Postsynaptic polypeptide neurotoxins isolated from the venoms of elapid and hydrophid snakes exhibit the ability to bind selectively to and inhibit different types of receptors that function in nerve signal transmission. On the basis of their amino acid-sequences and three-dimensional structures, these neurotoxins are clearly related, but nothing is yet known about the basis for their physiological receptor specificity. In this report, site-directed mutants of kappa-bungarotoxin, produced by an Escherichia coli expression system, are tested to determine the function of selected amino acid side chains in the interaction between toxin and receptor. Highly conserved residues at the bottom of the second loop (a region that has been shown to be a major point of contact with the receptor), particularly those residues at the junction between the beta-sheet and the end of the loop, were selected. The results demonstrate that a single amino acid substitution of the invariant arginine residue (Arg-40 to Ala-40) renders the toxin unable to inhibit nerve transmission in the chick ciliary ganglion up to a concentration of 10 muM. Significantly, the results also show that conversion to alanine of the nearby proline residue (Pro-42) found to be invariant in all kappa-neurotoxins, but not found in any potent alpha-neurotoxin, produces a toxin with full inhibitory capacity. However, the introduction of a lysine residue at this position (P-42-K), like that found in alpha-bungarotoxin, reduces activity significantly, Thus, in contrast to the alpha-neurotoxins, there seems to be a requirement for short uncharged side chains, such as proline or alanine, at this position in kappa-neurotoxins, but the specific presence of the imino acid proline in kappa-bungarotoxin is not strictly required and thus does not affect the orientation of the critical arginine side chain. These results suggest that this residue is at the interface of the toxin-receptor complex. The only residue always found in alpha-neurotoxins and never found in kappa-neurotoxins is Trp-32. It has long been hypothesized to form the basis for receptor specificity between these two families of postsynaptic neurotoxins. Placement of a tryptophan residue in kappa-bungarotoxin at the position where it is found as an invariant residue in all active alpha-neurotoxins also had no effect on kappa-bungarotoxin's ability to block nerve transmission in the ciliary ganglia. Thus, the neuronal receptor is fully capable of accommodating a bulky hydrophobic side chain at this position, and it does not contribute to the specificity of the toxin for the neuronal receptor. However, the absence of Trp-32 in kappa-neurotoxins may contribute to their low affinity for the muscle receptor, since it has been demonstrated that this residue probably plays a role in the interaction of alpha-neurotoxins with muscle receptors.