http://informahealthcare.com/phb ISSN 1388-0209 print/ISSN 1744-5116 online Editor-in-Chief: John M. Pezzuto Pharm Biol, Early Online: 1–5 ! 2015 Informa Healthcare USA, Inc. DOI: 10.3109/13880209.2014.957781
ORIGINAL ARTICLE
Potent protein tyrosine phosphatase 1B (PTP1B) inhibiting constituents from Anoectochilus chapaensis and molecular docking studies Jinyan Cai1, Lin Zhao2, and Weiye Tao3
Pharmaceutical Biology Downloaded from informahealthcare.com by University of Bristol on 02/25/15 For personal use only.
1
School of Pharmacy, Guangdong Pharmaceutical University, Guangzhou, China, 2School of Life Science and Bio-pharmaceutical, Guangdong Pharmaceutical University, Guangzhou, China, and 3School of Medical Information Engineering, Guangdong Pharmaceutical University, Guangzhou, China Abstract
Keywords
Context: Anoectochilus chapaensis Gagnep. (Orchidaceae), an indigenous and valuable Chinese folk medicine, has been used as an antidiabetic remedy. However, the bioactive constituents have not been reported. Objective: To explore potent protein tyrosine phosphatase 1B (PTP1B) inhibitors from the whole herbs of A. chapaensis for the treatment of diabetes. Materials and methods: The compounds were obtained by PTP1B bioactivity-guided isolation from the active fraction of ethonal extract of A. chapaensis, and elucidated by extensive spectroscopic methods and evaluated for their potential to inhibit PTP1B with a series of doses in dimethyl sulphoxide by a colorimetric assay in vitro. The Autodock program was used to dock the active compounds into the binding sites. Results: Fifteen compounds were identified; epifriedelanol, friedelane, 2a, 3b-dihydroxyolean12-en-23, 28, 30-trioic acid, dibutyl-phthalate, and 7-hydroxy-2-methoxy-9,10-dihydrophenanthrene-1,4-dione were isolated from the genera Anoectochilus for the first time. All 15 compounds were tested for their inhibitory activity against PTP1B in vitro. Nine active compounds exhibited potent inhibitory effect with IC50 values of 1.16–6.21 mM, which were comparable with the positive control suramin. The 3D-docking simulations showed negative binding energies of 7.4 to 8.5 kcal/mol and supported a high affinity to PTP1B residues in the pocket site, indicating that they may stabilize the open form and generate tighter binding to the catalytic sites of PTP1B. Discussion and conclusion: The results clearly demonstrated that the potential active constituents from A. chapaensis could inhibit PTP1B, which may be mainly attributed to a combination of triterpenoids and flavonoids.
Autodock program, Bioactivity-guided isolation, diabetes
Introduction Protein tyrosine phosphatase 1B (PTP1B) is a key negative regulator of insulin signaling, and mounting evidence has linked PTP1B to insulin resistance, obesity, and type 2 diabetes mellitus (T2DM) (Panzhinskiy et al., 2013). In the insulin-signaling pathway, PTP1B dephosphorylates the insulin receptor and the insulin receptor substrate IRS-1. Insulin signaling is turned off by dephosphorylation of the tyrosine residues on the activation loop of the IR abolishing its kinase activity. Deletion of PTP1B in the mouse resulted in enhanced insulin sensitivity as measured by improved glucose clearance in glucose and insulin tolerance tests (Elchebly et al., 1999; Klaman et al., 2000). So far, overwhelming evidence suggests
Correspondence: Jinyan Cai, School of Pharmacy, Guangdong Pharmaceutical University, Guangzhou, China. E-mail: caijy928@ 163.com
History Received 13 November 2013 Revised 6 June 2014 Accepted 11 August 2014 Published online 22 January 2015
that inhibiting PTP1B represents a highly promising approach to treat diabetes and obesity. PTP1B’s direct regulation of insulin and the leptin receptors makes it an ideal therapeutic target for T2DM and obesity (Feldhammer et al., 2013; Zhang & Zhang, 2007). The whole plant of Anoectochilus chapaensis Gagnep. (Orchidaceae), an indigenous and valuable Chinese folk medicine, has been used as a popular herbal drug in China and Vietnam (Anonymous, 1999). It is also called ‘‘king medicine’’ because of its diverse pharmacological effects. The whole dried plants have been widely used in China to treat diabetes, nephritis, etc. As a part of our ongoing effort to identify potent antidiabetic constituents from A. chapaensis, the PTP1B inhibitory activity of the extracts and isolates was examined. The ethyl acetate (EtOAc) fraction from ethanol (EtOH) extracts of A. chapaensis showed strong inhibitory bioactivity against PTP1B with an IC50 of 2.64 mg ml1. Here we report on the potent PTP1B inhibiting constituents from A. chapaensis and evaluate their binding site-directed details with the Autodock docking program.
Pharmaceutical Biology Downloaded from informahealthcare.com by University of Bristol on 02/25/15 For personal use only.
2
J. Cai et al.
Pharm Biol, Early Online: 1–5
Materials and methods
Extraction and isolation
General experimental procedures
Dry powdered herbs (1.6 kg) of A. chapaensis were refluxed with 95% (v/v) ethanol for 2 h, and each filtrate was concentrated to dryness in vacuo to render the total EtOH extract, which was suspended in distilled water and partitioned in sequence with petrol ether (PE), EtOAc and n-BuOH, thus yielding four fractions. The respective yields (%) of the PE fraction, EtOAc fraction, n-BuOH fraction and H2O fraction were 14.02, 10.90, 27.48, and 48.19%. The EtOAc fraction (10.90 g, IC50 ¼ 2.64 mg ml1) was chromatographed over silica gel (200–300 mesh) using PE-EtOAc mixtures of increasing polarity. Repeated chromatography with the same eluent over silica gel (400 mesh) afforded compounds 1 (9 mg), 2 (6 mg), 3 (14 mg), 4 (10 mg), 5 (8 mg), and 6 (25 mg); the 50% petroleum–EtOAc part (6.0 g, IC50 ¼ 1.71 mg ml1) was chromatographed over MCI gel using acetone–water (1:1 to 7:3) mixtures of decreasing polarity. Repeated chromatography with the same eluent over MCI gel and Sephadex LH-20 afforded compounds 7 (10 mg), 8 (125 mg), and 9 (10 mg). The 30% EtOH fraction (21.0 g, IC50 ¼ 1.45 mg ml1) was chromatographed over MCI gel using acetone–water (1:1 to 7:3) mixtures of decreasing polarity. Repeated chromatography with the same eluent over MCI gel and Sephadex LH-20 afforded compounds 10 (54 mg), 11 (13 mg), 12 (45 mg), 13 (32 mg), 14 (11 mg), and 15 (9 mg). Compounds 1–15 were identified (as in Figure 1) by comparing their physical and spectra data with the literature values, and the data of compound 5, a first isolated triterpene from Anoectochilus genus were listed below. 2a,3b-dihydroxyolean-12-en-23, 28, 30-trioic acid (5): white amorphous powder. The molecular formula, C30H44O8, was determined on the basis of negative HRESIMS data of [M-1] 531.2971. 1H-NMR (500 MHz, CD3OD), 13C-NMR (125 MHz, CD3OD), and HMBC data are listed in Table 1.
