Structures of Endocrine-Disrupting Chemicals Determine Binding to and Activation of the Estrogen Receptor α and Androgen Receptor

Abstract

Endocrine-disrupting chemicals (EDCs) can interact with nuclear receptors, including estrogen receptor α (ERα) and androgen receptor (AR), to affect the normal endocrine system function, causing severe symptoms. Limited studies queried the EDC mechanisms, focusing on limited chemicals or a set of structurally similar compounds. It remained uncertain how hundreds of diverse EDCs could bind to ERα and AR and cause distinct functional consequences. Here, we employed a series of computational methodologies to investigate the structural features of EDCs that bind to and activate ERα and AR based on more than 4000 compounds. We used molecular docking and molecular dynamics simulations to elucidate the functional consequences and validated structure−function correlations experimentally using a time-resolved fluorescence resonance energy-transfer assay. We found that EDCs share three levels of key fragments. Primary (20 for ERα and 18 for AR) and secondary fragments (38 for ERα and 29 for AR) are responsible for the binding to receptors, and tertiary fragments determine the activity type (agonist, antagonist, or mixed). In summary, our study provides a general mechanism for the EDC function. Discovering the three levels of key fragments may drive fast screening and evaluation of potential EDCs from large sets of commercially used synthetic compounds.

Machine Learning: New Ideas and Tools in Environmental Science and Engineering

“…Similarly, there are several attempts to interpret ML models to correlate factors with the chemical activity of EDCs, including unveiling features that make EDCs chemically active, determining the type of activity on the ERα or the AR, and how these features exert their functions. 45 Using ML techniques, researchers also found that the octanol−water partitioning coefficient (log K ow ) plays a dominant role in regulating plant uptake of organic contaminants, while their molecular weight plays a secondary role. 99 In recent years, data-driven analytics such as ML have become key tools for discovery in public health and ESE research to find hidden patterns and causal relationships and to identify key features, that is, chemical exposure or other social economic parameters, that are linked with health outcomes.…”

The VEGAHUB Platform: The Philosophy and the Tools

“…Tertiary fragments are then used to discriminate agonists, antagonists, or mixed agonist and antagonist modes of action. 27 Models for RARs and VDR are characterised by a lower sensitivity. A detailed documentation of each workflow and the associated fragments is available on the VEGAHUB website.…”

Rapid Screening of Human Transthyretin Disruptors through a Tiered in Silico Approach

“…Computational methods could be used to evaluate and screen chemicals on their potential endocrine activity, − which is significant in minimizing the adverse impacts of EDCs on human health and the environment and finally realizing the sustainable assessment and management of EDCs. To date, extensive predictive models were available for the end points related to disruption effects of nuclear receptors. − Even some models for predicting the disruption effects of nuclear receptors have been integrated into several pieces of publicly available software, e.g., VAGE () and OECD QSAR Toolbox (). However, the predictive models for the binding potency of compounds with nonreceptor mediated targets, e.g., hormone transporters, were much less than that of nuclear receptors .…”