1. The purpose of these investigations was to examine the role that the gamma 2 subunit plays in human GABA(A) receptor desensitization. Two different recombinant GABA(A) receptors (alpha 1 beta 3 and alpha 1 beta 3 gamma 2) were compared by measuring the relaxation of whole-cell currents during the application of GABA, isoguvacine or taurine. 2. At concentrations which trigger a maximum response (100-500 mu M GABA) the current relaxation usually fitted the sum of two exponentials. For alpha 1 beta 3 subunit receptors these values were tau(1) = 145 +/- 12 ms and tau(2) = 6.3 +/- 2.1 s (means +/- S.E.M). Receptors consisting of alpha 1 beta 3 gamma 2 subunits desensitized faster: tau(1) = 41.6 +/- 8.3 ms and tau(2) = 2.4 +/- 0.6 s. 3. The Hill dope, determined for each receptor subunit combination, was the same and greater than 1.0, implying two binding steps in the activation of both receptor subunit combinations. 4. For alpha 1 beta 3 subunit receptors the fast desensitization rates were unaltered by reducing the GABA concentration from the EC(100) (100 mu M) to the approximate EC(50) values (10-20 mu M), whereas for alpha 1 beta 3 gamma 2 subunit receptors a significant slowing was observed. The fast desensitization disappeared at agonist concentrations below the EC(50) for both subunit combinations. In contrast, the slow desensitization appeared at agonist concentrations near the EC(20). This rate was dependent on agonist concentration reaching a maximum near the EC(60) value of GABA. 5. The fast desensitization rates were unaltered by changing the holding potential of the cell during agonist application. However, for alpha 1 beta 3 gamma 2 subunit receptors the slow desensitization rate increased by approximately 15- to 20-fold over the range of voltages of -60 to +40 mV. This indicates that the gamma 2 subunit makes GABA(A) receptor desensitization voltage dependent. 6. Recovery from desensitization was also biphasic. The first recovery phase was faster for alpha 1 beta 3 gamma 2 than for alpha 1 beta 3 subunit receptors (0.13 vs. 0.03 s(-1), respectively). The second phase of recovery for the two receptors were the same (similar to 0.003 s(-1)). 7. There was only a poor correlation between agonist potency and the degree or time course of desensitization. Isoguvacine (EC(50) approximate to 10 mu M) induced biphasic relaxation for both alpha 1 beta 3 and alpha 1 beta 3 gamma 2 subunit receptors (tau(1) = 288.6 +/- 43.3 and 167 +/- 15 ms, and tau(2) = 8.0 +/- 1.9 and 4.4 +/- 0.4 s, respectively, for each subunit combination). Taurine (EC(50) approximate to 7 mM) usually induced monophasic relaxation for both subunit combinations (tau(2) = 7.1 +/- 1.6 and 23.0 +/- 6.6 s, respectively). 8. A computer model was developed to examine the effect of the gamma 2 subunit on the time course of a synaptic potential. It was found that the gamma 2 subunit theoretically prolongs the time course of a synaptic potential by inducing desensitization more rapidly. The subsequent relaxation of the desensitized receptors through the open state increases P-open (the probability that the GABA(A) receptor is in an open conducting state) altering the time course of the modelled potential. alpha 1 beta 3 subunit receptors do not desensitize sufficiently rapidly to induce this desensitized state and, therefore, are shorter in time course. These data imply that the physiological role of the gamma 2 subunit is to increase synaptic efficacy by prolonging P-open.