The Recovery From a Transient Inhibition of the Oxidative Metabolism of the Photoreceptors of the Drone (Apis Mellifera ♂)

The histology of the honey-bee drone retina suggests a remarkable metabolic compartmentation: the photoreceptors contain numerous mitochondria ranged at the periphery of the cell close to the cytoplasmic membrane, facing fin-like processes of the adjacent glial cells, which contain few mitochondria, and large quantities of glycogen beta particles (Perrelet, 1970). This suggests that the oxidative metabolism of the drone retina is largely confined to the photoreceptors (see Fig. 1).

Oxidative metabolism of the drone retina, as measured by O₂ consumption ⁠, increases when the retina is exposed to flashes of light (Tsacopoulos, Poitry & Bor-sellino, 1981; Tsacopoulos & Poitry, 1982). The drone retina cannot function without this ⁠: anoxia depolarizes the photoreceptors and suppresses the receptor potential within 2min (Baumann & Mauro, 1973). These effects of anoxia have also been observed in the photoreceptors of other arthropods (Wong, Wu, Mauro & Pak, 1976; Lantz & Mauro, 1978; Payne, 1981) but the molecular mechanism by which anoxia blocks the receptor potential is not clear. It is reasonable to suppose that suppression of by anoxia leads to a decrease of the ATP concentration in the photoreceptor, and thus to an inhibition of ATP-dependent processes, such as the working of the sodium pump. The present study examines the extent to which inhibition of the pump is involved in suppression of the receptor potential, and the role of secondary factors, such as an accumulation of K⁺ in the extracellular space.

For some experiments, the cut head preparation described by Fulpius & Baumann (1969) and slightly modified by Tsacopoulos al. (1981) was used. Briefly, the drone was decapitated, the antennae were cut and the head was fixed with the posterior surface up in a stainless steel microcuvette previously filled with melted wax (at 37–38°C) (Eicosan, Fluka, Buchs, Switzerland). A cut was made close to the posterior surface and parallel to the axes of the ommatidia with a new razor blade vibrating at 300 Hz. This exposed a layer of ommatidia which were then placed flush with the floor of a chamber (see Tsacopoulos et al. 1981) and which was superfused with oxygenated Ringer’s solution (composition in mmoll⁻¹: NaCl,280; KCl,10; CaCl₂, 1·6; MgCl₂.10; Tris,10; pH 7·4). Under these conditions only a retinal layer of 400 μm is oxygenated in darkness and therefore functioning (Tsacopoulos et al. 1981). When the retina is repetitively stimulated by light flashes this oxygenated layer is reduced to about 200 μm (Tsacopoulos & Poitry, 1982). In some other experiments, more particularly those in which was measured, slices 300 μm thick were made, as described by Tsacopoulos & Poitry (1982). The effect of anoxia on the receptor potential was identical in the two preparations. The Ringer’s solution was pumped to the chamber by a peristaltic pump with a flow rate of about 8 ml min⁻¹. Before reaching the chamber, the flow passed through a Teflon membrane thin layer gas exchanger (Dunnschicht Dialysator, L. Eschweiler and Co., Kiel, West Germany). This system was effective in quickly replacing O₂ by N2 (anoxia) in the Ringer’s solution (see Tsacopoulos et al. 1981).

Receptor potentials were recorded intracellularly with micropipettes using conventional techniques (Bader, Baumann & Bertrand, 1976). The local Po, was recorded with double-barrelled platinum microelectrodes as described previously (Tsacopoulos & Lehmenkuhler, 1977; Tsacopoulos et al. 1981).

The extracellular K⁺ concentration ([K⁺]_o) was measured with double-barrelled K⁺-sensitive microelectrodes as described by Coles & Tsacopoulos (1977, 1979) and Munoz, Deyhimi & Coles (1983).

Light from a 150 W xenon arc was focused on a diaphragm which could be occluded by an electromechanical shutter. The image of the diaphragm was focused on the retina with a microscope objective (Bader et al. 1976). The intensity at the retina was about 1·1 W cm⁻² and the standard flashes were 20–40 ms long.

