Discovery of new muscarinic acetylcholine receptor antagonists from Scopolia tangutica

Scopolia tangutica (S. tangutica) is a traditional Chinese medicinal plant used for antispasmodics, anesthesia, analgesia and sedation. Its pharmacological activities are mostly associated with the antagonistic activity at muscarinic acetylcholine receptors (mAchRs) of several known alkaloids such as atropine and scopolamine. With our recent identification of four hydroxycinnamic acid amides from S. tangutica, we hypothesized that this plant may contain previously unidentified alkaloids that may also contribute to its in vivo effect. Herein, we used a bioassay-guided multi-dimension separation strategy to discover novel mAchR antagonists from S. tangutica. The core of this approach is to use label-free cell phenotypic assay to first identify active fractions, and then to guide purification of active ligands. Besides four tropanes and six cinnamic acid amides that have been previously isolated from S. tangutica, we recently identified two new tropanes, one new cinnamic acid amide, and nine other compounds. Six tropane compounds purified from S. tangutica for the first time were confirmed to be competitive antagonists of muscarinic receptor 3 (M3), including the two new ones 8 and 12 with IC50 values of 1.97 μM and 4.47 μM, respectively. Furthermore, the cinnamic acid amide 17 displayed 15-fold selectivity for M1 over M3 receptors. These findings will be useful in designing lead compounds for mAchRs and elucidating mechanisms of action of S. tangutica.

cell phenotypic assay afforded by resonant waveguide grating (RWG) biosensor to first identify active fractions, and then to guide the purification of active compounds. Surface bound evanescent waves and tunable light source provided by the label-free screening device, RWG biochemical assay characterizes the process of dynamic mass redistribution (DMR) caused by probes interaction through refractive index variations 23 . The 384-well biosensor assay permits a holistic, pathway sensitive readout of receptor pharmacology with high throughput [24][25][26] . The noninvasive and holistic measurement of the label-free technique enables multiple assay formats to identify and elucidate the pharmacology of hit ligands or multiple targets all within a single screening campaign, especially for GPCRs 27,28 .
Herein, we applied the label-free cell phenotypic assay-guided preparation strategy to discover minor active alkaloids from S. tangutica. This workflow assisted us to isolate a series of pure compounds and identify their antagonistic activities on the endogenous M3 receptor in HT-29 cells 29 .

Identification of active fractions of alkaloid constituents from S. tangutica.
To perform the activity-guided purification, the alkaloids enriched from S. tangutica using the SPE method 19 were the first subject to separation on an XCharge C18 column. Results showed that the enriched alkaloids gave rises to a series of well separated and symmetric peaks even at an overloading amount on the column (Fig. 1a). Twenty-three fractions (F1 to F23) were collected sequentially according to visible peaks and these fractions have little peak overlapping (Fig. S1).
Given that S. tangutica is used to treat spasm and asthma, we screened these fractions on M3 receptor in HT-29 due to its high expression of M3 receptor endogenously and robust DMR signals after treatment with agonist 29 . The screening was performed via a two-step assay, of which the first step was to examine the agonistic activity of each fraction, and the second step to examine the ability of each fraction to block the DMR signal arising from the activation of M3. For instance, F8 triggers little DMR signal in HT-29 cells, similar to the control signals (Fig. 1b). However, the fraction almost completely blocks the DMR of 16 μ M acetylcholine, a non-selective agonist for muscarinic receptors (Fig. 1c), suggesting that F8 contains at least one M3 antagonist.
To illustrate the effect of all fractions in both cell lines, we produced a heat map of all fractions based on cluster analysis of all DMR responses obtained (Fig. 1d). Results show that F8 to F17 induce no clear DMR signals in HT-29, but have obvious inhibitory effects on the acetylcholine DMR, while F5, F6, F7 and F18 show partial inhibition. Histamine receptor (H receptor), another receptor also related to asthma, was also tested and A549 cell line was preferred for its endogenous expression 30 of H receptor, based on the fast proliferation and well adhering property of this cell line. As a result, nearly all fractions have little effect on the histamine DMR in A549. It suggests that these fractions F5 to F18 may contain M3 antagonists. Thus, these active fractions were chosen to purify compounds for investigating new M3 antagonists. For each fraction, real responses of both the fraction and the probe after the fraction pretreatment, each at six discrete time points post-stimulation (3, 6, 9, 15, 30, 45 min), were used for the cluster analysis. All fractions were assayed at 1.25 mg/L. The probe was acetylcholine (Ach) for M3 receptor in HT-29, and histamine (His) for histamine receptors in A549. The control was buffer. Color code is green, negative; red, positive; and black, zero response.
