Chemical Constituents of Muehlenbeckia tamnifolia (Kunth) Meisn (Polygonaceae) and Its In Vitro α-Amilase and α-Glucosidase Inhibitory Activities

The phytochemical investigation of Muehlenbeckia tamnifolia, collected in Loja-Ecuador, led to the isolation of nine known compounds identified as: lupeol acetate (1); cis-p-coumaric acid (2); lupeol (3); β-sitosterol (4) trans-p-coumaric acid (5); linoleic acid (6) (+)-catechin (7); afzelin (8) and quercitrin (9). The structures of the isolated compounds were determined based on analysis of NMR and MS data, as well as comparison with the literature. The hypoglycemic activity of crude extracts and isolated compounds was assessed by the ability to inhibit α-amylase and α-glucosidase enzymes. The hexane extract showed weak inhibitory activity on α-amylase, with an IC50 value of 625 µg·mL−1, while the other extracts and isolated compounds were inactive at the maximum dose tested. The results on α-glucosidase showed more favorable effects; the hexanic and methanolic extracts exhibited a strong inhibitory activity with IC50 values of 48.22 µg·mL−1 and 19.22 µg·mL−1, respectively. Four of the nine isolated compounds exhibited strong inhibitory activity with IC50 values below 8 µM, much higher than acarbose (377 uM). Linoleic acid was the most potent compound (IC50 = 0.42 µM) followed by afzelin, (+)-catechin and quercitrin.


Introduction
Diabetes is defined as a complex metabolic disorder that affects protein, carbohydrate and fat metabolism, all due to insulin insufficiency or insulin dysfunction. According to the World Health Organization (WHO), this disease affects more than 382 million people around the world [1]. Hyperglycemia and dyslipidemia factors, increased in diabetes, are involved in the cardiovascular failures that frequently affect patients. Diabetes is considered a public health problem, affecting people in both developed and developing countries, with no difference of race. According to the diabetes country profiles, diabetes mellitus accounts for 4% of total deaths in Ecuador, and ranks sixth in the population's death causes, followed by respiratory diseases [2]. For this reason, it is a priority issue for public health that needs the attention of both the government and academy. Folk medicine is well recognized in Ecuador, where almost 80% of the population uses plants as their primary source for medical treatment. Several plants are successfully used as antidiabetics in traditional medicine, and the WHO has recommended the assessment of these plants and their product derivatives for their potential use as oral therapy [3]. A scientific validation of medicinal plants is necessary in order to prove efficiency, potency and security. Although many plants have beneficial effects, they can

