Plasticity in the multifunctional buccal central pattern generator of Helisoma illuminated by the identification of phase 3 interneurons.

1. The mechanism for generating diverse patterns of buccal motor neuron activity was explored in the multifunctional central pattern generator (CPG) of Helisoma. The standard pattern of motor neuron activity, which results in typical feeding behavior, consists of three distinct phases of buccal motor neuron activity. We have previously identified CPG interneurons that control the motor neuron activity during phases 1 and 2 of the standard pattern. Here we identify a pair of interneurons responsible for buccal motor neuron activity during phase 3, and examine the variability in the interactions between this third subunit and other subunits of the CPG. 2. During the production of the standard pattern, phase 3 excitation in many buccal motor neurons follows a prominent phase 2 inhibitory postsynaptic potential. Therefore phase 3 excitation was previously attributed to postinhibitory rebound (PIR) in these motor neurons. Two classes of observations indicated that PIR was insufficient to account for phase 3 activity, necessitating phase 3 interneurons. 1) A subset of identified buccal neurons is inhibited during phase 3 by discrete synaptic input. 2) Other identified buccal neurons display discrete excitation during both phases 2 and 3. 3. A bilaterally symmetrical pair of CPG interneurons, named N3a, was identified and characterized as the source of phase 3 postsynaptic potentials in motor neurons. During phase 3 of the standard motor pattern, interneuron N3a generated bursts of action potentials. Stimulation of N3a, in quiescent preparations, evoked a depolarization in motor neurons that are excited during phase 3 and a hyperpolarization in motor neurons that are inhibited during phase 3. Hyperpolarization of N3a during patterned motor activity eliminated both phase 3 excitation and inhibition. Physiological and morphological characterization of interneuron N3a is provided to invite comparisons with possible homologues in other gastropod feeding CPGs. 4. These data support a model proposed for the organization of the tripartite buccal CPG. According to the model, each of the three phases of buccal motor neuron activity is controlled by discrete subsets of pattern-generating interneurons called subunit 1 (S1), subunit 2 (S2), and subunit 3 (S3). The standard pattern of buccal motor neuron activity underlying feeding is mediated by an S1-S2-S3 sequence of CPG subunit activity. However, a number of "nonstandard" patterns of buccal motor activity were observed. In particular, S2 and S3 activity can occur independently or be linked sequentially in rhythmic patterns other than the standard feeding pattern. Simultaneous recordings of S3 interneuron N3a with effector neurons indicated that N3a can account for phase-3-like postsynaptic potentials (PSPs) in nonstandard patterns. The variety of patterns of buccal motor neuron activity indicates that each CPG subunit can be active in the absence of, or in concert with, activity in any other subunit. 5. To explore how CPG activity may be regulated to generate a particular motor pattern from the CPG's full repertoire, we applied the neuromodulator serotonin. Serotonin initiated and sustained the production of an S2-S3 pattern of activity, in part by enhancing PIR in S3 interneuron N3a after the termination of phase 2 inhibition.