Electrophysiology and Luminescence of An Ophiuroid Radial Nerve

Propagation of nervous activity along echinoderm nerves is poorly understood. Studies on echinoid and asteroid radial nerve cords reveal summed nervous activity which propagates only short distances and always in a slow decremental fashion (Sandeman, 1965; Millott & Okumura, 1968; Binyon & Hasler, 1970; Podol’skii, 1972). However, these findings are not consistent with behavioural observations which indicate that the echinoderm radial nerve cord (RNC) transmits information in a rapid non-decremental fashion (Smith, 1950; Podol’skii, 1972).

The ophiuroid Ophiopsila Californica exhibits a visually observable luminescence which conducts the length of the arms in a rapid all-or-none manner. Furthermore, as will be demonstrated in this study, the luminescence is under the control of the RNC. The rapid conduction velocity and the all-or-none nature of the luminescence response are not compatible with mechanisms of nervous propagation previously described in asteroids and echinoids (Sandeman, 1965; Millot & Okumura, 1968; Binyon & Hasler, 1970; Podol’skii, 1972). These behavioural observations led to the search for nervous activity which could control rapidly propagated information such as the control of luminescence. In this paper, I present extracellular recordings from the ophiuroid RNC which demonstrate the presence of rapidly propagating, all-or-none, single unit spikes. Single unit spikes have not been reported for other classes of echinoderms. Morphological studies described here indicate that the single unit recordings in the ophiuroid may be attributed to the development of axons which are larger in diameter than those described in either asteroid or echinoid RNCs.

Ophiopsila Californica is a large brilliantly luminescent ophiuroid found subtidally off the Southern California coast. The length of the arms (up to 20 cm) makes this species particularly suitable for electrophysiological recordings. Collected specimens were maintained in a recirculating sea water system at 15 °C.

Luminescence propagation in the intact animal was recorded by use of multiple fibre optic light pipes which were positioned along the arm of a quiescent animal. The central disc or RNC of the brittlestar was then electrically stimulated with 100 to 500 ms duration pulses. The resultant luminescence propagated distally along the arm. Light signals were recorded with an EMI 9592B photomultiplier and amplified through a Philbrick P2AU operational amplifier. Signals were registered on either a Tektronix Type 565 oscilloscope or a Brush Type 260 chart recorder.

For determining the ionic requirements of luminescence in the intact arm, the animal was first incubated in calcium-free sea water (CaFSW) with 10 mm-EGTA for 90 min. Then pieces of arm were cut off and placed in a vial into which one or several of the following solutions were injected; 0·54 M-KCI, 0·54 M-NaCl, 0·36 M-CaCl₂, 0·36 M-MgCl₂, 0·36 M-MnCl₂, 0·36 M-BaCl₂, and 0·36 M-SrCl₂. The light production was monitored by a photomultiplier tube and displayed on a Brush chart recorder. Membrane fractions of luminescent tissue were prepared by thoroughly homogenizing an arm in either CaFSW or sodium and calcium-free water (Na-CaFSW). In Na-CaFSW the sodium was replaced with choline. Both solutions contained 10 mm-EGTA and 50 mm-Tris buffer at pH 8·0. The activity of the extracts was tested by monitoring light production while injecting various solutions into the extract. The ionophore A23187 was solubilized in 1% DMSO and used in concentrations of 10⁻⁴ M to 10⁻⁷ M.

For electrophysiological recording it was necessary to completely immobilize an animal in order to prevent autotomization. To do this an animal was first anaesthetized in CaFSW at 12 °C for 30 min. Then the animal was positioned oral side up to the bottom of the recording chamber. For immobilization, Lactona periphery wax was wedged all along the lateral sides of the arms. The segmentally organized ventral plates were then teased off to expose the RNC. The ventral plates could only be removed after treatment with CaFSW. Most electrical recordings were performed on the RNC without removing it from the arm. However, in a few cases the lateral nerve cords and dorsal hyponeural nerve cord were severed from the RNC and the RNC was removed from the arm. After an operation was completed, the arm was flushed with normal sea water (NSW) for at least one hour before the experiment was begun. The temperature of the recording chamber was maintained at 15·0 ± 1 °C unless otherwise specified.

