Efficient Synthesis and Docking Analysis of Selective CDK9 Inhibitor NVP-2

Abstract Graphical Abstract NVP-2 (1), a potent and selective inhibitor of cyclin-dependent kinase 9 (CDK9), showed potent antitumor activity in preclinical studies. In this work, we designed and adopted a convergent synthetic route to efficiently synthesize NVP-2 (1). The key intermediate (7) was synthesized from malononitrile (2) and 1-bromo-2-(2-bromoethoxy)ethane (3) by successive cyclization, reduction, nucleophilic substitution with 2-bromo-6-fluoropyridine, and Suzuki–Miyaura reaction with (5-chloro-2-fluoropyridin-4-yl)boronic acid. Another key intermediate (11) was synthesized from (S)-1-methoxypropan-2-ol (8) by reaction with TsCl, electrophilic substitution reaction with tert-butyl ((1r,4r)-4-aminocyclohexyl)carbamate, and then by deprotection of Boc. Finally, a substitution reaction by the key intermediates (7) and (11) to afford the target product NVP-2 (1). The reaction conditions of the whole synthesis process were simple and mild, free of harsh conditions such as the microwave reaction and dangerous reagents in the original patent, and realized the efficient synthesis of NVP-2. In addition, we analyzed the binding mode of NVP-2 in the active pocket of CDK9 to provide reasonable design ideas for subsequent discovery of novel CDK9 inhibitors.

To date, a few synthesis routes of NVP-2 and its key intermediates have been reported. However, there are some shortcomings, such as harsh reaction conditions, low yield, many side reactions, and potential safety hazards, in these reports that are not conducive to industrial production. 12,13 Therefore, in this work, we first optimized the synthetic route to realize the efficient synthesis of NVP-2. Since the binding mode between NVP-2 and CDK9 protein has not yet been reported in the literature, we analyzed the docking mode of NVP-2 in the active pocket of CDK9 to provide reasonable design ideas for subsequent discovery of novel CDK9 inhibitors.
On the basis of the literature route, 12,13 the synthetic route of key intermediate 11 is optimized and improved, shown in Scheme 2: inexpensive, readily available (S)-1methoxy propan-2-ol (8) was used as the starting material, and reacted with p-toluenesulfonyl chloride to efficiently obtain (S)-1-methoxy-2-propyl-p-toluenesulfonate (9). The mixture of intermediate (9) and tert-butyl ((1r,4r)-4-aminocyclohexyl)carbamate was heated at reflux to efficiently obtain the intermediate tert-butyl ((1R,4r)-4-(((R)-1methoxypropan-2-yl)amino)cyclohexyl)carbamate (10). Intermediate 10 was quickly reacted with hydrochloric acid solution to deprotect the Boc group, and gave (1r,4R)-N 1 -((R)-1-methoxypropan-2-yl)cyclohexane-1,4-diamine (11) hydrochloride salt. Intermediate 11 and the above key intermediate 7 undergo substitution reaction at a higher temperature to obtain the final product NVP-2. In the synthetic route in the literatures, trans-cyclohexanediamine was directly reacted with intermediate 9 at high temperature to obtain intermediate 11. 12,13 However, there are some disadvantages, such as byproduct formation, difficulties at purification, and low yield in this step. Also, the hazardous chemical reagent NaH was utilized in the synthesis of intermediate 9 in the literatures, posing significant safety concerns in industrial manufacturing. In our improved synthetic route, the raw material tert-butyl ((1r,4r)-4-aminocyclohexyl)carbamate was directly reacted with intermediate 9 at a mild temperature to efficiently obtain intermediate 10 (65% yield), which quickly deprotected the Boc group in the acid solution to obtain key intermediate 11 (95% yield). By our improved synthetic route, we can efficiently obtain the pure key intermediate 11. In addition, we used pyridine as the base instead of the dangerous reagent NaH, and greatly shortened the reaction time to efficiently synthesize the intermediate 9 with a high yield of 90%. In the last step to synthesize the NVP-2, the addition of inorganic base potassium carbonate could reduce the reaction time and increase the yield. During the entire NVP-2 reaction process, the whole operation is simple, efficient and safe, and apt to industrial production.

