Excitable Epithelial Cells in the Bioluminescent Scales of A Polynoid Worm; Effects of Various Ions On the Action Potentials and On the Excitation-Luminescence Coupling

Bioluminescence in different animal species can be the result of intracellular activities which are rapidly modulated (Henry & Michelson, 1978). However, very little information is available on the electrophysiological properties of light emitting cells. In Cnidaria, extracellular recordings showed that flashes were correlated with potential changes generated in epithelial or nervous conduction systems (Morin & Cooke, 1971 ; Anderson & Case, 1975; Bassot et al. 1978). Until recently, the demonstration of intracellularly recorded membrane activities in light emitting cells has only been obtained in the protozoan Noctiluca (Eckert, 1966) and in larval firefly (Oertel & Case, 1976).

In bioluminescent polynoid worms, the luminous activity is intracellular and originates from a single layer of epithelial cells modified as photocytes. These cells have been found suitable for electrophysiological investigations with microelectrodes (Herrera, 1977, 1979; Bilbaut, 1978 a, 1980). The photocytes constitute an homogeneous cell population, the photogenic area, localized in thin epithelial plates, the scales (elytre), which are inserted dorsally on the worm. These are excitable epithelial cells (Herrera, 1979; Bilbaut, 1980) which are electrically coupled by numerous gap junctions (Pavans de Ceccatty et al. 1972; Bilbaut, 1980). In the photogenic area a few photocytes would receive direct excitatory innervation (Herrera, 1979).

The photocytes are epithelial effector cells in which the luminescence is triggered by specific excitation of the cell membrane (Herrera, 1979; Bilbaut, 1980). Membrane electrogenesis in invertebrate effector cells often appears to require Ca ions. This has been demonstrated in the muscle fibres of crustaceans (Atwood, 1975, for review), annelids (Ito & Tashiro, 1970; Ito, Kuriyama & Tashiro, 1970) and insect larvae (Deitmer & Rathmayer, 1976; Fukuda, Furuyama & Kawa, 1977). An inward Ca current was also observed during the action potential produced in excitable glandular epithelial cells in a Cnidaria (Mackie, 1976). Furthermore, in the effector cells, the activation of effector mechanisms generally involves Ca ions which, in muscle fibres for example, are made available by the calcium influx of the action potential or can be released from intracellular store sites (Ebashi & Endo, 1968).

In the photocytes of the polynoid worm Acholoe, two kinds of propagated action potentials can be successively generated after intra- or extracellular stimulation of the preparation (Fig. 2a and 26). The firing threshold, the amplitude and the wave form of both responses have been shown to be different. One of these action potentials is specifically correlated with the light production (Bilbaut, 1980) which corresponds to a brief flash (Bilbaut & Bassot, 1977).

The present study intends to detail the ionic dependence of both action potentials produced in the photocytes in the polynoid worm Acholoe astericola (delle Ch.) and also to characterize the mechanism of the excitation-luminescence coupling. These results have been previously exposed in a short note (Bilbaut, 1978b). They indicate principally that (i) an inward Ca current occurs during the action potential correlated with the luminescence, (ii) the light production is dependent on the Ca entry into the photocytes. These data will be compared with the results obtained by Herrera (1979) in the bioluminescent scales of the polynoid worm Hesperonoe.

In addition, in the scales of the polynoid worms, the luminous epithelium is in continuity with an excitable and conducting non-photogenic epithelium (Herrera, 1979 ; Bilbaut, 1980). These common epithelial cells produce action potentials (Fig. 2 c). In Acholoe, their ionic dependence will also be examined in the present study.

Acholoe astericola is a commensal of the starfish Astropecten aurantiacus. The starfishes were collected by scuba-diving near the Marine Station of Banyuls-sur-Mer and kept in filtered natural sea water at a constant temperature (15 °C). For the experiments, the worms were carefully removed from the starfish and immobilized by cooling. Scales were detached from the worms using fine scissors.

The scale diameter is around 1·5–2 mm. The photogenic area lies in a part of the lower plane of the scales. It is surrounded by non-photogenic epithelial cells which also constitute the upper plane of the scales (Fig. 1). Because of the presence of an external collagenous cuticle which prevents the microelectrode penetration into the underlying epithelial cells, the scales were cut in narrow strips (0 · 5 mm wide). A fragment was transported into a dish in which it was maintained in a vertical position by two suction electrodes disposed one on each face (Fig. 1). Under these conditions, both epithelial layers were directly accessible to the microelectrodes.

