Clerodane Diterpenoids with Anti-hyperglycemic Activity from Tinospora crispa

Abstract Four new clerodane diterpenoids, tinosporols A–C (2–4) and tinosporoside A (5), together with six known analogues were isolated from the vines of Tinospora crispa. Their structures were established by extensive spectroscopic analysis. The relative configuration at C-12 in the known diterpenoid borapetoside E (1), the major component of the plant, was firstly established with the aid of molecular model. Compound 1 significantly reduced serum glucose levels at dose-dependent manners in alloxan-induced hyperglycemic mice and db/db type 2 diabetic mice. Graphical Abstract Electronic supplementary material The online version of this article (doi:10.1007/s13659-016-0109-3) contains supplementary material, which is available to authorized users.

. This plant is a prolific source of anti-hyperglycemic clerodane diterpenoids, among which borapetoside C was found to be an effective agent for the treatment of type 2 diabetes mellitus (T2DM) [2,3].
T2DM is a chronic metabolic disorder characterized by deregulation of glucose and lipid metabolism [4]. Currently, approximately 150 million people are suffered with diabetes worldwide, and this population could increase up to 300 million by 2025. The estimated prevalence of diabetes among a representative sample of Chinese adults was 11.6 % [5]. Numerous drugs, such as rosiglitazone (ROS) and metformin (MET), have been used in the treatment of T2DM. However, treatment with synthetic drugs has been reported to lead to various side effects [6]. Therefore, searching for better agents from herbs or natural products is necessary [7]. Traditional Chinese medicines (TCMs), which have been used by the Chinese to treat illnesses for thousands of years, are combination drugs comprising several different active compounds. TCMs are better at controlling complex disease systems such as diabetes and are less prone to causing drug resistance development [8]. As part of a BioBioPha [http://www.chemlib.cn] objective to assemble a large-scale natural products library valuable in the discovery of new drug leads from TCMs [9][10][11], phytochemical investigation on the vines of T. crispa led to the isolation of four new clerodane diterpenoids, tinosporols A-C (2-4) and tinosporoside A (5), together with six known analogues. The following describes the isolation and structural elucidation of compounds 1-5, and the antihyperglycemic activity of borapetoside E.
Compound 1 has spectral data (MS, 1 H and 13 C NMR) identical to those of the known compound borapetoside E [12], however, the stereochemistry at C-12 in this molecule was still undetermined. In the 1 H NMR spectrum (Table 1), two J values ( 3 J 12,11pro-R* = 10.0 Hz; 3 J 12,11pro-S* & 0 Hz) were contrasted with the coupling constants usually found in freely rotating systems (generally 5-8 Hz), indicative of that C(11)-C(12) bond bearing two large groups (butenolide and naphthane) can not rotate freely. In a Newman projection (Fig. 1), the trans relationship of H-11pro-R* and H-12 was implied by a large coupling constant ( 3 J 12,11pro-R* = 10.0 Hz), while the dihedral angle of H-11pro-S*/C-11/ C-12/H-12 was supposed to be around 90°, as indicated by the 3 J 12,11pro-S* value (&0 Hz). This conclusion was further supported by strong correlation of H-12$H-11pro-S* and no correlation of H-12$H-11pro-R*. In the MM2-optimized stereoview of 1 (Fig. 1)  Clerodane Diterpenoids with Anti-hyperglycemic Activity from Tinospora crispa 249 R*, and H 3 -20 are on the other side. Accordingly, the relative configuration at C-12 was established as S*. In a previous paper [7], it was reported that enzymatic hydrolysis of borapetoside D (6) yielded 1, the relative configuration at C-12 in 6 was therefore assigned as S*. Compound 2 was obtained as an inseparable C-16 epimeric mixture (3:1) where some of the signals appeared as duplicate in the NMR spectra (Tables 1, 2). Its molecular formula was determined as C 21 H 26 O 8 from HREIMS at m/ z 406.1623 [M] ? (calcd 406.1628), requiring 9 degrees of unsaturation. The IR spectrum suggested the presence of hydroxy (3411 cm -1 ) and carbonyl (1710 cm -1 ) functionalities. Comparison of its 13 C NMR data with that of rumphioside F (10) [7] revealed a remarkable resemblance except for the absence of a set of glucopyranosyl resonances and the signal assigned to C-12 being upfield (D = -6.1 ppm), which suggested that 2 was an aglycone of rumphioside F. Heteronuclear multiple bond connectivity (HMBC) correlations ( Fig. 1) from 12-OH to C-11/ C-12 and from H-12 to C-9/C-11/C-13/C-14 further verified the location of this hydroxy group. The configurations at C-5, C-6, C-8, and C-9 were consistent with those of rumphioside F (10) according to a careful analysis of the ROESY spectrum. Furthermore, the stereochemistry for C-12 was established by analysis of the proton coupling constants and ROESY data (Fig. 2). The anti relationship of H-11pro-R* and H-12 and the dihedral angle of H-11pro-S*/C-11/C-12/H-12 (ca.90°) is reasonable by two coupling constants ( 3 J 12,11pro-R* = 8.4/9.1 Hz; 3 J 12,11pro-S* & 0 Hz) recorded in CDCl 3 (Table 1). In the ROESY spectrum (Fig. 1)  . This group was attached at C-8 (d C 78.1) due to the chemical shift of C-8 in 3 showing a downfield shift (D = 4.8 ppm) relative to that of the corresponding hydroxyl-bearing carbon in tinocrispol A, which was also confirmed by the lack of esterification shift [2] of H-2 in 3 and weak HMBC correlation ( 4 J) from the acetyl proton (d H 2.12) to C-8. The acetyl group is a-oriented as indicated by a clear ROESY correlation between    (Tables 3,  4) displayed signals very similar to those of 7 [2]. The principal difference between them was the location of the glucopyranosyl moiety. This moiety was linked to C-2 as indicated by HMBC correlations from the anomeric proton (d H 4.44) to C-2 (d C 72.6). The relative configuration of 5 was determined to be the same as that of 7, based on detailed analysis of ROESY spectrum and coupling constants. Thus, the structure of 2 were established and given the trivial name tinosporoside A.
Based on the usage of T. crispa in TMCs and corrected structure characteristics of the compound 1, then the antihyperglycemia activities of compound 1 on hyperglycemic mice in vivo were evaluated. Following intravenously administrated compound 1 in alloxan-induced hyperglycemic mice which is a fast and classic type 1 diabetes mellitus (T1DM) model in pharmacological study [16]. In our study, alloxan injections markedly increased the serum glucose level from 7.0 ± 0.5 mM/L to 24.3 ± 3.7 mM/L (Fig. 3). Under this condition, compound 1 (dose of 20 mg/ kg) was able to induce a significant decrease of blood glucose down to 21.7 ± 6.7 mM/L after 5 times administration (25.9 ± 2.5 mM/L before compound treatment), whereas dose of 40 mg/kg of the compound caused a further down of glucose level to 15.9 ± 9.2 mM/L, indicating a tendency of dose-dependent effect of the compound on lowing blood glucose in this animal model. Furthermore, regarding that it is a valuable approach to extensively evaluate the drug effect on animals by evaluating its effect after drug withdrawing. Next we measured blood glucose levels on 3 or 5 days later after withdraw of compound 1. Surprisingly, the blood glucose level was even further down to 17.9 ± 3.8 mM/L after 3-days withdrawing for the group of 20 mg/kg treatment (compared to 21.7 ± 6.7 mM/L before drug withdrawing, Fig. 3), whereas glucose level after 3-days withdrawing in group of 40 mg/kg treatment maintained at 16.7 ± 4.7 mM/L which was similar to the level before drug withdraw (15.9 ± 9.2 mM/L, Fig. 3). Explaining for these results could be the maxim of glucose being recovered by the drug could be 16 mM/ L or so, so that 40 mg/kg treatment already achieved the maxim level in this approach. After withdrawing drug for 5 days, both glucose levels for groups of 20 or 40 mg/kg treatment fully recovered to the levels before drug treatments.
Furthermore, we asked that whether the effect of compound 1 on lowing blood glucose is specific for alloxaninduced hyperglycemic mice. Therefore we estimated the effect of the compound on db/db mice that is genetic animal model of which has been widely used in T2DM study. Compound 1 administration for 5 times slightly decreased glucose levels to 17.85 ± 2.83 mM/L of these mice at the dosage of 40 mg/kg (p = 0.041, compared to 21.93 ± 4.10 mM/L of model mice), whereas a lower dose (20 mg/ kg) didn't affect the glucose level (19.49 ± 3.91 mM/L, Fig. 4). Comparing with the strikingly decreased glucose (8.89 ± 2.80 mM/L) with MET administration, we concluded that T1DM mice are more sensitive than T2DM mice to compound 1.

