Evaluation of 4 H-4-chromenone derivatives as inhibitors of protein kinase CK 2

Protein kinase CK2 (Casein Kinase 2) is a ubiquitous serine/threonine protein kinase involved in various cell signal transduction pathways. Thus, CK2 is a new perspective target for anticancer drugs. The receptor-based virtual screening of 2000 compounds from combinatorial library of 4H-4-chromenones has been carried out in search for CK2-inhibitors. 90 compounds have been chosen for biological testing based on the score values calculated by DOCK 4.0 software. It has been revealed, that 3-(4-chloro-3,5dimethylphenoxy)-7-(4-methoxyphenylcarbonyloxy)-4-oxo-4H-chromene (12) and 7-(4-fluorophenylcarbonyloxy)-4-oxo-3-(4-phenylphenoxy)-4H-chromene (14) inhibit CK2 activity with / C 5 0 18.8 pM and ICso 22.4 fiM, respectively.

Introduction.Protein kinases are one of the largest enzymes family committed to the catalysis of protein phosphorylation, the most general and frequent me chanism controlling diverse aspects of cell life [1,2].
Protein kinase CK2 (Casein Kinase 2) is a ubiquitous protein serine/threonine kinase partici pating in the regulation of cell growth and prolife ration.
The negative regulation of the CK2 in cell control mechanisms (anti-apoptotic protecting function) has been also proven [3,4].Its activity is elevated in rapidly proliferating tissues as well as in the variety of tumors.CK2 may substantially contribute to car cinogenesis through its direct interaction with the cell-survival circuitry [3][4][5][6][7][8].Hence, CK2 is a promi sing target for anticancer drugs.
Molecular modeling techniques are intensively used in modern drug research and development.Strategies of the molecular modeling are based on the searching for structure-activity relationships of the chemical compounds for selection of promising com pounds against a biological target.Receptor-based approach uses methods of molecular docking, mole cular dynamics, energy minimization and molecular structure optimization to estimate ligand binding affinities to the receptor.The receptor spatial struc ture, obtained from X-ray, NMR or homology mo deling data is necessary for this approach.Six spatial structures of the inhibitor-CK2 complexes have been determined by X-ray technique [28,29,41,42].
The core of receptor-based approaches is a mo-lecular docking.Docking procedure consists of generation of the receptor-ligand complexes and estimation of favorable interactions by analyzing contributions of each component of the complex into total free energy of binding.Recently DOCK software was successfully applied in a number of screening activities [30][31][32][33]42].So, for virtual screening combinatorial library of 4H-4-chromenones we have used the system based on DOCK 4.0 package.

Materials and Methods.
Software background.To predict the affinity of diverse sets of compounds we have used our in-house screening system (Fig. 2).It uses the results on GAMESS AMI semi-empirical calculation [34], geometry optimization by GRO-MACS [35], and docking output of DOCK package [36].

