The molecular origin of receptor-based developmental and reproductive toxicity

Vajda and Beglov groups, in collaboration with the Boston University Superfund Research Project

Supported by:
NIEHS P42ES007381 Receptor-Based Developmental And Reproductive Toxicity Of Superfund Chemicals (PI: D. Ozonoff)

The Biomolecular Engineering Research Center is part of the Bioinformatics and Molecular Modeling Research Support Core, which offers computational tools, expertise, and services to a number of collaborative projects supported by the Boston University Superfund Research Program (BUSRP, see Our main contribution is the modeling interactions between xenobiotics and protein receptors, primarily nuclear receptors and cytochrome P450s, using structural bioinformatics and computational biology, including tools originally developed for structure-based drug design. In some cases the receptors are well characterized, and their three-dimensional structures have been determined either for the proteins themselves or their homologues. The availability of x-ray structures makes it possible to study receptor-ligand interactions at the atomic level using methods of molecular modeling and structural bioinformatics; and to screen databases of environmental chemicals for compounds most likely to bind to specific receptors. Inter-species differences in responses to various xenobiotics can be explained and predicted.

We have been working with a number of BUSRP groups. In collaboration with Drs. Stegeman  and Goldstone (Woods Hole Oceanographic Institution) we have developed a novel approach to the analysis of binding to cytochrome P450s (CYPs), a family of enzymes in the body involved in hormone synthesis and breakdown of hormones and toxic substances. The method generates ensembles of homology models used for docking, which also yields a distribution of ligand positions rather than a single state. The method was used to identify the structural bases underlying the inter-species catalytic differences between related cytochromes P450 (CYPs) in the metabolism of TCB, TCDD, and B[a]P. Our more recent collaboration focuses on the pregnane X receptor (PXR).

In collaboration with Dr. Waxman (BU Department of Biology) we have used computational solvent mapping to identify the hot spots in the ligand-binding domain of peroxisome proliferator activated receptor gamma (PPARγ), a key signaling molecule activated by a variety of environmental chemicals. We also studied the interactions between PPARγ and phthalate esters that are ubiquitous environmental contaminants. Several putative PPARγ-binding phthalates were identified using molecular docking and high throughput virtual screening, including compounds that were known to be agonists.

Fig. 1 A. Binding hot spots of PPAR-γ, represented by probe clusters, superimposed on the partial agonist 5-chloro-1-(4-chlorobenzyl)-3-(phenylthio)-1h-indole-2-carboxylic acid (ntzd), co-crystallized with PPAR-γ (PDB ID 2U5S). B. Docked position of tryphenyl phosphate (TPP), clearly binding to the same hot spot as the known partial agonist nTZDpa.

We currently work with Drs. Webster and Schlezinger (Boston University School of Public Health), who hypothesized that organophosphate flameretardants (OPFRs) and organophosphate plasticizers (OPPs) may have obesogenic effects due to their binding to PPARγ. Since both OPFRs and OPPs are present in the environment in relatively high concentrations, their binding to PPARγ would certainly be an issue. Computational solvent mapping of the PPAR-γ ligand binding domain (LBD)reveals two main binding regions within the binding site: (1) at the polar headgroup of thiazolidinediones (TZDs), strong agonists that interact with to the H12 helix, and (2) between the distal end of TZDs and the entrance of the binding site. The latter consists of three hot spots, in good agreement with the ring positions in several partial agonists such as 5-chloro-1-(4-chlorobenzyl)-3-(phenylthio)-1h-indole-2-carboxylic acid (also called nTZDpa, non-TZD partial agonist) (Fig. 1A)It was shown that a number of selective modulators, e.g., various indoles, activate PPAR-γ using a H12 independent mechanism. The selective modulators have around 25% of transcriptional activity of TZDs, and while they are less adipogenic and obesogenic than the TZDs, they still lead to body weight gain. The majority of organophosphate flame retardants such as triphenyl phosphate have structures with three rings (or ring-like structures) that overlap with the three hot spots (Fig. 1B). Since ligands that bind to a site generally overlap well with the binding hot spots, it is thus very likely that the organophosphates bind similarly to the non-TZD selective agonists. The latter are less obesogenic than the TZDs, and hence their effects are expected to be moderate. However, tri-substituted organotins such as triphenyltin that have also been used in flame retardants are known to be obesogenic, and their structures are similar to those of OPFRs, the similarity of binding is likely to imply that organophosphates are similarly obesogenic. The potential interactions between various organophosphates and PPARγ will be further studied by experimental and computational methods.