1. Electrical properties of cnidocytes isolated from the hydroid Cladonema and the scyphomedusa Chrysaora were examined using current- and voltage-clamp recording techniques.

  2. The stenoteles of Cladonema produced action potentials when depolarized above 0mV. The inward current that produced the action potential was a Na+ current. These cells also possessed an A-current and a K-current.

  3. Atrichous isorhizas from Chrysaora did not spike and did not have any inward currents. All cells examined had K-currents, some had A-currents also.

  4. Very few cnidocytes discharged during the course of the recordings, irrespective of the degree to which they were depolarized or hyperpolarized, or the presence or selective blockade of any ionic currents. When discharge did occur it could never be correlated with any obvious electrophysiological event.

  5. Recordings from cnidocytes in situ in tentacles of the siphonophore Physalia indicate that these cells do not spike. Their current/voltage relationships were linear. They too did not discharge in response to changes in membrane potential, suggesting that the failure of isolated cnidocytes to discharge cannot be attributed to the isolation procedure.

Cnidocytes, the major diagnostic feature of members of the phylum Cnidaria, are unique cells. For this reason and because of the medical problems caused by their stings, they have attracted considerable attention from biologists from various disciplines, resulting in an extensive literature on their structure, ultrastructure, development and discharge mechanisms (for reviews see Lenhoff & Hessinger, 1987; Mariscal, 1974, 1984).

Briefly, cnidocytes are cells whose interior is dominated by a large secretory product, the cnida, sometimes termed the capsule or cnidocyst. The cnida is a hardened capsule that contains an inverted tube. With appropriate stimulation, usually contact with some prey organism or potential predator, the operculum that seals the apical end of the cnida opens and the tube everts. Depending on the class of cnidocyte, the tube will either merely entangle the target or penetrate it and release a venom.

The sequence of steps linking application of the stimulus to discharge of the cnida is little understood. It is known that the appropriate stimulus has both a chemical and a mechanical component (Glaser & Sparrow, 1909 ; Mariscal, 1974) and it is generally believed that these stimuli are transduced by the ciliary apparatus that adorns the apical end of many cnidocytes. There are two classes of ciliary apparatus: ciliary cones, which are found only in anthozoans, and cnidocil apparatuses which occur in all other classes (Mariscal, 1974). It is not clear, however, whether all aspects of the stimulus are transduced by these presumed sensory structures or whether other cell types in the tissue are involved. Indeed, spirocytes lack a ciliary apparatus altogether, suggesting either that other less obvious structures on the cnidocyte are responsible for stimulus transduction or that the transduction is carried out by accessory cells that communicate with the spirocyte.

The reason for this uncertainty as to the site of the receptors stems from the fact that cnidarian cells are typically small and, because cnidarians are of the tissue level of organization, their tissues are complex. The tentacle of a typical cnidarian is composed of many cell types, some of which, such as neurones and epithelial cells, could interact with one another and with cnidocytes, thereby serving as a pathway for excitation of the cnidocytes. Indeed, there are several reports that the behavioural state of an animal can influence the sensitivity of the cnidocytes to discharge (Mariscal, 1973; Smith, Oshida & Bode, 1974). This implies a degree of endogenous control of the discharge, which is supported by several reports of neuro-cnidocyte synapses (Hufnagel, Kass-Simon & Lyon, 1985; Westfall, 1973a,b; Westfall, Yamataka & Enos, 1971). While the use of preparations of isolated cnidae circumvents some of the problems posed by whole tissues and has provided considerable information about several aspects of the biology of these cells (for a review of the methods, see McKay & Anderson, 1987), the isolated cnidae, by definition, are not bathed in cytoplasm and are not enclosed by a cell membrane. Since it is the cell membrane that interacts with the environment and with other cells, its presence is essential if a thorough understanding of the location of the receptors is to be achieved. Furthermore, both the cell membrane and the cell’s cytoplasm must be present if the entire pathway by which a stimulus to the cnidocyte gives rise to cnida discharge is to be fully understood.

Here we report on the electrophysiological properties of cnidocytes isolated from a hydroid and a scyphomedusa, and others in situ in a siphonophore. The study was undertaken primarily to examine the commonly perceived idea that cnida discharge might be triggered or otherwise controlled by electrophysiological events in the cell. This idea may have arisen from the observation that one of the most effective ways to evoke cnida discharge by isolated cnidae or intact tissues is with electrical stimulation (Holstein & Tardent, 1984). The present results indicate that although intact, isolated cnidae exhibit an array of voltage-dependent ionic currents, neither these nor any other obvious electrophysiological events evoke cnida discharge. To control against the possibility that these findings may reflect the fact that the cells were isolated, recordings were also obtained from cnidocytes in situ.

Recordings were obtained from cnidocytes of Physalia physalis (Hydrozoa: Siphonophora), Cladonema sp. (Hydrozoa) and Chrysaora quinquecirrha (Scyphozoa). All animals were maintained in the laboratory under appropriate conditions; Chrysaora and Physalia usually only survived for a few days and were used immediately after collection.

