Voltage-gated potassium channels (Kv) are targets for drugs of large chemical diversity. Although hydrophobic cations block Kv channels with Hill coefficients of 1, uncharged electron-rich (cationophilic) molecules often display Hill coefficients of 2. The mechanism of the latter block is unknown. Using a combination of computational and experimental approaches, we mapped the receptor for the immunosuppressant PAP-1 (5-(4-phenoxybutoxy)psoralen), a high-affinity blocker of Kv1.3 channels in lymphocytes. Ligand-docking using Monte Carlo minimizations suggested a model in which two cationophilic PAP-1 molecules coordinate a K ϩ ion in the pore with their coumarin moieties, whereas the hydrophobic phenoxyalkoxy side chains extend into the intrasubunit interfaces between helices S5 and S6. We tested the model by generating 58 point mutants involving residues in and around the predicted receptor and then determined their biophysical properties and sensitivity to PAP-1 by whole-cell patch-clamp. The model correctly predicted the key PAP-1-sensing residues in the outer helix, the P-loop, and the inner helix and explained the Hill coefficient of 2 by demonstrating that the Kv1.3 pore can accommodate two or even four PAP-1 molecules. The model further explained the voltagedependence of block by PAP-1 and its thousand-fold selectivity for Kv1.3 over non-Kv1 channels. The 23-to 125-fold selectivity of PAP-1 for Kv1.3 over other Kv1 channels is probably due to its preferential affinity to the C-type inactivated state, in which cessation of K ϩ flux stabilizes the tripartite PAP-1:K ϩ : PAP-1 complex in the pore. Our study provides a new concept for potassium channel block by cationophilic ligands.

No abstract

The voltage-gated potassium channel Kv1.3 constitutes an attractive pharmacological target for the treatment of effector memory T cell-mediated autoimmune diseases such as multiple sclerosis and psoriasis. Using 5-methoxypsoralen (5-MOP, 1), a compound isolated from Ruta graveolens, as a template we previously synthesized 5-(4-phenoxybutoxy)psoralen (PAP-1, 2) which inhibits Kv1.3 with an IC50 of 2 nM. Since PAP-1 is more than 1000-fold more potent than 5-MOP, we here investigated whether attaching a 4-phenoxybutoxy side-chain to other heterocyclic systems would also produce potent Kv1.3 blockers. While 4-phenoxybutoxy substituted quinolines, quinazolines and phenanthrenes were inactive, 4-phenoxybutoxy substituted quinolinones, furoquinolines, coumarins or furochromones inhibited Kv1.3 with IC50s of 150 nM to 10 µM in whole-cell patch-clamp experiments. Our most potent new compound is 4-(4-phenoxybutoxy)-7H-furo[3,2-g]chromene-7-thione (73, IC50 17 nM), in which the carbonyl oxygen of PAP-1 is replaced by sulfur. Taken together, our results demonstrate that the psoralen system is a crucial part of the pharmacophore of phenoxyalkoxypsoralen-type Kv1.3 blockers.

C2-Substituted silacyclopent-3-enes have been prepared in good yields in two steps starting from commercially available diallylsilane 9. Coupling between carbanion of 9 and aldehydes or epoxides followed by ring-closing metathesis of the corresponding substituted diallylsilanes led to various silacyclopent-3-enes having a beta- or a gamma-hydroxysilane moiety. Dihydroxylation and epoxidation of the sila-cycle then led stereoselectively to polyhydroxylated silacyclopentanes. These processes were shown to occur with opposite topicity, offering a complementary and stereocontrolled access to diastereomeric polyols having up to five contiguous stereogenic centers.