Synthesis and Evaluation of 5-(Heteroarylmethylene)hydantoins as Glycogen Synthase Kinase-3β Inhibitors

Glycogen synthase kinase-3 (GSK-3) is a serine/threonine kinase which plays a center role in the phosphorylation of a wide variety of proteins, generally leading to their inactivation. As such, GSK-3 is viewed as a therapeutic target. An ever-increasing number of small organic molecule inhibitors of GSK-3 have been reported. Phenylmethylene hydantoins are known to exhibit a wide range of inhibitory activities including for GSK-3β. A family of fourteen 2-heterocycle substituted methylene hydantoins (14, 17–29) were prepared and evaluated for the inhibition of GSK-3β at 25 μM. The IC50 values of five of these compounds was determined; the two best inhibitors are 5-[(4′-chloro-2-pyridinyl)methylene]hydantoin (IC50 = 2.14 ± 0.18 μM) and 5-[(6′-bromo-2-pyridinyl)methylene]hydantoin (IC50 = 3.39 ± 0.16 μM). The computational docking of the compounds with GSK-3β (pdb 1q41) revealed poses with hydrogen bonding to the backbone at Val135. The 5-[(heteroaryl)methylene]hydantoins did not strongly inhibit other metalloenzymes, demonstrating poor inhibitory activity against matrix metalloproteinase-12 at 25 μM and against human carbonic anhydrase at 200 μM, and were not inhibitors for Staphylococcus aureus pyruvate carboxylase at concentrations >1000 μM.

The structure of GSK-3β consists of an N-terminal β-strand domain and a C-terminal α-helix domain joined by a glycine-rich loop and hinge [18].The ATP-binding domain is at the interface of the β-strand and α-helix domains.The X-ray crystal structures of human GSK-3β with AMP-PNP, and the inhibitors staurosporine, indirubin-3-monoxime, and alsterpaullone reveal important interactions with the backbone carbonyl of Asp133 and the backbone amide NH and/or the carbonyl of Val135 (Figure 1) [19].An ever-growing number of small organic molecule GSK-3β inhibitors have been reported, with additional crystal structures or docking studies reinforcing these crucial interactions [20].

Chemistry
Four known (15 and 17-19) and ten new 5-(heteroarylmethylene)hydantoins (20-29) were prepare through the Knovenegel condensation of hydantoin with heteroarylaldhydes (Scheme 1).In most cases, the product precipitated from the ethanol/water reaction mixture; however, in the case of 21, 23, 24, and 25, several drops of conc.HCl were added to induce precipitation.Products 15 and 17-19 were identified by means of a comparison of their physical and/or spectral data to values from the literature [24,25,28].The structures of 6-substituted-5-methylenehydantoins 20-29, were assigned on the basis of their NMR spectral data.In particular, singlets in their 1 H NMR spectra in the range of δ 6.39-6.48ppm and signals in their 13 C NMR spectra in the range δ 165 and 101-105 ppm correspond to H6, C2, and C6, respectively [29].The exception to this is 24, where the signal for H6 appears at δ 6.61 ppm.This additional downfield shift is due to the proximity of the 3-chloro substituent.

Chemistry
Four known (15 and 17-19) and ten new 5-(heteroarylmethylene)hydantoins (20)(21)(22)(23)(24)(25)(26)(27)(28)(29) were prepare through the Knovenegel condensation of hydantoin with heteroarylaldhydes (Scheme 1).In most cases, the product precipitated from the ethanol/water reaction mixture; however, in the case of 21, 23, 24, and 25, several drops of conc.HCl were added to induce precipitation.Products 15 and 17-19 were identified by means of a comparison of their physical and/or spectral data to values from the literature [24,25,28].The structures of 6-substituted-5-methylenehydantoins 20-29, were assigned on the basis of their NMR spectral data.In particular, singlets in their 1 H NMR spectra in the range of δ 6.39-6.48ppm and signals in their 13 C NMR spectra in the range δ 165 and 101-105 ppm correspond to H6, C2, and C6, respectively [29].The exception to this is 24, where the signal for H6 appears at δ 6.61 ppm.This additional downfield shift is due to the proximity of the 3-chloro substituent.

