A helical capping motif in ShK toxin and its role in helix stabilization.

ShK toxin, a 35-residue polypeptide cross-linked by three disulfides, is a potent blocker of voltage-gated potassium channels and is of interest as a lead in the development of new immunosuppressant agents. ShK toxin contains two short stretches of alpha-helix, the first of which is preceded by a putative N-capping box encompassing residues Thr13 and Gln16. (1)H and (13)C NMR data support the presence of this structural motif, but the hydrogen bonds involving residues 13 and 16 in the solution structure of ShK toxin do not match the pattern expected for a conventional N-cap motif. They do, however, fit the pattern for the recently described ST-motif, class 4a (Wan and Milner-White (1999) Journal of Molecular Biology, 1999, Vol. 286, pp. 1651-1662). The (1)H NMR chemical shifts, nuclear Overhauser effects, and amide exchange rates of native ShK toxin are compared with those of three synthetic analogues with the substitutions Thr13 to Ala and Gln16 to Glu and Ala in order to determine the contribution of this motif to the structure and stability of ShK toxin. Disruption of the capping interactions destabilizes the helices, with the Thr13 to Ala substitution being much more disruptive than Gln16 to Ala, consistent with the lack of hydrogen bonding to the side chain of residue i + 4 in a class 4a ST-motif. Mutation of residues 13 and 16 has only a minor effect on potassium channel binding, probably because the disulfide bonding network minimizes the effect of loss of the capping motif on the overall structure. The implications of these findings for the design of ShK analogues are discussed.

Ionisation behaviour and solution properties of the potassium-channel blocker ShK toxin.

The effects of pH, temperature and polypeptide concentration on the solution structure and side chain interactions of ShK toxin, a potassium-channel-blocking polypeptide from the sea anemone Stichodactyla helianthus, have been investigated by means of one-dimensional and two-dimensional 1H-NMR spectroscopy. Resonance assignments have been obtained for most protons in the molecule, and for the alpha and beta carbon atoms. The lack of concentration dependence of the 1H chemical shifts and linewidths indicates that self-association is not significant and cannot account for the sheet-like structure near the N terminus. The structure is stable to high temperature, showing little change even at 353 K. This stability allowed backbone-amide temperature coefficients to be interpreted, and the correlation of these values with hydrogen bonds observed in the structures and with solvent exchange rates is discussed. pKa values have been measured for Asp5, His19 and Tyr23, and the contributions to these pKa values from other residues investigated using the analogues R11Q (denoting substitution of Argll with Gln), R11E, H19K, K22A, Y23A and K30A. These results show that Asp5 (pKa 2.8) makes an electrostatic interaction with Lys30, which may be partially responsible for the importance of these side chains in the folding of synthetic toxin. The phenolic pKa of Tyr23 is reduced to 8.7 in the native toxin, as a result of interactions with the positively charged side chains of Arg11 and to a lesser extent Lys22. Several hydrogen bonds between the Arg11 guanidino group and the Tyr23 phenolic group are found in the solution structures. As these three residues are implicated in the tight binding of ShK toxin to the T-lymphocyte voltage-gated potassium channel Kv1.3, their close interactions should be taken into account in models of binding of this toxin to the pore and vestibule of this and other potassium channels.