Cnidocytes were isolated from the various tissues enzymatically (McKay & Anderson, 1985, 1986). Chrysaora tentacles were treated with L-cysteine-activated papain, until microscopic examination revealed that sufficient cnidocytes had been released. This usually took 40–60 min. The released cnidocytes, which were heavier than the other cell types, were then separated from any debris and other cell types by density centrifugation in a Percoll-containing medium. The supernatants were discarded and the pelleted cnidocytes transferred to either sea water or saline for recordings.

The technique used for isolating cnidocytes from Cladonema was a simpler version of that given above. The excised hydranths were soaked in a Ca2+- and Mg2+- free saline (Table 1) containing 0·5mgml−1 of L-cysteine-activated papain. After 10 min they were transferred to the final recording solution (Cyanea saline or sea water) and triturated with a fine pipette until individual cnidocytes were released from the capitate tentacles. Once again, the cnidocytes were denser than other cell types and settled to the bottom of the dish.

Table 1.

Composition of media (in mmol l−1)

Composition of media (in mmol l−1)
Composition of media (in mmol l−1)

Recording techniques

Electrophysiological recordings were carried out using conventional microelectrode recording techniques and the whole-cell configuration of the patch-clamp technique (Hamill et al. 1981). Microelectrode recordings were obtained from cnidocytes in the tentacles of Physalia using microelectrodes filled with 3 mol l−1 KC1 with impedances in the range 15–25 MΩ. To obtain the recordings, a short piece of tentacle was excised from the animal, anaesthetized in a 1:1 mixture of isotonic (0·37 mol l−1) MgCl2 in sea water, then stretched out and pinned firmly to a layer of Sylgard (Dow Corning, Midland, MI) on the bottom of a Petri dish, using cactus spines (Opuntia sp.). The bathing medium was then replaced with normal sea water for recordings. Recorded signals were amplified with a conventional d.c. amplifier equipped with a Wheatstone bridge circuit for current injection, and displayed on a Nicolet 2090 digital oscilloscope. Lucifer Yellow injection was carried out in the manner described previously (Anderson & Schwab, 1981).

Whole-cell patch pipette recordings were obtained using Dagan 8900 patch-clamp amplifiers equipped with 0·1 GQ headstages. Pipettes were prepared from borosilicate glass (Boralux, Rochester Scientific, Rochester, NY) using a two-stage puller (Narashige PP83). They were coated with a layer of cured Sylgard but not fire-polished (Corey & Stevens, 1983), since we have found that, for unknown reasons, omitting the polishing results in better seals onto coelenterate cells. The coated electrodes were filled with one of the solutions given in Table 1. With these pipette solutions, the electrodes had impedances of 5–8 MΩ.

Once any offset currents had been neutralized, the pipette tip was pushed gently against the cell surface, usually near the basal end, and light suction applied. Seal formation was usually aided by the application of a negative potential to the interior of the pipette and, under these conditions, seal resistances in the range 1–10 GΩ could be obtained fairly routinely. Once an adequate seal had been obtained, capacitive transients introduced by the electrode were neutralized. Breakthrough into the intracellular configuration was achieved with additional suction.

For voltage-clamp recordings, compensation was made for 40–60% of the series resistance error by circuitry in the amplifier. For current-clamp recordings, the amplifier was switched to the current-clamp mode and sufficient current was injected into the cell to establish a membrane potential of −70 mV. Square current steps were then injected into the cell and any d.c. offsets attributable to pipette resistance neutralized with a Wheatstone bridge circuit in the amplifier. Membrane potential was then readjusted if necessary.

Current-clamp recordings obtained with patch pipettes were displayed in the same way as the microelectrode recordings. Voltage-clamp experiments were performed and the data digitized (80kHz), stored and manipulated by an IBM AT computer equipped with pClamp software (Axon Instruments, Burlingame, CA). With some cells and under some conditions, leakage currents were not linear at hyperpolarized membrane potentials, making their subtraction difficult. Such instances are noted. With the other cells, leakage and capacitive currents were removed from the records either by digital addition of currents generated by hyperpolarizing voltage steps one-half the amplitude of those used to generate ionic currents, or by scaling and subtracting the currents generated when cells that had been hyperpolarized to −120 mV were depolarized with voltage steps one-half the amplitude of those used to generate ionic currents.

Solutions

The compositions of the various solutions used in this study are given in Table 1. Artificial salines were based on a saline developed for the scyphomedusa Cyanea (Anderson & Schwab, 1984). The pH of all external solutions was adjusted to 7·4 with NaOH or HCl, that of all internal solutions was adjusted to 7-0 with KOH or CsOH. All experiments were conducted at room temperature (21–24°C).

Morphology of isolated cnidocytes

The isolation procedures produced a mixture of cnidocyte types, representing the variety found in the intact tissues. The morphology of the different types of cnidocyte was varied, as would be expected from the well-recognized variations in capsule morphology (Mariscal, 1974). Only the stenoteles of Cladonema and the atrichous isorhizas of Chrysaora (Fig. 1) were examined in any detail.

Fig. 1.

