Voltage-gated potassium channels (Ks) perform vital physiological functions and are targets in different disorders ranging from ataxia and arrhythmia to autoimmune diseases. An important issue is the search for and production of selective ligands of these channels. Peptide toxins found in scorpion venom named KTx excel in both potency and selectivity with respect to some potassium channel isoforms, which may present only minute differences in their structure. Despite several decades of research the molecular determinants of KTx selectivity are still poorly understood. Here we analyze MeKTx13-3 (Kalium ID: α-KTx 3.19) from the lesser Asian scorpion , a high-affinity K1.1 blocker (IC ~2 nM); it also affects K1.2 (IC ~100 nM), 1.3 (~10 nM) and 1.6 (~60 nM). By constructing computer models of its complex with K1.1-1.3 channels we identify specific contacts between the toxin and the three isoforms. We then perform mutagenesis to disturb the identified contacts with K1.1 and 1.2 and produce recombinant MeKTx13-3_AAAR, which differs by four amino acid residues from the parent toxin. As predicted by the modeling, this derivative shows decreased activity on K1.1 (IC ~550 nM) and 1.2 (~200 nM). It also has diminished activity on K1.6 (~1500 nM) but preserves K1.3 affinity as measured using the voltage-clamp technique on mammalian channels expressed in oocytes. In effect, we convert a selective K1.1 ligand into a new specific K1.3 ligand. MeKTx13-3 and its derivatives are attractive tools to study the structure-function relationship in potassium channel blockers.