Investigating the Binding Mode of Reversible LSD1 Inhibitors Derived from Stilbene Derivatives by 3D-QSAR, Molecular Docking, and Molecular Dynamics Simulation

Overexpression of lysine specific demethylase 1 (LSD1) has been found in many cancers. New anticancer drugs targeting LSD1 have been designed. The research on irreversible LSD1 inhibitors has entered the clinical stage, while the research on reversible LSD1 inhibitors has progressed slowly so far. In this study, 41 stilbene derivatives were studied as reversible inhibitors by three-dimensional quantitative structure–activity relationship (3D-QSAR). Comparative molecular field analysis (CoMFA q2 = 0.623, r2 = 0.987, rpred2 = 0.857) and comparative molecular similarity indices analysis (CoMSIA q2 = 0.728, r2 = 0.960, rpred2 = 0.899) were used to establish the model, and the structure–activity relationship of the compounds was explained by the contour maps. The binding site was predicted by two different kinds of software, and the binding modes of the compounds were further explored. A series of key amino acids Val288, Ser289, Gly314, Thr624, Lys661 were found to play a key role in the activity of the compounds. Molecular dynamics (MD) simulations were carried out for compounds 04, 17, 21, and 35, which had different activities. The reasons for the activity differences were explained by the interaction between compounds and LSD1. The binding free energy was calculated by molecular mechanics generalized Born surface area (MM/GBSA). We hope that this research will provide valuable information for the design of new reversible LSD1 inhibitors in the future.


S1. Compound Characterization Data
Reagents and solvents were purchased from commercial sources, when necessary, were purified and dried by standard methods. Melting points were determined on an X-5 micromelting apparatus and are uncorrected. 1 H NMR and 13 C NMR spectra were recorded on a Bruker Avance III HD 400 MHz and 100 MHz spectrometer at room temperature, using TMS as an internal standard. Chemical shifts were reported in ppm (d). Spin multiplicities were described as s (singlet), d (doublet), dd (double doublet), t (triplet), or m (multiplet). Coupling constants were reported in hertz (Hz). High resolution mass spectrometry (HRMS) was recorded on a Bruker MicrOTOF-Q III Micro mass spectrometer by electrospray ionization (ESI). Flash chromatography was performed on 200-300 mesh silica gel with the indicated solvent systems (Qingdao Haiyang Chemical, China).
General synthesis procedure for compounds III and V.
To a stirred solution of compound I (1.0 eq) and compound II or IV (1.05 eq) in dry DMF was added t-BuOK (3.0 eq) at 0℃. Then, the above reaction mixture was stirred for 0.5 h at room temperature. The mixture was poured into cold water, the resultant precipitate was filtered, washed with water, dried and purified by recrystallization from ethyl acetate to afford the pure product III or V.

S3. Structural validation
The Ramachandran plot obtained through the Procheck program was shown in Figure S2. The overall G-factor value of model is 0.02.
The torsion angles phi (φ) and psi (ψ) distributions of the Ramachandran plot of all non-glycine and non-proline residues, as shown in Table S2. Figure S2. Ramachandran plot. Most favored (red), additionally allowed (yellow), generously allowed (pale yellow) and disallowed regions (white color).
It can be seen from Fig.S2 that most residues were in the most favored region, a good quality model would be expected to have over 90% in the most favoured regions. From Table S2, 91.2% residues were in most favoured regions, while 8.5% residues were in additional allowed regions, and only 0.3% residues were in generously allowed regions, no one residue was in disallowed regions. These results showed that the model had good stereo-chemical quality. Table S2. Residues falling in the core region of the Ramachandran's plot.
Residues in most favoured regions 91.2% Residues in additional allowed regions 8.5% Residues in generously allowed regions 0.3% Residues in disallowed regions 0.0% Number of non-glycine and non-proline residues 100.0% At the same time, verify 3D and ERRAT were used to evaluate the quality of the model structure. Verify 3D judged the reliability of the model by comparing its 3D profile with its sequence. It is generally believed that at least 80% residues of the models with average 3D-1D score ≥ 0.2 are acceptable.
According to Fig.S3, 85.36% residues had an average 3D-1D score ≥ 0.2, indicated that the structure had a good environment profile. which mean that 90.229% residues in chain A were lower than the rejection limit of 95%. 89.516% residues in chain B were lower than the rejection limit of 95%. All the above evidences showed that the model used in this study had good quality and can be used in molecular docking and MD research.