1. The central control of sensory inputs from a proprioceptor [chordotonal organ (CO)] in the second joint [coxo-basipodite (CB)] of the fifth leg was studied in crayfish in vitro preparations (Fig. 1A). Simultaneous intracellular recordings from CBCO terminals (CBT) and postsynaptic motoneurons (MNs) were performed along with micropipette pressure ejection or bath application of gamma-aminobutyric acid (GABA), to study the presynaptic mechanisms at work in the CBT (Fig. 1B). 2. Two intracellular recordings were used to show that the spikes never overshoot, and that the more central the recording site within the neuropile, the smaller the spikes (Fig. 2). Only electronic conduction occurs, therefore, in the sensory afferents within the ganglion. 3. Pressure ejection of GABA close to the recording site of CBTs in the ganglion (Fig. 3A) gave rise to a membrane depolarization, the reversal potential of which was about -25 mV (Fig. 7), as well as to an increase in the membrane conductance (Fig. 3C) and a decrease in the orthodromic spike amplitude; moreover, it did not elicit either hyperpolarization, or any change in the membrane conductance of the postsynaptic MN (Fig. 3B), which indicates that pressure ejection of GABA affected only a restricted area around the CBT and not the postsynaptic MNs. 4. In CBT, spontaneous primary afferent depolarizations (PADs) occurred irregularly when the activity of the preparation was not rhythmic (Fig. 4A), and in bursts when the preparation displayed fictive locomotion (Fig. 4B). In the latter case, anti-dromic spikes were sometimes superimposed on PADs (Fig. 4D). The amplitude of the PADs was reduced when picrotoxin (PTX), a GABA antagonist, was applied (Fig. 5), which suggests that GABA may be involved in spontaneous PADs. The reversal potential of PADs was about - 25 mV (Figs. 6 and 7). 5. During simultaneous recordings from a CBT and a monosynaptically related MN, GABA applied by pressure ejection close to the CBT (Fig. 8A) completely suppressed the excitatory postsynaptic potentials (EPSPs) elicited by CBT spikes in the MN (Fig. 8, B and D). This was due to a presynaptic mechanism because no change in the membrane potential or membrane conductance was observed in the MN (Fig. 8C) and most of the CBTs associated with a given MN were affected (Fig. 9). 6. Simultaneously recording from a CBT and a monosynaptically related MN demonstrated that, during bouts of PADs, the spike amplitude decreased in proportion to the PAD amplitude (Fig. 10A). Moreover, in some experiments, where strong monosynaptic coupling between a CBT and a MN was observed, an orthodromic spike led to smaller EPSPs when it occurred at the same time as a PAD (Fig. 10, B and C). 7. Antidromic spikes superimposed on PADs were analyzed. Such antidromic spikes are generated in the CB sensory fibers at active sites out of the ganglion and indicate interactions between PADs and those sites. Nevertheless, in all experiments in which orthodromic spikes generated EPSPs in the corresponding MN, antidromic spikes evoked by PADs had no postsynaptic effects. 8. We conclude that GABA-mediated PADs are responsible for presynaptic inhibition in CBT. The involved mechanisms result in an increased attenuation of potentials along the CBTs, which completely or almost completely suppresses the electrotonic spike in the dendrite endings, where the synapses with the MNs are probably localized. This shunting mechanisms probably also affects antidromic spikes generated by PADs and is responsible for their having no postsynaptic effects.