Suction electrodes were used for stimulating and recording from the RNC. Polyethylene (Intramedic) tubing with a 1·9 mm O.D. and 1·3 mm I.D. was drawn under a flame to a final tip O.D. of 200 µm. This tip size corresponded to the typical diameter of the RNC. The electrode tip was lowered in a perpendicular manner to overlie but not touch a region of RNC which was to be recorded from. Then gentle suction was applied through the electrode by means of a syringe. The RNC lifted slightly to meet the electrode tip and a good seal was obtained. Usually the seal around the tip was maintained indefinitely and there was no need to apply further suction. Two immobile recording electrodes and a single mobile stimulating electrode were positioned along the arm in this manner. Rectangular pulses of 1−5 ms duration and 10−150 V were used to stimulate the RNC. The signals were amplified by a Tektronix Type 122 amplifier (0·80−10000 Hz bandwidth) in a single-ended mode, i.e. the bath electrode and the negative input of the amplifier were both tied to the ground lead of the amplifier. Recorded signals were registered on a Tektronix Type 565 oscilloscope and a Brush Type 260 chart recorder.

RNC tissue was prepared for electron microscopy by first removing the cord from the arm in CaFSW. The RNC was fixed in 1 % OsO₄ in NSW. The tissue was dehydrated through an acetone series and embedded in Spurr plastic. Thin sections were stained with lead citrate and uranyl acetate and were examined on a Siemens electron microscope.

Luminescent cells were found both within and surrounding the nervous tissue along the entire arm length (Brehm & Morin, 1977). Luminescence in the intact animal was elicited by either mechanical stimulation of the arms or direct electrical stimulation of the RNC. With weak stimulation the spread of luminescence in the RNC and surrounding tissue was restricted to only a few millimetres from the point of stimulation. However, with stronger stimulation the luminescence propagated the entire arm length without failure. The luminescence propagated at rates between 10 and 85 cm/s; this range of variation was seen within a single animal. The response to a single stimulus was a series of flashes which could be recorded along the entire arm (Fig. 1). Neither the number of flashes nor the individual flash intensities showed significant decrement over the length of the arm. The propagation of luminescence was blocked at any point where the RNC was severed.

Unfortunately, immobilization of an arm, a step which was a prerequisite for electrical recording, resulted in an inhibition of luminescence. Therefore, conditions adequate to provide electrical recordings blocked luminescence.

Luminescence in the animal was completely blocked when the animal was placed in CaFSW. Whole arm pieces containing RNC which were incubated in CaFSW with 10 mm-EGTA did not luminesce unless both KC1 and CaCl₂ were added. Extracts of the arm prepared in CaFSW containing 10 mm-EGTA still retained the capacity to luminesce when Ca was added back (Fig. 2). Addition of CaCl₂ in this case stimulated a slight amount of light production but KC1 was also required to elicit a large flash. Both the intact arm pieces and arm extracts flashed only once in response to addition of KC1 and CaCl₂. The extract responded to the Ca ionophore A23187 at a concentration as low as 10⁻⁷ M if CaCl₂ was added. If the ionophore was used, KC1 was not required to elicit a flash when CaCl₂ was added. SrCl₂ replaced CaCl in the presence of either KC1 or A23187. BaCl₂ stimulated a very slight light production and MgCl_a and MnCl₂ resulted in no light. Both the intact arm pieces and the extracts flashed in Na-CaFSW if CaCl₂ and KC1 or A23187 were added. The active fraction of the extract was particulate and all activity in a test tube spun down at 1000 g in 3 min. The dependence on the presence of either KC1 or A23187 indicated that the active fraction was composed of membranous-bound light-emitting molecules.