The NMR spectra were recorded on a Bruker Avance 500 DRX (500 MHz for 1H and 125 MHz for 13C) spectrometer (Rheinstetten, Germany) in deuterated solvents [dimethyl sulfoxide (DMSO)-d6, methanol (MeOD), and chloroform (CDCl3)]. High-resolution electrospray ionization mass spectrometry (HRESIMS) was obtained on a Bruker Bio TOF IIIQ mass spectrometer (Bruker Daltonics, Inc., Billerica, MA). The IR spectrum (KBr) was recorded on a Shimadzu FTIR spectrometer (Tokyo, Japan). Column chromatography was conducted using silica gel 60 (100–200 mesh, Qingdao Haiyang Chemical Co. Ltd., Qingdao, China), Sephadex LH-20 (20–150 mm, GE, USA), LiChro prep RP-18 (40– 63 mm, Merck, Darmstadt, Germany), and Diaion HP20 (250–850 mm, Sigma). All TLC was conducted on precoated Merck Kiesel gel 60 F254 plates (20 20 cm, 0.25 mm, Merck, Darmstadt, Germany), using 5% H2SO4 as a spray reagent. Melting points were determined using an XT-4 point apparatus (uncorrected). Optical rotations were determined on a Perkin-Elmer Model 343 polarimeter (Merck, Darmstadt, Germany). Chemicals and reagents PTP1B Tyrosine Phosphatase Drug Discovery Kit was purchased from Enzo Life Sciences Inc. (Lausen, Switzerland). All other chemicals and solvents used were purchased from E. Merck (Darmstadt, Germany) and SigmaAldrich (St. Gallen, Switzerland), unless otherwise stated. Plant material The herbs of A. chapaensis were collected in the month of September of 2010 from Yunnan province, southwest China. The plant was authenticated by Hongyan Ma, an associate professor in School of Traditional Chinese Medicine, Guangdong Pharmaceutical University. A voucher specimen (2010-DY1001) has been deposited in the herbarium of the School of Pharmacy, Guangdong Pharmaceutical University. Biological assay for the inhibition of PTP1B Human protein tyrosine phosphatase 1B (hPTP1B) activity was assayed at room temperature using the PTP1B Tyrosine Phosphatase Drug Discovery Kit, a colorimetric assay designed for screening inhibitors and modulators of PTP1B activity. The kit included human recombinant PTP1B (residues 1–322; MW ¼ 37.4 kDa), expressed in Escherichia coli. The protocol outlined by the manufacturer was followed rigidly. This kit is designed to perform endpoint assays in which each well contains a 100 ml reaction in Assay Buffer and is terminated by the addition of 25 ml of the phosphate detection reagent, BIOMOL REDÔ. The 100 ml ‘‘reaction’’ may consist either of PTP1B phosphatase acting on the phosphopeptide substrate or simply a dilution of the free Phosphate Standard (BML-KI470). Tested compounds were demonstrated to be pure as evidenced by NMR and TLC analysis. IC50 of the compounds which showed inhibition of over 50% were determined.
Molecular docking simulation in PTP1B inhibition-Autodock VINA The docking simulation of the PTP1B enzyme and tested molecules were successfully performed using structure modeling (MODELER-9V10) and a docking program (Autodock VINA), which have been used to estimate the conformation of the protein–ligand complex (Trott & Olson, 2010). Since this Autodock VINA program significantly improves the average accuracy of the binding mode predictions compared with the AutoDock program, the accuracy, repeatability, and reliability of the docking results can be improved for drug discovery. For the docking studies, the crystal structures of the protein targets [Protein Data Bank (PDB ID: 1NNY for human PTP1B)] were allocated from the protein sequence alignment. The 3D structures of test compounds were constructed and minimized using Chemsketch 3.5 and Omega 2.0 software (Open Eye Scientific Software, Santa Fe, NM) for 2D and 3D conformations, respectively. The predicted protein–ligand complexes were optimized and ranked according to the empirical scoring function, TM score (Structural alignment tool, sheba3.1, Open Eye Scientific Software, Santa Fe, NM), which
Identification of biologically active chemicals
DOI: 10.3109/13880209.2014.957781
3
HOOC COOH
H COOH
HO HO
H
1= O 3 = OH 4 = H2
R
HO
O
OH
Pharmaceutical Biology Downloaded from informahealthcare.com by University of Bristol on 02/25/15 For personal use only.
OH
O
6 R1=OH R2=H 7 R1=OCH3 R2=H 8 R1=OCH3 R2=Glc 9 R1=OCH3 R2=Glc-Rha
C
11 R=OH 12 R=H
R
OR2
OH 10
COOH 5 O
HO
R1
OCH3
HO
2
H
C
C
H
C
O
COOH
HO
O O 13
O H 3C
CHO 14
OH O 15
Figure 1. Structures of compounds 1–15.
Table 1. NMR data for compound 5 ( in ppm, J in Hz). No.
H
C
1 2 3 4 5 6
1.98m 3.66m 3.73d(10)
48.4t 69.2d 81.1d 55.2s 52.7d
7 8 9 10 11 12 13 14 15
1.65m 1.98m 1.73m 1.63m 1.73m 1.65m 5.32m 1.73m
HMBC (H!C)
No.