The preparations were stimulated extracellularly with either suction electrode. These were made from drawn polyethylene tubing filled with physiological saline solution in contact with silver chloride wire connected to a stimulus isolation unit. Square pulses were delivered to generate the action potentials in the epithelial cells. Intracellular stimulation was not used for the following reasons. In the epithelial cells intracellularly stimulated with double barrelled microelectrodes, the firing threshold for the spike generation was between – 20 and +10 mV partly masking the membrane responses (Bilbaut, 1980). When a second microelectrode was placed in another epithelial cell, propagated action potentials were recorded (Herrera, 1979; Bilbaut, 1980). However, attempts showed that it was difficult to hold the microelectrodes in two epithelial cells for more than 10 or 15 min. These different inconveniences led to the use of extracellular stimulation only.

The recording glass microelectrodes were pulled on a solenoid microelectrode puller (Narishige, PD5) and were filled with 3 M-KCI. Their initial resistance was 40 – 60 M Ω. They were inserted in an hydraulic microdrive micromanipulator (Narishige, MO 10) and connected to the input of a high-impedance preamplifier. Signals were displayed on an oscilloscope (Tektronix, 5103 N) and an oscillographic chart recorder (Hewlett Packard, 7402 A).

The normal saline had the following composition: NaCl, 500 mm; KC1, 10 mm; CaCl₂, 20 mm; MgCl₂, 12 mm and Tris (tris-(hydroxymethyl)-aminomethane)-HCl buffer (pH: 7-4), 10 mm. Na-free solutions were achieved by equimolar replacement of NaCl by Tris-, choline-, or lithium-chloride. Ca-free solution was obtained by replacing CaCI₂ with NaCl. In solutions where the Ca concentration was lowered, 2 and 10 mm-CaCl₂ were added to Ca-free solution and osmotically balanced with NaCl. The Mn saline was obtained by replacing 15 mm-NaCl in the normal solution by an equimolar amount of MnCl₂. In Sr or Ba solutions, CaCl₂ was substituted by equimolar amounts of SrCl₂ or BaCl₂. When used, tetrodotoxin (10^{− 6}g/ml) or tetraethylammonium chloride (10 mm) was added to the normal saline.

Luminescence was collected by an optic fibre light guide positioned over the preparation and connected to a photomultiplier tube (R.C.A. 1 P21). Anode current was displayed on the oscilloscope and chart recorder. In standard conditions, the successive stimulation of the bioluminescence in the scales progressively exhausted the luminous load of the photogenic area. No recovery of the bioluminescent capacities occurring in the isolated scales (Bilbaut & Bassot, 1977), the flash intensity decreased in proportion to the number of electrical pulses applied to the preparations. During the present experiments, the luminescence in the scale fragments was episodically stimulated with non-repetitive electrical pulses and only drastic changes in flash intensity and duration were considered as significant (± one half amplitude and duration compared to the control). The intensity of luminescent responses varied from one scale to another so the gain of the luminous activity records was arbitrarily chosen at the beginning of each experiment.

The experiments were performed at room temperature (20 °C). In all figures, the lower traces correspond to membrane electrical signals and the upper traces, when present, to luminescence recording. Asterisks indicate the stimulus artefacts. The duration of membrane responses was measured at one half repolarization.

The mean resting potential of both the photogenic and non-photogenic epithelial cells is – 71 ± 5 mV. When an extracellular stimulus (5 – 7 V ; 10 ms) is applied to the preparation, a simple action potential of 33 ± 8 mV amplitude and 74 ± 31 ms duration is recorded from the photocytes of the isolated scales (Fig. 2 a). If the stimulus intensity is increased (7 – 12 V; 10 ms) another response is generated in the same photocyte (Fig. 2 b). This is a 2-component action potential beginning with an overshooting spike (amplitude, 88 ±5 mV; duration, 17 ±3 ms) which is followed by a secondary depolarization (amplitude, 22 ± 5 mV; duration, 295 ± 135 ms). The luminous activity in the photogenic epithelium is typically correlated with only the 2-component action potential (Fig. 2b) although, in some cases, weak luminescence is associated with the simple action potential.

In the non-photogenic epithelium, extracellular electrical pulses (5 – 7 V; 10 ms) produce simple action potentials (amplitude, 60 ± 8 mV ; duration, 41 ± 3 ms) (Fig. 2 c).

In Na-free solutions, the resting potential of the photocyte membrane depolarized, by 4 – 6 mV when Na ions were replaced by Tris and 6 – 10 mV when replaced by choline.

The simple action potentials disappeared 30 – 40 min after incubation of the preparation in both Na-free solutions (not illustrated).