No spontaneous electrical activity was recorded from the RNC. A complex train of single unit spikes was recorded in response to a rectangular stimulus pulse when the stimulating and recording electrodes were separated by a distance greater than approximately 1 cm. Electrical recordings at distances less than 1 cm are treated in the following section (Fig. 3 A). The single unit spikes were multiphasic with a spectrum of amplitudes up to 75 µV and individual durations ranging from 10 to 20 ms (Fig-3B);

The spikes were propagated in an all-or-none manner with a distinct threshold. No change in amplitude occurred with increases in stimulus intensity (Fig. 3 B). However’ the number of spikes observed in response to a single pulse was graded with stimulus intensity. The duration of the spike train reached a maximum when a strong enough stimulus intensity was reached. At a 2 cm distance between stimulating and recording electrodes this maximal spike train had an average duration of 400 ms (s.D. + 100 ms). This duration stayed relatively constant with further increases in interelectrode distance (Fig. 4). An important finding was that even if the electrodes were separated by distances as great as 15 cm (the typical length of an entire RNC) a train was still recorded in undamaged preparations. The train at 15 cm distance often reflected a large number of units firing and the characteristics of the train were similar to that recorded at a 3 cm distance (Fig. 4) except for a substantially longer propagation time to reach the recording electrode.

Usually the spikes comprising a train showed too much variability in waveform and amplitude from one stimulus to the next to allow consistent identification of single units along the cord. This variability resulted from summation and cancellation of closely spaced units and also from large changes in the conduction velocity of units with repetitive stimuli (results will be presented later). In some cases, however, the largest spikes of a train could be individually traced along the RNC on the basis of waveform, conduction velocity, and a threshold which was somewhat lower than other units. In 31 cases examined units, were traced for distances up to 4·2 cm, which may represent the length of the longer RNC neurones. This distance was still only about the length of the RNC. In these cases the conduction velocity of the particular unit was computed on the basis of interrecording electrode distance. An average value of 35 cm/s with a range of 24−50 cm/s was obtained by this method. These values were similar to the 38 cm/s average and 15−78 cm/s range computed for the leading spike of the train on the basis of stimulating to recording electrode distance.

The conduction velocity of a unit increased in a graded manner with repetitive stimuli at intervals of 2 s or less (Fig. 5). Stimuli at 500 ms intervals were sufficient to establish and maintain a state of maximum conduction velocity. The conduction velocity was adjusted to any value between minimum and maximum by altering the stimulus interval. The absolute latency change with repetitive stimulation increased linearly with interelectrode distance (Fig. 6). The fact that the latency change increased with increased recording distance demonstrated that the change was not occurring at the site of stimulating electrode. When expressed as a change in conduction velocity the values were independent of recording distance. The average maximum increase in conduction velocity with repetitive stimulation was 18% and on one occasion a value of 37% was recorded. The change in conduction velocity with repetitive stimulation was independent of both stimulus strength and subthreshold repetitive stimuli.

The effect of repetitive stimulation on the number of units firing was also examined. The lowest stimulus intensity which would elicit any spikes was used and twin stimuli were delivered at intervals ranging from 30 s to 1 ms. The low stimulus intensity minimized the number of spikes to be examined. Between 30 s and 40 ms intervals no change in the number of spikes was usually observed in response to the second pulse. In a few exceptional cases, however, twin stimuli at 1 s intervals increased the number of units firing (Fig 5.). Intervals between 10 ms and 40 ms decreased the number of units responding to the second pulse, presumably as a result of the refractory period of the fibres.

At 15 °C the spikes propagated the entire RNC length even though the single units are probably much shorter than the RNC length. However, if the RNC was cooled to 4 °C, the units propagated no farther than 3·5−5· 8 cm in response to a single stimulus. This further supports an average length of 4·2 cm for the longer fibres as determined earlier on the basis of following single units. At 4 °C some spikes propagated the entire RNC length with repetitive stimulation at 1 s intervals and less. At 4 °C the number of spikes recorded at distances greater than 3·5−5·8 cm with repetitive stimulation was approximately equal to the number of spikes responding to a single stimulus at 15 °C. All studies involving repetitive stimuli were complicated by the fact that the RNC fatigued irreversibly with prolonged stimulation.

In every experiment attention was paid to the polarity of the RNC but it made no detectable difference at which end of the cord the recording and stimulating electrodes were placed. Also, the responses recorded from an intact RNC were identical in all respects to those recorded from an isolated RNC.

