Identification of novel β‐secretase inhibitors through the inclusion of protein flexibility in virtual screening calculations

Alzheimer's disease is a neurodegenerative disorder featuring both the aggregation of extraneuronal amyloid plaques and intraneuronal neurofibrillary tangles. The deposition of Aβ peptides in the brain plays a key role in the onset and progression of this disease, and β‐secretase (BACE‐1) is critical in this process. Thus, the inhibition of this enzyme is an attractive approach to anti‐Alzheimer therapy. So far, the rational design of new BACE‐1 inhibitors has been hampered by the intrinsic flexibility displayed by the enzyme which is not included in standard molecular modeling methods. Here, we incorporated protein flexibility into Virtual Screening (VS) in order to discover new BACE‐1 inhibitors in two ways. First, we used an ensemble of static Xray BACE‐1 structures representing the receptor flexibility in a VS of the NCI database. Also, the different conformations of BACE‐1 were combined to generate a unified description of the protein motion. To do this, a set of maps was computed for each protein conformation, combined in a single set using an energy weighting scheme and finally used in a second VS of the NCI database. The enrichment factors from these two VS approaches were used to assess their respective predictive power. The in vitro evaluation of the best ranking ligands finally led to the discovery of new leads which will be subjected to further structural optimization.