Light micrographs of isolated cnidocytes of the type used in this study. (A,B) Nomarski micrographs of stenoteles from Cladonema. Freshly isolated cells (A) usually possessed a very obvious cytoplasmic protrusion at the basal end (arrowhead) but this was later retracted and became only barely visible (B). Notice the very obvious cnidocil apparatus (c) and the spines (s) on the inverted tubule. (C) Nomarski micrograph of an atrichous isorhiza from Chrysaora. In these cells the tubule is irregularly coiled and gives the cnida a granular appearance. (D) A phase/fluorescence micrograph of part of a tentacle from Physalia. A Lucifer Yellow-filled cnidocyte is indicated with an arrowhead. Note that dye has not passed to any of the surrounding cells. Scale bars in all cases = 10 μm.

Fig. 1.

Light micrographs of isolated cnidocytes of the type used in this study. (A,B) Nomarski micrographs of stenoteles from Cladonema. Freshly isolated cells (A) usually possessed a very obvious cytoplasmic protrusion at the basal end (arrowhead) but this was later retracted and became only barely visible (B). Notice the very obvious cnidocil apparatus (c) and the spines (s) on the inverted tubule. (C) Nomarski micrograph of an atrichous isorhiza from Chrysaora. In these cells the tubule is irregularly coiled and gives the cnida a granular appearance. (D) A phase/fluorescence micrograph of part of a tentacle from Physalia. A Lucifer Yellow-filled cnidocyte is indicated with an arrowhead. Note that dye has not passed to any of the surrounding cells. Scale bars in all cases = 10 μm.

In all cases, the cnidocytes were easily distinguished from the surrounding tissue, and any debris, by the presence of the highly refractile cnida. The tube was usually visible within the cnida (Fig. 1) and in some cases, especially in the large stenoteles from Cladonema, the stereociliary complex at the apical end of the cell was easily discerned. Cnidocytes freshly isolated from Cladonema (Fig. 1A) usually had a long cytoplasmic extension at their basal end but these retracted within 15 min of isolation.

Electrophysiology of cnidocytes in situ

Microelectrode recordings were obtained from the large (28 μm in diameter) cnidocytes in the tentacles of Physalia. These recordings were obtained easily and were stable so long as the tentacle had been adequately pinned out. With sea water as the bathing medium, the mean resting potential of 15 cells was −58·8 ± 1·6 mV (±S.E.M.). Current/voltage (l/V) plots, obtained by injecting depolarizing and hyperpolarizing current into the cell through the bridge-balanced microelectrode were completely linear and showed no evidence of delayed rectification. The average input impedance was 36·5 ± 13·3 MΩ (±S.E.M. ; N = 3). The time constant of these cells, determined from the slope of a semi-logarithmic plot of the charging curve created by the injection of hyperpolarizing current, was 1·3–1·4 ms. Physalia cnidocytes did not discharge during these recordings, irrespective of the degree to which the cell was depolarized or hyperpolarized.

Because cnidocytes are quite small and the cell is chiefly occupied by the impenetrable capsule, there is always some question as to whether recordings are obtained from the cnidocytes or from some other cell type. To confirm the recording site, the impaled cells were filled with Lucifer Yellow (Fig. 1D). In every case, Lucifer Yellow was found in a cnidocyte, as defined by the presence of a cnida, and was restricted to that cnidocyte. There was no evidence of dye coupling to adjacent cells.

Electrophysiology of isolated cnidocytes

Whole-cell patch recordings were made from isolated cnidocytes. In all cases where current-clamp recordings were made the cells had negative resting potentials. However, the magnitude of that resting potential could not be measured accurately since although any offsets were neutralized immediately before seal formation, others resulting from dissimilarities between the pipette contents and the cytoplasm could not be subtracted and would have contributed to the measured membrane potentials. Unless otherwise stated, therefore, a resting potential of −70 mV was imposed on all cells. This value is slightly more negative than that of Physalia cnidocytes in situ (see above) but is consistent with those of cells in other marine cnidarians (for a review see Anderson & Schwab, 1982).

Stenoteles from Cladonema

The input impedances of these cells, as measured under current-clamp conditions with sea water as the bathing medium, were in the range 114–260 MΩ [mean = 181 ± 42 MΩ (±S.E.M.) N = 3] and were essentially linear over most of the voltage range except at subthreshold potentials close to 0 mV where there was usually a small amount of delayed rectification (Fig. 2A). Furthermore, with the largest hyperpolarizing steps the voltage records sometimes saturated. The membrane time constant (determined, as before, from the inverse slope of a semilogarithmic plot of the charging curve produced by the injection of a small step of hyperpolarizing current) ranged from 4 to 6 ms.

Fig. 2.

Intracellular, current-clamp recordings from Cladonema stenoteles. (A) Voltage response (upper traces) of a cell to injected current (lower traces). Depolarizing steps to positive membrane potentials evoked a single action potential. The current/voltage relationship of these cells was essentially linear over all negative membrane potentials although usually, as here, a small amount of delayed rectification occurred at potentials close to threshold. (B) A single action potential at higher gain and sweep speed, illustrating its waveform.

Fig. 2.

Intracellular, current-clamp recordings from Cladonema stenoteles. (A) Voltage response (upper traces) of a cell to injected current (lower traces). Depolarizing steps to positive membrane potentials evoked a single action potential. The current/voltage relationship of these cells was essentially linear over all negative membrane potentials although usually, as here, a small amount of delayed rectification occurred at potentials close to threshold. (B) A single action potential at higher gain and sweep speed, illustrating its waveform.