The ionic requirements for the all-or-none spikes were identical to the compound potential and are discussed in section B.

A graded potential, resulting from summed neuronal activity, was recorded when stimulating and recording electrodes were separated by less than 2 cm (Figs. 3 A, 4 and 7). The potential conducted decrementally, its amplitude decreasing logarithmically with recording distance. The amplitude was linearly related to stimulus intensity until saturation was achieved (Fig. 3 A). Amplitudes as large as 1-2 mV were recorded at very short interelectrode distances. The duration of the potential increased linearly from 18 ms at the shortest interelectrode distance to 250 ms at a 2 cm distance (Fig. 4). The compound potential was comprised of a small initial positive component which was partially obscured by the stimulus artifact at short distances and followed by a large negative component (Figs. 3 A, 4). The conduction velocity range was quite slow, 2·1−4·5 cm/s for the two components. The compound potential propagated in either direction along the RNC. The fibres giving rise to the potential had a very long refractory period between 100 and 400 ms.

All electrical activity was blocked in CaFSW (Fig. 7A). The electrical activity returned rapidly if either 10 mm-BaCl₂ or 10 mm-CaCl₂ was added to the CaFSW or if NSW was added (Fig. 7B). Electrical activity in Na-free sea water was unchanged over a 6 h period when the Na was replaced first by sucrose and later by isotonic choline chloride (Fig. 7C).

The RNC, a bundle of neurones, averaged 300 μm by 100 μm in cross-section and showed a swelling of diameter in each segment. The RNC was typically 15 cm in length. It was comprised of small, unsheathed fibres with a ventral margin of nucleated cell bodies (Fig. 8). The neurone cell bodies averaged 15µm in diameter and constituted about 75% of the cell numbers bordering the RNC.

The fibres in the RNC were entirely naked and had a consistent 125 A wide extracellular space between plasma membranes (Fig. 9). The fibre diameter ranged from 0·1 µm to at least 8·0 µm. It is not known whether the small fibres were axons or dendritic processes (Fig. 9B). The largest diameter fibres were at least 8·0 µm in diameter (Fig. 9 A) and they were fewer than 12 in number in any one cross-section of the entire RNC (Figs. 8, 9A). These exceptionally large diameter fibres were organized into bundles of 2 or 3 and they were probably axons based on their position and the occasional presence of both granular and agranular vesicles ranging from 550 Å to 775 Å (Figs. 8, 9 A). Of the remaining fibres, numbering many thousands, most had diameters less than 1·0 µm.

The compound potential described here from the ophiuroid RNC is similar to that recorded from the asteroid RNC (Binyon & Hasler, 1970; Podol’skii, 1972) and echinoid RNC (Sandeman, 1965 ; Millott & Okumura, 1968). The compound potential in ophiuroids, asteroids, and echinoids all show a response amplitude which is graded with stimulus intensity and a logarithmic decrease in amplitude with increased recording distance. Furthermore, the long refractory period of 100 ms-400 ms and the slow conduction velocity range of 2·1−4·5 cm/s reported here for the ophiuroid are also consistent with the compound potential in asteroids and echinoids. In all three classes of echinoderms examined, there are examples of decrementally controlled behavioural activities which are under RNC control. Papula retraction in asteroids (Bullock, 1965), spine movement in echinoids (Millott, 1966), and bioluminescence in ophiuroids all show decremental spread under particular stimulus conditions. However, there is still no compelling evidence to link such decrementally propagated behavioural activity to decrementally recorded electrical activity in the RNC. Moreover, the compound potential may have no functional significance whatsoever since the stimulus conditions are not physiological.