The overshooting amplitude of the 2-component action potential was unchanged although the overall amplitude decreased because the resting potential was less (Fig. 3 A, B). Simultaneously, the secondary depolarization was reversibly abolished (Fig. 3 A, B).

Lithium can sometimes be substituted for sodium in the spike generation of some excitable cells (Hille, 1970). When the scales were soaked in lithium solution, the membrane potential was depolarized by 10 – 15 mV. Occasionally, slow spontaneous membrane oscillations, around 5 – 6 mV in amplitude, were observed. The effects of lithium ions on the simple and on the 2-component action potentials (Fig. 3 C) were basically similar to those observed in Tris or choline solutions.

In Na-free solutions, especially choline and lithium solutions, the duration of the overshooting spike potential increased (Fig. 3 B, C) and reached 35 – 60 ms against 17 ms in normal saline. The depolarizing phase of the spike potential in Na-free solution was superposable to that observed in normal saline indicating that only the duration of the repolarizing phase was increased.

Flashes occurred in the various Na-free solutions and were accompanied by overshooting spikes alone (Fig. 3 A, B, and C). The strong decrease in flash intensity shown in Fig. 3B (choline solution) resulted from previous stimulation which progressively exhausted the bioluminescent load of the scale fragment. The increase in the flash duration (Fig. 3 A, B) appeared to be correlated with the increase in the spike duration. In lithium solution, the correlations between electrical and luminous astivities were not precisely established because of the weak intensity of the flash control (Fig. 3C). However, the records show that luminescence was produced when the lithium replaced sodium in the external solution (Fig. 3 C).

Tetrodotoxin which is recognized as a very effective blocker of sodium channels in many excitable cell membranes (Kao, 1966; Moore & Narahashi, 1967) did not block either the simple or the 2-component action potential when used at a concentration of 10⁻⁶ g/ml.

The electrical activity of the photocyte membranes was strongly modified by Ca-free solution and the effects evolved as follows:

1. 2– min after the change of solutions, the membrane potential began to depolarize and subsequently, the overshooting amplitude of the spike potential decreased drastically (Fig. 4). The secondary depolarization could be diminished.

   Fig. 4.

   View largeDownload slide

   Effects of Ca-free solution, (a) control, (b) 15 min after incubation of the preparation in Ca-free solution, the overshooting spike potential is reduced by 11 mV and the flash intensity decreases, (c) after 63 min in Ca-free solution, the cell membrane is depolarized by 14 mV and a weak membrane response subsists; no luminescence is produced, (d) recovery; the slight luminous intensity of the flash results from a glow of 23 s in duration which has been observed 30 min after soaking the scale fragment in Ca-free solution; this response has been accompanied by a depolarizing plateau potential as shown in Fig. 5 ; the light production during the plateau response has partly exhausted the bioluminescent load of the preparation.

   Fig. 4.

   View largeDownload slide

   Effects of Ca-free solution, (a) control, (b) 15 min after incubation of the preparation in Ca-free solution, the overshooting spike potential is reduced by 11 mV and the flash intensity decreases, (c) after 63 min in Ca-free solution, the cell membrane is depolarized by 14 mV and a weak membrane response subsists; no luminescence is produced, (d) recovery; the slight luminous intensity of the flash results from a glow of 23 s in duration which has been observed 30 min after soaking the scale fragment in Ca-free solution; this response has been accompanied by a depolarizing plateau potential as shown in Fig. 5 ; the light production during the plateau response has partly exhausted the bioluminescent load of the preparation.

2. when the spike amplitude was between 45 and 60 mV, brief (10 ms) electrical stimuli applied to the preparation produced particular membrane responses. These consisted in an initial spike followed by a long-lasting depolarizing plateau potential which could last as long as 180 s. The membrane potential of the plateau responses was often maintained at a steady value, equal to, or 4–5 mV lower than that of the initial spike (Fig. 5). In other cases, regular or disordered oscillations of the depolarizing plateau potential prolonged the initial spike (Fig. 6). The termination of the plateau phase was always marked by an abrupt repolarization (Fig. 5 and 6).

   Fig. 5.

   View largeDownload slide

   Responses produced during the replacement of the normal saline by Ca-free solution. Example of smooth depolarizing plateau potential correlated with a continuous luminescent activity; the initial membrane response consists of a spike which is associated with a flash (arrowheads).

   Fig. 5.

   View largeDownload slide

   Responses produced during the replacement of the normal saline by Ca-free solution. Example of smooth depolarizing plateau potential correlated with a continuous luminescent activity; the initial membrane response consists of a spike which is associated with a flash (arrowheads).

