A Transmesogloeal Conduction System in the Swimming Sea Anemone Stomphia

Sea anemones possess a simple nervous system. It consists of a network of nerve cells spread throughout the body, condensed in some areas to form tracts for rapid conduction. Ganglia, nerve cords and other structures associated with higher nervous systems are absent. Consequently, the anemones have attracted much attention as animals whose behaviour might be fully understood.

There are three systems in sea anemones that conduct diffusely in any direction from almost any point. They consist of a nerve net that conducts at 10–120 cm/s (the through-conducting nerve net; Pantin, 1935a, b) and two slower systems, the SS1 and the SS2 (McFarlane, 1969). Their properties have been reviewed elsewhere ( Lawn, 1976a); in the column at least, the SS1 seems to be ectodermal, whereas the through-conducting nerve net and the SS2 are endodermal. These systems help to coordinate behavioural activity over wide areas but are not designed to produce purely local effects. This is the function of a second nerve net, the interneural system (Pantin, 1935b), which controls local movements of the tentacles, the oral disc and the column. In the column, this system may consist of an endodermal network of multipolar nerve cells, and it has been suggested that such cells also act as a delayed initiation system (DIS) that connects with both the SS1 and the through-conducting nerve net (McFarlane & Jackson, 1976).

The picture that has emerged so far from our attempts to describe the sea anemone as a ‘behaviour machine’ (Pantin, 1965) is still incomplete. No provision has been made for the transfer of information from sensory cells in the ectoderm to effector cells in the endoderm. Such a route must exist if endodermal muscles are to respond lio chemical signals that contact the outside of the animal. Studies on the structure of the nervous system have shed little light on this subject and there are conflicting views on the existence of nervous connexions between ectoderm and endoderm ( Hertwig of Hertwig, 1879–80; Havet, 1901; Pantin, 1952; Batham, Pantin & Robson, 1960, 1961; Leghissa, 1965). The view most widely accepted is that the nervous system confines itself to epithelia, except in the tentacles and oral disc. Here, pathways cross the mesogloea (a supportive collagen matrix sandwiched between the epithelia) to serve ectodermal muscles, but no sensory connexions have been described (Batham, 1965). This implies that conduction from ectoderm to endoderm occurs only where these tissues meet, which in most sea anemones is at the base of the pharynx (Stephenson, 1928).

This paper describes an additional system that conducts across the mesogloea from ectoderm to endoderm. It is capable of transferring information from ectodermal receptors to endodermal effectors.

The sea anemones used in this study were Stomphia coccinea and S. didemon (Siebert, 1973). Both species were collected by benthic dredging; S. coccinea from Washington Sound, Washington at a depth of 100 m, and S. didemon from Barkley Sound, British Columbia at a depth of 40 m. The anemones were transported to Bamfield Marine Station where they were kept in running sea water at temperatures ranging from 9 to 12 °C. They were fed regularly and appeared to be healthy. Specimens of the starfish Dermasterias imbricala were collected intertidally from rocky shores in Barkley Sound. For experiments, the anemones were placed in tanks of running sea water kept at 11 °C. Polyethylene suction electrodes were used for recording and stimulating. A recording electrode was connected to a Tektronix AM 502 differential amplifier and the recorded activity was displayed on a Tektronix 5103 N storage oscilloscope. Some recordings were stored on magnetic tape for later display on a Brush 220 pen-recorder. Electrical stimulation was provided by a Devices Neurolog system. Shocks of 10 V, 5 ms duration were used in all experiments.

Animals that swam consistently to electrical stimulation of the column (Ross & Sutton, 1964; Lawn, 1976b) were selected. The recording electrode was placed on a tentacle, the best site for detecting electrical activity (Lawn & McFarlane, 1976). Stimulating electrodes were placed on different parts of the column according to the needs of the experiment. Approximately 30 animals were used and each was tested at least three times.

Sea anemones of the genus Stomphia were chosen for two reasons. Firstly, their swimming response (Stephenson, 1935) is repeatable and easy to see, and secondly, the conduction system that triggers the response (the SS1) is ectodermal (Lawn, 1976b) while the muscles that produce the swimming movements are endodermal (Sund, 1958). If it can be shown that activity in the SS1 produces swimming when its routes to the pharynx are blocked, then conduction across the mesogloea will have been demonstrated.

It is not possible to stimulate the SS1 alone simply by applying shocks to the outer surface of the animal. This also activates the through-conducting nerve net because current flows through the body wall and reaches the endoderm. If the column is cut carefully, however, a flap of ectoderm can be produced whose free end can be pulled away from the mesogloea. Shocks applied to this flap evoke nothing but SS1 pulses (Fig. 1 A). There is no danger of stimulating other systems directly.

