ABSTRACT
Two of the largest neurones in the terrestrial snail Achatina fulica are RPrl, located in the right parietal ganglion, and VI, located in the visceral ganglion. Several characteristics of these cells were studied in detail, including morphology, passive and active electrical properties, synaptic inputs, sensory inputs, motor outputs and sensitivity to transmitter substances. The results suggest that RPrl and VI form a bilaterally homologous pair of cells, yet they reside in asymmetrically placed ganglia. An explanation of this paradox is offered in the proposal that RPrl and V1 were formerly located in the bilaterally symmetrical intestinal ganglia. Their contemporary asymmetrical locations are accounted for by the hypothesis that, during the evolution of the pulmonate nervous system, the supraintestinal ganglion fused with the right pallial ganglion and the subintestinal ganglion fused with the visceral ganglion.
INTRODUCTION
The central nervous system of gastropod molluscs is composed of a collection of ganglia that have been subject to extensive changes during evolution. Comparative studies on the gross morphology of the nervous system in existing molluscan species, primitive and advanced, are helpful in reconstructing the history of ganglionic evolution (for reviews, see Bullock & Horridge, 1965; Kandel, 1979). A further source of information is the study of identifiable cells which are homologous in different molluscan species (Dorsett, 1974; Weiss & Kupfermann, 1976; Chase & Goodman, 1977; Dickinson, 1980). In these studies, it has been possible to infer common origins for homologous cells in hypothetical antecedent species. The concept of homologous cells can also be applied, in a somewhat different sense, to infer a common ganglionic origin for two cells which currently reside in different ganglia of the same, single species.
An instructive example of an intraspecific homology that has been used to reconstruct evolutionary changes in ganglionic organization is the pair of neurones R2-LP1 in the marine gastropod, Aplysia. The giant neurone R2, in the abdominal ganglion, and the giant neurone LP1, in the pleural ganglion, presently reside in unpaired asymmetrical ganglia. However, the two neurones share many specific properties including axonal distribution (Hughes & Tauc, 1961), synaptic inputs (Hughes, 1967) and transmitter biochemistry (Giller & Schwartz, 1971). The marked similarity of R2 and LP1 led Hughes & Tauc (1963) to propose that the two neurones are an homologous pair which were once situated in the bilaterally symmetrical pallial ganglia. According to this view, the asymmetry in the cells’ contemporary locations has occurred because of the asymmetrical fusion of the pallial ganglia with other ganglia. Specifically, the left pallial ganglion is thought to have fused with the left pleural ganglion, and the right pallial ganglion is thought to have fused with the supraintest-inal ganglion to form the right half of the contemporary abdominal ganglion (Eales, 1921; Hughes & Tauc, 1963; Kandel, 1979). As a consequence of these changes, the somata of R2 and LP1 have assumed asymmetrical positions, while the distribution of the cells’ axonal processes remains largely symmetrical.
In the pulmonate gastropod Achatina fulica, there exists a similar case of two giant neurones which closely resemble one another, yet are located in asymmetrical ganglia. The properties of these cells, RPrl in the parietal ganglion and VI in the visceral ganglion, are examined in detail in the present investigation. We have concluded that the neurones are an homologous pair whose contemporary locations can be accounted for by a history of asymmetrical fusion in earlier arrangements of the central nervous ganglia.
A preliminary report of some of these data has been published (Munoz, Pawson & Chase, 1982).
MATERIALS AND METHODS
Adult specimens of Achatina fulica (>150 daysold), weighing 20–60g, were taken from laboratory culture (P. A. Pawson & R. Chase, in preparation). The dissected tissue was pinned to a Sylgard base in a Petri dish and bathed in a saline solution which contained 67 mm-NaCl, 4mm-KCl, 13mm-MgC12, 11mm-CaCh, 1 gl−1 glucose and 5mm-Tris buffer. The pH was adjusted to 7·9 with HC1. For experiments in which it was necessary to block chemical transmission in the central ganglia, the ganglia were isolated from peripheral tissue with Vaseline partitions and perfused with a saline solution containing 26 mm-Mg and 0 mm-Ca.
Cells were penetrated with single-or double-barrel glass capillary electrodes filled with 2M-K-citrate and bevelled to 5–10 MΩ resistance. Conventional electrophysiological recording and stimulating techniques were used (for details see Chase & Goodman, 1977; P. A. Pawson & R. Chase, in preparation). Muscle movements were recorded using a Grass FT.03 force displacement transducer. The transducer signal was amplified and filtered through a Grass P16 amplifier. Putative synaptic transmitter substances, in concentrations of 10−4M, were dissolved in the high Mg saline (see above). These solutions were perfused through the bath at 6 ml min−1. The bath capacity was about 7 ml.