Molecular Docking
The docking of 15 and 17-29 into the ATP binding cleft of GSK-3β (pdb 1q41 [19]) was calculated with the Mcule online one-click program which uses the Vina docking algorithm (https://mcule.com,accessed on 8 February 2024).Vina uses a sophisticated gradient optimization method in its local optimizing procedure, and was designed to improve the speed of the execution and accuracy [30].The authors of this program describe the scoring function is 'more of a "machine learning" than directly physics-based in its nature'.Structures 15, 17-27, and 29 docked with modest affinity (-6.7 to -7.8 kcal/mol), where the best scored pose indicated hydrogen bonding interactions between the backbone carbonyl C=O and amide N-H of Val135 with the N-H at position 3 and C=O at position 4 of the hydantoin ring, respectively, as represented by the best docking pose for 23 (Figure 3A).Notably, there was no discernable hydrogen bonding to the backbone N-H of Asp133.In contrast, while the computational docking of 28 produced the best score, the top 4 poses reversed the orientation of the molecule such that the hydantoin ring is buried deep in the binding pocket where it does not make any interactions with the backbone amide group of Val135 (Figure 3B).
The predictive value of these computations may be limited.Indirubin-3-monoxime is the ligand in PDB 1q41.The top four docking results from the online Mcule program for indirubin-3-monoxime have the oxime OH oriented toward Val 135.This may be due to parameters which overemphasize the energy of this interaction.Notably, the docking of the methyl ether of indirubin-3-oxime resulted in an orientation of the ligand more closely matching the crystal structure.
of the methyl ether of indirubin-3-oxime resulted in an orientation of the ligand more closely matching the crystal structure.1).At 200 µM of 20-24, the activity of MMP-12 was <50% the activity of positive control; for compounds 21, 24, and 25 at 25 µM, the activity of MMP12 was 69-74% while the activity for the others was >80%.Thus, all compounds were more potent as inhibitors for GSK-3β than for MMP-12.A select subset of the 5-(heteroarylmethylene)hydantoins were evaluated as inhibitors of human carbonic anhydrase II (hCAII) at 200 µM (Table 1).Only 25 and 29 exhibited any inhibition (57% and 63% activity, respectively) while the others exhibited little or no inhibitory activity.Finally, this same subset of 5-(heteroarylmethylene)hydantoins were evaluated as inhibitors of isolated S. aureus pyruvate carboxylase (SaPC) in a fixed-time assay (Table 1) [27].No inhibitory activity was observed up to 1000 µM.

Inhibition of Matrix Metalloproteinase-12
Assays for human matrix metalloproteinase-12 activity were performed in a 96-well plate format for a total reaction volume of 100 µL, in an assay modified from Day and Cohen [32].All assay reagents were prepared and maintained at 22 • C until warmed to 37 • C. The omniMMP fluorogenic substrate and MMP-12 enzyme were prepared in assay buffer (50 mM HEPES, 10 mM CaCl 2 , 0.05% Brij-35, pH of 7.52), and small molecule effectors were prepared as a solution in 50% DMSO, 50% assay buffer.The final concentration of enzyme in the 100 µL enzymatic reaction was 0.0083 U/µL.A known inhibitor, N-[(4'-bromo [1,1'-biphenyl]-4-yl)sulfonyl]-L-valine (PD166793), was included where noted at a final concentration of 100 µM.OmniMMP fluorogenic substrate was added to the 100 µL enzymatic reaction at a final concentration of 4 µM.For all experiments, 30 µL of enzyme stock solution was added to each well in a 96-well, flat-bottom, black polystyrene microplate (SantaCruz Biotechnology, Dallas, TX, USA), followed by the addition of 10 µL of the effectors (50% assay buffer, 50% DMSO for the uninhibited reaction, PD166793 for the inhibition).The 96-well plate was then incubated in a plate reader at 37 • C for 30 min; concurrently, the substrate solution was also warmed to 37 • C for 30 min.The enzymatic reaction was initiated by adding 60 µL of the omniMMP fluorogenic substrate solution.The change in fluorescence was then monitored for 30 min with excitation and emission wavelengths at 320 and 400 nm, respectively, every 46 s for a total of 40 measurements.