When depolarized to 0mV, or more positive, these cells spiked (Fig. 2A,B). Only a single action potential was ever evoked, irrespective of the duration of the depolarization. The action potential was of relatively long duration, lasting 13–15 ms at half peak amplitude, and overshot 0 mV by, on average, 43·2 ±2·2mV (N = 6). The slope of the rising phase of the spike was about 28 Vs−1. Repolarization was considerably slower, particularly at the onset (2·6 V s−1), but towards the end of the spike the rate of repolarization more than doubled to 5·8 Vs−1.

When the same cells were examined under voltage-clamp conditions, a very reproducible family of currents was recorded. A typical family of total membrane currents produced by depolarization of a cell bathed in Cyanea saline using a pipette filled with normal patch solution (Table 1) is presented in Fig. 3. Total membrane currents consisted of an inward current, a fast, transient outward current and a steady-state outward current. These currents flowed simultaneously so the overall pattern was invariably complex. The inward current was sometimes obscured by the outward currents but was usually apparent as a brief inward-going current at the end of the first few suprathreshold depolarizing steps (Fig. 3A,B). Its reversal potential was undoubtedly affected by the outward currents flowing at the same time. The transient and steady-state outward currents were blocked by internal Cs+, TEA+ and 4-aminopyridine (4-AP) (Table 1), usually within 30s of breakthrough into the intracellular configuration, leaving the inward current in isolation.

Fig. 3.

Voltage-clamp recordings from Cladonema stenoteles. (A) Family of total membrane currents. Holding potential −70mV; first voltage step to −20mV with subsequent steps increasing by 7·5 mV. Total current consisted of an inward current, a transient and a steady-state outward current. External medium, saline; internal medium, normal patch solution. (B) l/V relationships of the peak transient outward current (open circles) and the currents recorded at the end of the voltage step (filled circles). The latter consists of both the inward current and the steady-state, outward current. (C) Inward currents recorded in isolation. Holding potential −70mV; first voltage step to −30mV with subsequent steps increasing by 10mV. External medium, sea water; internal medium, Cs+/TEA+ patch solution. (D) l/V relationship of the currents shown in C.

Fig. 3.

Voltage-clamp recordings from Cladonema stenoteles. (A) Family of total membrane currents. Holding potential −70mV; first voltage step to −20mV with subsequent steps increasing by 7·5 mV. Total current consisted of an inward current, a transient and a steady-state outward current. External medium, saline; internal medium, normal patch solution. (B) l/V relationships of the peak transient outward current (open circles) and the currents recorded at the end of the voltage step (filled circles). The latter consists of both the inward current and the steady-state, outward current. (C) Inward currents recorded in isolation. Holding potential −70mV; first voltage step to −30mV with subsequent steps increasing by 10mV. External medium, sea water; internal medium, Cs+/TEA+ patch solution. (D) l/V relationship of the currents shown in C.

Under these conditions inward current in the cells activated at −10 to −20mV, reached peak amplitude around +20 mV and reversed between +60 and +70 mV (mean = +65·7 ± 1·2 mV; N = 6) (Fig. 3C,D). At the end of the voltage step, a fast, inward tail current occurred. Inward current in these cells activated relatively quickly, reaching peak amplitude some 2 ms after the onset of the voltage step. With most cells, the inward current was a steady-state current but it occasionally showed a slight decay from its peak. This inward current was absent in the absence of extracellular Na+, even when [Ca2+]o was raised 10-fold, and was unaffected by the removal of extracellular Ca2+. Furthermore, the currents recorded in 40 mmol l−1 Ba2+ artificial sea water (ASW) (normal [Na+]o) were no different from those recorded in normal saline. The reversal potential of this current ( + 65·7 mV) was consistent with the equilibrium potential for Na+ ( + 64·6mV), calculated on the basis of the known extracellular and intracellular (pipette) Na+ concentrations. Furthermore, the amplitude of the current and its reversal potential decreased when [Na+]o was decreased. The exact relationship between [Na+]o and the reversal potential could not be determined accurately, however, since in low-Na+ solution, the leakage currents at hyperpolarized membrane potentials were far greater than those at more depolarized membrane potentials making leakage current subtraction and, therefore, accurate current measurement difficult. All inward current was blocked by extracellular Cd2+ (2–5 mmol l−1) but was insensitive to the traditional Na+ channel blocker tetrodotoxin (TTX) (1μmoll−1).

Outward current could be examined in isolation using a combination of normal patch solution and Cyanea saline (or sea water) with 5mmoll−1 Cd2+. There were two components to the outward current; a fast, transient outward current and a slower, steady-state outward current (Fig. 4A). The transient outward current usually activated at more negative membrane potentials than the steady-state current, but the difference was usually small and both usually activated between −20 and −30 mV (Fig. 4B). They were separated in the following manner. The outward current records shown in Fig. 4A were obtained from a cell clamped at a holding potential of −70 mV. Both components of the outward current were present. When the same cell was clamped at a holding potential of 0 mV, and the same voltage regime applied, the transient component disappeared, leaving only the steady-state current (Fig. 4C). This current activated at 0 mV and increased in amplitude with further depolarization (Fig. 4D). Digital subtraction of the records obtained at the two holding potentials provided a record of the transient current. This current activated some 5–10 mV below the steady-state, outward current and was smaller in amplitude.

