Intracellular microelectrodes were used to study a cluster of neurosecretory ‘Light Yellow’ Cells (LYC) in the central nervous system of Lymnaea stagnalis. LYC usually have a spontaneous firing pattern of bursts, lasting 10-600 s, alternating with periods of silence. Experiments on isolated single cells showed that the bursting activity has an endogenous origin.
Each action potential is followed by a depolarizing afterpotential (DAP), with an amplitude of about 10 mV, lasting several seconds. Bursts end with a subthreshold DAP. It is concluded that two pace-maker mechanisms are responsible for the bursting properties, one initiating and the other (the DAP) maintaining the burst.
The relationship between the electrical and the neurosecretory properties of the cells is discussed.
During the last decade several types of neurosecretory cell (N.S.C.) have been studied electrophysiologically. A limited number of these studies concerns N.s.c. in gastropod molluscs.
The electrical properties of N.s.c. in snails can differ considerably. Some are usually silent (Kupfermann & Kandel, 1970; Benjamin, Swindale & Slade, 1976), while others show various patterns of spontaneous activity (Frazier et al. 1967; Gainer, 1972; Benjamin & Swindale, 1975; Benjamin, 1978). One type of spontaneous activity described in gastropods consists of a fixed pattern of bursts, alternated by periods of hyperpolarization (e.g. cell R15 of Aplysia (Frazier et al. 1967) and cell 11 of Otala (Gainer, 1972)). Although synaptic inputs can modify the bursting activity of such cells, its basic pattern is endogenous (cf. Alving, 1968). Thompson & Smith (1976) have shown that depolarizing afterpotentials (DAPs) are involved in the generation of the successive spikes of the bursts of such cells.
During exploration of the central nervous system (C.N.S.) of Lymnaea stagnalis a cluster of cells which have unusual bursting characteristics was found in the right parietal ganglion. In some respects these neurones seemed similar to endogenously bursting cells in other snails. In contrast to these, however, they had very variable and comparatively long bursting cycles. The large number of bursting cells in the cluster was also conspicuous. Comparison with histological data of Wendelaar Bonga (1970) indicated that the cluster consists of neurosecretory ‘Light Yellow Cells’ (LYC; name based on histochemical staining). The present paper deals with the electrical properties of these cells and the intrinsic origin of the bursting pattern.
MATERIALS AND METHODS
Adult snails (shell heights 28–32 mm), bred under standard conditions (van der Steen, van den Hoven & Jager, 1969), were used. The C.N.S. was isolated and the outer sheath of the right parietal ganglion was partly dissected to expose the perineurium covering the LYC. The preparation was kept in a Ringer solution (30 mM-NaCl, 15 mM-KCl, 2 mM-MgCl2, 4 mM-CaCl2, 0·25 mM-Na2HPO4 and 18 mM-NaHC08; pH 8·1) at 17 ± 1 °C.
For LYC-isolation the C.N.S. was immersed for 15 min in Ringer solution to which 1–2 mg/ml Pronase (Calbiochem, B grade) was added. After washing in normal Ringer for 30 min the perineurium was dissected. A small part of the LYC-containing lobe was cut off and single somata, with short parts of the axons (20-100 /an), were isolated from it. Attempts to isolate LYC in a purely mechanical way were not successful.
The isolated LYC were transferred to a solution of 10% gelatin in Ringer (27 °C). The gelatin was coagulated by cooling. This facilitates the impalement of isolated neurones (cf.Kostenko, Geletyuk & Veprintsev, 1974). For histological purposes single cells were collected with glass capillaries and processed as described by Roubos (1973).
Intracellular recordings were made with microelectrodes (4–40 MΩ) filled with 3 M-KCI or 0·5 M-K2SO4. Current was injected via the same electrodes by means of a pulse generator or a direct current source. Pt electrodes were used for nerve stimulation. Preliminary experiments showed that mechanical contact with the nerve had to be avoided. Therefore, the electrodes were mounted in small Ringer-containing compartments in the bottom of the recording chamber, and they contacted the nerve via small Ringer-filled channels. Stimuli were delivered via a stimulus isolation unit.
Neuronal activity was stored on tape, displayed on a storage oscilloscope and photographed with a Polaroid camera.
Identification of the cluster of LYC
In the C.N.S. of L. stagnalis several types of N.S.C. occur (Wendelaar Bonga, 1970, 1971). The cells of some types lie in two or more homogeneous clusters which are identifiable by their general appearance, position and colour. In other cases identification is only possible by means of histochemical methods. According to the histochemical studies of Wendelaar Bonga (1970) the LYC lie in two clusters: one in the rostro-dorsal part of the visceral ganglion and the other in the ventral lobe of the right parietal ganglion. Each cluster contains about 40 cells. The LYC send axons into the nerves of the visceral and parietal ganglia and into two cerebral nerves (the left and right nuchal nerve).
