UHPLC-QTOF-MS/MS based phytochemical characterization and anti-hyperglycemic prospective of hydro-ethanolic leaf extract of Butea monosperma

Butea monosperma is one of the extensively used plants in traditional system of medicines for many therapeutic purposes. In this study, the antioxidant activity, α-glucosidase and α-amylase inhibition properties of freeze drying assisted ultrasonicated leaf extracts (hydro-ethanolic) of B. monosperma have been investigated. The findings revealed that 60% ethanolic fraction exhibited high phenolic contents, total flavonoid contents, highest antioxidant activity, and promising α-glucosidase and α-amylase inhibitions. The UHPLC-QTOF-MS/MS analysis indicated the presence of notable metabolites of significant medicinal potential including apigenin, apigenin C-hexoside C-pentoside, apigenin C-hexoside C-hexoside, apigenin-6,8-di-C-pentoside and genistin etc., in B. monosperma leave extract. Docking studies were carried out to determine the possible role of each phytochemical present in leaf extract. Binding affinity data and interaction pattern of all the possible phytochemicals in leaf extract of B. monosperma revealed that they can inhibit α-amylase and α-glucosidase synergistically to prevent hyperglycemia.

Material and Methods extract preparation. Fresh leaves were washed, paper dried, immediately quenched with liquid nitrogen and ground to fine powder. The powder was then lyophilized on Christ laboratory freeze dryer (Germany) at −68 °C under decreased pressure for 48 hours. The crude powder was soaked using ethanol-water solvent systems in different proportions (Pure H 2 O, C 2 H 5 OH 20%-100% with regular interval of 20% in each case) under suitable conditions followed by sonication using 150 Soniprep (UK). All six fractions were then vortexed for about 2 hours at ambient conditions. After filtration excess amount of solvent was evaporated using rotavapor under reduced pressure. Obtained fractions were lyophilized for further analysis at −68 °C for 48 hours. The percent yield of each fraction was calculated and stored at −80 °C for further use. total phenolic contents (tpc). TPC of understudy extracts were investigated by the Folin-Ciocalteu method 24 . Briefly, 100µL of every sample, after dissolving in CH 3 OH were mixed in 2% Na 2 CO 3 solution (2 mL). After incubation for 5 min, 100µL Folin Ciocalteu reagent was poured into sample mixture. It was then stayed for 30 min at room temperature (R T ) for development of color, followed by absorbance measurement at 750 nm through spectrophotometer. Outcomes were articulated as mg of gallic acid equivalent per gram dry extract (mg GAE/g DE) 25 .
Total flavonoid contents (TFC). TFC were estimated based upon already reported method 26 . Concisely, 50 mg of each crude sample mixture was soaked in 8mL of aqueous CH 3 OH (80%) followed by filtration using Whatmann no 42-filter paper. After that each sample fraction (300 µL), 30% CH 3 OH (3 mL), 0.5 molar NaNO 2 solution (125 µL) and 0.3 molar AlCl 3 .6H 2 O solution (125 µL) were mixed. Then further incubated for 5 min and added 1 mL of 1 molar NaOH. Measurement of absorbance was carried out at 510 nm by a spectrophotometer. Standard curve for TPC was drawn using rutin as standard and the results were presented as milligram of rutin equivalent per gram dry extract (mg RE/g DE).
DppH radical scavenging assay. Free radical inhibition potentials of the crude extracts were examined via 2,2-diphenyl-1-picryl-hydrazil (DPPH) using a previously reported method 27 . Concisely, 1 mL of 0.1 mM DPPH solution in CH 3 OH was added to 3-4 mL of all tested samples. After vigorous stirring the mixtures were kept undisturbed for 30 minutes at R T . Then absorbance measurement was done at 517 nm using spectrophotometer (UV-1700, Schimadzu, Japan). The BHA was taken as a standard antioxidant for comparison. The capability to inhibit the DPPH radicals was assessed using the following equation 28  β-carotene bleaching assay (BcB). Anti-oxidant efficiency of crude plant fractions can also be calculated in-vitro by evaluating the bleaching of β-carotene in presence of Linoleic acid. The β-carotene (2 mg) was added in CHCl 3 (10 mL) along with addition of linoleic acid (0.02 mL) and Tween 40 (0.2 mL) 30 . The 0.2 mL of each crude sample was added in prepared mixture. The positive control (BHA) was also run under same conditions. Then incubation was carried out for 15 min at R T and CHCl 3 was removed with the help of rotary evaporator at 39 °C followed by addition of 50 mL of H 2 O. The resulting mixtures were vortexed and absorbance was measured before and after incubation for 2 hours at 50 °C.