Fig. 4.

Outward currents in stenoteles from Cladonema. (A) Total outward currents consist of a fast, transient, outward current followed by a steady-state current. Holding potential −70 mV ; first voltage step to −20 mV, subsequent steps increasing by 7·5 mV. External medium, saline with 5 mmol l−1 Cd2+; internal medium, normal patch solution. (B) l/V relationship of the transient (open circles) and steady-state (filled circles) outward current. Peak steady-state outward current was arbitrarily measured at the end of the voltage step. (C) Steady-state outward current recorded in isolation from the same cell. Holding potential 0 mV; external and internal media as for B. (D) l/V relationship of the steady-state current (filled circles) from C. Data for the transient currents (open circles) that were isolated by digital subtraction of the currents in A and C are included also.

Fig. 4.

Outward currents in stenoteles from Cladonema. (A) Total outward currents consist of a fast, transient, outward current followed by a steady-state current. Holding potential −70 mV ; first voltage step to −20 mV, subsequent steps increasing by 7·5 mV. External medium, saline with 5 mmol l−1 Cd2+; internal medium, normal patch solution. (B) l/V relationship of the transient (open circles) and steady-state (filled circles) outward current. Peak steady-state outward current was arbitrarily measured at the end of the voltage step. (C) Steady-state outward current recorded in isolation from the same cell. Holding potential 0 mV; external and internal media as for B. (D) l/V relationship of the steady-state current (filled circles) from C. Data for the transient currents (open circles) that were isolated by digital subtraction of the currents in A and C are included also.

The stenoteles from Cladonema sometimes discharged during these recordings. However, discharge was very inconsistent and never coincided with any obvious electrophysiological event, be it a particular level of depolarization or the appearance, magnitude or absence of any particular ionic current. Overall, fewer than 10 % of the cells discharged during recordings.

Atrichous isorhizas from Chrysaora

Atrichous isorhizas from Chrysaora displayed very pronounced delayed rectification (Fig. 5A). For the cell shown in Fig. 5, the input impedance was 93 MΩ for hyperpolarizing current steps, but for depolarizing steps to 0mV or above this decreased to 76MΩ (Fig. 5B). Similar values were obtained from almost all cells examined. The time constant of these cells, measured as before, ranged from 4·0 to ·2ms (mean = 5·22 ± 0·6 ms). No action potentials were observed in these cells irrespective of the level of depolarization.

Fig. 5.

Current- and voltage-clamp recordings from atrichous isorhizas from Chrysaora. (A) Intracellular, current-clamp recordings. Voltage responses (upper traces) of a cell to injected currents (lower traces). Note that these cells did not spike and that the voltage responses to depolarizing and hyperpolarizing current were asymmetrical, indicating delayed rectification. (B) Data from A plotted graphically. The input impedance of this cell, as measured from the slope of the l/V relationship was 93 M Ω over all negative potentials but this decreased to 76M Ω upon depolarization to 0 mV or above. (C) Voltage-clamp recording of the steady-state, outward current in one cell. This was the only current identified in this cell. Holding potential −70mV; first step to −20mV, subsequent steps increasing by 10 mV. External medium, saline; internal medium, normal patch solution. (D) l/V relationship of the currents shown in C. (E) A transient and a steady-state outward current recorded from another cell. Holding potential −90mV; step depolarization regime and recording media as for C. (F) l/V relationships of the transient (open circles) and steady-state (closed circles) currents from E.

Fig. 5.

Current- and voltage-clamp recordings from atrichous isorhizas from Chrysaora. (A) Intracellular, current-clamp recordings. Voltage responses (upper traces) of a cell to injected currents (lower traces). Note that these cells did not spike and that the voltage responses to depolarizing and hyperpolarizing current were asymmetrical, indicating delayed rectification. (B) Data from A plotted graphically. The input impedance of this cell, as measured from the slope of the l/V relationship was 93 M Ω over all negative potentials but this decreased to 76M Ω upon depolarization to 0 mV or above. (C) Voltage-clamp recording of the steady-state, outward current in one cell. This was the only current identified in this cell. Holding potential −70mV; first step to −20mV, subsequent steps increasing by 10 mV. External medium, saline; internal medium, normal patch solution. (D) l/V relationship of the currents shown in C. (E) A transient and a steady-state outward current recorded from another cell. Holding potential −90mV; step depolarization regime and recording media as for C. (F) l/V relationships of the transient (open circles) and steady-state (closed circles) currents from E.

The records presented in Fig. 5C,E represent the types of currents recorded from these cells under voltage-clamp conditions. There were two types of outward current; one was present in all cells examined, the other was present in only a small proportion of the cells. This situation differs from that in Cladonema where both types occurred in all cells.

In all Chrysaora cnidocytes examined with pipettes filled with normal patch solution there was a steady-state outward current (Fig. 5C,D) that activated at membrane potentials close to 0 mV. This current had a sigmoidal waveform, reaching half peak amplitude in 2–3 ms and peak amplitude after 12–13 ms. The properties of this current were essentially the same as those of the steady-state, outward current recorded from stenoteles from Cladonema. This current was never recorded with patch pipettes filled with Cs+/TEA+ patch solution.

