Neuronal Feedback In Egg-Laying Behaviour of the Pond Snail Lymnaea Stagnalis

In Lymnaea stagnalis, the neuroendocrine cells controlling egg laying are the caudodorsal cells (CDCs), a group of about 100 electrotonically coupled neurones located in the cerebral ganglia and with axons forming a neurohaemal area in the intercerebral commissure (Wendelaar Bonga, 1970; de Vlieger et al. 1980). In the intact animal, CDC activity can be initiated by giving animals a clean water stimulus (CWS; consisting of transfer from dirty to clean water, Ter Maat et al. 1983) or can occur spontaneously. In both cases, CDC activity lasts about 60min and is always followed by egg-laying behaviour (Ter Maat et al. 1986, 1989). This begins within minutes of the CDCs starting to fire (Ter Maat et al. 1986) and consists of three phases: resting, turning and oviposition (Ter Maat et al. 1989).

Egg laying can also be induced by injecting the ovulatory hormone (caudodorsal cell hormone, CDCH) directly into the blood. Following such injections, animals do not, however, show a resting phase. Instead, they continue moving and then enter directly into the turning phase (about 60min after the injection; Ter Maat et al. 1989). Animals that are given injections but do not respond by laying eggs show no egg-laying behaviour. These results demonstrate that the presence of CDCH in the blood is alone insufficient to produce the fully coordinated pattern of egg-laying behaviour but, as the other CDC peptides that are normally released synchronously with the CDCH were absent, this does not contradict the hypothesis of Scheller and Axel (1984).

The same previous study (Ter Maat et al. 1989) also revealed a positive correlation between the number of eggs in the egg mass and the duration of the turning phase; the more eggs there were, the longer it lasted. This observation seemed to challenge the hypothesis that the actions of the released CDC peptides can account for the fully coordinated pattern of egg-laying behaviour. Instead, it suggested that input from the ovulated eggs is necessary for full expression and coordination of the egg-laying behavioural pattern of Lymnaea stagnalis. This idea is investigated in the present study, where lesions were made to disrupt the pathways likely to be involved in the transmission of sensory input from the reproductive tract to the neurones in the central nervous system (CNS) responsible for producing egg-laying behaviour. The turning-phase behaviour of animals, including, for the first time, measurements of shell rotation, was then analysed.

Adult laboratory-bred snails were used. Prior to experiments these were housed individually in perforated jars placed within an aquarium tank supplied with continuously running fresh water (20°C) and kept under a 12h:12h light:dark cycle. Animals were fed a daily ration of lettuce.

Fig. 1 shows the normal egg-laying behaviour of a single Lymnaea stagnalis following clean water stimulation (CWS, Ter Maat et al. 1983). Rasping activity, locomotion and shell position were quantified by replaying episodes of egg laying from video recordings. Rasps were counted. Shell turning was determined using a PC Vision Plus (Imaging Technology) videodigitizer and a PC/AT personal computer to capture and process a video frame every 15s. The angle between the line connecting the two tentacles and the line connecting the anterior and posterior ends of the shell was used to measure shell position relative to the head-foot. Locomotion was determined by using the video digitiser/computer system (described above) to measure the distance travelled (using the midpoint between the tentacles of the animal as a landmark) by the animal in each 15 s interval.

Approximately 15min after the CWS the animal entered the resting phase. During resting, there were few changes of shell position, no rasps and little locomotory activity. After remaining in the resting phase for approximately 40min, the animal entered the turning phase and started typical turning-phase behaviour. There were frequent rotations of the shell, the animal started to use its buccal mass to make rasping movements against the substratum and moved slowly within a localised area of the tank. Turning-phase behaviour occurred for approximately 60min and was followed by the oviposition phase (lasting about 10min), when the eggs were deposited on the substratum. During oviposition, the shell was held almost parallel to the head-foot, the level of rasping declined and the animal moved slowly forward. In the 30min period following the completion of oviposition, shell movements resumed and the rate of locomotion increased, but no rasping movements were made.