The pulses were contained in the middle of the column by scraping away two encircling bands of ectoderm above and below the flap (Fig. 1B). The effectiveness of these conduction blocks could be tested by applying shocks above and below the upper one while recording from the tentacles. The responses shown in Fig. 1 B verify that the upper block is working. The lower one could not be tested directly because pulses could not be recorded from the column or pedal disc. Indirect evidence that it does block conduction is cited later.

With the routes to the pharynx and pedal disc blocked, conduction across the mesogloea could now be tested. An anemone with an ectodermal flap and functioning blocks was attached to a mechanical transducer by a thread sewn through the upper portion of the column. The flap was stimulated with 10 shocks at a frequency of 1 per s, conditions that normally produce swimming (Lawn, 1976b). The anemone responded by extending its column which it then flexed several times from side to side (Fig. 2). These are typical features of the swimming response and both actions are produced by endodermal muscles.

The recording electrode attached to the tentacle confirmed that no SS1 pulses traversed the upper block. This means that the pulses could not reach the endoderm through the pharynx or through the nervous connexions in the oral disc. The pedal disc detached slowly and erratically during this experiment, in contrast with the smooth, rapid detachment seen in unoperated animals. This is the evidence that the lower block was working properly, for SS1 pulses must pass over the pedal disc to produce a smooth detachment (Lawn, unpublished observations).

This experiment shows that information from the SS1 reaches the endoderm by one route alone, and that is across the mesogloea of the column. The system that conducts this information will be termed the transmesogloeal system (TMS).

The extent of the TMS was determined by performing the same experiment on different parts of the column. The results indicated that conduction across the meso-gloea occurs over the entire column.

Flaps of ectoderm could not be cut successfully on the delicate oral and pedal discs. The oral disc, however, could be tested by stimulating the SS1 chemically. If the starfish Dermasterias imbricata is brought into contact with the tentacles of Stomphia, a train of pulses in the SS1 is produced and this triggers swimming (Lawn, 19766). The ectoderm was scraped away from the margin of the oral disc to prevent the pulses from reaching the column. The starfish was then touched against a tentacle and this produced a burst of SS1 pulses that triggered swimming. Although this shows that information can pass from ectoderm to endoderm in regions other than the column, it does not reveal the exact location of the routes. They could be in the tentacles, the oral disc or the pharynx, or perhaps all three. A similar approach could not be used on the pedal disc as its surface was insensitive to Dermasterias.

While these experiments provide information on the extent of the TMS in the column, they give no indication of how numerous the connexions are. The column was scraped to provide small islands of ectoderm, each of a different size and each supporting an ectodermal flap. Electrical stimulation of the flap produces pulses in the SS1 that are unable to spread beyond the margin of the island. Each island tested, including the smallest of 10 mm², was capable of triggering typical features of the swimming response. This indicates that the mesogloeal connexions are not scarce. In some instances, the only response was a single flexion whose origin was unrelated to the position of the flap.

The use of a flap rules out the possibility that current from the stimulating electrode is flowing directly through the mesogloea to excite the endoderm. In fact, direct stimulation of the naked mesogloea will not produce swimming. Instead, there is a local contraction of the underlying muscles or a general contraction due to excitation of the through-conducting nerve net. The failure of electrical stimulation to produce swimming in this case is puzzling. There may be, however, a simple explanation: removal of the ectoderm destroys the mesogloeal connexions.

The anemone did not respond when the starfish Dermasterias was placed against the mesogloea. This is not surprising for the chemoreceptors involved in the swimming response are ectodermal (Lawn, 1976b). Attempts to record electrical activity from the TMS have so far been unsuccessful.

The experiments described in this paper show that information from ectodermal receptors can reach the endoderm by transmesogloeal connexions. They also imply that this information may be conducted in other directions. For example, the positive response to stimulation of the upper and lower column suggests that one line of conduction in the column is longitudinal. The argument for this rests on the assumption that the pacemaker for the swimming flexions resides in a zone encircling the middle third of the column ( Robson, 1961 b). If swimming is to occur, pulses originating from other areas must be conducted longitudinally to reach this pacemaker. The tests on the islands of ectoderm imply that yet another line of conduction exists in the column. This follows from the observation that a flexion may originate at any point on the circumference, not only at the point of stimulation. The evoked pulse must therefore travel around the circumference to reach the muscle concerned. Such experiments, however, do not reveal the exact location of either pathway: longitudinal and circumferential conduction could occur in the mesogloea or the endoderm. The pathways could be located only if it were possible to block conduction in the endoderm while leaving mesogloeal connexions intact.

