Use of artificial intelligence in structure—affinity correlations of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) receptor ligands

Publisher Website
Google Scholar
Abstract
The Computer-Automated Structure Evaluation (CASE) Program, an expert system that automatically selects relevant descriptors for structure-activity relationships, has been used to analyze the binding of various ligands to the tetrachloro-dibenzo-p-dioxin (TCDD) receptor or Ah receptor. Two databases were analyzed. One database contained 136 polycyclic aromatic hydrocarbons (PAH), substituted dibenzo-p-dioxins, dibenzofurans and biphenyls whose binding affinities were measured by a sucrose density gradient technique. The other 87 compound database contained PAH, nitro-PAH, halo-PAH and N-heterocycles. Their binding affinities were measured by the electrofocusing assay. Within each training set significant correlations between the affinity for the TCDD receptor and relevant molecular fragments identified by the CASE program were observed. Among the halogenated aromatic hydrocarbons, fragments containing lateral halogens and a longitudinal hydrogen appeared important for TCDD receptor binding. The fragments of PAH and heterocyclic compounds that were most activating with respect to TCDD receptor binding were found to contain the classical ‘bay’ region and were in fact identical to the fragments found previously to be related to carcinogenkity. It was found that the activating fragments from PAH and heterocyclic compounds were different from those found within the halogenated compounds such as dibenzo-p-dioxins, dibenzofurans and biphenyls. One interpretation of the data is that two different recognition sites may be involved in Ah receptor binding.

This publication has 0 references indexed in Scilit: