Calvin Yu-Chian Chen

TCM Database@Taiwan: The World's Largest Traditional Chinese Medicine Database for Drug Screening In Silico

Rapid advancing computational technologies have greatly speeded up the development of computer-aided drug design (CADD). Recently, pharmaceutical companies have increasingly shifted their attentions toward traditional Chinese medicine (TCM) for novel lead compounds. Despite the growing number of studies on TCM, there is no free 3D small molecular structure database of TCM available for virtual screening or molecular simulation. To address this shortcoming, we have constructed TCM Database@Taiwan (http://tcm.cmu.edu.tw/) based on information collected from Chinese medical texts and scientific publications. TCM Database@Taiwan is currently the worlds largest non-commercial TCM database. This web-based database contains more than 20,000 pure compounds isolated from 453 TCM ingredients. Both cdx (2D) and Tripos mol2 (3D) formats of each pure compound in the database are available for download and virtual screening. The TCM database includes both simple and advanced web-based query options that can specify search clauses, such as molecular properties, substructures, TCM ingredients, and TCM classification, based on intended drug actions. The TCM database can be easily accessed by all researchers conducting CADD. Over the last eight years, numerous volunteers have devoted their time to analyze TCM ingredients from Chinese medical texts as well as to construct structure files for each isolated compound. We believe that TCM Database@Taiwan will be a milestone on the path towards modernizing traditional Chinese medicine.

Ligand-Based Dual Target Drug Design for H1N1: Swine Flu- A Preliminary First Study

Abstract In March and April, 2009, an outbreak of H1N1 influenza in Mexico had led to hundreds of confirmed cases and the death toll had risen to 160. The worldwide spread of H1N1 has been attracting global attention and arising an overwhelming fear. So far, the vaccine and remedy has been in urgent need. In this study, a QSAR model and pharmacophore map of neuramini- dase (NA) type 1 (N1) contained two hydrogen bond acceptor features, one hydrogen bond donor feature, and one positive ionizable feature. NCI database was employed in virtual screen by the N1 pharmacophore map features. After screening, compounds were obtained and then docked into haemagglutinin type 1 (H1) to find out the candidate drugs for dual target of both N1 and H1. The candidate, NCI0353858, selected via virtual screening and docking, might be functional to this worldwide disease; consequently, further clinical investigations and scientific application are urgently demanded. We realize the proposed ligand does not have much validity without conducting a study on the stability of the protein-ligand complex by MD simulations and binding free energy, and such a study is underway and will be reported later in this journal. Nevertheless, the present study is clear, consistent and could give a rational explanation for the binding mode of the best selected ligand.

Identification of Potent EGFR Inhibitors from TCM Database@Taiwan

Overexpression of epidermal growth factor receptor (EGFR) has been associated with cancer. Targeted inhibition of the EGFR pathway has been shown to limit proliferation of cancerous cells. Hence, we employed Traditional Chinese Medicine Database (TCM Database@Taiwan) (http://tcm.cmu.edu.tw) to identify potential EGFR inhibitor. Multiple Linear Regression (MLR), Support Vector Machine (SVM), Comparative Molecular Field Analysis (CoMFA), and Comparative Molecular Similarities Indices Analysis (CoMSIA) models were generated using a training set of EGFR ligands of known inhibitory activities. The top four TCM candidates based on DockScore were 2-O-caffeoyl tartaric acid, Emitine, Rosmaricine, and 2-O-feruloyl tartaric acid, and all had higher binding affinities than the control Iressa®. The TCM candidates had interactions with Asp855, Lys716, and Lys728, all which are residues of the protein kinase binding site. Validated MLR (r2 = 0.7858) and SVM (r2 = 0.8754) models predicted good bioactivity for the TCM candidates. In addition, the TCM candidates contoured well to the 3D-Quantitative Structure-Activity Relationship (3D-QSAR) map derived from the CoMFA (q2 = 0.721, r2 = 0.986) and CoMSIA (q2 = 0.662, r2 = 0.988) models. The steric field, hydrophobic field, and H-bond of the 3D-QSAR map were well matched by each TCM candidate. Molecular docking indicated that all TCM candidates formed H-bonds within the EGFR protein kinase domain. Based on the different structures, H-bonds were formed at either Asp855 or Lys716/Lys728. The compounds remained stable throughout molecular dynamics (MD) simulation. Based on the results of this study, 2-O-caffeoyl tartaric acid, Emitine, Rosmaricine, and 2-O-feruloyl tartaric acid are suggested to be potential EGFR inhibitors.