   Fig. 6.

   View largeDownload slide

   Other responses obtained under the same conditions as Fig. 5. Example of oscillatory depolarizing plateau potential accompanied by flashes (arrowheads) which emerge from a continuous luminescent background.

   Fig. 6.

   View largeDownload slide

   Other responses obtained under the same conditions as Fig. 5. Example of oscillatory depolarizing plateau potential accompanied by flashes (arrowheads) which emerge from a continuous luminescent background.

3. if stimuli were repeated immediately after a long plateau potential had been recorded, other plateau spikes were evoked but their amplitude decreased continuously. After 50–60 min in Ca-free solution, the resting potential was around – 50 mV and only brief and weak membrane responses were produced when the preparations were strongly stimulated (Fig. 4). The normal activities of the photocyte membrane recovered after returning to the standard saline solution (Fig. 4).

Luminous activity continued in Ca-free solution during the decline of the spike potential amplitude. Flash intensity appeared to decrease as the action potential amplitude decreased (Fig. 4). However, because of the progressive exhaustion of the bioluminescent products caused by previous and episodic stimulation of the scale fragments, the correlations between spike amplitude and flash intensity cannot be readily established.

Luminescent activity also occurred during the steady or oscillatory plateau potential. In the first case (steady plateau), the luminescence consisted of a brief flash associated with the initial spike followed by a prolonged glow (Fig. 5). The duration of the light production was often as long as the depolarizing membrane response. In some cases, when the duration of the depolarizing plateau potential lasted several minutes, the luminescence ended before the abrupt repolarization of the cell membrane. In the second case (oscillating plateau), the luminescence was characterized by successive flashes which correlated with the oscillations of the depolarizing plateau potential (Fig. 6). A particularity of this response was that flashes emerged from a continuous and steady luminescent background (Fig. 6).

When the spike responses were completely abolished in Ca-free solution, no luminescence was observed whatever the stimulus strength (Fig. 4). After returning to the standard saline, flashes again accompanied the 2-component action potentials (Fig-4).

As would be expected from the above experiments, the amplitude of the overshooting spike of the 2-component action potentials is correlated with the external Ca concentration.

Following a tenfold reduction in external Ca concentration (20 mm to 2 mm), the amplitude of the spike potential decreased by 27–30 mV. In addition, the amplitude and duration of the secondary depolarization were sometimes modified. In some cases, the initial spike was prolonged by an oscillatory depolarizing plateau potential generally shorter (1–5 s) than those observed during the substitution of the normal saline by Ca-free solution as previously reported.

For a twofold change in external Ca concentration (20 mm to 10 mm), the decrease in the overshooting spike amplitude was 10–12 mV. The relationship between the overshooting amplitude changes and external Ca concentration (Fig. 7) is consistent with theoretical values given by Nernst’s equation for Ca electrodes which are respectively, 29 mV for a tenfold change and 8 7 mV for a twofold change at 2 °C.

The resting potential of the photocyte membrane and the amplitude of the simple action potential were not affected in either Ca solution.

After incubation in 2 mm-Ca solution, luminescent activity continued. Luminous intensities appeared to decrease as the overshooting spike amplitude decreased. When the oscillations of the depolarizing plateau potentials occurred they were accompanied by flashes which emerged from a persistent luminescent background.

The results presented above seem to indicate that Ca ions are involved in the mechanism of spike generation of photocyte membranes. Mn ions are known to block inward Ca currents in different excitable cells (Fatt & Ginsborg, 1958; Hagiwara & Nakajima, 1966; Baker, Meves & Ridgway, 1973; Kleinhaus, 1976).

In the presence of 15 mm-Mn, the amplitude of the overshooting spike potential decreased progressively (Fig. 8) and, in some cases, the secondary depolarization was attenuated. Incubation of the preparation in the Mn solution for 15–20 min led to complete abolition of the overshooting spike. However, photocyte membrane electrogenesis was never totally inhibited since, as shown in Fig. 8, the simple action potentials were revealed to be Mn-resistant and could be generated in this solution.

Luminescence could not be evoked under these conditions even when strong electrical pulses depolarized the cell membrane to between –10 and +15 mV for 600 ms (Fig. 9). The effects of the Mn ions on the action potentials and luminescence were reversible (Fig. 9).

Divalent cations such as strontium and barium have been shown to compete with Ca ions as charge carriers in the spike generation mechanism of various excitable cells (Fatt & Ginsborg, 1958 ; Hagiwara, Fukuda & Eaton, 1974; Ito & Tashiro, 1970). The effects of these cations on the excitation of the photocyte membrane was examined.