General Experimental Procedures
Optical rotations were measured on a Jasco P-1020 automatic digital polarimeter. UV spectra were obtained in an HPLC (Agilent 1200, DAD). IR spectra were obtained using a Bruker Tensor 27 FT-IR spectrometer with KBr pellets. NMR spectra were acquired with a Bruker Avance

Plant Material
The vines of T. crispa were collected in April 2012 in Simao, Yunnan Province, China. The sample was identified by Mr. Yu Chen of Kunming Institute of Botany, Chinese Academy of Sciences. The voucher specimen (No. BBP20120452) was deposited at BioBioPha Co., Ltd.
Fraction F (90 g) was separated by silica gel CC (CHCl 3 /MeOH, 5:1?3:1) into fractions F1 and F2. Compound 6 (162 mg) was obtained from F2 by MCI CC (MeOH/H 2 O, 5?25 %).    These animals were housed in a temperature and humiditycontrolled room with a 12-hour light/dark cycle. Animals were given ad libitum access to food and water throughout the study. All of the procedures were performed in accordance with the Institute Ethical Committee for Experimental Animal Use. For inducing hyperglycemic mice, 8-week-old ICR mice were intravenously injection with freshly prepared alloxan (60 mg/kg). Serum glucose level was measured with glucose kit (Changsha Sinocare Inc., China). The mice with plasma glucose level greater than 11.1 mM/L were considered diabetic. Animals were treated with the compound 1 or MET twice daily for 2.5 days at the doses 20, 40 or 200 mg/kg as indicated in the figures.

Statistical Analysis
The data were presented as mean ± SEM, subjected to one-way analysis of variance followed by Dunnett's multiple comparison tests using GraphPad Prism version 5.0 software. *P \ 0.05, **P \ 0.01.