H
C2, 3, 5, 10 C25 C2, 23, 24
16 17 18 19 20
1.55m
C6, 10
24.7t 33.6t 40.5s 49.4d 39.2s 24.3t 123.6d 145.0s 42.9s 28.9t
21 C5 C8, 10, 11, 25, 26 C9, 11, 14, 18
estimates the binding free energy of the ligand–receptor complex. Statistics All results were presented as mean ± SEM of triplicate samples.
Results and discussion Using the PTP1B bioassay as a guide, the chromatography of the EtOAc fraction afforded 15 known compounds (Figure 1), identified by comparison of their physical and spectral data with those of reported values as friedelin (1), sorghumol (2), epifriedelanol (3), friedelane (4), 2a, 3b-dihydroxyolean-12en-23, 28, 30-trioic acid (5), quercetin (6), isorhamnetin (7), isorhamnetin-3-O-b-D-glucoside (8), isorhamnetin-3-Ob-D-rutinoside (9), vanillic acid (10), caffeic acid (11), p-hydroxy-cinnamic acid (12), dibutyl-phthalate (13),
22 23 24 25 26 27 28 29 30
2.80m 1.94m,1.65m 1.98m 1.31m 1.65m 1.12s 1.04s 0.80s 1.20s 1.18s
C 21.8t 44.7s 44.0d 43.5t 47.1s 31.3t 35.1t 180.7s 12.9q 17.5q 17.6q 26.5q 181.5s 30.5q 180.7s
HMBC (H!C) C17 C12, 13, 14, 17, 19 C29 C17, 19, 20 C17 C3, C1, C7, C8,
4, 5, 23 5, 9, 10 8, 9, 14 13, 14, 15
C20, 21, 30
p-hydroxy benzaldehyde (14), and 7-hydroxy-2-methoxy9,10-dihydrophenanthrene-1,4-dione (15). Compounds 3, 4, 5, 13, and 15 were isolated from the genera Anoectochilus for the first time. All 15 compounds were evaluated for their potential to inhibit PTP1B activity. The results showed that 1–9 inhibited PTP1B activity with average IC50 values of 1.16–6.21 mM (Table 2). They were comparable with that of suramin (IC50 ¼ 11 ± 1 mM) (McCain et al., 2004), the positive control in the PTP1B Tyrosine Phosphatase Drug Discovery Kit. The Autodock docking program was employed to dock the compounds into the binding sites of the crystallographic ˚ from the structure of PTP1B, with all residues defined as 4 A ligand in the original complex. In particular, a semi-empirical free energy force field was employed to predict binding of protein–ligand complexes of a known structure and the binding energies for both the bound and unbound states (Morris et al., 2009). Recently, structure-based simulation of
4
J. Cai et al.
Pharm Biol, Early Online: 1–5
the PTP1B inhibitors has been performed for developing novel therapeutic drugs with selectivity and cell permeability (Barr, 2010). The main structural features of PTP1B have been well established and consist of 435 amino-acid residues, including residues 30–278, which comprise the
Pharmaceutical Biology Downloaded from informahealthcare.com by University of Bristol on 02/25/15 For personal use only.
Table 2. IC50 values of PTP1B inhibiting activity and binding energy of compounds 1–9.
Compounds
IC50 (mM)
Binding energy (kcal/mol)
Friedelin (1) Sorghumol (2) Epifriedelanol (3) Friedelane (4) 2a,3b-dihydroxyolean-12-en-23, 28, 30-trioic acid (5) Quercetin (6) Isorhamnetin (7) Isorhamnetin-3-O-b-D-glucoside (8) Isorhamnetin-3-O-b-D-rutinoside (9)
6.21 ± 0.02 3.50 ± 0.01 3.75 ± 0.14 4.60 ± 0.05 2.65 ± 0.03
8.1 7.4 8.3 8.5 8
5.63 ± 0.04 1.75 ± 0.02 1.16 ± 0.03 1.20 ± 0.05
7.5 7.4 7.5 7.8
catalytic domain containing the catalytic residue Cys215, the secondary phosphate-binding loop (P-loop) mediated by residues His214–Arg221, and the WPD loop identified as residues Thr177–Pro189 (Wang et al., 2009). The 3D docking simulations of compounds 1–9 with some structural difference of the pentacyclic triterpenes and flavonols were predicted using the Autodock program to evaluate the binding site-directed inhibition of PTP1B (Table 2 and Figure 2). A reported highly selective ligand of C40H37N3O10 (3-({5[(N-acetyl-3-{4-[(carboxycarbonyl)(2-carboxy phenyl)amino]1-naphthyl}-L-alanyl)amino]pentyl}oxy)-2-naphthoic acid) was simultaneously employed for the positive control. X-ray data confirmed that the inhibitor bound with the catalytic site in the native, ‘‘open’’ conformation. It displayed excellent potency and good selectivity over many other phosphatases (Szczepankiewicz et al., 2003). It displayed a negative binding energy of 7.8 kcal/mol. The tested compounds 1–9 showed negative binding energies of 7.4 to 8.5 kcal/ mol (Table 2), comparable with that of the positive reference. The results support and provide evidence of the
Figure 2. 3D molecular docking models for PTP1B protein’s active pocket site of selected potent candidates: (A) isorhamnetin-3-O-b-D-rutinoside (9), (B) friedelane (4), (C) 3b-dihydroxyolean-2-en-23,28,30-trioic acid (5), and (D) epifriedelanol (3).
Identification of biologically active chemicals
Pharmaceutical Biology Downloaded from informahealthcare.com by University of Bristol on 02/25/15 For personal use only.