Structure-Based and Ligand-Based Drug Design for HER 2 Receptor

Abstract Human epidermal growth factor receptor 2, HER2, is a commonly over-expressed tyrosine kinase receptor found in many types of carcinoma. Despite that there are several HER2 inhibitors, namely Iressa, Tarceva and Tykerb, currently in clinical trials, all can cause several side effects. In this study, both structure-based and ligand-based drug design were employed to design novel HER2 inhibitors from traditional Chinese medicine (TCM). The HER2 structure model was built in homology modeling based on known receptors of the same family. Docking and de novo evolution experiments were performed to identify candidates and to build derivatives. A training set of 32 compounds with inhibitory activities to HER2 was used to formulate the pharmacophore hypotheses that were subsequently used to examine candidates obtained from the docking study. Hydrogen bond interactions, salt-bridge formations and pi-stacking were observed between the ligands and Phe731, Lys753, Asp863 and Asp808 of HER2 protein. Combining results from both docking and pharmacophore mapping analysis, CLC015-5, CLC604-11 and CLC604-18 were well accepted and consistent in both approaches and were considered as the most potential HER2 inhibitors.

Two birds with one stone? Possible dual-targeting H1N1 inhibitors from traditional Chinese medicine.

The H1N1 influenza pandemic of 2009 has claimed over 18,000 lives. During this pandemic, development of drug resistance further complicated efforts to control and treat the widespread illness. This research utilizes traditional Chinese medicine Database@Taiwan (TCM Database@Taiwan) to screen for compounds that simultaneously target H1 and N1 to overcome current difficulties with virus mutations. The top three candidates were de novo derivatives of xylopine and rosmaricine. Bioactivity of the de novo derivatives against N1 were validated by multiple machine learning prediction models. Ability of the de novo compounds to maintain CoMFA/CoMSIA contour and form key interactions implied bioactivity within H1 as well. Addition of a pyridinium fragment was critical to form stable interactions in H1 and N1 as supported by molecular dynamics (MD) simulation. Results from MD, hydrophobic interactions, and torsion angles are consistent and support the findings of docking. Multiple anchors and lack of binding to residues prone to mutation suggest that the TCM de novo derivatives may be resistant to drug resistance and are advantageous over conventional H1N1 treatments such as oseltamivir. These results suggest that the TCM de novo derivatives may be suitable candidates of dual-targeting drugs for influenza.

What is the effective component in suanzaoren decoction for curing insomnia? Discovery by virtual screening and molecular dynamic simulation

Abstract The reliable structure of gamma aminobutyric acid type A (GABA-A) receptor was built based on several criteria. According to zolpidem and GABA binding conformations, the key residues that were indicated to be the determination of binding were consistent with our simulation. Investigation of the major effective constituents from suanzaoren to modulate the GABA-A was the aim of the study. Jujuboside A, which was indicated to be the effective constituent from suanzaoren, had no blood-brain barrier (BBB) penetration and was unable to bind at both binding sites due to its large volume. In addition, the glycoside groups on jujuboside A were easily to be hydrolyzed. In contrast, jujubogenin, which was hydrolyzed from jujuboside A, had the most compatible binding conformation. In addition, jujubogenin formed two HBs with the key residue β2-Thr226 and β2-Tyr229 at the GABA binding site. Moreover, it gained the comparably highest scoring values among suanzaoren constituents. Furthermore, the Adsorption, Distribution, Metabolism, Excretion, and Toxicity (ADMET) descriptor predicted that jujubogenin have good BBB penetration. Consequently, we suggested jujubogenin to be the effective suanzaoren constituent to mediate the GABA-A receptor.

