We have recently shown that an arylaminopyridazine derivative of GABA, SR 95103 [2-(3-carboxypropyl)-3-amino-4-methyl-6-phenylpyridazinium chloride], is a selective and competitive GABA-A receptor antagonist. In order to further explore the structural requirements for GABA receptor affinity, we synthesized a series of 38 compounds by attaching various pyridazinic structures to GABA or GABA-like side chains. Most of the compounds displaced [3H]GABA from rat brain membranes. All the active compounds antagonized the GABA-elicited enhancement of [3H]diazepam binding, strongly suggesting that all these compounds are GABA-A receptor antagonists. None of the compounds that displaced [3H]GABA from rat brain membranes interacted with other GABA recognition sites (GABA-B receptor, GABA uptake binding site, glutamate decarboxylase, GABA-transaminase). They did not interact with the Cl- ionophore associated with the GABA-A receptor and did not interact with the benzodiazepine, strychnine, and glutamate binding sites. Thus, these compounds appear to be specific GABA-A receptor antagonists. In terms of structure-activity, it can be concluded that a GABA moiety bearing a positive charge is necessary for optimal GABA-A receptor recognition. Additional binding sites are tolerated only if they are part of a charge-delocalized amidinic or guanidinic system. If this delocalization is achieved by linking a butyric acid moiety to the N(2) nitrogen of a 3-aminopyridazine, GABA-antagonistic character is produced. The highest potency (approximately equal to 250 times bicuculline) was observed when an aromatic pi system, bearing electron-donating substituents, was present on the 6-position of the pyridazine ring.