Scientific RepoRts | 7:46067 | DOI: 10.1038/srep46067 Purification of alkaloid compounds from active fractions of S. tangutica. We employed multi-dimension HPLC to purify and identify active compound(s) from each active fraction. First, XCharge SCX (SCX) column was used as the second dimensional liquid chromatography, due to the difference in separation selectivity from the first dimension. For instance, compared to the XCharge C18 column in the first dimension separation (Fig. 2a), F8 exhibited different chromatography behavior on the SCX column, leading to further fine separation of this fraction (Fig. 2b). Thus, we performed systematical preparation of all active fractions (F5 to F18) and obtained 111 secondary fractions in total using the SCX column. We further performed purity analyses of all these secondary fractions on both XCharge C18 and SCX columns. Results showed that out of the 111 fractions, only five were pure.
Second, we performed the third dimensional separation to obtain pure compounds from the fractions. An XCharge C18 column in a different mobile phase system from the first dimensional separation was used to realize further purification and desalting process synchronously. In the third dimension, peaks were systematically collected, including minor peaks like F8-2-P4, the fourth peak of F8-2 (Fig. 2c), the purity of which was confirmed by MS (Fig. 2d). As a result, we obtained 255 final fractions in total. Purity analyses show that 136 fractions have purity higher than 90%. For the remaining 119 fractions, their insufficient quantities make it difficult to further purify compounds of high structural similarity, including homologues and isomers, such as F16-7-P6 (Fig. S3) and F16-7-P7 (Fig. S4). Furthermore, among the pure fractions, duplications between adjacent fractions result in repeated isolation of the same compound, such as the compound norhyoscyamine that was obtained as a secondary fraction F12-8 and a final fraction F12-10-P1.
Third, MS and Nuclear Magnetic Resonance (NMR) were used to determine chemical structures of active compounds. Two new tropane compounds and one cinnamic acid amide were structurally characterized in detail and known compounds were elucidated in Supporting information.  Fig. 2f shown. This compound was characterized as N1-p-dihydrocoumaroyl-N10-dihydrocaffeoyl spermidine. The specific assignment of the three new compounds was supplied in Supporting information.
The twelve tropane ligands isolated from S. tangutica are atropine analogs and share high similarity in structure, permitting structure-activity relationship analysis. Among these compounds, the potency of 1 is comparable as that of 3, indicating that the introduction of an epoxide ring at R 1 and R 2 does not damage the antagonistic potency of the backbone. Centered with 1, bonding hydroxyl at R 5 (4) decreases its potency by about 8 folds. The introduction of hydroxyl at R 5 and removal of methyl at R 3 (7) reduce the activity by approximately 25 folds. Replacing hydroxyl at R 4 with glucose group (8) exhibits about 50-fold decrease in inhibitory activity. Centered with 3, the introduction of hydroxyl at R 5 (10) and removal of methyl at R 3 (11) slightly reduce activity, while addition of hydroxyl at R 1 (2) greatly reduce activity to partial inhibition level. Placing hydroxyl at both R 1 and R 5 is 6, which is nearly inactive on M3 receptor. Overall, the modifications of R 1 and R 2 are vital to regulate the activity of hyoscyamine (3). Introduction of the epoxide ring at R 1 and R 2 (1) have little effect, but presence of hydroxyl at R 1 or R 2 (2) or both (5) damage inhibitory activity tremendously. Removal of methyl at R 3 (3 vs 11 and 4 vs 7) reduced activity fewer than 5 folds. This structure-activity relationship will provide guidance for the design of lead compounds for M3 receptor.
Pharmacological characterization of cinnamic acid amides. We examined the pharmacology of eight cinnamic acid amides on M3 receptor in HT-29 using the two-step DMR assay. As a result, among these cinnamic acid amides, only 17 is active to antagonize M3 receptor, as it stimulates little response in HT-29 cells and dose-dependently inhibited the acetylcholine DMR ( Fig. 4a and Fig. 4b). The two-dimensional purification and mass spectrum of this active compound were shown in Fig. S5. Considering that 17 is a novel chemical ligand displaying M3 antagonistic activity, we examined its selectivity over M1 receptor. Result showed that 17 also dose-dependently inhibited the acetylcholine DMR (Fig. 4c) in the M1-transfected CHO cells. The dose-dependent inhibition on M3 and M1 were presented in Fig. 4d and IC 50 values obtained were 15.57 ± 5.09 μ M Competitive antagonism of alkaloid compounds. We applied a co-stimulation DMR assay to determine whether active alkaloid compounds are competitive antagonists or not. Here, acetylcholine at a series of concentrations were prepared in the presence of compound 2, 3, and 11, each at a fixed dose, and then used to co-stimulate HT-29 cells. The three compounds display diverse pharmacological actions. The control without existence of compound is in the left and both 3 and 11 dose-dependently shift the acetylcholine dose curve to right ( Fig. 5a and 5b), suggesting that both act as a competitive antagonist for M3 receptor. In contrast, compound 2 (anisodamine) exhibited complicated effect; at a low dose (10 nM) it increases the potency of acetylcholine, but at a high dose (1 μ M) it decreases the potency of acetylcholine (Fig. 5c). Further investigation of the biochemical mechanism of these compounds will be of great interest.