Insights into designing the dual-targeted HER2/HSP90 inhibitors.

Heat shock protein 90 (HSP90) and human epidermal growth factor receptor 2 (HER2) are two key cancer markers actively involved in several signal pathways for cancer cell growth. In this study, we focused on the designing of dual-targeted HSP and HER2 inhibitors. Comparative molecular field analysis (CoMFA), comparative molecular similarity indices analysis (CoMSIA), and pharmacophore analysis were employed for generating the activity prediction models. The results of CoMFA model showed highly predictive r(2) value with 0.922 and 0.885 in HSP90 and HER2, respectively. In CoMSIA model, the r(2) values were 0.967 and 0.936 in HSP90 and HER2, respectively. The contour maps of both targets showed that there were similar regions of bulky favored area. Additionally, the Hypogen results for HER2 showed high cost difference as 59.13 and r-value as 0.909. At the C2 position of the benzene ring, the HER2 model favored steric bulkier substitutes more than HSP90. The Hypogen results for HSP90 also showed reliable values in cost difference, 85.82 and r-value, 0.902. Overall, we investigated the significances of QSAR models and pharmacophore features for designing the HER2/HSP90 dual-targeted inhibitors.

Weighted Equation and Rules—A Novel Concept for Evaluating Protein-Ligand Interaction

Abstract In this study, a novel methodology for evaluating protein-ligand interaction and quantitated the traditional Chinese medicine (TCM) by Yin-Yang theory are proposed and investigated by a case report of the human epidermal growth factor receptor 2 (HER2)-ligand. Inhibitors (n = 176) of HER2 from references with a broad range of activities (IC50) were employed to the docking program to calculate the binding affinities. The docking score of twelve scoring functions versus actual pIC50 plot were regressed. According to the weighted rules, the coefficient of determinations (R2) from the regression analysis of each scoring function and pIC50 were chosen as the weights in the weighted equation. The R2 (0.5858) of weighted score (WS) versus actual pIC50 was statistically higher than that of the consensus score (CS) (R2 = 0.2441). The WS method lies in combining the scoring functions from different algorithms to evaluate the sum of binding affinities that is more comprehensive than any single scoring function can achieve. The WS calculated by equation successfully shows a statically significant correlation with good predictability. Thus, this methodology might provide a persuasive virtual screening criterion to evaluate the protein-ligand interaction and quantitative analysis of the functions for Chinese medicine in the future.

A Novel Strategy for Designing the Selective PPAR Agonist by the “Sum of Activity” Model

Abstract Peroxisome proliferator-activated receptors α, δ, and γ are a collection of ligand-activated transcription factors crucial in lipid and glucose homeostasis. The involvement of these receptors in lipid metabolism makes them perfect therapeutic target for treating obesity and stroke. In this study, ‘sum of activity’ model was employed to design multi-target agonists. We used a new strategy to design agonists that fit both α and δ but not γ to avoid side effect. The CoMFA and CoMSIA models were used to explore the pharmacophore features by constructing three individual models: (a) α-model, (b) δ-model and (c) γ-model, and two sum models: (d) α, δ- model, and (e) α, δ, γ- model. The CoMFA model yielded a significant cross validation value, q2, of 0.729 and non-cross validation value, r2, of 0.933 in the alpha;, δ-model. The CoMSIA studies yielded the best predictive models with q2 of 0.622 in A + S and with r2 of 0.911 in the α, δ-model. Finally, we proposed that distinct features shown in models (a), (b), (d) but not (c) and (e) should be accounted in designing weight-controlling drugs.

How to design a drug for the disordered proteins

Structural disorders of proteins make drug design a difficult task. The gel-like state of intrinsically disordered protein (IDP) or intrinsically disordered regions (IDRs) remains a big puzzle for drug designers. Here, we propose a novel concept for drug design by understanding protein disintegration and protein-protein interaction (PPI) using molecular dynamics (MD) simulation and propose a possible approach for overcoming current obstacles in IDP drug design.