It remains to fit these observations to what is known about the anatomy of Stomphia. What are the structures that form the TMS ? In other anemones it has been proposed that the nervous system is confined to the epithelia. Hence, the base of the pharynx, the cinclides, the points where the retractor and parietobasilar muscles insert onto the pedal disc (Robson, 1965) - all have been suggested as sites where ectoderm meets endoderm and all could serve to transfer information between the epithelia. But Stomphia has no cinclides, and the experiments with the conduction blocks show that the pharynx and pedal disc are not the only sites where information can be transferred.

Robson (1963) has studied the neuroanatomy of Stomphia and describes a network of large multipolar neurones whose cell bodies are evenly scattered throughout the endoderm of the column. They extend in one plane between the circular muscle and the endodermal epithelium and appear to be fully interconnected with the endodermal nervous system. But one observation may be of prime importance. It was stated that an occasional process from a multipolar neurone would pass into the mesogloea towards the ectoderm. Such processes could represent the structural units of the TMS.

One should be cautious, however, with speculations on the architecture of the anemone nervous system. In a recent study on the neuroanatomy of Stomphia, Peteya (1976) pointed out that the identification of nervous elements in primitive animals presents great problems. For example, the interpretation of structures observed under the light microscope or the electron microscope is highly dependent on the type of fixative used. This is certainly true for the anemones and probably applies to other coelenterates too. To compound this problem further, contemporary definitions of the neurone and the synapse start to break down when applied to simple nervous systems and the distinction between neurones and non-nervous fibre systems may become obscure. Peteya (1976) describes large multipolar neurones in the column endoderm of Stomphia, thereby endorsing Robson’s observations. No mention is made, however, of processes that arise from these neurones and penetrate the mesogloea. Peteya recognizes a total of five fibre types in the mesogloea, none of which could be labelled as nervous. He concludes that there is, in fact, no clearly demonstrated case of nerve fibres crossing the mesogloea at any point to interconnect ectoderm and endoderm and states that the question concerning the presence or absence of mesogloeal neurones remains the greatest uncertainty arising from studies of the actinian nervous system. Now that transmesogloeal conduction has been demonstrated in Stomphia, it is clear that a thorough re-examination of the mesogloea will be necessary. Until this has been done, processes from the large multipolar neurones must remain as the most promising candidates for the TMS.

Small multipolar cells occur more sparsely in the column endoderm of Calliactis parasitica (Robson, 1965) but McFarlane & Jackson (1976) suggest that they may form the structural bases for the DIS in this anemone. Conduction from endoderm to ectoderm, one of the characteristic features of the DIS, has yet to be demonstrated in Stomphia. It is possible, however, that the TMS and the DIS are the same system operating in different directions. In both cases, the link between the epithelia would be provided by the processes of the multipolar nerve cells. If this interpretation is correct, the term ‘transmesogloeal system’ provides what is probably a more accurate description of the arrangement, because ‘delayed initiation’ relates to conduction in one direction alone. In either case, electrical activity has not been recorded from the system directly.

Ideas about the significance of the TMS are mainly speculative at this stage. It provides an alternative pathway for conducted electrical activity and its role in transferring information from ectodermal receptors to endodermal effectors is obviously a vital one. But the distribution of the connexions is puzzling. Why are they spread so widely around the column? This might reflect merely a primitive arrangement; a fairly unspecialized network of connexions intimately linked with ectoderm and endoderm. If this were so, the TMS would meet the requirements of a primary nervous system (see references in Batham, Pantin & Robson, 1961), one that might represent a starting-point in the evolution of more specialized arrangements.

The numerous connexions across the mesogloea would make sense for an animal that possesses great powers of regeneration. If a Stomphia is bisected, each half will give a typical swimming response when touched by a Dermasterias (Wilson, 1959; Robson, 1961a). The wide distribution of connexions may also serve to short-circuit the conduction routes in the tentacles. This would help to decrease the reaction time of the anemone when its column is touched by a Dermasterias or bitten into by the nudibranch mollusc Aeolidia papillosa ( Robson, 1961b).

It seems likely that transmesogloeal systems will be found in other anemones where they too will function as conduction routes linking ectodermal receptors to endodermal effectors. The discovery of this pathway, considered with the recent discoveries of slow conduction systems (McFarlane, 1969) and inhibitory mechanisms in the through-conducting nerve net (Lawn, 1976c, d), endorses the view that the earlier picture of the ‘behaviour machine’, while invaluable, was oversimplified.

This work was done during the tenure of a post-doctoral fellowship at the Bamfield Marine Station. I thank the Western Canadian Universities Marine Biological Society for their support. It is a pleasure to thank Dr D. M. Ross for his continued interest and for the provision of funds to complete the manuscript.