Equimolar substitution of external Ca by Sr ions maintained the electrical activities of photocytes. In most cases, the first electrical stimulus evoked repetitive firing (Fig. 10). This consisted of successive overshooting spikes generated at various frequencies, lower than 10 spikes per second. The duration of bursts was between 2 and 41 s. Under these conditions, spike amplitude was comparable to that recorded in the normal saline. During these responses, the successive spikes never repolarized completely. Around – 30 mV a slow depolarization occurred and gave rise to a new spike (Fig. 10). These bursts of action potentials ended with a slow repolarization. When the stimulation was repeated immediately after a burst, the duration of the new bursts decreased continuously so that after 4 or 5 stimuli, single 2-component action potentials were elicited. In Sr solution, the duration of single or repetitive overshooting spikes was around 50–60 ms, sometimes reaching 400–500 ms, against 17 ms in normal saline.

When Sr ions were substituted for Ca ions in the external medium, luminescent activities of photocytes were maintained. Repetitive flashes occurred, one for one, with the action potentials in the bursts (Fig. 10). These flashes sometimes summated but, when the repetitive spikes were widely spaced, the flashes were emitted separately and no luminescent background was recorded (Fig. 10). In most preparations, the luminous intensity of flashes was not modified in Sr solution. However, in some (4 of 17), the flash intensity decreased to 1/100 of its initial value while the spike potential amplitude remained unchanged (Fig. 11). After returning to normal saline luminous intensity of flashes again increased as shown in Fig. 11.

Ba ions rapidly depolarized the photocyte membrane by 15 mV. 1–2 min after soaking the preparation in this solution, repetitive spontaneous depolarizing oscillations about 12–16 mV in amplitude and 50–100 ms in duration were sometimes observed. In most cases, repetitive overshooting spikes were spontaneously generated (Fig. 12). Spikes could occur in successive bursts. Repetitive membrane firing always ended by a persistent depolarizing plateau potential and the repolarization of the cell membrane was rarely observed even after 3 or 4 min of maintained depolarization.

In Ba solution, luminous activities were correlated with overshooting rhythmic spikes (Fig. 12) and the flashes often summated at the beginning of the bursts. In scale fragments, luminescence was always exhausted before the complete depolarization of the photocyte membrane producing the persistent plateau potential.

In excitable cells, the repolarizing phase of the action potentials is generally produced by activation of an outward K current, T.E.A. ions are known to prolong the duration of membrane depolarization in many excitable cells (Hagiwara & Watanabe, 1955; Fatt & Ginsborg, 1958) and in some preparations, this effect has been demonstrated to result from blockade of the outward K current (Hille, 1970). In the present study, the effects of T.E.A. ions on the electrical and luminescent activities of the photocytes were examined.

2–3 min after T.E.A. ions (10 mm) were added to the external solution, the duration of the overshooting spike potential of photocytes was lengthened (Fig. 13). Furthermore, in most tested cells, the amplitude of the spike increased by 6–16 mV while that of the secondary depolarization increased by 5–8 mV (Fig. 13). After 5–10 min of incubation, the overshooting spike evoked by a brief electrical stimulus (10 ms) was followed by a long-lasting depolarizing plateau potential which could be maintained around zero potential during 9–15 s before repolarization. Rhythmic positive oscillations of between 1 and 15 mV were invariably superimposed on the plateau (Fig. 13).

The intensity and duration of flashes appeared to increase as the amplitude and duration of spike potential increased (Fig. 13). During the plateau potential responses, successive flashes were elicited. Each occurred synchronously with positive deflection of the membrane potential (Fig. 13). Summation of luminous intensity was also observed when the frequency of plateau potential oscillations was higher that 5/s. Neither luminescent glowing, nor luminescent background was produced during the T.E.A.-induced depolarizing plateau potential.

As previously reported, simple action potentials were recorded in the common epithelial cells. Their amplitude was around 60 mV.

When Na ions were replaced with Tris, the spike amplitude was strongly reduced. However, the membrane responses did not completely disappear and weak depolarizations (10–15 mV in amplitude) continued 60–65 min after soaking in Na-free solution (Fig. 14). Simultaneously, in photocytes of the same preparation the simple action potential and the secondary depolarization of the 2-component action potential were abolished. Tetrodotoxin (10⁻⁶ g/ml) did not affect the spike generation.