DOI: 10.3109/13880209.2014.957781
potent significant inhibition results in vitro. Friedelane (4) (8.5 kcal/mol) and isorhamnetin-3-O-b-D-rutinoside (9) (7.8 kcal/mol) showed the best affinity to PTP1B’s pocket site among the tested triterpenoids and flavonoids. Considering docking results of the PTP1B–compounds complex, the tested inhibitors were stably posed in similar pocket domains of PTP1B residues, including Trp179 and Gly183 in the WPD loop; Asp48, Ser216, Ala217, Gly220, ˚ Arg221, and Gln266 in the pocket site, all of which were 4 A from the inhibitor in the original complex. In particular, the WPD loop plays a much more important role in the specificity and affinity of the inhibitors. In the presence of the ‘‘open’’ conformation of the WPD-loop, the binding pocket of PTP1B is easily accessible to the substrate. After substrate binding, the WPD loop closes over the active site, forming a tight binding pocket for the substrate. The WPD loop closes onto the substrate and thereby positions the thiolate of Cys215 for nucleophilic attack of the phosphotyrosine. It has been observed that the conformational and dynamic features of WPD-loop play a key role in providing a smooth entrance for the inhibitors moving into the binding pocket as well as a favorable microenvironment to stabilize them (Wang et al., 2009). Potent, yet highly selective, PTP1B inhibitory agents can be acquired by targeting the area defined by residues Lys41, Arg47, and Asp48 (Sun et al., 2003). Moreover, the binding energies of 3, 4, 5, and 9 were 8.3, 8.5, 8, and 7.8 kcal/mol, respectively, indicating high affinity to PTP1B residues. Detailed binding mode analysis with docking simulation showed that the inhibitors can be stabilized by the simultaneous establishment of multiple hydrogen bonds and van der Waals contacts in the pocket site (Figure 2). Structure–activity relationship study of these sugar-substituted flavonoid derivatives demonstrated that the PTP1B inhibitory activity was strongly influenced by the carbohydrate moiety at the C-3 position. The rutinosyl substituent at the C-3 position of isorhamnetin skeletons (9) slightly increases the proximity to the Trp179 and Gly183 residues in the WPD loop, resulting in a relatively small difference with 7 and 8 in the binding energy and affinity. They also interacted with Ser216, Ala217, Gly220, and Arg221 in the P loop, indicating that the inhibitors may reduce the mobility of the WPD loop toward a more rigid conformation, which inhibits WPD loop closure and prevents substrate binding (Popov, 2011). Moreover, the triterpene (5) with polar groups may have advantages with regard to the occupation of the anionic charged P loop at a physiologic pH (Park et al., 2009). In addition, the carboxylic group of oleanolic acid (5) at C-23 and hydroxyl groups at C-2 and C-3 generated extensive hydrogen bonds with Ser216, Ala217, and Gly220 in the P-loop, which significantly contributed to its high affinity. Polar moieties substituted triterpenes (5) and flavonoids (7–9) isolated from A. chapaensis showed negative binding energies and a high proximity to PTP1B residues, including Trp179 and Gly183 in the WPD loop; Asp48, Ser216, Ala217, Gly220, Arg221, and Gln266 in the pocket site. These findings may lead to the design of polar moieties substituted triterpenes or flavonoids as PTP1B inhibitors that interact with both catalytic and allosteric sites to achieve selectivity. Furthermore, the triterpenoid backbone of
5
friedelane (4) interacted with non-polar residues based on the van der Waals contacts, and it may have a high affinity with the pocket site through hydrophobic interaction. In conclusion, nine potential inhibitors of PTP1B were identified by experiments and the docking simulation, considering that carboxylic and hydroxyl groups to form the above-mentioned broad arrangement of hydrogen bonds. Both enzymatic test and docking results demonstrate that the contributions of functional groups with a polar moiety may cause high affinity to PTP1B and a lower binding energy. Therefore, polar moieties substituted triterpenes or flavonoids as PTP1B inhibitors hold promise as therapeutic agents for the treatment of diabetes and related disorders.
Declaration of interest The authors declare that they have no conflict of interest. The authors thank National Natural Science Foundation of China (No. 81001628) and Guangdong Natural Science Foundation (No. S2013010014771) for financial support.
References Anonymous. (1999). Editorial Committee of the Flore of China of Chinese Academy of Science. Flora Republicae Popularis Sinicae, vol. 17. Peking: Science Press. Barr AJ. (2010). Protein tyrosine phosphatases as drug targets: Strategies and challenges of inhibitor development. Future Med Chem 2: 1563–76. Elchebly M, Payette P, Michaliszyn E, et al. (1999). Increased insulin sensitivity and obesity resistance in mice lacking the protein tyrosine phosphatase-1B gene. Science 283:1544–8. Feldhammer M, Uetani N, Miranda-Saavedra D, Tremblay ML. (2013). PTP1B: A simple enzyme for a complex world. Crit Rev Biochem Mol Biol 5:430–45. Klaman LD, Boss O, Peroni OD, et al. (2000). Increased energy expenditure, decreased adiposity, and tissue-specific insulin sensitivity in protein-tyrosine phosphatase 1B-deficient mice. Mol Cell Biol 20:5479–89. McCain DF, Wu L, Nickel P, et al. (2004). Suramin derivatives as inhibitors and activators of protein-tyrosine phosphatases. J Biol Chem 279:14713–25. Morris GM, Huey R, Lindstrom W, et al. (2009). AutoDock 4 and AutoDock-Tools 4: Automated docking with selective receptor flexibility. J Comput Chem 30:2785–91. Panzhinskiy E, Ren J, Nair S. (2013). Pharmacological inhibition of protein tyrosine phosphatase 1B: A promising strategy for the treatment of obesity and type 2 diabetes mellitus. Curr Med Chem 21:2609–25. Park H, Bhattarai BR, Ham SW, Cho H. (2009). Structure based virtual screening approach to identify novel classes of PTP1B inhibitors. Eur J Med Chem 44:3280–4. Popov D. (2011). Novel protein tyrosine phosphatase 1B inhibitors: Interaction requirements for improved intracellular efficacy in type 2 diabetes mellitus and obesity control. Biochem Biophys Res Co 410: 377–81. Sun JP, Fedorov AA, Lee SY, et al. (2003). Crystal structure of PTP1B complexed with a potent and selective bidentate inhibitor. J Biol Chem 14:12406–14. Szczepankiewicz BG, Liu G, Hajduk PJ, et al. (2003). Discovery of a potent, selective protein tyrosine phosphatase 1B inhibitor using a linked-fragment strategy. J Am Chem Soc 125:4087–96. Trott O, Olson AJ. (2010). AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization and multithreading. J Comput Chem 31:455–61. Wang JF, Gong K, Wei DQ, et al. (2009). Molecular dynamics studies on the interactions of PTP1B with inhibitors: From the first phosphatebinding site to the second one. Protein Eng Des Sel 6:349–55. Zhang S, Zhang ZY. (2007). PTP1B as a drug target: Recent developments in PTP1B inhibitor discovery. Drug Discov Today 10: 373–81.