Data were collected for RPrl and VI in each animal. Cell diameters were measured under a dissecting microscope using a micrometer eyepiece. The major (a) and minor b) axes of the cell somata were measured and the equivalent diameters [d = (a2b)1/3] were calculated (Mirolli & Talbott, 1972). Current pulses were delivered through one side of a double-barrelled electrode, and voltage responses were recorded through the other side to determine rheobase current, threshold voltage, action potential amplitude, input resistance and time constant. Rheobase current was defined as the minimum amount of depolarizing current necessary to evoke a spike. The threshold voltage was defined as the difference between the resting membrane potential and the membrane potential at the time that an evoked spike was initiated. Action potential amplitudes were measured from resting potential to peak depolarization. The input resistance was calculated as the slope of the current-voltage curve in the linear portion of the hyperpolarizing response. Time constant values were derived from voltage transients that minimally displaced the cell membrane and fell in the linear portion of the cell’s current-voltage curve (Gorman & Mirolli, 1972). The time constants were calculated as the negative reciprocal of the slope of ln (Vmax—Vt) versus time (Merickel, Eyman & Kater, 1977). The data are presented throughout as means ± standard error (S.E.) of the mean.
Horseradish peroxidase (HRP: Sigma Type VI or Boehringer Type I) was injected into cells through glass capillary electrodes using air pressure (7-14kPa). The HRP was dissolved as a 2% solution in 0·1 M-K-citrate. Filled cells were left for 2 h at room temperature or 24 h at 4 °C to allow diffusion of HRP into fine processes. The HRP was reacted with benzidine dihydrochloride and H2O2 according to histochemical procedures described by Muller & McMahan (1976). The ganglia containing cells filled with HRP were viewed as whole mounts. Peripheral tissue was cut on a cryostat as 60–90μm sections following histological processing.
Lucifer Yellow CH (generously supplied by W. Stewart) was pressure injected into cells to test for dye coupling (Stewart, 1978). Electrode tips were filled with 1% (w/v) Lucifer Yellow CH in 0·01 M-LiCl. Injected tissue was processed according to Stewart (1978) and viewed under a Leitz Orthoplan microscope equipped with incident excitation fluorescence.
RESULTS
Morphology
The two giant neurones, RPrl and V1, are respectively located on the dorsal aspects of the right parietal and visceral ganglia (Fig. 1). They were readily identifiable on the basis of several criteria including the size and position of their somata, the action potential waveform and the nature of the spontaneous electrical activity.
RPrl and V1 are among the largest cells in the nervous system of Achatina fulica. The equivalent soma diameter of RPrl ranged from 194–345 μm while that of V1 ranged from 219–375 μm. Within a given animal the soma diameters were closely matched, with a mean difference of 24 ± 6μm (N = 16). V1 was consistently larger than RPrl (Chi-square, P<0·05).
When HRP is injected into RPrl and VI in the same animal, each cell appears as the mirror image of the other (Fig. 2A). Both cells have extensive axon branching in the right parietal, visceral and left parietal ganglia and somewhat reduced branching bilaterally in the pleural and pedal ganglia. Both cells send axons out the right parietal, medial parietal, columellar and left parietal nerves and have extensive projections in the multiple right and left pedal nerves. The pedal nerve projections heavily favour the ipsilateral side. Thus, both cells have many axon branches that project away from the suboesophageal ring, always in the posterior direction. Neither cell sends an axon into either pair of connective nerves linking the ring to the cerebral ganglia.
The peripheral destinations of HRP-filled axons were examined in serially sectioned material. The projections of RPrl and V1 in the pedal nerves travel to the foot musculature posterior to the pedal ganglia. The axons in the right parietal, medial parietal, columellar and left parietal nerves innervate the mantle region. In the case of the projections in the right parietal nerve, the axons of both RPrl and V1 were followed into mantle tissue of similar histological appearance where they branched and appeared to terminate (Fig. 2B, C).
Electrical properties
Neither cell produced spontaneous action potentials in the isolated preparation. The resting membrane potentials were closely matched. The resting potential of RPr1 was −58·3 ± l·5mV (N = 13) and that of V1 was −59·0± l·5mV (N = 14).
The passive membrane properties of RPrl and VI were remarkably similar. Fig. 3 shows typical passive responses to injection of current pulses as recorded from the two cells in a single animal. The charging transients (Fig. 3A, B) had a slow time course typical of molluscan giant neurones. The current-voltage curves (Fig. 3C) show that both cells exhibited anomalous rectification when hyperpolarizing responses were 15–20 mV below the resting membrane potential. The mean values for input resistance and time constant (Table 2) showed no significant difference in these parameters for the two cells (t-test; P> 0·05).
