The stimulation frequency-noradrenaline release relationship was studied in the vas deferens and the cerebral cortex of NMRI mice, mice in which the alpha (2A)-, the, alpha (2B)-, the alpha (2C)- or both the alpha (2A)- and the alpha (2C)-adrenoceptor gene had been disrupted (alpha 2AKO, alpha 2BKO, alpha 2CKO and alpha 2ACKO), and the wildtype mice from which the knockout animals had been generated. Tissue pieces were preincubated with H-3-noradrenaline and then superfused and stimulated electrically with a constant number of pulses (30 in vas deferens and 50 in brain cortex) at frequencies between 0.03 and 100 Hz. The frequency-evoked tritium overflow curves ascended monophasically in the vas deferens of wildtype and NMRI mice. Disruption of the alpha (2B)-adrenoceptor gene caused no change. In the vas deferens of alpha 2CKO mice, the overflow evoked by low frequencies (0.3 and 1 Hz) was slightly increased. In the vas deferens of alpha 2AKO and alpha 2ACKO mice, the evoked overflow was increased to a greater extent. Rauwolscine (1 muM) caused a marked increase of the evoked overflow of tritium from the vas deferens of NMRI, wildtype, alpha 2BKO and alpha 2CKO mice. Rauwolscine also increased the evoked overflow of tritium from the vas deferens of alpha 2AKO and alpha 2ACKO mice, but to a smaller extent. The gene disruptions and rauwolscine slightly steepened the slope of the vas deferens frequency-overflow curve. In the brain cortex of wildtype. and NMRI mice, the frequency-evoked tritium overflow curves were U-shaped. In the brain cortex of alpha 2BKO and alpha 2CKO mice, the evoked overflow was slightly reduced. In the brain cortex of alpha 2AKO and alpha 2ACKO mice, in contrast, the evoked overflow was increased. Rauwolscine (1 muM) caused a marked increase of the evoked overflow of tritium from the brain cortex of NMRI, wildtype, alpha 2BKO and alpha 2CKO mice. Rauwolscine also increased the evoked overflow of tritium from the brain cortex of alpha 2AKO and alpha 2ACKO mice, but to a smaller extent. The gene disruptions and rauwolscine flattened the U shape of the brain cortex frequency-overflow curve. It is concluded that alpha (2)-autoinhibition is one factor that shapes the frequency-noradrenaline release relationships in the mouse vas deferens and cerebral cortex. The autoreceptors are mainly alpha (2A) and to a minor extent, and well detectable in the vas deferens only, alpha (2C). When both the alpha (2A)- and the alpha (2C)-adrenoceptor have been deleted, alpha (2B)-adrenoceptors may be expressed as autoreceptors in noradrenergic neurons. It seems possible that alpha (2C)-autoreceptors depress mainly release at low (around 1 Hz) whereas alpha (2A)-autoreceptors depress mainly release at high (around 10 Hz) frequencies.