Results and Discussion
The phytochemical study of M. tamnifolia led to the isolation and identification of nine compounds, whose structures are shown in Figure 1. The spectral properties of these known compounds, including 1 H-NMR and 13 C-NMR data, were identical to those previously described in the literature. Among the species of this genus, (+)-catechin (7) and quercitrin (9) have been reported only in M. platyclada [16]. Lupeol acetate (1), cis-p-coumaric acid (2), lupeol (3), β-sitosterol (4), trans-p-coumaric acid (5), linoleic acid (6) and afzelin (8) are reported here for the first time in this genus.
The 1 H-NMR spectrum of compound 1 shows the presence of an acetyl group as a singlet at 2.05 ppm, typical of the methyl bonded to a carboxyl group. The 13 C-NMR shows a signal at 173.8 ppm, consistent with the carbon atom of an ester group, and a signal at 21.4 ppm corresponding to the methyl group of an acetyl derivative. The remaining 1 H-and 13 C-NMR signals are similar to those obtained for compound 3 [17].
The 1 H-NMR spectrum of compound 2 ( Figure 1) displayed two doublets, one at 7.63 (2H, d, J = 8.8 Hz) and 6.79 (2H, d, J = 8.8 Hz) ppm, typical of a 1,4-substituted aromatic ring. Two doublets at 6.83 (1H, J = 12.5 Hz) and 5.82 (1H, J = 12.5 Hz) ppm revealed the presence of a cis olefin coupling, corresponding to H-7 and H-8 in the proposed structure. The above spectral features are in close agreement to those observed for cis-p-coumaric acid [18].  The 1 H-NMR spectrum of compound 3 shows six singlets, corresponding to tertiary methyl groups, between 0.75 and 1.02 ppm and a singlet at 1.25 ppm, typical of a methyl group in an isopropenyl system. Two olefinic protons at 4.68 and 4.56 ppm are consistent with the methylene group of the same propylenic system. The 13 C-NMR data shows the characteristic signals of C-3 at 79.2 ppm, C-20 at 151.1 ppm and C-29 at 109.5 ppm [17].
The 1 H-NMR spectrum of compound 4 shows two signals, two methine groups at 3.51 and another at 5.35 ppm, consistent with H-3 and H-6 of a steroid nucleus. The 13 C-NMR spectrum showed 29 carbon atoms including an oxymethine carbon signal at 72.0 ppm, two olefinic carbons at 140.9 and 121.9 ppm and two methylene carbon signals at 34.1 and 26.2 ppm, corresponding to C-23 and C-22 [19].
The 1 H-NMR spectrum acquired for compound 5 (Figure 1) was almost identical to that of cis-p-coumaric acid (2), suggesting a close structural similarity between these two compounds. The difference is found in two doublets (J = 16.0 Hz) centered at 7.64 and 6.29 ppm which suggest the trans-derivative of compound 2 [20].
Compound 6 was characterized using GC-MS. Peak identification was carried out by comparing the recorded mass spectra with those present in the instrument library (WILEY 7n.l).
The 1 H-NMR spectrum of compound 7 shows three signals corresponding to aromatic protons at 6.72, 6.76 and 6.83, corresponding to a 1,3,4-trisubstituted ring. The spectrum also shows the signals of two aromatic proton signals at 5.85 and 5.93 ppm, typical of a 1,2,3,5-tetrasubstituted ring. Two oxygenated methine protons signals are located at 3.97 and 4.56 ppm, consistent with H-3 and H-2 in the structure of catechin, respectively. Moreover, two methylenic diastereotopic proton signals can be observed at 2.55 and 2.87 ppm, which is consistent with position 4 of catechin. The 13 C-NMR spectrum and DEPT experiment show five oxygenated quaternary aromatic signals at 157.8, 157.6, 156.9, 146.2 and 146.2 ppm, two non-oxygenated quaternary aromatic signals at 132.2 and 100.8 ppm, and five CH aromatic signals at 120.0, 116.1, 115.2, 96.3 and 95.5 ppm. Two oxygenated methine carbons signals at 82.8 and 68.8 ppm and one methylene carbon signal at 28.5 ppm are also present. The metabolite has been identified as the dextrorotatory enantiomer by polarimetry [21].
The 1 H-NMR spectrum of compound 8 shows the proton signals of a para-disubstituted benzene ring at 7.77 and 6.93 ppm, those of a 1,2,3,5-tetrasubstituted benzene ring at 6.38 and 6.20 ppm, that of a typical anomeric proton doublet at 5.38 ppm and the signals of a sugar moiety between 4.20 and 2.00 ppm. Furthermore, the characteristic doublet of the rhamnose methyl group is present at 0.92 ppm. The 13 C-NMR spectrum showed 21 signals, including the one of an α,β-unsaturated ketone at 179.6 ppm. Six signals of quaternary aromatic and olefinic oxygenated carbon atoms are present at 165.9, 163.2, 161.6, 159.3, 158.6 and 136.1 ppm, while two quaternary aromatic non-oxygenated atoms correspond to signals at 122.6 and 105.73 ppm. Other CH aromatic carbon atoms are represented by signals at 131.9 (positions 2 and 6 ), 116.5 (positions 3 and 5 ), 100.1 and 94.9 ppm. The rhamnose moiety anomeric carbon atom corresponds to the signal at 103.5 ppm [22,23].
The 1 H-NMR spectrum of compound 9 shows the proton signals of a 1,3,4-trisubstituted benzene ring at 7.33, 7.31 and 6.91 ppm, and the remaining signals are in complete agreement with those previously reported for compound 8 [22].

α-Amylase and α-Glucosidase Inhibition Activity
In the α-amylase assay only the hexanic extract was active at the maximum dose tested (625 µg/mL with a sample solution of 10 mg/mL). The other extracts and compounds exhibited inhibition percentages less than 35% at the tested conditions and were not included in this report.
For the α-glucosidase assay, both hexanic and methanolic extracts exhibited strong inhibitory activity and only four compounds were active. Three of them belonging to the flavonoid family and one belonging to the essential fatty acids group exhibited strong inhibitory activity, with IC 50 values below than 8 µM (Table 1). Their inhibitory activities were even 48.5-to 754-fold higher than the commercial drug Acarbose (IC 50 = 377 µM). The remaining five compounds exhibited no inhibitory activity over α-glucosidase at the maximum dose tested (less than 1% of enzyme inhibition) and only trans-p-coumaric acid exhibited 27% of enzyme inhibition. The activity of Afzelin (Kaempferol 3-O-rhamnoside) was 20-fold higher compared to the activity reported by Ajish et al. [24], who obtained an IC 50 value of 81.16 µM. The method described previously in the methodology measures the pNP released during 60 min which allows us to obtain more reliable data by measuring enzymatic catalysis for a long period of time instead of short periods or even end point measurements. Additionally, using a high amount of enzyme can greatly affect the IC 50 value obtained. Lupeol Acetate - Quercitrin * -7.77 D Acarbose 10 377 Note: * Max concentration tested: 10 mg/mL for extracts and 1 mM for compounds.
Moreover, the flavonoids have demonstrated to be moderate to strong inhibitors of α-glucosidase, with IC 50 values below 100 µM [24][25][26][27][28]. Linoleic acid was demonstrated to be the most potent compound, followed by afzelin, (+)-catechin and quercitrin. PUFA (polyunsatured fatty acids) have been demonstrated to be good inhibitors of glucosidase, exhibiting different types of inhibition and potency, as demonstrated by Liu et al. [29], where C18 unsaturated fatty acids where more active versus longer fatty acid chains with more double bonds (or even no double bonds). The α-glucosidase inhibitors acting toward enzymes in the intestine have been shown to effectively delay the blood glucose level, lowering glucose absorption.