where, Ao was absorbance of sample before incubation, Co was the absorbance of control before incubation, At was absorbance of sample after incubation and Ct was absorbance of blank after incubation.
the α-amylase inhibitory activity. The in-vitro anti-diabetic potential of each extract was assessed by measuring their inhibition against starch hydrolyzing enzyme, α-amylase. For this purpose, about 1% of the sample extract was mixed with potato starch (25 mL) along with continuous stirring. Then the enzyme (100 mg) was added to starch solution, stirred and incubated for 1 hour at 38 °C. After that enzymatic activity was stopped by addition of dinitrosalicylic acid in NaOH (2 mL). The sample mixture was then subjected to centrifugation for a while and glucose contents were calculated in the obtained clear solution. Absorbance was noted at 540 nm spectrophotometrically. A test for positive control (acarbose) was also carried out and percent inhibition was evaluated by formula given below 31 .
amylase inhibition (%) (Ab As) where, Ab was absorbance of blank and As was absorbance of Sample.
the α-glucosidase inhibition assay. The α-glucosidase inhibitory activity of fractions was performed by following method 32 . The crude hydroethanolic leaf extracts of B. monosperma were dissolved in 0.1 molar phosphate buffer (pH = 6.9) containing carbohydrate hydrolyzing enzyme. After incubation at 37 °C for 10 min, reaction was started by adding 10 µL of p-nitrophenol-α-D-glucopyranoside in buffer. Re-incubation of mixtures was carried out at 25 °C for 5 min and absorbance was noted at 405 nm and compared with acarbose. All the measurements were made in triplicate and percentage inhibition was calculated 33,34 .
α − = − × glucosidase inhibition (%) (Ab As) As 100 where, Ab was absorbance of blank and As was absorbance of sample  35 . Preparation of ligands, downloaded enzymes, 3D protonation, energy minimization and determination of binding site was carried out by our previously reported methods 35,36 . The view of the docking results and analysis of their surface with graphical representations were done using MOE and discovery studio visualizer 37 .
www.nature.com/scientificreports www.nature.com/scientificreports/ Statistical analysis. The statistical analysis was performed to evaluate the significance level of difference in means by applying one way Analysis of Variance (ANOVA) through Minitab 17.0 software. The standard deviation (SD±) was also calculated for triplicate values.

Results and Discussion
extract yield. Yields of aqueous and hydroethanolic extracts of B. monosperma are shown in Fig. 1. The different solvent systems selected for extraction influenced the extract yields from leaves. The maximum extract yield (19.79 ± 0.49%) was obtained with 60% ethanol. It was considerably different from extract yields achieved by 100% ethanol (15.53 ± 0.20%) and 20% ethanol (14.37 ± 0.13%). That is why 60% ethanol was considered as a suitable choice for the optimum extract yields from leaves of B. monosperma. The statistical analysis revealed that extract yield for 60% ethanol was significantly higher than other fractions (ρ < 0.05). tpc and tfc. Plants, both edible and non-edible are rich source of secondary metabolites including phenolic and flavonoids. These metabolites play a vital role in many activities including anti-oxidant activity 38 . Findings regarding TPC and TFC are summarized in Table 1. The results showed that 60% extract exhibited maximum yield of TPC (125.25 ± 1.25 mg GAE/g DE) and TFC (65.15 ± 0.55 mg RE/g DE), respectively. Aqueous extract exhibited lowest yield of TPC (67.85 ± 1.25mg GAE/g DE) and TFC (27.74 ± 0.74 mg RE/g DE), respectively. The TPC and TFC are well known for antioxidant potential to reduce oxidative stress to an acceptable level. The statistical analysis indicated that the TPC and TFC by 60% ethanol for both plants were significantly higher than the other solvent systems used (ρ < 0.05). The solvent polarity played a vital role in enhancing the TPC and TFC yields in respective extracts. The ethanol is comparatively safe and non-destructive solvent for the purpose of extraction. The addition of aqueous phase in ethanol probably the decisive factor to enhance the extraction efficiency which was reflected in form of high phenolic and flavonoid levels.