In the presence of 15 mm-Mn ions, the spike potential amplitude could be decreased by 8–10 mV and the maximum rate of rise was sometimes modified (control: 4 mV/ ms; Mn solution: 1·8 mV/ms) (Fig. 15).

Effects of Ca and other ions on the spike generation of the simple action potentials produced in the common epithelial cells were not examined.

The results presented in this study indicate that, in Na-free solution, the simple action potential and the secondary depolarization of the 2-component action potential are reversibly abolished. Simultaneously, the overall amplitude of the spike potential decreases in parallel with the resting potential of the photocyte membrane. Consequently, one may infer that (i) the simple action potential and the secondary depolarization are Na-dependent, (ii) the overshooting spike potential is Na-independent.

Lithium was shown to be unable to replace Na ions as charge carriers during the action potential of the photocyte membrane. However, the capacity of lithium to replace sodium in the depolarizing mechanism of excitable membranes is variable. For example, action potentials in frog atrial myocardium consist of fast Na-dependent T.T.x.-sensitive depolarization followed by slow Na-Ca-dependent T.T.X.-insensitive plateau potential (Rougier et al. 1969). It has been shown that in this preparation, lithium can replace sodium as charge carriers during the earlier depolarizing phase but not during the plateau potential (Chesnais et al. 1971).

Ca ions do not appear to be involved in the ionic mechanism of the simple action potential. Indeed, the amplitude of the simple action potential is unaffected by changes in external Ca concentration and the responses which continue in Ca-free solution probably correspond to the Na-electrogenesis of the photocyte membrane.

On the other hand, the overshooting spikes of the 2-component action potentials are abolished in Ca-free solution and are blocked in the presence of Mn ions. For a tenfold change in external Ca concentration, the slope of the overshooting potential change is in the range of 30 mV. In addition, Sr and Ba ions can substitute for Ca ions and can maintain the electrical activity of the photocyte membranes. These results suggest that the depolarizing phase of the overshooting spike is dependent on a transient inward Ca current through the cell membrane.

Ba ions have been shown to produce various effects on the electrical activities of the photocyte membranes. One of these effects, the capability for spontaneous spike generation is probably caused by the depolarization of the cell membrane which can be sufficient to reach the firing threshold. In the muscle fibres of the earthworm, such a mechanism was proposed by Ito et al. (1970) to account for Ba-induced spontaneous spike generation. Furthermore, in addition to their ability to replace Ca ions as charge carriers, Ba ions are known to decrease or block outward K currents (Werman & Grundfest, 1961; Hagiwara & Naka, 1964; Sperelakis, Schneider & Harris, 1967). The repetitive firing of the photocyte membrane observed in Ba solution may be produced by the progressive decrease of K conductance which maintains the membrane potential at a sufficient level to generate rhythmic spikes. The persistent depolarizing plateau potential which subsequently prolongs the membrane response would result from complete blocking of the outward K current by Ba ions. Another explanation can also be proposed. Indeed, Ca ions have been shown to increase K conductance in invertebrate neurones (Meech, 1972). In the present experiments, the Ba ions replacing the Ca ions would be unable to activate the K efflux with the result that the cell membrane fails to repolarize normally, and the spikes become very long. These observations also indicate that the repolarizing phase of the spike potential is probably dependent on an outward K current which confirms the photocyte membrane responses observed in the presence of T.E.A. ions.

From the effects of various ions on the electrical activities of the photocyte membranes, it is concluded that: (i) the simple action potential depends on an inward Na current which is T.T.X.- and Mn-resistant, (ii) the 2-component action potential is a Ca-Na-dependent T.T.x.-insensitive action potential in which two distinct ionic mechanisms are involved. During the overshooting spike potential, an inward current appears to be carried by Ca ions only while the secondary potential appears to depend on an influx of Na ions into the cells.

In the polynoid worm Hesperonoe (Herrera, 1979), two different membrane responses were generated after intracellular current injection (pulse duration: 50ms) into photocytes. The first was a Na-dependent action potential (amplitude: 43 mV). At a more positive firing threshold, another response which correlated with luminescence was elicited. In isolated scales, this began with one or several overshooting spikes (overshooting amplitude: 13 mV) which were followed with persistent depolarizing plateau potential. The plateau duration could reach 3 or 4 min. Herrera (1979) reported that the initial overshooting spikes were Ca-dependent and that the long-lasting plateaus were probably maintained by a persistent Ca influx across the photocyte membrane. However, when the scales remained attached to the body of the worm, Herrera (1979) observed that the membrane responses accompanying the luminescence were different. These were 2-component action potentials comparable to those described in the isolated scales of the polynoid worm Acholoe (Bilbaut, 1980). In Hesperonoe, the ionic dependence of the action potential produced in the scales connected to their segment was not studied by Herrera (1979).

