.30 (0.03) a

21

Raffinose

4.0 (0.3) a

6.6 (0.6) c

Maltotriose

0.031 (0.003) a

0.057 (0.006) b

Digalactosyl-methyl-scyllo-inositol

0.045 (0.004) b

Digalactosyl-methyl-scyllo-inositol

0.23 (0.04) a

18

22 23 24 25

0.04 (0.02) b 0.50 (0.09) b 0.014 (0.002) a

0.046 (0.005) a

Trisaccharide

0.08 (0.01) a

Digalactosyl-scyllo-inositol

0.013 (0.002) a

Trigalactosyl-glycerol

0.038 (0.003) a

31

DGMI

0.033 (0.004) a

0.06 (0.02) a,b

32

Stachyose

13.2 (1.3) a,b

11.5 (1.6) a

35

Trigalactosyl-methyl-scyllo-inositol 0.12 (0.02) a

0.27 (0.06) a,b,c

36

Trigalactosyl-methyl-scyllo-inositol 0.3 (0.1) a

0.9 (0.2) b,c

37

Maltotetraose

0.15 (0.03) a

0.24 (0.04) b

38

Trigalactosyl-inositol

0.004 (0.007) a

0.030 (0.01) a,b

39

Trigalactosyl -inositol

0.024 (0.006) a

0.04 (0.02) a

40

Verbascose

50.4 (9.9) a

54.5 (1.7) a,b

45

Ajugose

0.6 (0.2) a

1.5 (0.3) a

Total α-Gos

68.2 (9.6) a

74.1 (4.3) a

Total free inositols

4.0 (0.3) a

5.3 (0.8) a

Total glycosyl-cyclitols

1.9 (0.6) a

3.5 (0.7) a,b

Total non-bioactive carbohydrates

12.4 (0.6) a

13.0 (2.1) a

30

a-d

0.31 (0.06) a

0.07 (0.02) a,b

0.07 (0.02) a,b

0.058 (0.006) b

0.045 (0.006) b

0.024 (0.005) a

0.021 (0.003) a

1.1 (0.1) b 0.45 (0.03) c 6.3 (0.8) b,c 0.097 (0.005)

1.0 (0.1) b

0.8 (0.1) a,b

0.39 (0.05) b,c

0.330 (0.004) a,b

5.4 (0.3) a,b

5.1 (0.6) a,b

0.06 (0.02) b

0.051 (0.004) b

0.036 (0.008) 0.024 (0.003) a

0.041 (0.007) b

0.8 0.63 (0.05) b,c (0.1) c 0.019 (0.004) 0.02 (0.01) a

0.45 (0.07) b 0.011 (0.003) a

a

Digalactosyl -inositol

29

0.27 (0.17) a

b

0.03 (0.01) a

28

0.28 (0.06) a,b

c

Digalactosyl -inositol 27

0.31 (0.06) a,b

a

Digalactosyl -methyl-scyllo-inositol 0.012 (0.004) a

26

0.34 (0.05) b 0.43 (0.14) a 0.08 (0.03) b 0.09 (0.01) c 0.024 (0.004)

0.02 (0.01) a

0.014 (0.002) 0.018 (0.006) a

0.023 (0.008) a

a

0.059 (0.002) a

0.071 (0.005) 0.05 (0.02) a

0.057 (0.004) a

a

0.096 (0.005) a

0.113 (0.007) 0.08 (0.03) a

0.108 (0.006) a

a

0.023 (0.004) a

0.047 (0.008) 0.04 (0.01) b

0.022 (0.003) a

b

0.046 (0.006) a

0.056 (0.007) 0.051 (0.009) a

0.03 (0.02) a

a

0.08 (0.02) b 15.4 (1.9) b,c 0.37 (0.09) c 1.3 (0.4) c 0.27 (0.01) b 0.07 (0.03) b 0.04 (0.01) a 62.2 (2.7) a,b 2.0 (0.9) a 85.9 (6.4) a,b 6.6 (1.1) a 5.2 (0.9) b 16.7 (2.3) a,b

0.08 (0.03) b

0.05 (0.01) a,b

13.4 (1.3) a,b

16.3 (2.1) c

0.37 (0.07) c

0.19 (0.02) a,b

1.0 (0.2) b,c

0.65 (0.07) a,b

0.24 (0.05) b

0.226 (0.005) a,b

0.05 (0.02) a,b

0.022 (0.008) a

0.04 (0.02) a

0.033 (0.005) a

54.9 (13.0) a,b

74.0 (6.1) b

0.9 (0.2) a

0.9 (0.6) a

74.5 (14.8) a

96.3 (9.4) b

7.8 (2.0) a

4.7 (0.25) a

4.3 (0.8) b

3.2(0.5) a,b

20.4 (2.9) b

14.2 (1.7) a

Different letters indicate significant differences (P < 0.05) among the samples for carbohydrate concentrations.* Standard deviation in parenthesis.

40

41

Highlights



An exhaustive characterization of bioactive LMWC in mung bean was carried out



Several cyclitol glycosides were detected for the first time in mung bean



Bioactive carbohydrates were successfully extracted by MAE



MAE was an efficient and a faster alternative than SLE to extract bioactive LMWC

42