All studies on echinoderm RNC’s (Sandeman, 1965; Millott & Okumura, 1968; Binyon &Hasler, 1970; Podol’skii, 1972) with the possible exception of one (Takahashi, 1964) have dealt exclusively with this graded decremental neural activity, yet behavioural manifestations of through conducted RNC activity exists in at least two of the three classes of echinoderms. The control of ambulacral pedicle closure (Podol’skii, 1972) and tube foot-pointing direction (Sandeman, 1965) in asteroids, and as shown in this paper, ophiuroid luminescence can all be rapidly through conducted over the surface of the animal. In all three cases the propagation of the response is blocked by severing the RNC. Through-conducted spikes which might be responsible for the control of such activity have been recorded in this study on the ophiuroid. The spikes propagate in all-or-none manner over long distances as opposed to the graded activity which cannot be recorded over distances greater than 2 cm in most studies. The large size of the single unit spikes, the short refractory period, and the rapid conduction velocity indicate that these units are unique from those giving rise to the graded activity.

At the physiological temperature of 15 °C, single unit spikes were recorded over the full 15 cm length of the RNC. However, if the preparation was cooled to 4 °C all recordable electrical activity was blocked within 6 cm from the stimulating electrode. Under these conditions repetitive stimuli are necessary to enable spikes to be recorded over the entire RNC length. A likely interpretation of these findings is that the cold temperature reduces the efficiency of synaptic transmission as it does at the frog neuromuscular junction (Katz & Miledi, 1965). At physiological temperatures, the synapses are transmissive and allow through-conducted information to propagate the entire RNC length in response to a single stimulus pulse. These results indicate that a linear series of neurones comprise the through-conduction system of the ophiuroid RNC. Furthermore, the neurones are capable of all-or-none propagation and are perhaps as long as 6 cm. This length is supported by experiments where single units, identified on the basis of threshold, conduction velocity, and waveform, were followed over a 4·2 cm distance.

Apparent increases in conduction velocity averaging 18% were recorded from single units in response to repetitive stimulation. Moreover, the conduction velocity of the spikes could be set to any value within lower and upper limits simply by adjusting the stimulus frequency. The results could be explained by changes in either synaptic delay or in actual conduction velocity of the neurones. Conduction velocity increases up to 20% with repetitive stimulation have been observed in annelid giant fibres (Bullock, 1951) and in the hemichordate ventral nerve cord (Pickens, 1970). The mechanisms of the conduction velocity increases in these preparations is not understood. However, in the coelentrate Calamactis it has been suggested that apparent conduction velocity increases of 16% are due to changes in the speed of synaptic transmission (Pickens, 1974). In the ophiuroid, changes in synaptic efficacy would not be expected to result in latency changes as large as 60 ms (Fig. 6). The mechanism of the conduction velocity increases with repetitive stimulation remains unresolved in the ophiuroid.

The largest fibres in the RNC reach at least 8 µm (Figs. 8, 9A). This is in contrast with the bulk of the RNC fibres which are considerably smaller, in some cases as small as 0·1 µm (Fig. 9 B). The large fibres could represent either axon hillock regions, dendritic processes, muscle tails lacking myofilaments, processes from epithelial cells or axons. The large fibres are usually seen in a region devoid of somas arguing against an axon hillock region. Both muscle tails and epithelial cell processes have been demonstrated in echinoderm RNCs (for review see Pentreath & Cobb, 1972) and care must be taken not to confuse them with actual neuronal elements. In the case of Ophiopsila, however, the large fibres are found deep within the RNC where muscle tails and epithelial processes would not be expected to penetrate. Furthermore, the presence of small granular and agranular vesicles in the fibres suggests that the fibres are in fact not dendrites, muscle tails, or epithelial cell processes, but rather are axons. These axons are not homologous to the ‘giant’ axons in the hyponeural nerve of the ophiuroid Ophiothrix (Pentreath & Cottrell, 1971). The hyponeural nerve overlies the RNC. The axons described in this study are scattered deep within the RNC and are found in association with the RNC, not in the hyponeural nerve cord, in every cross-section examined.

The size of the largest axons in this ophiuroid is substantially larger than any axons previously described in echinoderm RNCs. In asteroids, Kawaguti (1965) reports the largest axons to be 0·5 µm and Cobb (1970) describes axonal varicosities reaching up to 3·0 µM. In echinoids the findings of maximum diameter are 1·0 µm and 2·5 µm by Kawaguti, Kamishima & Kobashi (1965) and Cobb (1970) respectively. The large axons in Ophiopsila are probably responsible for stimulus induced through conduction of luminescence. The large size of these axons may explain why single unit electrical activity can be recorded in the ophiuroid but apparently not in the asteroid or echinoid. The bulk of the ophiuroid axons and presumably all axons in asteroid and echinoid RNCs are too small to record from unless excited simultaneously. Simultaneous stimulation of RNC axons leads to the compound potential discussed earlier.