Consequently, the results indicate that in Acholoe as well as Hesperonoe (Herrera, 1979) the initial overshooting spike potentials correlated with the light production are dependent on an inward Ca current which flows across the photocyte membrane through voltage sensitive channels.

In invertebrates, the glandular excitable epithelial cells in the Cnidarian Hippo-podius exhibit 2-component action potentials in which an initial overshooting Na-dependent spike is followed by a Ca-dependent secondary potential. This latter potential was related to the glandular secretory activity (Mackie, 1976). In some crustacean muscle fibres, composite membrane responses were also reported. In Callinectes (Suarez-Kurtz, 1979), Ca-dependent graded potential changes were prolonged with Na-dependent depolarizing afterpotential. The author showed that the afterpotential amplitude could be modulated by the intracellular Ca concentration changes consecutive to the Ca-dependent response.

In photocytes of the polynoid worm Acholoe, and in agreement with Herrera’s observations (1979) in Hesperonoe, there is good evidence to show that the Na-dependent action potential and the Ca-dependent overshooting spike potential result from separate mechanisms. For instance, Ca-dependent spikes are produced in Na-free solutions. Na-dependent action potentials are unaffected in low external Ca concentration and are observed in the presence of Mn ions which block the Ca-dependent spikes. Lastly, in standard saline, the firing membrane threshold of the Na-dependent action potential is lower than that generating the Ca-dependent spike (Bilbaut, 1980). Thus, when the photocyte membrane response correlated to the luminescence is produced, Na-dependent and Ca-dependent inward currents are simultaneously activated. This means that during the overshooting spikes, Na ions flow into photocytes through specific membrane channels. The Na entry into photo-cytes during the Ca-dependent spike generation would also explain the origin of the secondary depolarization of the 2-component action potential. Indeed, the mean duration of the Ca-dependent spike is 17 ms; the duration of the Na-dependent action potential is around 74 ms. Therefore, it is proposed that during the electro- genic event which accompanies luminescence, a fast Ca-dependent spike is superposed on a slow Na-dependent action potential. Thus, the Na-dependent secondary depolarization and the Na-dependent action potential would be one and the same.

Light production is unaffected by Na-free solutions and in effect, flashes are associated only with the Ca-dependent spikes. All the experiments that lead to the blocking of the Ca-dependent spikes also block the luminescent activity. In Ca-free solution or in the presence of Mn ions, no light is produced even when the photocyte membrane is strongly depolarized. This suggests that the light production is not likely to be potential dependent.

Additional evidence may be drawn from the observations of electrical and luminescent responses produced in T.E.A. solution. Indeed, long-lasting depolarizing plateau potentials maintained between –10 and +15 mV are obtained in the presence of T.E.A. ions which are known to block the outward K current (Hille, 1970). Luminescence recorded in these circumstances consists of separate successive flashes which are associated with positive deflections of the membrane plateau potential. If the light production was potential dependent, a continuous luminous activity at least as long as the plateau potential duration would probably be observed. Flashes were separately emitted and no luminescent background was recorded during the T.E.A.-induced plateau potential. Furthermore, in the polynoid worms Hesperonoe (Herrera, 1979) and Harmothoe (Nicolas, Moreau & Guerrier, 1978) no luminescent activity was produced in whole scales depolarized with high external K concentration when Ca ions were omitted from the saline solution.

Consequently, as previously reported in Acholoe (Bilbaut, 1978 b) and in agreement with Herrera’s conclusions (1977, 1979) from Hesperonoe, the activation of the luminescence in the photocytes of polynoid worms is dependent on Ca entry into cells.

In Acholoe, the depolarizing plateau responses, which are generated during the substitution of the normal saline with Ca-free solution, are always accompanied by persistent glowing. This signifies that the plateau potentials are maintained by a continuous inward ionic current which can be carried by Ca ions. Indeed, since luminescence appears to depend so strongly on Ca entry into the photocytes, one may assume that light production gives evidence of Ca movement across the photocyte membrane. The plateau potentials recorded in Ca-free solution probably occur because small amounts of Ca subsist in the saline solution. Similar responses were episodically observed in 2 mm external Ca solution.