The cell bodies studied in the present experiments are situated in a distinct cluster in the right parietal ganglion, at the same location as the LYC described by Wendelaar Bonga (1970). The cells are roughly spherical (diameters 50–80 μm) and pale yellow to yellow-orange in colour. They cover the main part of the lateral region and a smaller part of the median region of the ventral lobe (Fig. 1).
Ultrastructural studies on cells isolated from this cluster confirmed that they are the LYC described by Wendelaar Bonga (1970): the cells contain, and apparently synthesize, neurosecretory granules having a diameter and an appearance typical for LYC (Fig. 2).
Discharge pattern of LYC in situ
Activity was recorded from about 100 LYC. Most were from the lateral part of the ganglion, the others from the median part. No obvious differences were found between cells from different locations.
All LYC were spontaneously active. In 70% of the cells this activity consisted of a typical cyclic pattern of bursts of action potentials: burst-fast hyperpolarizing shift - slow depolarization - fast depolarizing shift up to the spike threshold - burst - and so forth (e.g. Figs. 3, 5 A 1, 5 B). This activity pattern persisted for several hours. The ‘resting’ membrane potential varied from 50 to 60 mV (at the peak of hyperpolarization) to 40– 45 mV (during a burst, measured as the value half-way between spike threshold and spike after-hyperpolarization).
Fig. 3 shows a recording of 1 h, beginning 30 min after penetration of the cell (there is a clear ‘penetration effect’: just after impalement the cells are always active, never hyperpolarized, and the initial 2 or 3 cycles are always somewhat longer than the later cycles). The bursts differ considerably in duration (from 10 s to 10 min). In the cell shown in Fig. 3, the bursts varied from 2 min (last burst) to 9 min (2nd and 3rd burst). The number of spikes within a burst varies from ten to several hundreds. Most cells show predominantly long bursts, some, however, fire only short bursts. In the latter cells the interburst periods tend also to be short.
During long portions of a burst the cells show a regular firing rate (at some constant value between 1/2 and 2 spikes s−1). This, however, is usually interrupted by gradual increases and/or decreases of the spike frequency. The mean spike frequencies of the bursts of individual cells often differ (Fig. 3). During the interburst periods the membrane potential occasionally shows slow, irregular changes (5 mV or less), not resembling ordinary synaptic potentials (Fig. 3; see also Fig. 8).
Fig. 4 illustrates the shape of the action potentials of the LYC. The amplitudes are about 75 mV and the spikes last 20–50 ms (at 50% spike height). The spikes have a shoulder during repolarization. This feature becomes more pronounced (the spike duration can increase considerably) during the initial portion of the burst and during the increases of spike frequency mentioned above. The increase in duration of the action potentials is often accompanied by a decrease in spike amplitude. The latter is especially clear during a rise of the spike frequency (Fig. 3). The action potentials are followed by hyperpolarizing afterpotentials of about 7 mV (arbitrarily measured as the difference between the membrane potential just before the burst and the undershoot). These afterpotentials slightly decrease during the first stage of the bursts (Fig. 4).
Fig. 5 shows the effect of current injection on the activity of an LYC. During hyperpolarization with currents of 1 nA or more, all spike activity is inhibited (Fig. 5 A 2). With weak depolarizing currents (about 0·4 nA) the cells fire regularly; spike activity may completely stop after a variable period of firing, after which no hyperpolarization is observed (Fig. 5 A 4). With stronger depolarizing currents (1 nA or more) an inactive, depolarized state develops almost immediately (Fig. 5 A 5). During depolarization the spikes (if present) have a relatively low amplitude (and also a long duration) when compared to spikes in a bursting cell without current injection (see Figs. 5 A 4 and 5 A 3, respectively).
The effects of current application show that the spike and burst mechanisms involved are considerably affected by currents as small as 1 nA or even less. The influence of current injection on the firing pattern suggests that endogenous processes are, at least in part, responsible for the observed firing pattern.
A role of endogenous factors is also suggested by experiments in which initially the cell was, by current injection, brought into a hyperpolarized or a depolarized state; if then the amount of current was diminished in small steps, at a certain current step a shift in depolarizing or hyperpolarizing direction, respectively, was observed (Fig. 6 A–C). These shifts were similar in appearance to the shifts occurring during normal bursting at the onset and the end of a burst, respectively. So, apparently, current injection can bring about the same shifts as those occurring spontaneously during normal bursting.
