STRUCTURE-ACTIVITY STUDIES ON SCORPION TOXINS THAT BLOCK POTASSIUM CHANNELS

Scorpion venoms contain toxins that block different types of potassium channels. Some of these toxins have affinity for high conductance Ca2 +-activated K+ channels and for dendrotoxin-sensitive voltage-dependen t K+ channels. The structural features that determine the specificity of binding to different channel types are not known, We investigated t his using natural and synthetic scorpion toxins. We have tested the ef fects of charybdotoxin (CTX) and two homologues (Lqh 15-1 and Lqh 18-2 ), iberiotoxin (IbTX), and kaliotoxin (KTX) from the scorpions Leiurus quinquestriatus hebreus, Buthus tamulus and Androctonus mauretanicus mauretanicus, respectively, and synthetic variants of CTX, namely CTX( 2-37), CTX(3-37), CTX(4-37), and CTX(7-37), on a Ca2+-activated K+ cur rent (I-K-Ca) at a mammalian motor nerve terminal, and on the binding of a radiolabelled dendrotoxin, I-125-DpI, to voltage-dependent K+ cha nnels on rat brain synaptosomal membranes. The native toxins contain 3 7-38 amino acid residues, they are over 30% identical in sequence (CTX and IbTX are 68% identical), and they have similar three-dimensional conformations, All toxins, except IbTX, displaced I-125-DpI from its s ynaptosomal binding sites: Lqh 18-2 (K-i = 0.25 nM), KTX (K-i = 2.1 nM ), CTX (K-i = 3.8 nM), CTX(2-37), (K-i = 30 nM), Lqg 15-1 (K-i = 5O nM ), CTX(3-37) (K-i = 60 nM), CTX(4-37) (K-i = 50 nM), CTX(7-37) (K-i = 105 nM). IbTX had no effect at 3 mu M. When variants of CTX with delet ions at the N-terminal portion were tested for their activity on I-K-C a on motor nerve terminals in mouse triangularis sterni nerve-muscle p reparations, CTX(3-37) and CTX(4-37) were ineffective at 100 nM; and C TX7-37 was ineffective at up to 1 mu M. IbTX and CTX (100 nM) complete ly blocked I-K-Ca, but KTX (100 nM) did not affect the nerve terminal I-K-Ca. Different residues appear to be important for interactions of the toxins with different K+ channels. IbTX did not displace dendrotox in binding, but it did block I-K-Ca, whereas KTX was as active as CTX against dendrotoxin binding but it did not affect the I-K-Ca of the mo tor nerve terminals. The N-terminal section of the toxins appears to b e particularly involved in block of I-K-Ca at the motor nerve terminal : it is truncated in the inactive synthetic CTX variants; and it is po sitively charged at lysine-6 in KTX (which is inactive), but negativel y charged in IbTX and neutral in CTX. Phenylalanine at position 2 seem s to be essential: there is a marked loss in activity between CTX(2-37 ) and CTX(3-37), and KTX has valine-2. Phenylalanine may be important in the beta-sheet region of charybdotoxin and iberiotoxin. For binding to dendrotoxin sites, the inactive IbTX lacks the conserved asparagin e residues at positions 4 and 30, and contains additional negatively c harged residues at positions 4, 6 and 24. The side-chains of these res idues are on the opposite face of the molecule from the positively cha rged residues in the beta-sheet region (namely Arg 25, Lys 27 and Arg 34).