DppH radical scavenging activity. The potential to scavenge DPPH radical was calculated in terms of IC 50 value. The IC 50 value represents the concentration which inhibit a chemical or biochemical process by 50% in vitro and is frequently used to express the results of in vitro assays 39 . The IC 50 values for various fractions are presented in Fig. 2. The IC 50 value of BHA (standard) was computed to be 35.47 ± 1.24 µg/mL. The IC 50 value of 54.847 ± 0.6 µg/mL was calculated for 60% ethanolic extract followed by the 80% (IC 50 = 66.08 ± 0.58 µg/mL), 40% (IC 50 = 71.47 ± 0.92 µg/mL), 100% (IC 50 = 80.16 ± 1.33 µg/mL) and 20% ethanol extract (IC 50 = 86.4 ± 1.35 µg/mL). The pure aqueous extract depicted the minimum antioxidant properties as indicated by its IC 50 value (99.76 ± 1.24 µg/mL). It was concluded that antioxidant potential of all the extracts might be due to high phenolic and flavonoid contents as polyphenols are well recognized natural antioxidants 40,41 . The IC 50 value of DPPH scavenging by 60% ethanolic extract was significantly higher than the remaining fractions (ρ < 0.05). However, no extract could match the IC 50 value exhibited by BHA as shown by statistical analysis (ρ < 0.05). The DPPH radical scavenging is widely adopted method to evaluate the antiradical potential of plant extracts. The DPPH scavenging shown by 60% ethanolic extract was comparable to the most recently reported inhibition of aqueous extract of Strychno spotatorum (IC 50 = 50.22 ± 2.21 μg/mL) but less than crude methanol extract of Adiantum capillus (IC 50 = 39.02 μg/mL) 42,43 . The results depicted the 60% ethanolic extract as the most potent antioxidant fraction. tAp assay. Antioxidant potential of crude leaf extracts was judged by noting the variation in oxidation state of molybdenum (Mo) from +6 to +5 by extracts. This reduction resulted in formation of green color complex  www.nature.com/scientificreports www.nature.com/scientificreports/ which absorbed at 695 nm. The results are presented as Fig. 3. The 60% crude ethanolic fraction showed highest antioxidant capacity having TAP of 205.25 ± 2.05 mg ASE/g DE which was considerably higher than the ascorbic acid (90.2 ± 1.1 mg ASE/g DE). The pure water extract exhibited the lowest TAP value (112.15 ± 1.11 mg ASE/g DE) among all extracts. The TAP value of 60% ethanolic extract was statistically significant when compared with other extracts (ρ < 0.05). The anti-oxidant capacity of 60% ethanolic extract of B. monosperma was also significantly higher than formerly reported n-butanol extract of Anchomanes difformis (90mg ASE/g DE). These results suggested that 60% ethanolic leaf extract of B. monosperma was a rich source of antioxidants 44 . Beta carotene linoleic acid assay. The peroxide inhibition for bleaching of β-carotene for 60% extract was 75.44 ± 1.05%. Comparative investigation indicated that 60% ethanolic extract exerted the most prominent antioxidant potential among all fractions (Fig. 4). However, no extract could match the inhibition percentage exhibited by BHA (ρ < 0.05). This discriminatory behavior could be because of variable dissemination of bioactives in extracts. The antioxidant capability of 60% ethanolic extract of B. monosperma leaves was significantly prominent than recently reported inhibition percentage of Bromelia laciniosa ethanolic extract which was 17.88 ± 3.135% 45 . A past report demonstrated that inhibitory potential in bleaching of β-carotene by plant extracts was dose dependent. The high concentrations of extracts might be more effective because of higher contents of bioactive components predominantly phenolic and flavonoids. These were probably responsible for anti-radical and anti-oxidant prospective of plants 46 .   The results of inhibitory effects of B. monosperma leaf extracts against αglucosidase are represented as Fig. 6. Enzyme inhibition was influenced by extracts obtained under various solvent compositions designed for extract preparation. The maximum enzyme inhibition was shown by 60% ethanolic extract (IC 50 = 55.7 ± 1.30 µg/mL) compared to other extracts and found significantly higher than the values exhibited by other extracts (ρ < 0.05). The α-glucosidase inhibition potential of 60% ethanolic leaf extract was much higher than previously reported inhibitory activity of ethanolic extract of Melia azedarach L. and aqueous extract of Cissus cornifolia leaves with IC 50 values of 3444.11μg/mL and 75.31 ± 9.34 μg/mL respectively 48,49 . The high α-amylase and α-glucosidase inhibitory properties by hydroethanolic extracts of B. monosperma were probably due to presence of some significantly effective phytochemicals. The inhibition of these dietary enzymes by extracts provided an appropriate choice which might be able to low the intestinal glucose absorption leading to decline in postprandial glucose level inside living system 50 .