Most LYC show the bursting pattern described above. About 30% of the cells, however, exhibit other electrical properties. Occasionally spontaneous depolarization leads to a complete and lasting abolition of spike activity (Fig. 7). In such cases normal bursting can be temporarily restored by the injection of a weak hyperpolarizing current. The aforementioned shift of the membrane potential can also be seen in such spontaneously depolarized cells, if the cell is hyperpolarized (Fig. 6D). Other bursting cells develop a prolonged, silent, hyperpolarized state. A small depolarizing current can reinitiate bursting at this time, but, again, only temporarily (Fig. 8).
In a small number of cells still other activity patterns were found: irregular bursting alternated by short, strong depolarizing waves; bursting without noticeable interburst hyperpolarization; and oscillations of the membrane potential with occasional spikes.
In general, the cells of one preparation show the same type of activity. In a few cases a transition from one type to another was observed. No relation was found between the time of the day (the experiments were carried out between 10.00 a.m. and 18.30 p.m.) and the type of activity.
Effects of evoked spikes on the bursting pattern of LYC in situ
An action potential can be evoked by the injection of a short current pulse via the recording electrode and via the right internal pallial nerve (most LYC have an axon in this nerve; unpublished results). Both methods of stimulation have similar effects (synaptic events did not affect the results described below). For technical reasons antidromic stimulation was applied in most experiments.
When LYC are stimulated during a burst an extra spike is produced (Fig. 9 A). The next spontaneous spike appears with a delay similar to the normal interspike periods of the burst. This suggests that the spikes of a burst are produced by an endogenous mechanism (cf. Frazier et al. 1967). When LYC are stimulated during an interburst period an extra spike is evoked followed by a new burst (Fig. 9B, upper trace). Apparently such a spike can activate the bursting mechanism.
A spike, evoked immediately after a burst does not trigger a new burst. It is, however, followed by a depolarizing afterpotential (DAP) with an amplitude of 5–12 mV and a duration of several seconds (Fig. 9C). DAPs of successive spikes summate (Fig. 9 C). After a number of spikes (provided the spike intervals are sufficiently short) the cell suddenly starts bursting (Fig. 9 C). The most likely explanation for this is that the summated DAP finally becomes supra-threshold and evokes an action potential, again followed by a supra-threshold DAP, etc.
This implies that the gradual interspike depolarizations of the bursts represent the initial phase of the DAPs (Fig. 9B). Towards the end of a burst the interspike depolarizations gradually become less steep (Fig. 9 D), which means that, assuming the slope of the DAP is related to its amplitude, the DAPs progressively decrease. The last action potential of a burst is always followed by a subthreshold DAP (Fig. 9D).
Shortly after a burst the DAPs of evoked spikes are smaller than later in an interburst period (Fig. 9B). This explains why one action potential can sometimes initiate a burst, whereas in other instances a train of action potentials is needed.
From these results it is concluded that the DAPs constitute the endogenous mechanism responsible for the continued generation of spikes during a burst.
Discharge pattern of isolated LYC
In in situ LYC, potential changes resembling synaptic potentials were hardly ever observed. This indicates that under the experimental conditions LYC seldom received synaptic inputs, which implies that the whole bursting pattern of the cells, including burst initiation, might have an endogenous origin.
To test this assumption, single isolated cells were studied. Fig. 10 A shows that the activity of these cells is similar to that observed in LYC in situ. The bursting pattern is more regular (cf. Fig. 3) and the action potentials last slightly longer but they are of normal height and shape. DAPs and interburst hyperpolarizations are present albeit reduced (Fig. 10B). Probably due to the reduction of the DAPs, more stimuli are needed to trigger a burst (Fig. 10 C).
These observations demonstrate that the bursts of the LYC are continued and terminated endogenously; this establishes the endogenous nature of the DAP already suggested in the previous section. The observations also show that the bursts are initiated endogenously.
Ultrastructural studies have shown that the LYC exhibit a continuous synthesis and release of neurosecretory material (Wendelaar Bonga, 1970, 1971). The present observations demonstrate that they usually display a spontaneous bursting activity. These ultrastructural and neurophysiological data correspond well with the generally accepted view that in N.s.c. the electrical and the release activities are interrelated. Also other types of N.s.c. of L. stagnalis, which show a continuous release of neurosecretory material (Wendelaar Bonga, 1970, 1971) have a strong tendency towards spontaneous activity (Benjamin & Swindale, 1975), whereas cells with a more discontinuous release do not (Benjamin et al. 1976). The supposition that the electrical and release activities of N.S.C. are interrelated is sustained by recent in vivo observations on Aplysia, which show that egg-laying is preceded by repetitive spike activity in the egg-laying hormone producing neurones, the ‘bag’ cells (Pinsker & Dudek, 1977).
