Multi-structure docking analysis of BACE1 crystal structures and non-peptidic ligands

Area of Science:

• Computational chemistry
• Drug discovery
• Structural biology

Background:

• Designing non-peptidic inhibitors for Beta-site Amyloid Cleaving Enzyme 1 (BACE1) is a key goal in Alzheimer's disease research.
• Molecular docking is a widely used computational method for this purpose, but its accuracy depends heavily on the selection of appropriate BACE1 crystal structures.

Purpose of the Study:

• To develop a robust computational protocol for identifying effective non-peptidic BACE1 inhibitors.
• To evaluate the impact of using multiple BACE1 crystal structures versus a single structure in docking studies.
• To identify specific BACE1 crystal structures that enhance the accuracy of predictive models.

Main Methods:

• A multi-structure docking protocol was employed, docking 615 ligands into 150 BACE1 structures.
• Docking scores were post-processed using exploratory data analysis, regression, and discriminant analysis.
• Principal Component Analysis (PCA) was used to cluster BACE1 structures.

Main Results:

• Single-structure docking yielded low accuracy in predicting BACE1 inhibitor activity.
• Multi-structural docking scores, when post-processed with chemometrics, generated highly accurate predictive models.
• BACE1 structures with PDB IDs 4B70, 4DVF, and 2WEZ demonstrated high accuracy in discriminating active from inactive compounds.
• Discriminant structures were found to be centrally located within PCA-based clusters of BACE1 structures.

Conclusions:

• The selection of appropriate BACE1 crystal structures is critical for successful molecular docking.
• A multi-structural docking approach combined with chemometric analysis provides a more accurate method for designing BACE1 inhibitors.
• Specific BACE1 crystal structures (4B70, 4DVF, 2WEZ) can serve as predominant models for docking studies aiming to discover non-peptidic BACE1 inhibitors.