Ecological and Structural Drivers Underlying Species-Specific Estrogen Receptor α Responses to Estrogenic Chemicals in Mammals

Species-specific sensitivity to estrogenic chemicals complicates cross-species extrapolation in environmental risk assessment, yet its mechanistic basis remains poorly understood. Here, we investigated how ecological traits and amino acid variations in estrogen receptor alpha (ERα) mediate responses to estrogenic chemicals across mammals using an integrated New Approach Methodologies (NAMs) framework combining bioinformatics, in silico structural modeling, machine learning (ML), and in vitro transactivation assays. Two-way clustering of 169 mammalian ERα sequences and 107 ecological traits revealed distinct groupings associated with diet, habitat, and reproductive strategies. Homology model analysis showed that omnivorous species tend to possess larger ligand-binding pockets than carnivores and herbivores. In vitro assays of ERα from six mammalian species demonstrated substantial interspecies variability in response to natural and synthetic estrogenic chemicals. A cetacean-specific substitution from Asn to Ser at position 349 (corresponding to Asn349 in human ERα), which alters hydrogen bonding with Tyr537 in helix-12, was associated with aquatic lifestyles and altered receptor responsiveness. Additionally, ML models identified brain-to-body mass ratio as a candidate predictor of ERα-ligand binding domain sequence variation and ERα-mediated in vitro transactivation sensitivity across species. These findings identify ecological and structural drivers of species-specific estrogen sensitivity and provide a mechanistic framework for improving cross-species prediction of chemical effects in environmental risk assessment.