UHpLc-Q-tof-MS/MS analysis. UHPLC-QTOF-MS/MS was used for metabolite profiling of 60% etha-
nolic leaf extract. Full chromatogram of 60% sample is shown as Fig. 7. The mass spectrums along with structures of identified compounds are indicated as Fig. 8. The detail of compounds with their typical fragments (m/z) is given in Table 2.
Proposed fragmentation pattern of the identified compounds are shown in Fig. 9. Compound (1a) was appeared at retention time (t R ) 12.146 min having molecular ion peak [M-H] − at 269 m/z and its characteristic fragment ion was observed at151 m/z. [M-H-C 8 H 6 O] −51 . Further fragmentation of precursor ion produced daughter ions at 225 m/z and 117 m/z due to neutral loss of CO 2 and C 7 H 4 O 4 52,53 and at m/z 117 in MS spectrum (Fig. 9a). The appearance of these peaks in the chromatogram may be due to cross ring (C-ring) (C-ring) bonds breakage in deprotonated flavonoid molecule 54 which confirmed compound (1a) as apigenin.
Compounds (2b), (3c) and (4d) were recognized as C-glycosylated derivatives . These type of compounds are characterized by the loss of typical fragment loss because of the breakage of sugar pyranos ring, namely −120 amu and −90 amu in case of hexocides 55,56 .
Compound (8h) was detected at t R 9.359 min, giving molecular ion [M-H− at 533 m/z as the most intense ion. The base peak at 443 m/z, resulted by neutral loss of 90 amu from precursor ion [M-H-C 2 H 4 O 2 ] − , which suggested that this compound was C-linked glycoside. Moreover, since mass of deprotonated (8h) was 264 amu more than that of apigenin so it was clearly shown that compound contained two pentose units (132+132amu). Hence (8h) was characterized as apigenin-6,8-di-C-pentoside (Fig. 9h) 67,68 .
Apigenin has gained interests since last few decades as a valuable health promoting agent in view of its low inherent toxicity. Apigenin is associated with strong antioxidant and antidiabetic properties. This fact supports the utilization of apigenin rich source in folk medicine for the treatment of DM 69,70 . The methanolic leaf extract of Achillea sivasica presented most potent antioxidant properties with IC 50 0.22 μg/mL, probably because of the highest phenolic and flavonoid contents including apigenin-C-hexoside-C-pentoside, apigenin-C-hexoside-C-hexoside, apigenin-8-C-glucoside, coumaric acid hexoside derivative and so forth 71 .
Genistein and two other isoflavone namely daidzein, and glycitin of soybean were previously reported as strong inhibitors of α-glucosidase in dose-dependent manner 72 .
The investigation on impact of phenolic acids on glucose uptake was carried out in an insulin resistant cell cultured model. It was reported that vanillic acid improved glucose uptake capacity amongst studied phenolic acids. Moreover, it was reported that a significant decrease occurs in serum insulin level, triglycerides and free fatty acids in rats fed on high fat diet upon consumption of vanillic acid. The study confirmed the protective effect of vanillic acid against hyper-insulinemia, hyperlipidemia and hyperglycemia. These results additionally proposed the capability of vanillic acid in preventing the progress of DM 73 . A recent study reported the antioxidant behavior, α-glucosidase and α-amylase inhibitory action of Hyophorbe lagenicaulis leaf extracts. The phytochemical responsible for the antioxidant and enzyme inhibitory properties in leaf extract of Hyophorbe lagenicaulis were identified as kaempferol, rutin, hesperetin 5-O-glucoside, kaempferol-coumaroyl-glucoside, luteolin 3-glucoside, Isorhamnetin-3-O-rutinoside, trimethoxyflavone derivatives and citric acid 74 . Another investigation reported the strong antioxidant and α-glucosidase inhibitory potential of apigenin rich leaf extract of Cycas revoluta 75 . www.nature.com/scientificreports www.nature.com/scientificreports/ The findings of current work regarding secondary metabolite identification indicated the high value compounds including apigenin derivatives and vanillic acid, associated with substantial biological attributes.

Molecular docking studies.
