The voltage-gated potassium channel in T lymphocytes, Kv1.3, an important target for immunosuppressants, is blocked by picomolar concentrations of the polypeptide ShK toxin and its analogue ShK-Dap22. ShK-Dap22 shows increased selectivity for Kv1.3, and our goal was to determine the molecular basis for this selectivity by probing the interactions of ShK and ShK-Dap22 with the pore and vestibule of Kv1.3. The free energies of interactions between toxin and channel residues were measured using mutant cycle analyses. These data, interpreted as approximate distance restraints, guided molecular dynamics simulations in which the toxins were docked with a model of Kv1.3 based on the crystal structure of the bacterial K+-channel KcsA. Despite the similar tertiary structures of the two ligands, the mutant cycle data imply that they make different contacts with Kv1.3, and they can be docked with the channel in configurations that are consistent with the mutant cycle data for each toxin but quite distinct from one another. ShK binds to Kv1.3 with Lys22 occupying the negatively charged pore of the channel, whereas the equivalent residue in ShK-Dap22 interacts with residues further out in the vestibule, producing a significant change in toxin orientation. The increased selectivity of ShK-Dap22 is achieved by strong interactions of Dap22 with His404 and Asp386 on Kv1.3, with only weak interactions between the channel pore and the toxin. Potent and specific blockade of Kv1.3 apparently occurs without insertion of a positively charged residue into the channel pore. Moreover, the finding that a single residue substitution alters the binding configuration emphasizes the need to obtain consistent data from multiple mutant cycle experiments in attempts to define protein interaction surfaces using these data.