Some other bursting gastropod neurones are also N.S.C. (e.g. cell R15 of Aplysia; Kupfermann & Weiss, 1976; cell 11 of Otala; Gainer, 1972). Since isolated LYC (with only a short part of the axon) still show spontaneous bursting, the action potentials must be initiated in or near the cell body; this favours the hypothesis that secretion is triggered by centrifugal spike activity (cf. Finlayson & Osborne, 1975). Gainer (1972) suggested that bursting might serve as a facilitatory mechanism for release.
The bursts of the LYC terminate with a subthreshold depolarization and are followed by a rapid hyperpolarization; after a slow decay the hyperpolarization leads to a quick depolarization which triggers a new burst. Neurones with a similar type of activity have been identified in several other snails (Aplysia:Strumwasser, 1965; Frazier et al. 1967; Helix-. Kerkut & Meech, 1966; Sakharov & Salánki, 1969; Tritonia:Willows & Hoyle, 1968; Otala:Gainer, 1972; Helisoma:Kater & Kaneko, 1972; Archidoris and Anisodoris: Thompson & Smith, 1976; Achatina: Takeuchi, Matsumoto & Sakai, 1977; Chase & Goodman, 1977). These cells also have other electrophysiological properties in common with the LYC. (1) Bursting is due to an endogenous mechanism (cf. Alving, 1968; Gainer, 1972; Kater & Kaneko, 1972). (2) The spikes have a long duration which increases considerably during a burst (cf. Strumwasser, 1968; Kater & Kaneko, 1972). (3) When current is injected into a cell, sudden shifts of the membrane potential are observed at certain current steps (Fig. 6); these shifts are similar to those occurring during normal spontaneous bursting. In some cases such a ‘threshold’ phenomenon has been shown to be related to a ‘negative resistance characteristic’ of the steady-state current-voltage relationship (e.g. Wilson & Wachtel, 1974). (4) In the LYC depolarizing afterpotentials are involved in the development of repetitive activity. Thompson & Smith (1976) observed the same in several bursting neurones of other snails.
The slow decay of hyperpolarization which, in bursting pace-maker cells, triggers the fast depolarizing shift before the first spike, has often been thought to be the pacemaker potential of the entire burst (e.g. Waziri, Frazier & Kandel, 1965). Evidently, however, the DAP constitutes a second pace-maker process, since it initiates the spikes within a burst. The DAP thus represents a process comparable to the pace-maker potential in beating pace-maker cells. It might be that the fast depolarizing shift preceding the first spike is a DAP-like process (Thompson & Smith, 1976). After blocking the spike activity of bursting pace-maker cells by treatment with tetrodotoxin, slow oscillations of the membrane potential remain (Strumwasser, 1968; Mathieu & Roberge, 1971), probably representing ‘burst pace-maker potential’ oscillations. This is not inconsistent with the presence of an additional ‘spike pace-maker potential’ or DAP, since in the absence of spike activity also the DAPs will disappear.
Probably the membrane properties of the LYC are similar to those of other bursting pace-makers(e.g. Smith, Barker & Gainer, 1975; Johnston, 1976; Heyer & Lux, 1976; Thompson & Smith, 1976).
LYC differ from other bursting pace-makers in that they exhibit bursts of comparatively long duration (several minutes). In this respect they are intermediate between common endogenous bursters, which exhibit short bursts (duration: about 10 sec) and continuously beating cells. LYC are also special in showing a great variability of burst parameters. Moreover, the cells do not only show bursting activity but can exhibit vaiious other types of activity (cf. Chase & Goodman, 1977). These observations suggest that the activity of the LYC is modulated by (as yet unknown) exogenous and/or endogenous factors.
The author thanks Dr T. A. de Vlieger for his stimulating suggestions during the experiments, Prof. Dr J. Lever, Drs T. A. de Vlieger, C. Janse and B. E. C. Plesch for helpful comments on earlier draughts of the manuscript, Dr E. W. Roubos for electron- microscopic work, Ing. M. L. J. van Vilsteren for technical assistance and Mr S. Paniry and Mrs P. F. M. Ohr for their help in the preparation of the manuscript.