To determine whether neuronal input from the reproductive tract is important for the full expression of turning-phase behaviour, lesions were made (under MgCl₂ anaesthesia) to disrupt the neuronal pathways between the reproductive tract and the CNS (see Fig. 2A). Previous anatomical studies (Elo, 1938) suggested that nerves from the parietal or visceral ganglia were most likely to innervate the reproductive tract. Lesions (N=5 for each group) were therefore made of both pleuroparietal connectives, the two nerves (the internal and external right mantle nerves) of the right parietal ganglion and all four nerves (the genital, anal, cutaneous mantle and intestinal nerves) of the visceral ganglion. These latter nerves were first all lesioned simultaneously and then, in separate groups of animals (N=5 for each), lesioned individually. Sham-operated animals (N=10) served as controls. Egg laying was then induced (on postoperative day 3) either by injection of CDCH [20pmol of synthetic CDCH (Eurosequence, Groningen) in 10 μl of dibutylamine and 40 μl of Hepes saline] into the foot sinus, or by giving animals a CWS. The first of these methods was used for the group of animals with bilateral lesions of the pleuroparietal connectives, for the group of animals where all four visceral nerves were cut together and for five of the controls. The CWS was used for all other groups of animals and for the remaining five controls). The reason that CDCH injections were used to induce egg laying in the groups of animals with lesions of the pleuroparietal connectives or all visceral nerves was that previous work (Geraerts et al. 1984) suggested that they would not respond to the CWS. The turning-phase behaviour of the animals during the 40min immediately prior to the start of oviposition was analyzed. Total rasps, average deviation from normal shell position and average speed were determined over this period.

Rasps, shell position and locomotion all showed considerable variability between animals. Therefore, in animals receiving the CWS, the three measures of these activities were also determined during 15min following CWS, which is before the start of the CDC discharge (Ter Maat et al. 1989). In injected animals, the 30min period prior to the injection was used. The rates of rasping, shell position and locomotion during these baseline periods correlated strongly with each of the rates exhibited during turning-phase behaviour. They were consequently used to correct rasping, shell position and locomotion of individuals during the turning phase (analysis of covariance, using the pre-egg-laying score as the covariate). The measures of all experimental groups, corrected for the covariate, were then compared with controls. Because this involves eight comparisons for each measure, the Bonferroni protection level (P<0.016; see Fig. 2) is given, instead of the P-value of the individual tests (P<0.002 at worst).

As shown in Fig. 2B, some of these lesions had a dramatic effect on turning-phase behaviour. Animals with lesions of the pleuroparietal connectives, all visceral nerves, or the intestinal nerve showed significantly less shell turning and fewer rasping movements, and a significantly higher rate of locomotion, than both groups of control animals. The higher locomotory level was because animals moved around the whole tank until the eggs started to leave the vagina. The other lesions did not significantly affect any of the three turning-phase behavioural patterns. The almost complete abolition of turning-phase behaviour after bilateral lesions of the pleuroparietal connectives, all visceral nerves or the intestinal nerve suggests that sensory input is important for the expression of turning behaviour. The results suggest that the information from the reproductive tract enters the CNS mainly via the intestinal nerve and is then transferred to the anterior ganglia of the CNS via the pleuroparietal connectives.

To investigate whether the intestinal nerve was sufficient for the production of turning-phase behaviour, it was left intact while the three other visceral nerves were cut (N=5). After egg laying had been induced, by injection of CDCH, these animals showed normal levels of all three turning-phase behavioural patterns in the last 40min of the turning phase. This demonstrates that the intestinal nerve is, in addition to being necessary for normal turning, also sufficient for the production of normal turning behaviour when left intact.

In Fig. 3 the time course of the egg-laying behaviour of individual control and experimental animals is shown. Rasping and locomotion were quantified for the full egg-laying episode and shell position was scored for the hour prior to the end of oviposition.

In Fig. 3A, the CWS was used to induce egg laying (animals were transferred to clean water at the start of the plots). The control animal showed typical turning-phase behaviour, whereas the animal with the intestinal nerve cut showed lower levels of rasping and shell turning, despite laying a larger egg mass. When the intestinal nerve was the only visceral nerve left intact and egg laying was induced with an injection of CDCH, the three behavioural patterns were similar in both timing and levels to that of the control animal (Fig. 3B).

In summary, it was found that lesioning pathways between the reproductive tract of the animal and the CNS prevents Lymnaea stagnalis from entering the turning phase of overt egg-laying behaviour. This shows that the animal uses neuronal feedback to achieve the proper sequence of the egg-laying behaviour pattern. This means that sequential release of peptides by the CDCs or differences in the time courses of the effects of the various peptides need not be invoked to explain the orderly course of egg-laying behaviour. Thus, as with many other complex stereotyped behaviours (e.g. locust flight, Wendler, 1974; walking in cats and cockroaches, Forssberg et al. 1975; Wong and Pearson, 1976; lamprey swimming, Grillner et al. 1981), input into the CNS is necessary for expression of the full behavioural pattern.