Plant Material
The leaves of M. tamnifolia were collected in the flowering stage in the locality of "Cerro Gañil", province of Loja, Ecuador, in February 2015, at 2987 m.a.s.l. (Meters above sea level, coordinates 686617 N; 9602180 E). The plant material was identified by Nixon Cumbicus and, voucher specimens (PPN-PL-006) are deposited at the Herbarium HUTPL at the Universidad Técnica Particular de Loja.

Extraction and Isolation of Compounds
The dried leaves of M. tamnifolia, (500 g) were macerated for 1 h in a series of solvents according to a rule of increasing polarity: hexane (Hex), ethyl acetate (EtOAc) and methanol (MeOH), each solvent 6 L three times. The obtained solutions were filtered and concentrated under reduced pressure, to get three total extracts: 6.25 g (Hex) 6.35 g (EtOAc), and 63.34 g (MeOH).
The hexane extract (3 g) was submitted to silica column chromatography, with an extract/silica ratio of 1:55. The column was eluted according to a gradient of increasing polarity, from hexane-ethyl acetate 90:10 to 100% ethyl acetate, obtaining a total of 16 fractions (MM01-5/MM16-5).
The fraction MM04-40 was purified on reversed phase column chromatography, with an isocratic system of MeOH-H 2 O 90:10, to obtain compound 9 (15 mg).

Measurement of α-Amylase Inhibitory Activity
The inhibitory activity is measured according to the method reported by Xiao et al. [30] with slight modifications. Acarbose was used as the positive control. Starch solution was prepared by dissolving 1 g in 50 mL of 0.4 M NaOH and heated for 5 min at 100 • C. After cooling in iced water, the solution pH was adjusted to 7 with 2 M HCl and H 2 O was added to complete 100 mL. Sample solutions were prepared by dissolving 10 mg in 1 mL MeOH:H 2 O (50:50). Several dilutions in PBS (SIGMA-P4417) were made in case of getting complete enzyme inhibition. PBS solution (35 µL), substrate (35 µL) and sample (5 µL) solutions were mixed in a 96-well microtiter plate, and the mixture was pre-incubated at 37 • C for 1 min. Then, 20 µL of a 50 µg/mL α-amylase solution was added to each well. The plate was incubated for 15 min. The reaction was terminated by addition of 50 µL of 0.1 M HCl and then 150 µL of 0.5 mM iodine solution (0.5 mM I 2 and 0.5 mM KI) were added. The absorbance was measured in a microplate reader (EPOCH 2, Biotek Instruments Inc. Winoosky, VT, USA) at 580 nm. Inhibitory activity (%) was calculated according to the formula described by Kusano et al. [31] as follows: where ABS1 is the absorbance of the incubated solution containing sample, starch and amylase, ABS2 is the absorbance of incubated solution containing sample and starch, ABS3 is the absorbance of incubated solution containing starch and amylase, and ABS4 is the absorbance of incubated solution containing starch. IC 50 value was calculated by curve fitting of data (GraphPad Prism 5.0, GraphPad Software, Inc., La Jolla, CA, USA). Acarbose was used as positive control.

α-Glucosidase Inhibition Assay
α-Glucosidase inhibitory activity was determined using a 96-well microtiter plate with p-nitrophenyl-α-D-glucopyranoside as the substrate, according to the method describe by Tao et al. [32], with slight modifications. Sample solutions were prepared by dissolving 10 mg in 1 mL MeOH:H 2 O (50:50). Several dilutions in PBS were made in case of getting complete enzyme inhibition. First, 75 µL of PBS (SIGMA-P4417) was mixed with 5 µL of the sample and 20 µL of the enzyme solution (0.15 U/mL in PBS pH 7.4), then it was pre-incubated at 37 • C for 5 min prior to the initiation of the reaction by adding the substrate. After pre-incubation, 20 µL of PNPG (5 mM in phosphate buffer, pH 7.4) was added and then incubated at 37 • C. The amount of p-nitrophenol (p-NP) released was measured in an EPOCH 2 (BIOTEK ® ) microplate reader at 405 nm for 60 min, recording the absorbance every 5 min. The results were expressed as inhibition percentage by means of the formula described by Choi et al. [33] as follows: where A 0 is the absorbance recorded for the enzymatic activity without inhibitor (control), and A s is the absorbance recorded for the enzymatic activity in presence of the inhibitor (sample). IC 50 value was calculated by curve fitting of data (GraphPad Prism 5.0). Acarbose was used as positive control.