1. The GABA modulating and GABA-mimetic actions of the general anaesthetic etomidate were examined in voltage-clamp recordings performed on Xenopus laevis oocytes induced, by cRNA injection, to express human recombinant gamma-aminobutyric acidA (GABAA) receptor subunits. 2. Currents mediated by recombinant receptors with the ternary subunit composition alpha x beta y gamma 2L (where x = 1,2,3 or 6 and y = 1 or 2), in response to GABA applied at the appropriate EC10, were enhanced by etomidate in a manner that was dependent upon the identity of both the alpha and beta subunit isoforms. 3. For the beta 2-subunit containing receptors tested, the EC50 for the potentiation of GABA-evoked currents by etomidate (range 0.6 to 1.2 microM) was little affected by the nature of the alpha subunit present within the hetero-oligomeric complex. However, replacement of the beta 2 by the beta 1 subunit produced a 9-12 fold increase in the etomidate EC50 (6 to 11 microM) for all alpha-isoforms tested. 4. For alpha 1, alpha 2 and alpha 6, but not alpha 3-subunit containing receptors, the maximal potentiation of GABA-evoked currents by etomidate was greater for beta 2- than for beta 1-subunit containing receptors. This was most clearly exemplified by receptors composed of alpha 6 beta 1 gamma 2L compared to alpha 6 beta 2 gamma 2L subunits, where a maximally effective concentration of etomidate potentiated currents evoked by GABA at EC10 to 28 +/- 2% and 169 +/- 4% of the maximal GABA response, respectively. 5. For alpha 1 subunit-containing receptors, the potency and maximal potentiating effect of either pentobarbitone or propofol was essentially unaffected by the beta subunit isoform contained within the receptor complex. The potency of the anaesthetic neurosteroid 5 alpha-pregnan-3 alpha-ol-20-one was marginally higher for beta 1 rather than the beta 2 subunit-containing receptor, although its maximal effect was similar at the two receptor isoforms. 6. The GABA-mimetic action of etomidate was supported by beta 2- but not beta 1-subunit containing receptors, whereas that of pentobarbitone or propofol was evident with either beta isoform. For beta 2-subunit containing receptors, both the agonist EC50 and the maximal current produced by etomidate were additionally influenced by the alpha isoform. 7. It is concluded that the subtype of beta-subunit influences the potency with which etomidate potentiates GABA-evoked currents and that the beta isoform is a crucial determinant of the GABA-mimetic activity of this compound. The nature of the alpha-subunit also impacts upon the maximal potentiation and activation that the compound may elicit. Such pronounced influences may aid the identification of the site that recognises etomidate. More generally, these results provide a clear example of structural specificity in anaesthetic action.