The design of selective HDAC inhibitors is of interest in development of molecular scapels to define the biological role of these enzymes, and ultimately, less toxic drugs. Comparative Binding Energy Analyses1 (COMBINE), were conducted on HDAC complexes to highlight the discriminant interactions and generate predictions of target selectivity. HDAC/inhibitor complexes (93 total) compiled from 12 experimental (HDAC-2, -4, -7, -8) and homology modeled (HDAC-1, -3, -5, -6A, -6B, -9, -10, -11) HDAC deacetylase domains and 9 HDAC inhibitors for which the experimental activities were available from a single paper. Experimental ligand/HDAC complexes were retrieved from the PDB, while the other were obtained from four homology-model automated servers. Compound with unknown bound conformation were directly modeled into the HDAC catalytic sites and the resulting complexes were energy minimized using AMBER. After minimization a multi-sequence alignment of enzymes was performed by the Modeller2 program; the complexes were submitted to a COMBINE-like procedure generated by the combination of the AutoGrid3 program with the R4 environment language. Interaction energies were calculated on a per-residue basis considering electrostatic (ELE), steric (STE) and desolvation (DRY), while statistical PLS5-based models were build and validated using in-house R scripts; this allowed identification of those residues responsible for both inhibitory activity and selectivity. A COMBINE model was obtained showing r2, q2 LOO, q2 LSO-5 and SDEPCV of 0.77, 0.72, 0.72 and 0.87, respectively. To test the predictive ability of the model, 10 HDAC inhibitors with enzyme-inhibition data for different isoforms were used to build an external test set of 73 complexes; the activities and isoform selectivity of the HDACs inhibitors were correctly predicted in most cases. To our knowledge, this is the first general attempt to obtain a structure-based approach to rationalize the development of novel compounds endowed with both high potency and selectivity
Histone Deacetylase Inhibitors. The Development of a Model for Target Prediction / L., Silvestri; Ballante, Flavio; G. R., Marshall; Ragno, Rino. - (2011). (Intervento presentato al convegno CDDD Meeting, Computationally Driven Drug Discovery Meeting tenutosi a L'Aquila nel 21-23 Novembre 2011).
Histone Deacetylase Inhibitors. The Development of a Model for Target Prediction
BALLANTE, FLAVIO;RAGNO, Rino
2011
Abstract
The design of selective HDAC inhibitors is of interest in development of molecular scapels to define the biological role of these enzymes, and ultimately, less toxic drugs. Comparative Binding Energy Analyses1 (COMBINE), were conducted on HDAC complexes to highlight the discriminant interactions and generate predictions of target selectivity. HDAC/inhibitor complexes (93 total) compiled from 12 experimental (HDAC-2, -4, -7, -8) and homology modeled (HDAC-1, -3, -5, -6A, -6B, -9, -10, -11) HDAC deacetylase domains and 9 HDAC inhibitors for which the experimental activities were available from a single paper. Experimental ligand/HDAC complexes were retrieved from the PDB, while the other were obtained from four homology-model automated servers. Compound with unknown bound conformation were directly modeled into the HDAC catalytic sites and the resulting complexes were energy minimized using AMBER. After minimization a multi-sequence alignment of enzymes was performed by the Modeller2 program; the complexes were submitted to a COMBINE-like procedure generated by the combination of the AutoGrid3 program with the R4 environment language. Interaction energies were calculated on a per-residue basis considering electrostatic (ELE), steric (STE) and desolvation (DRY), while statistical PLS5-based models were build and validated using in-house R scripts; this allowed identification of those residues responsible for both inhibitory activity and selectivity. A COMBINE model was obtained showing r2, q2 LOO, q2 LSO-5 and SDEPCV of 0.77, 0.72, 0.72 and 0.87, respectively. To test the predictive ability of the model, 10 HDAC inhibitors with enzyme-inhibition data for different isoforms were used to build an external test set of 73 complexes; the activities and isoform selectivity of the HDACs inhibitors were correctly predicted in most cases. To our knowledge, this is the first general attempt to obtain a structure-based approach to rationalize the development of novel compounds endowed with both high potency and selectivityI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