ORIGINAL ARTICLE
Potent protein tyrosine phosphatase 1B (PTP1B) inhibiting constituents from Anoectochilus chapaensis and molecular docking studies Jinyan Cai1, Lin Zhao2, and Weiye Tao3
Pharmaceutical Biology Downloaded from informahealthcare.com by University of Bristol on 02/25/15 For personal use only.
1
School of Pharmacy, Guangdong Pharmaceutical University, Guangzhou, China, 2School of Life Science and Bio-pharmaceutical, Guangdong Pharmaceutical University, Guangzhou, China, and 3School of Medical Information Engineering, Guangdong Pharmaceutical University, Guangzhou, China Abstract
Keywords
Context: Anoectochilus chapaensis Gagnep. (Orchidaceae), an indigenous and valuable Chinese folk medicine, has been used as an antidiabetic remedy. However, the bioactive constituents have not been reported. Objective: To explore potent protein tyrosine phosphatase 1B (PTP1B) inhibitors from the whole herbs of A. chapaensis for the treatment of diabetes. Materials and methods: The compounds were obtained by PTP1B bioactivity-guided isolation from the active fraction of ethonal extract of A. chapaensis, and elucidated by extensive spectroscopic methods and evaluated for their potential to inhibit PTP1B with a series of doses in dimethyl sulphoxide by a colorimetric assay in vitro. The Autodock program was used to dock the active compounds into the binding sites. Results: Fifteen compounds were identified; epifriedelanol, friedelane, 2a, 3b-dihydroxyolean12-en-23, 28, 30-trioic acid, dibutyl-phthalate, and 7-hydroxy-2-methoxy-9,10-dihydrophenanthrene-1,4-dione were isolated from the genera Anoectochilus for the first time. All 15 compounds were tested for their inhibitory activity against PTP1B in vitro. Nine active compounds exhibited potent inhibitory effect with IC50 values of 1.16–6.21 mM, which were comparable with the positive control suramin. The 3D-docking simulations showed negative binding energies of 7.4 to 8.5 kcal/mol and supported a high affinity to PTP1B residues in the pocket site, indicating that they may stabilize the open form and generate tighter binding to the catalytic sites of PTP1B. Discussion and conclusion: The results clearly demonstrated that the potential active constituents from A. chapaensis could inhibit PTP1B, which may be mainly attributed to a combination of triterpenoids and flavonoids.
Autodock program, Bioactivity-guided isolation, diabetes
Introduction Protein tyrosine phosphatase 1B (PTP1B) is a key negative regulator of insulin signaling, and mounting evidence has linked PTP1B to insulin resistance, obesity, and type 2 diabetes mellitus (T2DM) (Panzhinskiy et al., 2013). In the insulin-signaling pathway, PTP1B dephosphorylates the insulin receptor and the insulin receptor substrate IRS-1. Insulin signaling is turned off by dephosphorylation of the tyrosine residues on the activation loop of the IR abolishing its kinase activity. Deletion of PTP1B in the mouse resulted in enhanced insulin sensitivity as measured by improved glucose clearance in glucose and insulin tolerance tests (Elchebly et al., 1999; Klaman et al., 2000). So far, overwhelming evidence suggests
Correspondence: Jinyan Cai, School of Pharmacy, Guangdong Pharmaceutical University, Guangzhou, China. E-mail: caijy928@ 163.com
History Received 13 November 2013 Revised 6 June 2014 Accepted 11 August 2014 Published online 22 January 2015
that inhibiting PTP1B represents a highly promising approach to treat diabetes and obesity. PTP1B’s direct regulation of insulin and the leptin receptors makes it an ideal therapeutic target for T2DM and obesity (Feldhammer et al., 2013; Zhang & Zhang, 2007). The whole plant of Anoectochilus chapaensis Gagnep. (Orchidaceae), an indigenous and valuable Chinese folk medicine, has been used as a popular herbal drug in China and Vietnam (Anonymous, 1999). It is also called ‘‘king medicine’’ because of its diverse pharmacological effects. The whole dried plants have been widely used in China to treat diabetes, nephritis, etc. As a part of our ongoing effort to identify potent antidiabetic constituents from A. chapaensis, the PTP1B inhibitory activity of the extracts and isolates was examined. The ethyl acetate (EtOAc) fraction from ethanol (EtOH) extracts of A. chapaensis showed strong inhibitory bioactivity against PTP1B with an IC50 of 2.64 mg ml1. Here we report on the potent PTP1B inhibiting constituents from A. chapaensis and evaluate their binding site-directed details with the Autodock docking program.
Pharmaceutical Biology Downloaded from informahealthcare.com by University of Bristol on 02/25/15 For personal use only.
2
J. Cai et al.
Pharm Biol, Early Online: 1–5
Materials and methods
Extraction and isolation
General experimental procedures
Dry powdered herbs (1.6 kg) of A. chapaensis were refluxed with 95% (v/v) ethanol for 2 h, and each filtrate was concentrated to dryness in vacuo to render the total EtOH extract, which was suspended in distilled water and partitioned in sequence with petrol ether (PE), EtOAc and n-BuOH, thus yielding four fractions. The respective yields (%) of the PE fraction, EtOAc fraction, n-BuOH fraction and H2O fraction were 14.02, 10.90, 27.48, and 48.19%. The EtOAc fraction (10.90 g, IC50 ¼ 2.64 mg ml1) was chromatographed over silica gel (200–300 mesh) using PE-EtOAc mixtures of increasing polarity. Repeated chromatography with the same eluent over silica gel (400 mesh) afforded compounds 1 (9 mg), 2 (6 mg), 3 (14 mg), 4 (10 mg), 5 (8 mg), and 6 (25 mg); the 50% petroleum–EtOAc part (6.0 g, IC50 ¼ 1.71 mg ml1) was chromatographed over MCI gel using acetone–water (1:1 to 7:3) mixtures of decreasing polarity. Repeated chromatography with the same eluent over MCI gel and Sephadex LH-20 afforded compounds 7 (10 mg), 8 (125 mg), and 9 (10 mg). The 30% EtOH fraction (21.0 g, IC50 ¼ 1.45 mg ml1) was chromatographed over MCI gel using acetone–water (1:1 to 7:3) mixtures of decreasing polarity. Repeated chromatography with the same eluent over MCI gel and Sephadex LH-20 afforded compounds 10 (54 mg), 11 (13 mg), 12 (45 mg), 13 (32 mg), 14 (11 mg), and 15 (9 mg). Compounds 1–15 were identified (as in Figure 1) by comparing their physical and spectra data with the literature values, and the data of compound 5, a first isolated triterpene from Anoectochilus genus were listed below. 2a,3b-dihydroxyolean-12-en-23, 28, 30-trioic acid (5): white amorphous powder. The molecular formula, C30H44O8, was determined on the basis of negative HRESIMS data of [M-1] 531.2971. 1H-NMR (500 MHz, CD3OD), 13C-NMR (125 MHz, CD3OD), and HMBC data are listed in Table 1.