The luminescent tissue in Ophiopsila is contained within the RNC and surrounding tissue (Brehm & Morin, 1977). The luminescent cells within the RNC are restricted to the ganglionic regions and do not make direct connexions between segments. Therefore, the propagation of luminescence along the RNC cannot be provided exclusively by the luminescent cells. Furthermore, the size of the luminescent cells are comparable to the smallest fibres in the RNC (Brehm & Morin, 1977). The luminescence, however, propagates along the RNC at rates up to 85 cm/s. It seems unlikely that 0·4 µm diameter luminescent cells within the RNC could provide this rapid propagation. Luminescent cells in adjacent ganglia must therefore, be connected by large, rapidly propagating, all-or-none neurones of the type recorded from in this study. The largest neurones have never been observed to be luminescent. Since simultaneous recording of electrical activity and luminescence was not practical, it was not determined how such control is achieved.

The finding of a calcium requirement for the compound potential in the RNC of asteroids (Binyon & Hasler, 1970) and echinoids (Millott & Okumura, 1968) is supported in this study on the ophiuroid RNC. All compound and single unit activity in in the ophiuroid is reversibly lost in a preparation where the calcium is removed, whereas substitution of sodium with sucrose or choline has no effect. Calcium spikes from neurones have been identified in the somas and presynaptic terminals of molluscan neurones. In the case of Aplysia giant neurones the calcius spike is restricted to the soma and does not invade the axon (Junge & Miller, 1974) For this reason, removal of calcium from the sea water bathing the neurones does not block axonal propagation in that preparation. In the ophiuroid, extracellular stimulation is probably stimulating the axons directly. One would not expect propagation to be lost unless the axons themselves exhibited a calcium requirement for axonal propagation. Alternatively, low calcium might block propagation by blocking synaptic transmission. However, evidence is presented which suggests that the axons are much longer than a few millimetres which is the distance over which all recordable activity is lost in CaFSW. The requirement of calcium for neuronal propagation in echinoderms is a subject which warrants intracellular investigation.

The luminescent cells within and surrounding the RNC also exhibit a calcium requirement to flash. Addition of KC1 to depolarize the cells or electrical stimulation of the RNC leads to multiple flashes in the intact arm or arm pieces. When the arm is incubated in CaFSW, however, luminescence cannot be elicited with KC1 or electrical stimulation. Membrane fractions of dissociated luminescent tissue will also flash if KC1 or the calcium ionophore A23187 are added in the presence of either Ca or Sr. Neither Mn nor Mg will substitute for the Ca but Ba will activate a slight amount of luminescence. If the fractions are prepared in sodium-free sea water the extract will still flash in response to KC1 and Ca addition. Therefore, the luminescent tissue, like the RNC neurones, exhibits a strict calcium requirement and lacks a sodium requirement. The similar ionic requirements between luminescent cells and RNC neurones are particularly interesting in the light of morphological coexistence of the two within the RNC (Brehm & Morin, 1977). Also, the luminescent tissue in the RNC and surrounding tissue, when examined by light microscopy, are morphologically indistinguishable from axons within the RNC (Brehm & Morin, 1977). It seems likely then, that the luminescent cells evolved from neurones thus providing a calcium conductance to activate the luminescent molecules within the cell. This development would also place the luminescent cells in close proximity to the rapidly propagating neurones which activate the luminescent cells all along the RNC.

I thank Dr James Morin for his guidance and support during the course of this project. I am deeply indebted to Bibbi Wolowski and Herman Kabe for their expert assistance and time which they devoted to the morphological aspects of the work. Finally, I gratefully acknowledge Laurinda Jaffe, Dr Albert Herrera, and Gail Mandel for their critical reading of the manuscript and helpful discussions concerning this project.