Illustrated in is the protocol we
Illustrated in is the protocol we applied for the screens of novel furoic acids as ACL inhibitors. We constructed our furan carboxylate library that contained 1446 2-furoic met inhibitor derivatives and 501 3-furoic acid derivatives by performing substructural searches of the ZINC database. We limited the molecular weight (MW) of the molecules to less than 500Da. Three-dimensional (3D) conformations of each furoic acid were generated from the smiles ZINC representation using the Openbabel (v2.3.2) software suite. Instead of using the generic 3D structures in the ZINC database, Openbabel was applied for generating 3D conformers to ensure correct compound protonation states at physiological pH. The 3D conformers were prepared for docking using Autodocktools (v1.5.6) for Autodock Vina and Openbabel (v2.3.2) for Glamdock. Although docking analyses indicated that the 2- and 3-furoic acids could occupy the CBD, we thought that the new furan derivatives might bind to the protein differently and therefore chose a large docking area for virtual screening. We applied the search grids to a 46Å×46Å×46Å cube centered around the CBD of ACL for Autodock Vina and used the citrate cocrystal structure binding orientation as the center search point for Glamdock. The Glamdock search space is of a similar volume to that of Vina. Compounds were then evaluated using the scoring functions within the Autodock Vina and Glamdock, and ranked according to the top docking score for each algorithm to generate two separate hit lists. Docking scores for compounds in the top 10% of each docking algorithm were found to be 2–3 times lower in energy than those of the 2-/3-furoic acid cores. Due to the known unreliability of docking scores to predict absolute potency, we chose to use them as an indicator of relative potency between compounds in our furoic acid library. The compounds that ranked among the top 10% of each list were then identified as ‘dual virtual hits’. We examined the docking poses of the ‘dual virtual hits’ by both Vina and Glamdock, and found the furoic acid moiety to occupy the CBD, which supports our hypothesis that the CBD was likely the target of the furoic acids. The twenty four (24) commercially available ‘dual virtual hits’ were then purchased from chemical suppliers and tested for their inhibitory activity against the ACL enzyme. We first tested the 24 virtual hits at 10 and 100μM in the ADP Glo ACL enzymatic assay that measures directly the ADP concentration formed from the enzymatic reaction. The assay was performed using ADP-Glo luminescence assay reagents. It measures ACL activity by quantification of the amount of ADP produced by the enzymatic reaction. The luminescent signal from the assay is correlated with the amount of ADP produced and is proportionally correlated with the level of ACL activity. The assay is highly sensitive and suitable for rapidly screening of large number of compounds (CV value<10%, S/B ratio>25, Z’-factor>0.7). Under our assay conditions (pH 8.0, 10mM MgCl, 5mM DTT, 100μM ATP, 200μM CoA, 200μM sodium citrate, and 30ng ACL), the IC of the known ACL inhibitor BMS-303141 varied from 0.3 to 0.5μM, which is within the range of the reported IC values of 0.94μM by a C radioisotope assay and 0.13μM by the traditional maleate dehydrogenase catalyzed reduction assay., summarizes the structures of the 24 virtual hits and their ACL inhibitory activities tested at two concentrations (10μM and 100μM). We categorized the chemical scaffolds into 4 subtypes: 4-substituted-2-furoic acids (Type-A: A1-8), 5-substituted-2-furoic acids (Type-B: B1-2), 5-sulfonamido-benzofuran-3-carboxylic acid (Type-C: C1-11), and 5-benzyloxy-benzofuran-3-carboxylic acid (Type-D: D1-3). Based on the activity criteria of >20% inhibition at 10μM and >80% at 100μM, we considered 11 compounds as confirm hits, which represents a hit rate of 45.8%. Besides, seven (7) additional compounds showed >50% inhibition at 100μM, which further demonstrated the high success rate of our method.