The NMR spectra were recorded on a Bruker Avance 500 DRX (500 MHz for 1H and 125 MHz for 13C) spectrometer (Rheinstetten, Germany) in deuterated solvents [dimethyl sulfoxide (DMSO)-d6, methanol (MeOD), and chloroform (CDCl3)]. High-resolution electrospray ionization mass spectrometry (HRESIMS) was obtained on a Bruker Bio TOF IIIQ mass spectrometer (Bruker Daltonics, Inc., Billerica, MA). The IR spectrum (KBr) was recorded on a Shimadzu FTIR spectrometer (Tokyo, Japan). Column chromatography was conducted using silica gel 60 (100–200 mesh, Qingdao Haiyang Chemical Co. Ltd., Qingdao, China), Sephadex LH-20 (20–150 mm, GE, USA), LiChro prep RP-18 (40– 63 mm, Merck, Darmstadt, Germany), and Diaion HP20 (250–850 mm, Sigma). All TLC was conducted on precoated Merck Kiesel gel 60 F254 plates (20 20 cm, 0.25 mm, Merck, Darmstadt, Germany), using 5% H2SO4 as a spray reagent. Melting points were determined using an XT-4 point apparatus (uncorrected). Optical rotations were determined on a Perkin-Elmer Model 343 polarimeter (Merck, Darmstadt, Germany). Chemicals and reagents PTP1B Tyrosine Phosphatase Drug Discovery Kit was purchased from Enzo Life Sciences Inc. (Lausen, Switzerland). All other chemicals and solvents used were purchased from E. Merck (Darmstadt, Germany) and SigmaAldrich (St. Gallen, Switzerland), unless otherwise stated. Plant material The herbs of A. chapaensis were collected in the month of September of 2010 from Yunnan province, southwest China. The plant was authenticated by Hongyan Ma, an associate professor in School of Traditional Chinese Medicine, Guangdong Pharmaceutical University. A voucher specimen (2010-DY1001) has been deposited in the herbarium of the School of Pharmacy, Guangdong Pharmaceutical University. Biological assay for the inhibition of PTP1B Human protein tyrosine phosphatase 1B (hPTP1B) activity was assayed at room temperature using the PTP1B Tyrosine Phosphatase Drug Discovery Kit, a colorimetric assay designed for screening inhibitors and modulators of PTP1B activity. The kit included human recombinant PTP1B (residues 1–322; MW ¼ 37.4 kDa), expressed in Escherichia coli. The protocol outlined by the manufacturer was followed rigidly. This kit is designed to perform endpoint assays in which each well contains a 100 ml reaction in Assay Buffer and is terminated by the addition of 25 ml of the phosphate detection reagent, BIOMOL REDÔ. The 100 ml ‘‘reaction’’ may consist either of PTP1B phosphatase acting on the phosphopeptide substrate or simply a dilution of the free Phosphate Standard (BML-KI470). Tested compounds were demonstrated to be pure as evidenced by NMR and TLC analysis. IC50 of the compounds which showed inhibition of over 50% were determined.
Molecular docking simulation in PTP1B inhibition-Autodock VINA The docking simulation of the PTP1B enzyme and tested molecules were successfully performed using structure modeling (MODELER-9V10) and a docking program (Autodock VINA), which have been used to estimate the conformation of the protein–ligand complex (Trott & Olson, 2010). Since this Autodock VINA program significantly improves the average accuracy of the binding mode predictions compared with the AutoDock program, the accuracy, repeatability, and reliability of the docking results can be improved for drug discovery. For the docking studies, the crystal structures of the protein targets [Protein Data Bank (PDB ID: 1NNY for human PTP1B)] were allocated from the protein sequence alignment. The 3D structures of test compounds were constructed and minimized using Chemsketch 3.5 and Omega 2.0 software (Open Eye Scientific Software, Santa Fe, NM) for 2D and 3D conformations, respectively. The predicted protein–ligand complexes were optimized and ranked according to the empirical scoring function, TM score (Structural alignment tool, sheba3.1, Open Eye Scientific Software, Santa Fe, NM), which
Identification of biologically active chemicals
DOI: 10.3109/13880209.2014.957781
3
HOOC COOH
H COOH
HO HO
H
1= O 3 = OH 4 = H2
R
HO
O
OH
Pharmaceutical Biology Downloaded from informahealthcare.com by University of Bristol on 02/25/15 For personal use only.
OH
O
6 R1=OH R2=H 7 R1=OCH3 R2=H 8 R1=OCH3 R2=Glc 9 R1=OCH3 R2=Glc-Rha
C
11 R=OH 12 R=H
R
OR2
OH 10
COOH 5 O
HO
R1
OCH3
HO
2
H
C
C
H
C
O
COOH
HO
O O 13
O H 3C
CHO 14
OH O 15
Figure 1. Structures of compounds 1–15.
Table 1. NMR data for compound 5 ( in ppm, J in Hz). No.
H
C
1 2 3 4 5 6
1.98m 3.66m 3.73d(10)
48.4t 69.2d 81.1d 55.2s 52.7d
7 8 9 10 11 12 13 14 15
1.65m 1.98m 1.73m 1.63m 1.73m 1.65m 5.32m 1.73m
HMBC (H!C)
No.