Discussion
What is critical to isolate new bioactive alkaloids from traditional Chinese medicinal plants is to have a high-resolution separation system and an effective activity-guided protocol. The separation system requires high separation selectivity, resolution and symmetric peak shape so that alkaloid products can be pure enough for structure determination and activity identification. This is particularly essential for isolation and identification of minor active alkaloids that have great potential to be new ligands, compared to abundant constitutes. S. tangutica is a promising TCM to discover novel bioactive drugs from its alkaloid constituents in large-scale multi-dimension preparation. We here applied a large-scale multi-dimensional preparation approach to systematically isolate compounds from S. tangutica. Considering the diversity of chemical constituents in a medicinal plant, bioassay-guided separation is a good strategy to improve efficiency and precision in the discovery of active ligands. We here employed an unbiased and label-free cell phenotypic profiling approach to guide the isolation of minor active compounds from S. tangutica.
As a result, we have isolated for the first time five tropane alkaloids 7, 8, 9, 10 and 12, and demonstrated their antagonistic activity on M3 receptor. Among them, two glucuronide conjugated scopolamine analogs (8 and 12) are similar to metabolites of scopolamine and anisodine, and 7 is one of the metabolites of scopolamine 31,50 , implicating that the metabolites of muscarinic receptor antagonists may also contribute to their in vivo effects. Of note, there are a great number of other minor or trace compounds remaining as a mixture of isomers or homologues and in low quantity in S. tangutica. Further study of them is important to discover new active ones. These results presents the power of the label-free cell phenotypic profiling-guided preparation protocol to isolate and XCharge C18 and XCharge SCX columns were purchased from Acchrom Co. (Dalian, China). Na 2 SO 4 and NaH 2 PO 4 were from Sinopharm Chemical Reagent Co. (Beijing, China). Formic acid (FA) and phosphoric acid were bought from J&K Chemical Co. (Shanghai, China). Acetonitrile (ACN) and ethanol were purchased from Fulltime corporation in Anhui province. The mass spectrum information of purified compounds was obtained using Agilent 1290 Infinity UPLC and 6540 UHD Q-TOF and Mass Hunter software was used to process mass spectra data. The structures of compounds were identified with Bruker 400 MHz NMR spectrometer and Bruker 500 MHz NMR spectrometer. Optical rotation values were tested using Jasco P-1020 polarimeter (Japan); infrared spectrums (IR) were obtained by Bruker Tensor27 FTIR spectrometer; ultraviolet (UV) spectrums were acquired with Shimadzu UV-2401PC spectrometer (Japan).
Dynamic mass redistribution (DMR) assays. All DMR assays were performed using an Epic ® BT system. HT-29 cells and A549 cells were seeded in Epic ® 384-well biosensor microplate (Corning) with a density of 32000 cells per well and 15000 cells per well, respectively. Then the microplate was cultured for 22 hrs or 15 hrs in the corresponding cell medium to form a confluent monolayer for HT-29 or CHO-M1 cells, respectively. After being washed, the cells were maintained with assay buffer and incubated for 1 hr.
For profiling of the fractions, a 2-min baseline was first established, followed by adding twenty three fractions at 1.25 mg/L and recording the fraction-induced DMR signals for 1 hr. Then, after 2-min baseline was re-established, acetylcholine at 16 μ M and histamine at 8 μ M were added for HT-29 and A549 cell line, respectively. The DMR responses were recorded for another 1 hr. For the IC 50 determination of pure compounds, a 2-min baseline was first established. Then compounds at varied doses were added manually and the DMR response was recorded for 1 h. Afterwards a 2-min baseline was re-established, and acetylcholine at 16 μ M and acetylcholine at 4 μ M were added to HT-29 and CHO-M1, respectively. DMR response was monitored for another 1 h.
For co-stimulation assays, a 2-min baseline was first established. Acetylcholine dose series in the absence and presence of a given ligands at fixed doses were used to stimulate the HT-29 cells.
Data analysis. Data process and analysis were performed on Microsoft excel 2010 and GraphPad Prism 6.02 (GraphPad Software Inc., San Diego, CA, USA). All IC 50 values reported were shown as mean ± standard deviation in duplicate from two dependent experiments (n = 4). The heat map was completed by Cluster 3.0 and Treeview after processing in Microsoft excel 2010.