Inhibition of Human Carbonic Anhydrase II
Assays for human carbonic anhydrase II activity were performed in a 96-well plate format for a total reaction volume of 100 µL, in an assay modified from Day and Cohen [32].All assay reagents and solutions were prepared separately and maintained at 22 • C until warmed to 30 • C. All substrates, effectors, and enzyme were freshly dissolved and diluted in assay buffer containing 50 mM Tris (pH 8.0).The final concentration of the human carbonic anhydrase II enzyme was 200 nM.Where noted, acetazolamide was included at a final concentration of 10 µM.The substrate, p-nitrophenyl acetate, was added at a final concentration of 504 µM.For all experiments, 20 µL of enzyme stock solution was added to each well in a 96-well, flat-bottom, polystyrene microplate (Santa Cruz Biotechnology, Dallas, TX, USA), followed by the addition of 10 µL of the effectors (assay buffer for the uninhibited reaction, acetazolamide for the inhibition).The 96-well plate was then incubated in a plate reader at 30 • C for 10 min; concurrently, the substrate solution was also warmed to 30 • C for 10 min.The enzymatic reaction was initiated by adding 70 µL of the substrate solution (p-nitrophenyl acetate).The absorbance values were then measured at 405 nm every 30 s over a period of 20 min for a total of 40 measurements.

Inhibition of S. aureus Pyruvate Carboxylase (SaPC)
Malate dehydrogenase-coupled enzyme assays were performed at 22 • C in a 96-well plate format for a total reaction volume of 200 µL.Assay conditions consisted of 100 mM Tris (pH 7.8), 7 mM MgCl 2 , 150 mM KCl, and 0.5% Triton x-100.First, 20 µL of compound was added to obtain the desired final concentration; 20 µL of malate dehydrogenase (MDH) was added such that the final concentration in the assay was 20 U/mL; and lastly, 140 µL of substrates was added to initiate the reaction (HCO 3− , ATP, and NADH to a final assay concentration of 15 mM, 2.5 mM, and 0.25 mM, respectively).To account for the possible compound inhibition of MDH, the procedure was modified such that 20 µL oxaloacetate was added to a final concentration of 30 mM, in place of PC.Reagents were dispensed manually by a hand-held, multi-channel micropipette, and absorbance measurements were recorded at 340 nM with a Molecular Devices SpectraMax i3x Multi-mode plate reader.

Conclusions
Several 5-(heteroarylmethylene)hydantoins were found to be single-digit micromolar inhibitors of GSK-3β, with 25 and 27 as the most potent ones.Computational docking predicted that these bind in the ATP binding domain, with hydrogen bonding between the hydantoin functionality and the backbone C=O and N-H of a valine 135 in this domain.The GSK-3β inhibitory activities of the compounds reported in this manuscript are in the same range as previously reported for substituted 5-phenylmethylenehydantoins (1-8) [23], and they are considerably less potent than the highly potent inhibitors stauroporine (IC 50 = 15 nM), indirubin-3-monoxime (IC 50 = 22 nM), and alsterpaullone (IC 50 = 4 nM).Developing more potent inhibitors based on this scaffold will likely require the introduction of additional functionality which can take advantage of hydrogen bonding and/or lipophilic functionality present within the binding pocket.Certain of these compounds exhibited an inhibition of MMP-12 at 200 µM, and further inhibitor design may be needed to improve the selectivity for GSK-3β over MMP-12; there was essentially no inhibitory activity observed for hCAII or SaPC.

Supplementary Materials:
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ph17050570/s1, 1 H and/or 13 C NMR spectra of synthesized compounds.Funding: This research was funded by the National Institutes of Health, grant number GM131356.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author due to legal reasons.