H
C2, 3, 5, 10 C25 C2, 23, 24
16 17 18 19 20
1.55m
C6, 10
24.7t 33.6t 40.5s 49.4d 39.2s 24.3t 123.6d 145.0s 42.9s 28.9t
21 C5 C8, 10, 11, 25, 26 C9, 11, 14, 18
estimates the binding free energy of the ligand–receptor complex. Statistics All results were presented as mean ± SEM of triplicate samples.
Results and discussion Using the PTP1B bioassay as a guide, the chromatography of the EtOAc fraction afforded 15 known compounds (Figure 1), identified by comparison of their physical and spectral data with those of reported values as friedelin (1), sorghumol (2), epifriedelanol (3), friedelane (4), 2a, 3b-dihydroxyolean-12en-23, 28, 30-trioic acid (5), quercetin (6), isorhamnetin (7), isorhamnetin-3-O-b-D-glucoside (8), isorhamnetin-3-Ob-D-rutinoside (9), vanillic acid (10), caffeic acid (11), p-hydroxy-cinnamic acid (12), dibutyl-phthalate (13),
22 23 24 25 26 27 28 29 30
2.80m 1.94m,1.65m 1.98m 1.31m 1.65m 1.12s 1.04s 0.80s 1.20s 1.18s
C 21.8t 44.7s 44.0d 43.5t 47.1s 31.3t 35.1t 180.7s 12.9q 17.5q 17.6q 26.5q 181.5s 30.5q 180.7s
HMBC (H!C) C17 C12, 13, 14, 17, 19 C29 C17, 19, 20 C17 C3, C1, C7, C8,
4, 5, 23 5, 9, 10 8, 9, 14 13, 14, 15
C20, 21, 30
p-hydroxy benzaldehyde (14), and 7-hydroxy-2-methoxy9,10-dihydrophenanthrene-1,4-dione (15). Compounds 3, 4, 5, 13, and 15 were isolated from the genera Anoectochilus for the first time. All 15 compounds were evaluated for their potential to inhibit PTP1B activity. The results showed that 1–9 inhibited PTP1B activity with average IC50 values of 1.16–6.21 mM (Table 2). They were comparable with that of suramin (IC50 ¼ 11 ± 1 mM) (McCain et al., 2004), the positive control in the PTP1B Tyrosine Phosphatase Drug Discovery Kit. The Autodock docking program was employed to dock the compounds into the binding sites of the crystallographic ˚ from the structure of PTP1B, with all residues defined as 4 A ligand in the original complex. In particular, a semi-empirical free energy force field was employed to predict binding of protein–ligand complexes of a known structure and the binding energies for both the bound and unbound states (Morris et al., 2009). Recently, structure-based simulation of
4
J. Cai et al.
Pharm Biol, Early Online: 1–5
the PTP1B inhibitors has been performed for developing novel therapeutic drugs with selectivity and cell permeability (Barr, 2010). The main structural features of PTP1B have been well established and consist of 435 amino-acid residues, including residues 30–278, which comprise the
Pharmaceutical Biology Downloaded from informahealthcare.com by University of Bristol on 02/25/15 For personal use only.
Table 2. IC50 values of PTP1B inhibiting activity and binding energy of compounds 1–9.
Compounds
IC50 (mM)
Binding energy (kcal/mol)
Friedelin (1) Sorghumol (2) Epifriedelanol (3) Friedelane (4) 2a,3b-dihydroxyolean-12-en-23, 28, 30-trioic acid (5) Quercetin (6) Isorhamnetin (7) Isorhamnetin-3-O-b-D-glucoside (8) Isorhamnetin-3-O-b-D-rutinoside (9)
6.21 ± 0.02 3.50 ± 0.01 3.75 ± 0.14 4.60 ± 0.05 2.65 ± 0.03
8.1 7.4 8.3 8.5 8
5.63 ± 0.04 1.75 ± 0.02 1.16 ± 0.03 1.20 ± 0.05
7.5 7.4 7.5 7.8
catalytic domain containing the catalytic residue Cys215, the secondary phosphate-binding loop (P-loop) mediated by residues His214–Arg221, and the WPD loop identified as residues Thr177–Pro189 (Wang et al., 2009). The 3D docking simulations of compounds 1–9 with some structural difference of the pentacyclic triterpenes and flavonols were predicted using the Autodock program to evaluate the binding site-directed inhibition of PTP1B (Table 2 and Figure 2). A reported highly selective ligand of C40H37N3O10 (3-({5[(N-acetyl-3-{4-[(carboxycarbonyl)(2-carboxy phenyl)amino]1-naphthyl}-L-alanyl)amino]pentyl}oxy)-2-naphthoic acid) was simultaneously employed for the positive control. X-ray data confirmed that the inhibitor bound with the catalytic site in the native, ‘‘open’’ conformation. It displayed excellent potency and good selectivity over many other phosphatases (Szczepankiewicz et al., 2003). It displayed a negative binding energy of 7.8 kcal/mol. The tested compounds 1–9 showed negative binding energies of 7.4 to 8.5 kcal/ mol (Table 2), comparable with that of the positive reference. The results support and provide evidence of the
Figure 2. 3D molecular docking models for PTP1B protein’s active pocket site of selected potent candidates: (A) isorhamnetin-3-O-b-D-rutinoside (9), (B) friedelane (4), (C) 3b-dihydroxyolean-2-en-23,28,30-trioic acid (5), and (D) epifriedelanol (3).
Identification of biologically active chemicals
Pharmaceutical Biology Downloaded from informahealthcare.com by University of Bristol on 02/25/15 For personal use only.