Docking Analysis
NVP-2 competitively binds to the ATP active pocket of CDK9 protein to exert potent CDK9 inhibitory activity. However, the binding mode between NVP-2 and CDK9 protein has not been reported. Revealing and predicting the binding mode between NVP-2 and CDK9 protein could provide reasonable design ideas for the discovery of novel CDK9 inhibitors. Therefore, we applied a computer docking technology to simulate the interaction between NVP-2 and CDK9 protein.
The predicted binding mode of NVP-2 in the ATP active pocket of CDK9 protein is shown in ►Fig. 2. In the ATP active pocket of CDK9 protein (►Fig. 2A, B), it contains a very important hinge region and some hydrophobic pockets. In the hinge region, Cys106 residue can form vital hydrogen bonds with many small-molecule CDK9 inhibitors, including flavopiridol, variolins, and others. The weakening or disappearance of this key hydrogen bond force will greatly reduce the inhibitory activity of CDK9 inhibitors. Asp104 is also an important amino residue located in the hinge region; some CDK9 inhibitors could form dual hydrogen bonds with Cys106 and Asp104 residues. 6 From the binding mode of NVP-2 in the CDK9 active pocket (►Fig. 2C, D), it can be seen that the structure of NVP-2 can be well accommodated in the CDK9 active pocket, and the pyridine amino group in the NVP-2 structure can form dual hydrogen bonds with the Cys106 residue in the hinge region. This strong dual hydrogen bonding force endows NVP-2 with potent CDK9 inhibitory activity. In addition, the cyan group in the NVP-2 structure extends into the hydrophobic pocket of the CDK9 protein to enhance the interaction, while the cyclohexylamino fragment in the NVP-2 structure extends into the solvent-exposed region, suggesting that structural modification of this part can maintain or improve the inhibitory activity against CDK9. The above docking analysis provided valuable clues for modification and discovery of novel CDK9 inhibitors based on the structure of NVP-2.

Conclusion
NVP-2, a very potent and selective ATP-competitive CDK9 inhibitor, may be a promising antitumor agent to enter clinical research in the near future. In this work, we optimized the reagents and conditions of the synthetic route and realized the efficient synthesis of NVP-2. The whole synthesis process was simple and mild, and overcame the harsh conditions such as the microwave reaction and dangerous reagents. In addition, we analyzed the docking mode of NVP-2 in the active pocket of CDK9, and provided valuable clues for modification and discovery of novel CDK9 inhibitors based on the structure of NVP-2.

Experimental Section
Chemistry Unless otherwise noted, all solvents and reagents were commercially available (Bidepharm, J&K Scientific, Sinopharm) and used without further purification. All reactions were monitored by thin-layer chromatography on 0.25 mm silica gel plates (60 GF-254) and visualized with UV light, or iodine vapor. 1 H NMR and 13 C NMR were generated in DMSOd 6 or CDCl 3 on Varian Mercury 400, 500, or 600 MHz NMR spectrometers. Chemical shifts were reported in parts per million (ppm). Multiplicity of 1 H NMR signals was reported as single (s), double (d), triplet (t), quartet (q), and multiplet (m). ESI-MS (electrospray ionization mass spectra) were determined on an API 4000 spectrometer.

Synthesis of 4-(Aminomethyl)tetrahydro-2H-pyran-4carbonitrile (5)
To the solution of intermediate (4, 1.0 g, 7.34 mmol) in ethanol (35 mL) was slowly added sodium borohydride (0.83 g, 22.0 mmol) under ice bath conditions. The reaction mixture was removed from the ice bath and stirred at 60°C for 4 hours. The mixture was concentrated under reduced pressure and the resulting residue was diluted with ethyl acetate (100 mL), then washed with water and brine (150 mL Â 3). The organic phase was dried over anhydrous sodium sulfate, filtered, and concentrated to dryness under a vacuum evaporator to give a crude product (5), which was used directly in the next step. Low-resolution mass spectrometry (LRMS) m/z calcd. for

Docking Simulation
The binding mode between NVP-2 and the CDK9 active pocket was completed by the conventional computer docking software Schrödinger-Maestro (https://www.schrodinger.com/products/ maestro, version 11.0, Schrödinger software package, New York, United States, 2020). The structural optimization of CDK9 protein (PDB number: 4BCF) was completed by the Protein Preparation Wizard sector in Maestro. The structure characterization of CDK9 protein was performed by adding hydrogens, creating disulfide bonds, filling in missing side chains, deleting water, restrained minimization, and other optimization operations (other parameters are kept unchanged). Subsequently, the receptor grid box (10 Â 10 Â 10 Å) was obtained based on the optimized structure of CDK9 protein. Continue to use LigPrep in Maestro to process the NVP-2 structure to generate various three-dimensional molecular configurations, and then used the Ligand Docking sector to complete the docking simulation between NVP-2 and CDK9 protein.