Lesions of the pleuroparietal connectives caused a greater change in turning behavioural patterns than lesions of the intestinal nerve, so the possibility that there are other peripheral pathways by which the sensory input reaches the CNS cannot be ruled out. The intestinal nerve would, however, seem to be the most important pathway. This is because (i) lesions of other visceral and right parietal nerves had no significant effect on turning behaviour and (ii) normal turning behaviour occurred when the intestinal nerve was left intact and the other visceral ganglion nerves were lesioned.

An important consideration, however, is whether the effects seen after lesioning the intestinal nerve could have been due to some cause other than the disruption of a sensory pathway between the reproductive tract and the CNS. Notably, the fact that the intestinal nerve contains the axons of heart motor neurones (Buckett et al. 1990) may have confounded our results because cutting it might disrupt cardiovascular regulation and thus alter egg-laying behaviour. We do not believe, however, that this can explain the present results. Nor is it likely that lesions of the intestinal nerve would disrupt the motor pathways producing the behavioural patterns of the turning phase. This is because the motor neurones innervating the muscles used during rasping, shell turning and locomotion are located in the buccal, cerebral and pedal ganglia (Benjamin et al. 1985; Jansen and Ter Maat, 1985; Winlow and Haydon, 1986). Furthermore, animals could still rasp (and eat normal amounts of food), turn their shell and move after the lesion.

The resting phase of egg-laying behaviour shows little variation in duration, irrespective of the number of eggs deposited. It is thought to be under peptidergic control (Jansen and Ter Maat, 1985). When the CDCs fire a discharge, an identified interneurone (the ring neurone) becomes active (Jansen and Bos, 1984). This neurone has synaptic connections with pedal ganglion neurones involved in the control of locomotion and shell position, and it might cause an inhibition of locomotion and the adoption of the posture typical of the resting phase (Jansen and Ter Maat, 1985). Although peptidergic control may be sufficient to coordinate resting, it is less likely to be able to coordinate turning fully. The most probable function of turning behaviour is cleaning of the substratum where the eggs will be deposited (Ter Maat et al. 1989). Given that the size of individual egg masses can vary from less than 10 to more than 200 eggs, it would seem plausible that, by utilising neuronal input from the reproductive tract to control turning-phase behaviour, animals can prepare an area of the substratum appropriate for the number of eggs to be deposited with more precision than if the turning phase were solely under peptidergic control. An interesting question is whether the input from the reproductive tract is in itself sufficient to cause turning-phase behaviour or whether its effects are conditional on the presence of the peptides that are released during the CDC discharge, which may have modulatory effects on the neuronal circuits underlying the behaviour pattern.

During the present study, animals laid eggs in a clean glass chamber. It was observed that the egg masses laid by animals that did not show normal turning-phase behaviour were much less securely attached to the substratum than is normally the case (in some cases, even becoming detached from the substratum at the end of the oviposition phase). This would seem to provide additional support for the previous suggestion that the adaptive role of turning-phase behaviour is to clean the substratum and to facilitate good adhesion of the egg mass.

In Aplysia californica, the discharge of the neurosecretory bag cells (which are similar to the CDCs of Lymnaea stagnalis and release an ovulatory hormone and other peptides; Kupfermann, 1967; Rothman et al. 1983) is sufficient to initiate the appetitive phase of egg-laying behaviour, but, as in Lymnaea stagnalis, sensory input from the ovulated eggs (via pathways yet to be determined) is necessary for the occurrence of consummatory egg-laying behavioural patterns (Cobbs and Pinsker, 1982; Ferguson et al. 1986). It may be a general feature of the egg-laying behavioural pattern of all gastropod molluscs that input from the genital tract is necessary for the full expression of the behaviour. Further comparative studies are necessary, however, before this can be stated with confidence.

We thank Professor T. A. de Vlieger and members of his group for comments on a previous draft of this manuscript and Ms T. Laan for help in preparing the manuscript. G.P.F also thanks A. D. G. de Beer for much additional support.