The second outward current recorded, observed in 13% of all Chrysaora cells examined, was an early, transient, outward current (Fig. 5E,F). This current was partly obscured by the steady-state, outward current and attempts to separate them as before, using depolarized holding potentials, were unsuccessful. However, these experiments were done before those with the Cladonema stenoteles, and did not use holding potentials less negative than −20mV. With the Cladonema stenoteles a holding potential of 0mV was necessary to remove all the transient current. The transient outward current in the Chrysaora atrichous isorhizas activated at more negative potentials (−10 to −20mV) than the steady-state current (Fig. 5F) and, once again, was never observed when the pipettes contained Cs+/TEA+ patch solution.

No inward currents were recorded with Cs+/TEA+ patch solution. Once again, with some cells the leakage current was non-linear, particularly at hyperpolarized potentials where it was especially large. The reason for this is unclear.

These cells sometimes discharged during recordings but did so less frequently then stenoteles from Cladonema. Once again, discharge was not correlated with any obvious electrophysiological event.

The results presented here indicate that apparently functional cnidocytes with intact membranes can be isolated from a variety of cnidarian tissues. A weakened membrane would manifest itself in a high leakage to earth during whole-cell recordings. This was not the case here; upon breakthrough into the intracellular configuration the resistance to earth was essentially the same as that previously, and attributable primarily to seal resistance. In addition, if the membranes were damaged, trans-membrane ionic currents would have been absent since the cells would presumably be unable to maintain the electrochemical gradients that form the driving forces for those currents.

Voltage-activated ionic currents, as exhibited by cnidocytes, are found in many cell types (for a review, see Hille, 1984). These ionic currents were examined in detail in the stenoteles from Cladonema and to a lesser degree in the atrichous isorhizas from Chrysaora. In each case, their properties were very similar to those of ionic currents in cells from other cnidaria (Dunlap, Takeda & Brehm, 1987; Hagiwara, Yoshida & Yoshii, 1981; Stein & Anderson, 1984, 1985) and other organisms (Hagiwara, 1983; Hille, 1984). The isolated stenoteles and atrichous isorhizas possess two outward currents, the first a transient current, the second a steady-state current. Their properties, specifically their sensitivity to internal TEA+, Cs+ and 4-AP and their voltage dependencies, are consistent with those of two very common K+ currents, the transient A-current and the steady-state delayed rectifier or K-current. In other cell types the A-current is a voltage-activated and voltage-inactivated current and thus inactivates completely during a prolonged depolarization (for a review see Rogawski, 1985). These properties allow it to be separated from the other voltage-activated K+ current, IK, merely by clamping the cell at depolarized holding potentials as was done here. Given these similarities, it is reasonable to classify the transient outward current in these cells as an A-current.

The waveform and properties of the steady-state, outward current are the same as those of the delayed rectifier or K-current. This classification is given added weight here by the finding that, in atrichous isorhizas from Chrysaora, this current activates at the same potential as the delayed rectifier recorded under current-clamp conditions (Fig. 5A). Thus, it is reasonable to classify the steady-state, outward current as a K-current.

The action potential of stenoteles from Cladonema is presumably a Na+ spike. Inward current in these cells was unaffected by the removal of Ca2+ but was absent in Na+-free saline and its amplitude and reversal potential were Na+-dependent. Thus, even though the waveform of the inward current in these cells (Fig. 3C) is not typical of that of fast, transient, Na+ currents in other excitable cells (Hille, 1984), including neurones in cnidarians (Anderson, 1986), and was insensitive to the classical Na+ channel blocker TTX, this inward current is a Na+ current. The insensitivity to TTX is not surprising since Na+ currents in other cnidarian tissues are not affected by TTX (Anderson, 1986; Anderson & Schwab, 1982).

The function of this action potential is not clear. It is clearly not required for cnidocyte discharge since the cells never discharged when action potentials occurred and the Physalia and Chrysaora cnidocytes did not produce action potentials. However, epithelial tissues of many hydrozoans are epithelial conduction systems (for a review, see Anderson, 1980). The cells are electrogenic and action potentials travel intercellularly by way of gap junctions. While gap junctions are not known to be present on cnidocytes and, judging by the lack of dye coupling between Physalia cnidocytes and other cell types in their tentacles (Fig. 1D), are probably not present on mature Physalia cnidocytes, their presence on cnidocytes in Cladonema cannot yet be excluded. It is possible, therefore, that the electrogenic capabilities of cnidocytes from Cladonema may be to allow them to participate in the propagation of epithelial action potentials. These action potentials would not, in themselves, evoke cnida discharge since discharge was never seen following the appearance of an action potential. However, the passage of the action potentials might affect the threshold for discharge or some other property of these cells.

If the action potential is indeed a consequence of these cells being part of an epithelial conduction system it is interesting to note that action potentials from epithelial cells of another hydroid, Obelia, are quite different from those described here, being Ca2+-dependent and considerably longer (Dunlap et al. 1987).

