TY  - JOUR
AU  - Wienen-Schmidt, Barbara
AU  - Schmidt, Denis
AU  - Gerber, Hans-Dieter
AU  - Heine, Andreas
AU  - Gohlke, Holger
AU  - Klebe, Gerhard
TI  - Surprising Non-Additivity of Methyl-Groups in Drug-Kinase Interaction
JO  - ACS chemical biology
VL  - 14
IS  - 12
SN  - 1554-8937
CY  - Washington, DC
PB  - Soc.
M1  - FZJ-2019-05295
SP  - 2585-2594
PY  - 2019
AB  - Drug optimization is guided by biophysical methods with increasing popularity. In the context of lead structure modifications, the introduction of methyl groups is a simple but potentially powerful approach. Hence, it is crucial to systematically investigate the influence of ligand methylation on biophysical characteristics such as thermodynamics. Here, we investigate the influence of ligand methylation in different positions and combinations on the drug–kinase interaction. Binding modes and complex structures were analyzed using protein crystallography. Thermodynamic signatures were measured via isothermal titration calorimetry (ITC). An extensive computational analysis supported the understanding of the underlying mechanisms. We found that not only position but also stereochemistry of the methyl group has an influence on binding potency as well as the thermodynamic signature of ligand binding to the protein. Strikingly, the combination of single methyl groups does not lead to additive effects. In our case, the merger of two methyl groups in one ligand leads to an entirely new alternative ligand binding mode in the protein ligand complex. Moreover, the combination of the two methyl groups also resulted in a nonadditive thermodynamic profile of ligand binding. Molecular dynamics (MD) simulations revealed distinguished characteristic motions of the ligands in solution explaining the pronounced thermodynamic changes. The unexpected drastic change in protein ligand interaction highlights the importance of crystallographic control even for minor modifications such as the introduction of a methyl group. For an in-depth understanding of ligand binding behavior, MD simulations have shown to be a powerful tool.
LB  - PUB:(DE-HGF)16
C6  - pmid:31638770
UR  - <Go to ISI:>//WOS:000504806100012
DO  - DOI:10.1021/acschembio.9b00476
UR  - https://juser.fz-juelich.de/record/866049
ER  -