QSAR and mechanistic interpretation of estrogen receptor binding

Area of Science:

• Computational Chemistry
• Toxicology
• Molecular Modeling

Background:

• Understanding chemical interactions with the human estrogen receptor (hER) is crucial for assessing endocrine disruption potential.
• Existing modeling approaches require enhancement to accurately capture the nuances of hER-ligand binding.
• A comprehensive dataset of diverse chemicals is necessary for robust model development.

Purpose of the Study:

• To develop and validate a multi-dimensional COmmon REactivity PAttern (COREPA) modeling approach for predicting chemical binding to hER.
• To identify distinct chemical interaction types with hER based on reactivity patterns and molecular descriptors.
• To establish predictive models for various relative binding affinity (RBA) ranges.

Main Methods:

• Utilized a training set of 645 chemicals, including steroid and environmental compounds.
• Implemented multivariate COREPA analysis with optimized conformational generation and local hydrophobicity description.
• Analyzed reactivity patterns based on nucleophilic site distances to define interaction types (A-B, A-C, A-B-C/AD).

Main Results:

• Identified three distinct interaction types (A-B, A-C, A-B-C/AD) governing chemical binding to hER.
• Developed COREPA models for specific RBA ranges (>10%, 10-0.1%, 0.1-0.01%), incorporating prefiltering requirements and classification nodes.
• Observed reduced distinction between interaction types at lower RBA ranges, necessitating class-specific modeling (e.g., phenols, phthalates).

Conclusions:

• The enhanced multi-dimensional COREPA approach provides a powerful framework for predicting hER binding affinity and mechanism.
• The identified interaction types and affinity-based models offer valuable insights into structure-activity relationships for hER ligands.
• The ultimate model, structured as a battery of local models, enables accurate prediction across diverse chemical structures and activity levels.