The possibility that these cells form part of an epithelial conduction system is given added support by the observation that the atrichous isorhizas of Chrysaora, a scyphozoan, did not spike. Epithelial conduction has never been described in any scyphozoan (Anderson, 1980). However, action potentials were never observed during recordings from cnidocytes in Physalia and yet epithelial conduction is very prevalent in the siphonophores. Furthermore, inward currents with a waveform similar to those of the stenoteles of Cladonema have been recorded from atrichous isorhizas isolated from the acrorhagi of the sea anemone Anthopleura (McKay & Anderson, 1987), another class not known to possess epithelial conduction, ft is not known, however, whether the inward current in these cells gives rise to an action potential.

Perhaps the most remarkable feature of the outward currents recorded here is their speed of activation. The A-currents in both the stenoteles and the atrichous isorhizas peaked within 1 ms. A-currents in photocytes from Obelia peak 20–30 ms after the onset of a voltage step (Dunlap el al. 1987), and A-currents in many other cell types have a similar time course (Rogawski, 1985). In addition, the K-currents achieved 80–90 % of its final amplitude within 2–3 ms. The reason for these very high rates of activation is unclear. In the case of Cladonema one possibility is that it may merely reflect the properties of a very fast, epithelial conduction system in this animal and that the ionic currents in other epithelial cells in the animal are equally fast. Alternatively, while it is clear that these ionic currents are not responsible for cnida discharge, they may serve to counteract rapidly any tendency for the cell to depolarize in response to inappropriate stimuli - in effect, a method of voltageclamping the cells in situ. Cnidocyte discharge is extremely rapid; stenoteles in the tentacles of Hydra discharge completely within 3 ms (Holstein & Tardent, 1984), with the initial pre-discharge events occurring within the first few microseconds. However, in those experiments discharge was effected by electrical stimulation and the normal transductory processes of the cells were presumably bypassed. Since the process of stimulus transduction in intact cells has not been examined in any detail, its time course is not known but it is reasonable to assume that it must be fast, of the order of a few milliseconds, since the whole process of cnidocyte discharge has presumably evolved to minimize the delay between contact with the target and discharge. Thus, any mechanism used to prevent untimely discharge would have to operate on a similar time scale to be successful. The ionic currents in these cells operate within this time scale so may serve this purpose. Alternatively, since the typical stimulus for discharge has both a mechanical and a chemical component, the high activation rate of the A-current in particular may be a way of ensuring that only when the two stimuli are applied within a few milliseconds of each other does the cnidocyte respond. If the interval between the two components of the stimulus is too great, any depolarization evoked by the first stimulus would be counteracted by the A-current before the second component is received.

One major conclusion to be derived from this work is that cnida discharge is not triggered by imposed changes in membrane potential alone, nor by the presence or selective blockade of any of the voltage-activated ionic currents available to the cell. This finding is somewhat surprising since electric shock is commonly used to discharge in situ cnidocytes and isolated cnidae (Holstein & Tardent, 1984). However, the magnitudes of the electrical shocks used in those cases may be completely non-physiological and the discharge may merely represent breakdown of the cell or cnida membranes. Some cnidocytes did discharge during the course of our recordings but the incidence was low and could never be correlated with any obvious electrophysiological event. It is arguable that the isolation procedure had interfered with some component of the discharge machinery. This is unlikely, however, since cnidocytes in the tentacles of Physalia also failed to discharge in response to changes in membrane potential. Furthermore, the cnidocytes retained the ability to discharge and discharge was perfectly normal, as judged visually.

The requirements for discharge of the intact cnidocyte remain obscure. The normal stimulus for cnidocyte discharge is known to have both a mechanical and a chemical component (Glaser & Sparrow, 1909; Mariscal, 1974) and the receptors that transduce these stimuli are generally thought to be associated with the cnidocil apparatus, more because this structure ‘looks’ like a sensory structure than for any other reason. Given the fact that membrane potential changes do not evoke discharge, one reasonable model for the control of cnidocyte discharge is that transduction of the applied stimuli leads either to the opening of a discrete population of highly specific, voltage-insensitive ion channels or to the release of a second messenger. The ions or second messengers would then somehow interact with the discharge mechanism and produce cnida discharge.

The solution to these and other questions on the biology of these cells should come with further work now that appropriate and viable preparations have been developed.

This work was supported by grant BNS 85-06193 from the National Science Foundation.