Receptor preparation.
A receptor molecule has been minimized in water with GROMACS molecular dynamics simulation package (GROMACS force field, steepest descent algorithm, 1000 steps, em_tolerance = 100, em_step = 0,001).Active site spheres were calculated with DOCK sphgen software.31 spheres from the biggest cluster of 37 spheres were selected to fill receptor active site.Six spheres were deleted manually since their positions were outside of the active site cavity.Connolly MS (http://www.netsci.org/Science/Compchem/featurel4.html) and Grid programs from DOCK package were used to generate receptor Connolly surface and energy grids.Surface and grid calculations were performed with parameter settings as in [37], except for grid spacing that was set to 0.3.
Ligand database processing.. Ligand molecules have been processed with SCREENER in-house software (the preprocessing of input ligand database file, converting 2D structures to 3D), GROMACS MD package (fast energy minimization of the ligands by GROMOS 96 force field), and GAMESS QM package (complete energy minimization by AMI semi-empirical method, calculation of partial charges).Our own program TOPBUILDER has been used to generate GROMACS-formatted molecular topologies, control ligand energy minimization in GROMACS, and assign atom partial charges calculated in GAMESS.
Flexible docking.DOCK program has been used for receptor-ligand flexible docking.DOCK input parameters have been set as in [37] with some exceptions to increase the calculations accuracy: the minimum of heavy atoms in the anchor was set to 6, the maximum number of orientations was set to 1000, and the «all atoms* model has been chosen.The docking scores have been obtained in the range from -14 up to -51 kcal/mol and used for selection of potential inhibitors.The compounds with scores less than -42 kcal/mol have been chosen as promising.The 90 best-scored candidates have been inspected for sterical clashes and unfavorable contacts and taken for the kinase assay analysis.
Chemical synthesis of the library of 4H-4-chromenone derivatives has been performed by modification of the reported methods [38][39][40].Structure and purity of the synthesized compounds have been confirmed by 1 H NMR spectroscopy.The spectra have been obtained with Varian VXR-300 NMR spectrometer at 300 MHz.
Biological testing.The selected compounds have been tested using the kinase in vitro assay.The volume of reaction was 30 fil (buffer: 20 mM Tris-HC1, pH 7.5 at 25 °C; 50 mM KC1; 10 mM MgCl 2 ).Each reaction contained: 1 fig of peptide substrate, nearly 500 /iM final; 10 units of CK2 human, recombinant, from «BioLabs» (USA), concentration 500000 unit/ml supplied in water buffered solution (0.02 /¿1 of purchased solution was added for 1 reaction point).The reaction master mix was prepared without ATP and aliquoted in 1.5 ml tubes at room tem perature.
The stock solutions of inhibitors were prepared in DMSO, the concentration of inhibitor was 5 mM.The concentration of DMSO in the reaction did not exceed 3 %.At a higher concentrations, DMSO inhibits the CK2 activity more than 10 %.
ATP solution was prepared separately.For each sample 0.05 mCi of y-[P 32 ]ATP was taken (specific activity of 100 fiCi/fiU).
The total concentration of labeled and unlabeled ATP was 40 fiM.The reaction was started with adding ATP mix.The time of reaction was 20 min at 30 °C.The reaction was stopped by adding 20 ^1 of 0.5 M orthophosphoric acid, reaction mixture was loaded on the 20 mm filter discs of the cellulose phosphate paper («Whatman», Great Britain).Filters were washed three times with 0.075 M orthophos phoric acid at room temperature and dried.
For detection of products, dried filters were counted by Cherenkow's method on the LKB gammacounter. 1 ^1 of DMSO was added to the reaction volume instead of the inhibitor stock solution for a positive control (blank).As a negative control we used Quercetine, the known inhibitor of CK2, in final concentration of 0.55 ftM to inhibit the CK2 activity to 50 %.The sorption control was used from time to time as a reaction mix without enzyme, the non specific sorption did not exceed 5 % of the lowest counts.
Results and Discussion.We have carried out the computer receptor-based virtual screening of 2000 compounds of combinatorial library of 4H-4-chromenone derivatives for design of CK2-inhibitors based on DOCK, GAMESS, GROMACS, and TOPBUIL- M) (Fig. 3).
We performed visual inspection of obtained comp lexes and tried to indicate interactions caused activity of the inhibitors.The ligand-receptor hydrogen bon ding that usually makes major contributions into ligand affinity have not been observed in both comp lexes.But it must be noted, that active site of rigid receptor does not structurally optimal for the binding of such type of ligands, and hydrogen bonds can be formed upon ligand-receptor complex fluctuations in solvent.Nevertheless, DOCK software selected these ligands as promising on the base of hydrophobic contacts.So it is reasonably to suppose that hydro phobic interaction makes the most important contri bution to the stabilization of the inhibitors 12 and 14 in the CK2 active site (Fig. 4).The main hydrophobic contact for the inhibitor 14 (Fig 4, A) is a clamping of 4'-phenylphenilen residue into hydrophobic cave formed by amino acid residues Phell3, Ile95, Ilel74, Trpl76.The additional contribition into complex stability is a stacking between 4'-phenyl residue of 3-hydroxyphenyl group of ligand and residue Phell3.Isoflavone core and oxygen of its carboxyl group are also involved in weak hydrophobic and electrostatic interactions with residues Asnll8, Metl63 and Leu45.
In complex with the inhibitor 12 (Fig 4 , B) the stacking with Phell3 is not observed.But stability of the complex is probably achieved due to 4H-4chromenone ring which situated deeper in the cleft and three methyl groups that enforce fixation of the ligand.The selectivity of the compounds is doubtful.There is no any strongly marked interaction with the key CK2 residues Val66 and He 174 that would cause selectivity of effective CK2 inhibitors [25].
Conclusions.The 90 compounds from the com binatorial library of 2000 4H-4-chromenones have been selected using computer receptor-based virtual screening.It is revealed, that 14 from them (15 % of selected compounds) inhibit CK2 activity with IC 50 < < 33 /<M.The two compounds 3-