To further strengthen our in vitro results, we also performed molecular docking studies using Molecular Operating Environment (MOE 2016.08). Before docking studies of phytoconstituents of leaf extract of B. monosperma, we performed docking studies on a validation set of the already reported flavones, flavanones and isoflavanone ( Table 3). The docking studies on validation set was carried out under the assumption that the predicted binding affinities along with their reported in vitro activity for porcine pancreatic α-amylase will be predictive of possible role of each phytochemical component in the synergistic effect 76 . www.nature.com/scientificreports www.nature.com/scientificreports/ Three-dimensional structure of porcine pancreatic α-amylase (PPA) complexed with acarbose was downloaded from Protein Data Bank (PDB code 1OSE). For α-glucosidase, docking studies were carried out on homology modelled α-glucosidase reported by our research group 35 . The binding energy data of the validation set for porcine pancreatic α-amylase is given in Table 3. All the compounds are found to show a relationship between binding affinity and IC 50 value, except for apigenin, which showed weaker binding energy than expected from in vitro experiment. The binding cleft of α-amylase lies deep near its center and consists of Asp197, Glu233 and Asp300. While, the active site consists of several aromatic residues and side chains. Aromatic residue present are: Trp58,   Table 3. Binding affinity data and in vitro results of known inhibitors (validation set) of porcine pancreatic αamylase.
Trp59, Tyr62, His101, Pro163, Ile235, Tyr258, His299, His305 and Ala307. The side chains of Arg61 Asp165, Lys200 and Asp236 are also important. Three-dimensional (3D) binding pose of all superposed compounds of validation set is shown in Fig. 10a. The interaction plot showed that these inhibitors form hydrogen bond interactions with key active site residues as well as residues of the binding cleft. Although, apigenin showed weak binding affinity, it forms hydrogen bonding interactions with Asp197 and Asp300. A hydrophobic π-π stacking interaction was also observed between Trp59 and 4-hydroxyphenyl ring (Fig. 10e). Acarbose (7) with IC 50 value 5.3 μM and binding affinity of −9.5683 kcal/mol establishes hydrogen bonding interactions with all important residues. Two-dimensional interaction plot of all compounds is shown in Fig. S-1 (Supporting Information). The bioactive compounds identified through UHPLC-QTOF-MS/MS based phytochemical characterization were subjected to docking simulations to determine their binding affinities. The binding affinity data of the compounds is presented in Table 4. The results showed that binding affinities range from −4.7156 to −9.5683 kcal/mol with porcine pancreatic α-amylase. Three-dimensional (3D) binding pose of all the identified bioactive compounds are shown in Fig. 10b. The interactions of the ligands with active site amino acid residues of enzyme are shown in Table 4. Two-dimensional (2D) interaction plot of all compounds is shown in Fig. S-2 (Supporting Information). The Fig. 10c-f showed the binding poses of genistein, apigenin 7-O-diglucuronide and apigenin-6,8-di-C-pentoside (Compound 1, 3 and 6 in Table 4). The binding-pose of compound 7 (Apigenin-C-hexoside-C-pentoside isomer, Table 4) superposed on native ligand is shown in Fig. 11a. The 3D binding interaction of establishes hydrogen bond interactions with Asp197, Lys200, Glu240 and Gly304.

No. Compound
Binding Affinity (α-Amylase) Interacting residues of PPA.  Table 4. Binding affinity data and ligand interactions shown by possible isolated phytochemicals against porcine pancreatic α-amylase.
The structural interactions of plant based phytochemicals with active sites of α-amylase and α-glucosidase have been reported in some studies. The blockage of active site region of dietary enzymes by secondary metabolites of plants might be a decisive factor behind the enzymatic activity loss. The energy binding calculations regarding activity loss of dietary enzymes reported in previously published literature indicated a close relationship between enzyme inhibition activities of phytochemicals and acarbose 75,77 . conclusions In current work, antidiabetic and antioxidant potential of hydro-ethanolic leaf extracts of B. monosperma were evaluated. The extract yields, TPC and TFC suggested the 60% ethanol as most effective solvent composition for optimum extraction. The 60% ethanolic extract was proved as most efficient fraction with maximum antioxidant and α-glucosidase inhibitory potential. The UHPLC-Q-TOF-MS/MS analysis revealed the presence of secondary metabolites of medicinal importance. The findings of molecular docking based on binding affinity data and interaction pattern of phytochemicals in leaf extract of B. monosperma revealed that they can inhibit α-amylase and α-glucosidase synergistically to prevent hyperglycemia.    Table 5. Binding affinity data and ligand interactions shown by possible isolated phytochemicals against yeast α-glucosidase.