Anderson
,
P. A. V.
(
1980
).
Epithelial conduction: its properties and function
.
Prog. Neurobtol
.
15
,
161
203
.
Anderson
,
P. A. V.
(
1986
).
Voltage clamp analysis of a fast, TTX-insensitive sodium current in coelenterate neurons
.
Soc. Neurosci. Abstr
.
12
,
43
.
Anderson
,
P. A. V.
&
Schwab
,
W. E.
(
1981
).
The organization and structure of nerve and muscle in the jellyfish Cyanea capillata (Coelenterata; Scyphozoa)
.
J. Morph
.
170
,
383
399
.
Anderson
,
P. A. V.
&
Schwab
,
W. E.
(
1982
).
Recent advances and model systems in coelenterate neurobiology
.
Prog. Neurobiol
.
19
,
591
600
.
Anderson
,
P. A. V.
&
Schwab
,
W. E.
(
1984
).
An epithelial cell-free preparation of the motor nerve net of Cyanea (Coelenterata; Scyphozoa)
.
Biol. Bull. mar. biol. Lab., Woods Hole
166
,
396
408
.
Corey
,
D. P.
&
Stevens
,
C. F.
(
1983
).
Science and technology of patch-recording electrodes
.
In Single Channel Recording
(ed.
B.
Sakmann
&
E.
Neher
), pp.
53
68
.
New York, London
:
Plenum Press
.
Dunlap
,
K.
,
Takeda
,
K.
&
Brehm
,
P.
(
1987
).
Activation of a calcium-dependent photoprotein by chemical communication through gap junctions
.
Nature, Land
.
325
,
60
62
.
Glaser
,
O. C.
&
Sparrow
,
C. M.
(
1909
).
The physiology of nematocysts
.
J. exp. Zool
.
6
,
426
486
.
Hagiwara
,
S.
(
1983
).
Membrane Potential-dependent Ion Channels in Cell Membrane: Phylogenetic and Developmental Approaches
.
New York
:
Raven Press
.
Hagiwara
,
S.
,
Yoshida
,
S.
&
Yoshii
,
M.
(
1981
).
Transient and delayed potassium currents in the egg cell membrane of the coelenterate Renilla koellikeri
.
J. Physiol., Lond
.
318
,
123
141
.
Hamill
,
O. P.
,
Marty
,
A.
,
Neher
,
E.
,
Sakmann
,
B.
&
Sigworth
,
F. J.
(
1981
).
Improved patchclamp techniques for high-resolution recording from cells and cell-free membrane patches
.
Pflügers Arch. ges. Physiol
.
391
,
85
100
.
Hille
,
B.
(
1984
).
Ionic Channels of Excitable Cells
.
Sunderland, MA
:
Sinauer Press
.
Holstein
,
T.
&
Tardent
,
P.
(
1984
).
An ultrahigh-speed analysis of exocytosis: nematocyst discharge
.
Science
223
,
830
832
.
Hufnagel
,
L. A.
,
Kass-Simon
,
G.
&
Lyon
,
M. K.
(
1985
).
Functional organization of battery cell complexes in tentacles of Hydra attentuate
.
J. Morph
.
184
,
323
341
.
Lenhoff
,
H. M.
&
Hessinger
,
D.
(eds) (
1987
).
The Biology of Nematocysts
.
Orlando, FL
:
Academic Press
.
Mckay
,
M. C.
&
Anderson
,
P. A. V.
(
1985
).
Properties of single cnidocytes isolated from sea anemones
.
Am. Zool
.
25
,
31a
.
Mckay
,
M. C.
&
Anderson
,
P. A. V.
(
1986
).
Electrophysiology of isolated and in situ cnidocytes of coelenterates
.
Am. Zool
.
26
,
70a
.
Mckay
,
M. C.
&
Anderson
,
P. A. V.
(
1987
).
On the isolation and properties of cnidocytes and cnidae
.
In The Biology of Nematocysts
(ed.
H. M.
Lenhoff
&
D.
Hessinger
).
Orlando, FL
:
Academic Press
(in press).
Mariscal
,
R. N.
(
1973
).
The control of nematocyst discharge during feeding by sea anemones
.
Publ. Seto mar. biol. Lab
.
20
,
696
702
.
Mariscal
,
R. N.
(
1974
).
Nematocysts
.
In Coelenterate Biology: Reviews and New Perspectives
(ed.
L.
Muscatine
&
H. M.
Lenhoff
), pp.
129
178
.
New York, San Francisco, London
:
Academic Press
.
Mariscal
,
R. N.
(
1984
).
Cnidaria: Cnidae
.
In The Biology of the Integument
(ed.
J.
Bereiter-Hahn
,
A. G.
Matoltsy
&
K. S.
Richards
), pp.
57
68
.
Berlin
:
Springer-Verlag
.
Rogawski
,
M. A.
(
1985
).
The A-current: how ubiquitous a feature of excitable cells
.
Trends Neurosci
.
8
,
214
219
.
Smith
,
S.
,
Oshida
,
S.
&
Bode
,
H.
(
1974
).
Inhibition of nematocyst discharge in Hydra fed to repletion
.
Biol. Bull. mar. biol. Lab., Woods Hole
147
,
186
202
.
Stein
,
P. G.
&
Anderson
,
P. A. V.
(
1984
).
The physiology of single giant smooth muscle cells isolated from the ctenophore Mnemiopsis
.
Biophys. J
.
45
,
233a
.
Stein
,
P. G.
&
Anderson
,
P. A. V.
(
1985
).
Ionic currents in an isolated smooth muscle cell
.
Biophys. J
.
47
,
466a
.
Westfall
,
J. A.
(
1973a
).
Ultrastructural evidence for neuromuscular systems in coelenterates
.
Am. Zool
.
13
,
237
246
.
Westfall
,
J. A.
(
1973b
).
Ultrastructural evidence for a granule-containing sensory-motor interneuron in Hydra httoralis
.
J. Ultrastruct. Res
.
42
,
268
282
.
Westfall
,
J. A.
,
Yamataka
,
S.
&
Enos
,
P. D.
(
1971
).
Ultrastructural evidence for polarized synapses in the nerve net of Hydra
.
J. Cell Biol
.
51
,
318
323
.