The electrical excitability of RPrl and V1 was also similar. Measurements of rheo-base current, threshold voltage and action potential amplitude (Table 2) showed no significant difference between the cells (t-test; P>0·05). Both cells had prominent negative afterpotentials of 15–20mV following a spike (Figs 7, 8).
Inputs
There was a high degree of synchrony in the spontaneously occurring EPSPs recorded in RPrl and V1 (Fig. 4). Spontaneous IPSPs were rare, but those evoked by nerve stimulation were likewise synchronous (Fig. 5F). The EPSPs tended to be larger in RPrl than in V1 (Fig. 4), probably owing to the larger soma size of V1 which could result in a greater electrotonic attenuation of synaptic potentials propagating to the somatic electrode from synaptic sites in the neuropil.
Electrical stimulation of any of the cerebropleural or cerebropedal connective nerves led to synchronous, compound EPSPs in RPrl and VI (Fig. 5A-D). The response was always larger in the cell ipsilateral to the stimulated connective. A lateral bias was also seen in the latency of response following mechanical stimulation of the foot. The cell located ipsilateral to the site of stimulation consistently reached threshold before the contralateral cell (Table 1).
Electrical stimulation of any of the peripheral nerves joining the suboesophageal ganglion also led to synchronous, compound EPSPs. Repeated nerve stimulation resulted in depression of the compound EPSP (Fig. 5). The rate of depression varied as a function of the nerve which was being stimulated and the recent history of stimulation, but the course of depression was always parallel in the two cells. Even trial to trial variations in response amplitude were closely matched (Fig. 5E). Stimulation of the cutaneous pedal nerves led to synchronous IPSPs in RPrl and VI, as well as depressing compound EPSPs (Fig. 5F).
The giant neurone V2, located in the visceral ganglion, is a source of monosynaptic input to RPrl and V1 (P. A. Pawson & R. Chase, in preparation). Intrasomatic stimulation of this neurone evoked synchronous, facilitating EPSPs in both post-synaptic cells (Fig. 6). The time course of facilitation was very similar.
A possible explanation for the synchrony of electrical events in RPrl and V1 might be that the two cells are electrically coupled. Our experiments, however, do not support such an interpretation. Current injection into either cell produced no detectable change of membrane potential in the other cell. Nor were there coupling potentials associated with spiking activity. Finally, no dye coupling was observed following injection of Lucifer Yellow into either cell. Together, these results argue against the presence of electrical coupling between RPrl and VI (Stewart, 1978).
The preponderant occurrence of synchronous synaptic potentials in RPrl and V1 therefore implies that the majority of inputs to each cell derive from presynaptic neurones which also provide inputs to the other cell, i.e. RPrl and V1 are parallel pathways of information transfer.
Outputs
Intrasomatic electrical stimulation of either RPrl or VI resulted in movements of the foot (Fig. 7A) and the mantle (Fig. 7B). Strong muscular contractions of the foot could be evoked even when chemical synaptic transmission in the centrafganglia was suppressed by the use of a Ringer solution containing twice the normal Mg2+ concentration and zero Ca2+ (Fig. 7A). Since injections of Lucifer Yellow revealed no transneuronal migration of dye, and thus presumably no electrical synapses with follower cells (Stewart, 1978), the evidence suggests that both RPrl and V1 are responsible for taking a motor command to the periphery.
Repeated activity in RPrl or V1 led to short-term facilitation and long-term depression of the behavioural response. Facilitation was demonstrated by the fact that a single spike in either cell was ineffective in producing any observable movement, whereas two spikes separated by an interval of 1 s or less evoked a pronounced response (Fig. 8). Depression was seen with repeated bursts of spikes elicited at 1 min−1. Repeated stimulation led to a progressive decrease in response strength (Fig. 9A) accompanied by an increase in response latency (Fig. 9B).
Responses to transmitters
Substances previously identified as synaptic transmitters in molluscan ganglia were added to the saline perfusion while recording simultaneously from RPrl and V1 In every case, the two cells responded in a qualitatively similar manner to the introduced chemicals. Acetylcholine produced a synchronous hyperpolarization (Fig. 10A). Addition of serotonin, GABA, or octopamine led to a depolarization (Fig. 10B, C, D). Neither glycine nor glutamic acid had any detectable effect (Fig. 10E, F) at the concentration used.
DISCUSSION
We have described a remarkable similarity in the properties of two molluscan neurones. The large number of properties which are shared by the two cells (summarized in Table 2) is unlikely to be fortuitous, especially since some of these properties, e.g. soma size, HRP morphology and motor drive to peripheral muscle, are unusual or unique.
