McRobb FM et al. (2014), In silico analysis of the conservation of human...

XB-ART-48345

Environ Sci Technol 2014 Jan 01;483:1964-72. doi: 10.1021/es404568a.

Show Gene links Show Anatomy links

In silico analysis of the conservation of human toxicity and endocrine disruption targets in aquatic species.

McRobb FM , Sahagún V , Kufareva I , Abagyan R .

Abstract
Pharmaceuticals and industrial chemicals, both in the environment and in research settings, commonly interact with aquatic vertebrates. Due to their short life-cycles and the traits that can be generalized to other organisms, fish and amphibians are attractive models for the evaluation of toxicity caused by endocrine disrupting chemicals (EDCs) and adverse drug reactions. EDCs, such as pharmaceuticals or plasticizers, alter the normal function of the endocrine system and pose a significant hazard to human health and the environment. The selection of suitable animal models for toxicity testing is often reliant on high sequence identity between the human proteins and their animal orthologs. Herein, we compare in silico the ligand-binding sites of 28 human "side-effect" targets to their corresponding orthologs in Danio rerio, Pimephales promelas, Takifugu rubripes, Xenopus laevis, and Xenopus tropicalis, as well as subpockets involved in protein interactions with specific chemicals. We found that the ligand-binding pockets had much higher conservation than the full proteins, while the peroxisome proliferator-activated receptor γ and corticotropin-releasing factor receptor 1 were notable exceptions. Furthermore, we demonstrated that the conservation of subpockets may vary dramatically. Finally, we identified the aquatic model(s) with the highest binding site similarity, compared to the corresponding human toxicity target.

PubMed ID: 24392850
PMC ID: PMC3951377
Article link: Environ Sci Technol
Grant support: [+]

Species referenced: Xenopus tropicalis Xenopus laevis
Genes referenced: crh pomc

Article Images: [+] show captions

References [+] :

Abagyan, Biased probability Monte Carlo conformational searches and electrostatic calculations for peptides and proteins. 1994, Pubmed

	Figure 1. Variations in sequence conservation across the sequence of the AR for D. rerio, P. promelas, T. rubripes, X. laevis, and X. tropicalis compared to the human AR (binding site residues highlighted in cyan). All sequences were window averaged across 25 residues. Abbreviations: AF1/2, activation function 1/2; DBD, DNA binding domain; LBD, ligand binding domain.14
	Figure 2. Sequence similarity (percentage and color) and sequence identity (number of identical residues/number of aligned residues is shown in parentheses) for the 28 toxicity target proteins of (a) the full sequence and (b) the ligand-contact residues conserved for 80% of the cocrystallized ligands. White spaces indicate that no ortholog was identified (often due to an incomplete proteome).
	Figure 3. (a) Sequence similarity (percentage and color) and sequence identity (number of identical residues/number of aligned residues is shown in parentheses) for the three A2AR subpockets (white spaces indicate that no ortholog was identified). A2AR crystal structures (gray ribbons), all cocrystallized ligands (mesh), and subpocket (solid surface); (b) subpocket 1 (agonist-bound structures), (c) subpocket 2 (the endogenous agonist-bound structure), and (d) subpocket 3 (antagonist-bound structures).