A new method that uses liquid chromatography to determine the binding activity of a class of inhibitors of nicotinic acetylcholine receptors has led to the first computer-based model of the inner lumen of a ligand-gated ion channel, according to researchers at the National Institutes of Health.
Last year, Irving Wainer and Krzystof Jozwiak of the National Institute on Aging at NIH published a paper describing their chromatographic method for studying the activity of noncompetitive inhibitors for nicotinic receptors in Analytical Chemistry. The method, according to Wainer, reduces the time it takes to get functional data from a week or more using rubidium uptake or patch-clamp techniques to only a few hours. While competitive ligands such as nicotine are relatively well characterized because they bind on the outside surface of the receptor, “the trick has been that a lot of compounds also bind inside the central lumen, and these are extremely difficult to identify and to characterize,” said Wainer. “There hasn’t really been any good computational method to determine and describe the binding to this non-competitive inhibitor binding of the central lumen and to design new ones.”
Armed with the new chromatographic technique, however, Wainer and Jozwiak were able to generate enough data to build a molecular model of the inner surface of the ion channel and a method of docking a ligand into the active site.
The model will allow researchers to virtually screen chemical libraries against the ion channel to predict the pharmacological importance of the interaction as well as screen for unexpected interactions and toxicities of a drug candidate due to off-target interactions, according to the NIH team.
The nicotinic receptors are only one subgroup of the ligand-gated ion channel superfamily. The model, therefore, “opens up all of these other ion channels because [Jozwiak] found inside the lumen a couple of binding sites that were not known before, which are key to some of the non-competitive inhibition, and we should be able to expand this quite rapidly into the whole family,” Wainer said. “We’ve now described the first tree, and the rest of the forest should be simple to fall in line.”
Wainer said that the model is sensitive enough to computationally predict the functional differences between stereoisomers that bind to a receptor. In one case, he noted, the model predicted a 300 calorie difference in binding energy — out of a total binding energy of around 2 kilocalories — between dextromethorphan and its enantiomer levomethorphan. The prediction indicated the dextromethorphan was more stable, so the complex would remain intact longer and the functional blockage would last longer — an observable functional difference, Wainer noted.
Using the model, Jozwiak has designed three new molecules that NIH expects to have synthesized this fall. “We know from the computer what their stability should be, and we will then run both the functional studies and the chromatographic studies and I’m fairly confident that we will see the predicted results,” Wainer said.
The model is currently available for licensing from NIH. (Contact: Cristina Thalhammer-Reyero, [email protected] mail.nih.gov.)