While the properties of RPrl and VI are, for the most part, very similar, they also differ in a consistent manner with respect to the laterality of their inputs and outputs. The relevant data are the following: (1) electrical stimulation of the cerebral connectives evokes compound EPSPs in both cells, but the response in the ipsilateral cell is always larger than in the contralateral cell (Fig. 5A-D), (2) mechanical stimulation of the foot causes the ipsilateral cell to reach threshold before the contralateral cell (Table 1) and (3) the axon collaterals of each cell favour the ipsilateral side of the body (Figs 1,2). Together, these findings suggest that RPrl and V1 might be considered as a bilaterally homologous pair of atypically large neurones similar to the giant metacerebral cells in gastropod molluscs (Kandel & Tauc, 1966; Weiss & Kupfer-mann, 1976), the Retzius cells in leech abdominal ganglia (Lent, 1973), or the Mauth-ner cells in fish and amphibian medulla (Faber & Korn, 1978). Seemingly at odds with this interpretation, however, is the fact that RPrl and VI have cell bodies which are located in two ganglia which are not bilaterally symmetrical.
The present pair of cells is similar to the pair R2-LP1 in Aplysia, because in both cases symmetrical cellular properties co-exist with asymmetrical locations. Hughes & Tauc (1963) attempted to explain this paradox in Aplysia by proposing that the two cell bodies were originally located in symmetrically paired pallial ganglia at an ancestral stage in the evolution of the species. An analogous argument can be made in the case of the neurones RPrl and V1 in Achatina.
The major scientific problem in the evolution of the gastropod nervous system is to explain why the condition of chiastoneury (crossing of the visceral connectives) is present in all members of the most primitive subclass, the Prosobranchia, and in some of the Opisthobranchia, but not in the Pulmonata (with the exception of Chilina, see below). While it is generally believed that the characteristic euthyneury (uncrossed arrangement) of the opisthobranchs resulted from a process of detorsion coincident upon a posterior migration of the mantle complex, euthyneury in the pulmonates does not appear to have followed detorsion (for reviews, see Bullock & Horridge, 1965; Kandel, 1979). Rather, in the pulmonates there was apparently a drastic shortening of the connectives and a fusion of ganglia (Bargmann, 1930). There are wide disagreements in the literature, however, as to the details of this inferred historical process. The interpretation which we provide below is based on a selection and coalescence of earlier ideas, consistent with the experimental results described in the present paper.
The nervous system of the hypothetical prosobranch ancestor to modern pulmonates was characterized, in part, by a circumoesophageal ring of ‘head ganglia’ comprising the bilaterally paired pleural, pedal and cerebral ganglia. Chiastoneury was present, with the supraintestinal ganglion on the left side of the body and the subintestinal ganglion on the right side (Fig. 11A). The most posterior portion of the body was innervated by the unpaired visceral ganglion. As the visceral nerve loop elongated, two new ganglia appeared (Brace, 1977). These new ganglia, termed pallial because they innervate the mantle, are present in forms which are considered transitional between the prosobranchs and the pulmonates (Fig. 1,11B). Subsequently, there occurred a shortening of the connectives that brought the supraintestinal ganglion to the right Bide and the subintestinal ganglion to the left side. At the same time, shortening of the connectives led to a fusion of ganglia. The primitive pulmonate Chilina (Fig. 11C) represents a presumed transitional stage in this process (Haeckel, 1913). In Chilina, the supraintestinal ganglion has fused with the right pallial ganglion to form the bi-lobed right parietal ganglion. The subintestinal ganglion has moved close to the visceral ganglion but has not yet fused with it. In the final stage of pulmonate evolution, attained by the stylommatophoran pulmonates such as Achatina (Fig. 11D), a further shortening of the connectives resulted in the fusion of the subintestinal and visceral ganglia (Pelseneer, 1901; Krull, 1934). Thus, of the five posterior ganglia of early pulmonates, only the left pallial ganglion (now called left parietal) has not undergone fusion.
Of particular interest in the above account is the assumption that the two intestinal ganglia fused asymmetrically. While the supraintestinal ganglion fused with the palliaganglion on the right, the subintestinal ganglion fused with the visceral ganglion in the centre. It is our contention that the similarity of the neurones RPrl and V1 is a vestige of an earlier bilateral symmetry in the locations of the sub-and supraintestinar ganglia. The cells can be called an homologous pair because RPrl was originally located in the supraintestinal ganglion and VI was located in the subintestinal ganglion. While this argument is speculative, it is nonetheless consistent with certain previous interpretations of the evolution of the pulmonate nervous system (Barg-mann, 1930; Krull, 1934; Brace, 1977). Our data on identifiable cells therefore support the evolutionary view presented, which previously has been supported only by observations on the geometric arrangement of the ganglia and the innervation of the ganglia by peripheral nerves.
ACKNOWLEDGEMENTS
This work was supported by a grant given to R. Chase from the Natural Sciences and Engineering Research Council of Canada.