Predicting the Location of Channel Blocker Binding Sites By Simultaneously Applying Two Inhibitors.

Andrew Thompson2, Sarah Lummis2, Gavin Jarvis1. 1School of Pharmacy, Queens University Belfast, Belfast, UK, 2Department of Biochemistry, University of Cambridge, Cambridge, UK

When two channel blockers are applied separately they have similar inhibitory effects on ion conductance regardless of the locations of their binding sites within the channel. We hypothesise that if two blockers are applied simultaneously, the extent of channel inhibition will differ according to whether they bind to shared or separate sites. To identify the binding sites of novel compounds, we developed an analytical method that utilises the observed extent of inhibition by two compounds applied individually to predict the expected extent of inhibition when the same compounds are applied simultaneously, according to whether they act at discrete or overlapping binding sites.

We define two hypothetical models for predicting the effect of a pair of channel blockers applied simultaneously when they have: (1) Separate binding sites: the binding of the two blockers is independent and both compounds may bind at the same time; (2) Shared binding sites: the blockers effectively compete with each other such that only one or other compound can bind at any one time. We show that the simultaneous application of two channel blockers would result in different levels of blockade depending on whether they bind according to Model 1 or Model 2. The difference in inhibition between the two Models = [In1In2(1 – In1)(1 – In2)]/[1- In1In2], where In1 is the level of current inhibition caused by blocker 1 alone and In2 the level of inhibition caused by blocker 2 alone. The Models predict that the maximum difference would occur when In1 and In2 are each ∼60%. Under these circumstances, the predicted level of dual inhibition for Model 1 is 84% and for Model 2 it is 75%. Hence, the maximum possible observable difference is approximately 9% of the maximal conductance.

We tested our analysis using diltiazem (DZM), bilobalide (BB) and ginkgolide B (GB) which have known binding sites in the 5-HT3 receptor channel (Thompson, 2011). 5-HT3A was expressed in Xenopus oocytes and responses measured using Two Electrode Voltage Clamp (Thompson, 2011). Channels were activated with a supra-maximal concentration of 5-HT and inhibition measured for each of the compounds alone and in combination.

To simulate compounds that bind at the same site, DZM was applied as two independent solutions causing inhibition of 81.4±3.0% (mean±sd: n = 7). Model 2 more accurately predicted the extent of inhibition (80.8±1.7%: p = 0.85, ANOVA) than Model 1 (89.1±1.4%: p = 0.03, ANOVA). Similar results were obtained when GB and BB, which compete for the same site, were co-applied. When BB and DZM were co-applied the observed level of inhibition (90.8±2.7%) was more closely predicted by Model 1 (85.2±1.3%: p = 0.04, ANOVA) than Model 2 (76.7±1.4%: p<0.001, ANOVA), indicative of their non-overlapping binding sites (Thompson, 2011).

In conclusion, we show that it is possible to distinguish between blockers that bind at shared or separate locations in channels. The approach is simple and may be of particular value where other pharmacological methods (e.g., radioligand binding) are not applicable, or where there are only limited quantities of a test compound.

Thompson A.J. et al., (2011) Mol Pharmacol, 80, 183-190.