DOI: 10.3109/13880209.2014.957781
potent significant inhibition results in vitro. Friedelane (4) (8.5 kcal/mol) and isorhamnetin-3-O-b-D-rutinoside (9) (7.8 kcal/mol) showed the best affinity to PTP1B’s pocket site among the tested triterpenoids and flavonoids. Considering docking results of the PTP1B–compounds complex, the tested inhibitors were stably posed in similar pocket domains of PTP1B residues, including Trp179 and Gly183 in the WPD loop; Asp48, Ser216, Ala217, Gly220, ˚ Arg221, and Gln266 in the pocket site, all of which were 4 A from the inhibitor in the original complex. In particular, the WPD loop plays a much more important role in the specificity and affinity of the inhibitors. In the presence of the ‘‘open’’ conformation of the WPD-loop, the binding pocket of PTP1B is easily accessible to the substrate. After substrate binding, the WPD loop closes over the active site, forming a tight binding pocket for the substrate. The WPD loop closes onto the substrate and thereby positions the thiolate of Cys215 for nucleophilic attack of the phosphotyrosine. It has been observed that the conformational and dynamic features of WPD-loop play a key role in providing a smooth entrance for the inhibitors moving into the binding pocket as well as a favorable microenvironment to stabilize them (Wang et al., 2009). Potent, yet highly selective, PTP1B inhibitory agents can be acquired by targeting the area defined by residues Lys41, Arg47, and Asp48 (Sun et al., 2003). Moreover, the binding energies of 3, 4, 5, and 9 were 8.3, 8.5, 8, and 7.8 kcal/mol, respectively, indicating high affinity to PTP1B residues. Detailed binding mode analysis with docking simulation showed that the inhibitors can be stabilized by the simultaneous establishment of multiple hydrogen bonds and van der Waals contacts in the pocket site (Figure 2). Structure–activity relationship study of these sugar-substituted flavonoid derivatives demonstrated that the PTP1B inhibitory activity was strongly influenced by the carbohydrate moiety at the C-3 position. The rutinosyl substituent at the C-3 position of isorhamnetin skeletons (9) slightly increases the proximity to the Trp179 and Gly183 residues in the WPD loop, resulting in a relatively small difference with 7 and 8 in the binding energy and affinity. They also interacted with Ser216, Ala217, Gly220, and Arg221 in the P loop, indicating that the inhibitors may reduce the mobility of the WPD loop toward a more rigid conformation, which inhibits WPD loop closure and prevents substrate binding (Popov, 2011). Moreover, the triterpene (5) with polar groups may have advantages with regard to the occupation of the anionic charged P loop at a physiologic pH (Park et al., 2009). In addition, the carboxylic group of oleanolic acid (5) at C-23 and hydroxyl groups at C-2 and C-3 generated extensive hydrogen bonds with Ser216, Ala217, and Gly220 in the P-loop, which significantly contributed to its high affinity. Polar moieties substituted triterpenes (5) and flavonoids (7–9) isolated from A. chapaensis showed negative binding energies and a high proximity to PTP1B residues, including Trp179 and Gly183 in the WPD loop; Asp48, Ser216, Ala217, Gly220, Arg221, and Gln266 in the pocket site. These findings may lead to the design of polar moieties substituted triterpenes or flavonoids as PTP1B inhibitors that interact with both catalytic and allosteric sites to achieve selectivity. Furthermore, the triterpenoid backbone of
5
friedelane (4) interacted with non-polar residues based on the van der Waals contacts, and it may have a high affinity with the pocket site through hydrophobic interaction. In conclusion, nine potential inhibitors of PTP1B were identified by experiments and the docking simulation, considering that carboxylic and hydroxyl groups to form the above-mentioned broad arrangement of hydrogen bonds. Both enzymatic test and docking results demonstrate that the contributions of functional groups with a polar moiety may cause high affinity to PTP1B and a lower binding energy. Therefore, polar moieties substituted triterpenes or flavonoids as PTP1B inhibitors hold promise as therapeutic agents for the treatment of diabetes and related disorders.
Declaration of interest The authors declare that they have no conflict of interest. The authors thank National Natural Science Foundation of China (No. 81001628) and Guangdong Natural Science Foundation (No. S2013010014771) for financial support.
References Anonymous. (1999). Editorial Committee of the Flore of China of Chinese Academy of Science. Flora Republicae Popularis Sinicae, vol. 17. Peking: Science Press. Barr AJ. (2010). Protein tyrosine phosphatases as drug targets: Strategies and challenges of inhibitor development. Future Med Chem 2: 1563–76. Elchebly M, Payette P, Michaliszyn E, et al. (1999). Increased insulin sensitivity and obesity resistance in mice lacking the protein tyrosine phosphatase-1B gene. Science 283:1544–8. Feldhammer M, Uetani N, Miranda-Saavedra D, Tremblay ML. (2013). PTP1B: A simple enzyme for a complex world. Crit Rev Biochem Mol Biol 5:430–45. Klaman LD, Boss O, Peroni OD, et al. (2000). Increased energy expenditure, decreased adiposity, and tissue-specific insulin sensitivity in protein-tyrosine phosphatase 1B-deficient mice. Mol Cell Biol 20:5479–89. McCain DF, Wu L, Nickel P, et al. (2004). Suramin derivatives as inhibitors and activators of protein-tyrosine phosphatases. J Biol Chem 279:14713–25. Morris GM, Huey R, Lindstrom W, et al. (2009). AutoDock 4 and AutoDock-Tools 4: Automated docking with selective receptor flexibility. J Comput Chem 30:2785–91. Panzhinskiy E, Ren J, Nair S. (2013). Pharmacological inhibition of protein tyrosine phosphatase 1B: A promising strategy for the treatment of obesity and type 2 diabetes mellitus. Curr Med Chem 21:2609–25. Park H, Bhattarai BR, Ham SW, Cho H. (2009). Structure based virtual screening approach to identify novel classes of PTP1B inhibitors. Eur J Med Chem 44:3280–4. Popov D. (2011). Novel protein tyrosine phosphatase 1B inhibitors: Interaction requirements for improved intracellular efficacy in type 2 diabetes mellitus and obesity control. Biochem Biophys Res Co 410: 377–81. Sun JP, Fedorov AA, Lee SY, et al. (2003). Crystal structure of PTP1B complexed with a potent and selective bidentate inhibitor. J Biol Chem 14:12406–14. Szczepankiewicz BG, Liu G, Hajduk PJ, et al. (2003). Discovery of a potent, selective protein tyrosine phosphatase 1B inhibitor using a linked-fragment strategy. J Am Chem Soc 125:4087–96. Trott O, Olson AJ. (2010). AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization and multithreading. J Comput Chem 31:455–61. Wang JF, Gong K, Wei DQ, et al. (2009). Molecular dynamics studies on the interactions of PTP1B with inhibitors: From the first phosphatebinding site to the second one. Protein Eng Des Sel 6:349–55. Zhang S, Zhang ZY. (2007). PTP1B as a drug target: Recent developments in PTP1B inhibitor discovery. Drug Discov Today 10: 373–81.
