Specificity of cholesterol and analogs to modulate BK channels points to direct sterol–channel protein interactions

The activity (Po) of large-conductance voltage/Ca(2+)-gated K(+) (BK) channels is blunted by cholesterol levels within the range found in natural membranes. We probed BK channel-forming α (cbv1) subunits in phospholipid bilayers with cholesterol and related monohydroxysterols and performed computational dynamics to pinpoint the structural requirements for monohydroxysterols to reduce BK Po and obtain insights into cholesterol's mechanism of action. Cholesterol, cholestanol, and coprostanol reduced Po by shortening mean open and lengthening mean closed times, whereas epicholesterol, epicholestanol, epicoprostanol, and cholesterol trisnorcholenic acid were ineffective. Thus, channel inhibition by monohydroxysterols requires the β configuration of the C3 hydroxyl and is favored by the hydrophobic nature of the side chain, while having lax requirements on the sterol A/B ring fusion. Destabilization of BK channel open state(s) has been previously interpreted as reflecting increased bilayer lateral stress by cholesterol. Lateral stress is controlled by the sterol molecular area and lipid monolayer lateral tension, the latter being related to the sterol ability to adopt a planar conformation in lipid media. However, we found that the differential efficacies of monohydroxysterols to reduce Po (cholesterol≥coprostanol≥cholestanol>>>epicholesterol) did not follow molecular area rank (coprostanol>>epicholesterol>cholesterol>cholestanol). In addition, computationally predicted energies for cholesterol (effective BK inhibitor) and epicholesterol (ineffective) to adopt a planar conformation were similar. Finally, cholesterol and coprostanol reduced Po, yet these sterols have opposite effects on tight lipid packing and, likely, on lateral stress. Collectively, these findings suggest that an increase in bilayer lateral stress is unlikely to underlie the differential ability of cholesterol and related steroids to inhibit BK channels. Remarkably, ent-cholesterol (cholesterol mirror image) failed to reduce Po, indicating that cholesterol efficacy requires sterol stereospecific recognition by a protein surface. The BK channel phenotype resembled that of α homotetramers. Thus, we hypothesize that a cholesterol-recognizing protein surface resides at the BK α subunit itself.