1. The desensitizing acetylcholine (ACh) response of bovine chromaffin cells maintained in culture was examined using rapid agonist applications (of 2 s duration) which imposed nominal drug concentrations within 50 ms. This study was aimed, firstly, at identifying which of the alpha 3, alpha 4 and alpha 7 subunits known to be present in these cells is predominant in the ACh-evoked response and secondly, on the effects on these neuronal nicotinic ACh receptors (AChR) of cyclothiazide (CT), an agent acting as a modulator of a gating desensitization site on other ligand-gated channels. 2. Locally applied 100 microM ACh evoked peak currents (IACh) of -1.5 +/- 0.1 nA (n = 83) at a holding potential of -60 mV. The ACh dose-response curve yielded an estimated EC50 of 60 microM. This current was not sustained but desensitized during the application period; it displayed strong inward rectification, but desensitized equally whether the evoked current was inward or outward going. The latter observation excludes alpha 4 as a major contributor to the recorded current. Because the response was almost insensitive to a 1 microM alpha-bungarotoxin pretreatment (IACh = -1.2 +/- 0.1 nA; n = 6), and because 1, 1-dimethyl-4-phenylpiperazinium (DMPP) works as a potent agonist (peak current = -1.9 nA, n = 2 for 100 microM DMPP), the alpha 7 subunit is also a minor contributor to the response. Taken together, these observations suggest a dominant alpha 3 type of response. 3. Triple exponential fits were used to describe the characteristics of the ACh-evoked currents; one component to fit the rising phase, with 2 components to describe the decay phase. The decay times were 100 ms and 4 s for the fast and slow components respectively. The rate of the slow decay component increased systematically with recording time, approximately doubling from its initial value within 20-40 min. Furthermore there was a gradual rundown of the response, seen first as a loss of the late component of the current, measured at 2 s, with the peak current amplitude decreasing later in the recording.4. CT, when coapplied with ACh, produced a dose-dependent inhibition of the ACh-evoked peak current. The effect showed little voltage-dependency with 100 microM CT producing 46 +/- 5% (s.d.; n = 3)and 47 +/- 8% (s.d.; n = 7) inhibition at -100 and -60 mV respectively. At + 60 mV, inhibition was estimated to be 26 +/- 7% (s.d.; n = 3).5. After pre-exposure of the cells to CT by bath application, 10 and 30 microM CT produced poorly reversible 20 +/- 9% (n = 7) and 42 +/- 5% (n = 4) inhibitions of the peak current respectively. There were no discernible effects on the fitted decay constants at any CT concentration tested, although an increased inhibitory effect of CT was observed at higher concentrations (100 microM) on the amplitude of the late component measured at 2 s.6. Similar effects were observed in conditions chosen to isolate the alpha 3 type of receptor: namely when using DMPP as an agonist, or after a-bungarotoxin pretreatment.7. The 2,3-benzodiazepine, GYKI 53655, is known to antagonize the action of CT on AMPA receptors.Coapplication of 50 microM GYKI 53655 with ACh (100 microM) produced a 29 +/- 4% inhibition of the peak ACh-evoked current and 44 +/- 6% inhibition of its amplitude at 2 s (n = 4). This response was fully reversible. Brief applications of both CT (100 microM) and GYKI 53655 (50 microM) with ACh via the microperfusion system produced a fully reversible inhibition that was not significantly different from the values obtained with either CT or GYKI 53655 alone, with 37 +/- 6% inhibition of peak and 61 +/- 9%inhibition of the amplitude at 2 s (n = 3).8. The results obtained suggest that the CT effect is to impede recovery from a slow desensitization,with a more pronounced effect with longer CT applications. Globally, CT favours the 'rundown state' of the neuronal nicotinic AChR.