However, another hypothesis can be considered. The plateau responses could be maintained by a Na influx activating the luminescence by a mechanism of Na-induced release of Ca. Such a mechanism has been demonstrated in the exocytotic release of insulin from pancreatic cells (Lowe et al. 1976) and was proposed to account for part of the contraction in frog twitch muscle fibres (Caillé, Ildefonse & Rougier, 1978). Although the excitation-luminescence coupling mechanism in photocytes will be discussed in the next section, the hypothesis that a purely inward Na current would maintain the photocyte plateau potentials and activate the luminescence appears, however, inconsistent with the following data : (i) the plateau amplitude decreases in relation to the duration of incubation in Ca-free solution ; after 50-60 min in this solution, the plateau responses totally disappear and only the Na-dependent action potentials can be elicited, (ii) in the presence of Mn ions, the Ca-dependent spikes are blocked and no luminescence occurs when the Na-dependent Mn-resistant action potentials are generated. In normal saline, the Na-dependent action potentials are sometimes accompanied by slight luminescent activities; these can result from weak Ca amounts which flow into the photocytes during the Na membrane electrogenesis.

The lack of persistent luminescent background during T.E.A.-induced plateau potentials seem to indicate that no continuous Ca influx occurs. The plateau potentials produced in the presence of T.E.A. ions appear principally to be caused by blocking of K membrane channels.

Ca ions have also been demonstrated to serve several functions in excitable cells in addition to serving as charge carriers (Reuter, 1975) during the membrane excitation. For example, Ca ions are involved in transmitter release at the presynaptic nerve terminals (Katz & Miledi, 1969; Hubbard, 1973), secretory processes in nervous and non-nervous tissues (Rubin, 1974) and in contractile protein activation (Szent-Gyorgyi, 1975). The bioluminescent photoprotein aequorin extracted from the jellyfish Aequorea (Shimomura, Johnson & Saiga, 1962) is also known to emit light in the presence of weak Ca concentration (Shimomura, Johnson & Saiga, 1963) and is thereby used as a sensitive Ca probe when injected into cells (Blinks, Prendergast & Allen, 1976).

In the photocytes of the polynoid worm Acholoe, different hypotheses can account for the mechanism by which Ca ions couple the membrane excitation to luminescence. If the bioluminescent metabolites correspond to an aequorin-like photoprotein, Ca entry consecutive to the photocyte spike potential would be able to directly activate the light production. Direct coupling has been reported in frog atrial myocardium (Vassort & Rougier, 1972). In some Crustacea, the Ca influx occurring during electrogenesis of the muscle fibre membrane would be sufficient to cause an appreciable activation of contractile proteins (Hagiwara, Takahashi & Junge, 1968 ; Ashley & Ridgway 1970). However, a similar hypothesis retained by Herrera (1979) to explain the excitation-luminescence coupling in the worm Hesperonoe is not consistent with the ultrastructural organization of photogenic granules, the intracellular sources of the bioluminescence (Bassot & Bilbaut, 1977). The photogenic granules are paracrystalline networks of tight tubules of endoplasmic reticulum (Bassot, 1966) which can establish close apposition with the photocyte plasma membrane (Pavans de Cecatty et al. 1977). Tubular compartments can be considered as possible sequestering sites. Thus, the luminescent activation could result from an indirect Ca coupling according to the following alternatives: (i) a Ca-induced release of Ca as described in skinned muscle fibres (Ford & Podolsky, 1972; Endo, 1975; Fabiato & Fabiato, 1978), (ii) a Ca-induced transfer of metabolites between the intratubular and the extratubular compartments; this hypothesis would put forward that each of the two compartments sequesters one particular substance (Bassot, 1966) such as either an enzyme or a substrate.

The action potentials produced in the non-photogenic epithelial cells appear to be principally dependent on an inward Na current which is T.T.X.-insensitive. The membrane activity which subsists in the Na-free solution and the effects of Mn ions which reduce the amplitude and change the rate of rise of the spikes, are at present unexplained.

In the scales of the polynoid worm Hesperonoe (Herrera, 1979) the non-photogenic epithelial cells also produce simple non-overshooting action potentials which are the same as those elicited in the photogenic epithelial cells. As previously mentioned in the present article, these were Na-dependent, T.T.x.-insensitive, Mn resistant action potentials. However, Herrera (1979) extrapolating to zero potential the relationship between external Na concentration changes and the spike amplitude (slope: 61 mV) found an abnormally high intracellular Na concentration (around 1-3 M). The author concluded that an ion other than Na could be involved in the ionic mechanism of these action potentials.

This work was supported by the C.N.R.S. (L.A. 244). I thank Professor O. Rougier and Dr M. Ildefonse for critical reading of the manuscript. I am grateful to F. Hemming for correction of the English text and A. Bosch for his technical assistance.