Recordings were digitized at 10 Hz, and were then displayed on a graphics terminal (Tektronix 4006) and stored on magnetic disc by use of a microcomputer (IMSAI 8080 microcomputer system, San Leonard, CA). The kinetics of the extra O₂ consumption induced by a single flash of light were calculated using a Fast Fourier Transform routine as described by Tsacopoulos & Poitry (1982). The voltage signal from the K⁺ microelectrodes was transformed by the computer to a potassium signal on the basis of the calibration curve of each microelectrode.

Recordings were made from photoreceptors which were exposed to light flashes every 4 s to induce depolarizing receptor potentials. When O2 was replaced by N2, a level of near zero was achieved in about 15 s. About 15 s later there was a depolarization of the resting potential, and the amplitude of the receptor potential diminished (Fig. 2). Depolarization was completed, and receptor potential was abolished in about 2min, as found previously (Baumann & Mauro, 1973). When O₂ was restored, the membrane potential quickly recovered to the pre-anoxia level followed by a progressive recovery of the amplitude of the receptor potential. These effects of anoxia were qualitatively the same in all cells studied.

Hypoxia with local around 10 mmHg only slightly depolarized the photoreceptor without any appreciable effect on the receptor potential. This is in accord with the observation that in the superfused drone retina is independent of the local ⁠, down to a ⁠, as low as 10mmHg (Tsacopoulos et al. 1981).

The delay between the onset of anoxia and the effects upon the photoreceptor (Fig. 2), which we term the resistance of the photoreceptor to anoxia, would indicate that the metabolic needs of the photoreceptor during this period were met by anaerobic metabolism, as in the brain (McIlwain & Bachelard, 1971), or by an energy store. Glycolysis does not appear to be involved because extracellular pH measurements (using pH-sensitive microelectrodes) and biochemical analysis of the superfusate have shown that anoxia does not cause either acidification of the extracellular space or accumulation of lactate in the drone retina (Coles & Tsacopoulos, 1985).

To investigate whether an ATP store was depleted during anoxia, the effect of light intensity upon the time required for anoxia to block the receptor potential was examined. It has previously been shown that the amplitude of and Δ [Na⁺]I elicited by a single flash of light are linearly related to the logarithm of the intensity of the light (Tsacopoulos & Poitry, 1982; Coles & Orkand, 1982).

When the light intensity was reduced by 3 log units from the maximum (Fig. 3C) the duration of the anoxia necessary to block the photoresponses was increased by about 1-5 times. Also, in the case of stimulation with low intensity flashes (Fig. 3C), the membrane depolarization and the reduction of the amplitude of the receptor potential had clearly different time courses.

The time required for anoxia to block the photoresponse depended also on how recently the photoreceptors had been exposed to a previous anoxia. In the experiment shown in Fig. 4, the first exposure to anoxia produced 90% inhibition of the receptor potential in 90s (Fig. 4A). The retina was then reoxygenated for different periods before further anoxic exposures. After 5 min of reoxygenation, anoxia produced 90% inhibition in 45s (Fig. 4B), whereas after 16min recovery, an anoxic period of 94s was required (Fig. 4C), similar to that for the first exposure (Fig. 4A).

Resting potential and membrane potential returned to similar values after all three exposures, indicating that these parameters do not determine the resistance to anoxia.

To examine whether ionic perturbations are induced in determining the resistance of the photoreceptors to anoxia, we measured changes of the extracellular K⁺ concentration ([K⁺]_o) as described by Coles & Tsacopoulos (1979). In the typical experiment shown in Fig. 5, the retina was stimulated with strong flashes of light presented every 2 min. Each flash induced a transient increase of [K⁺]_o (Fig. 5A) (see Tsacopoulos et al. 1983). In physiological conditions [K⁺]_o reached a peak of about 40 mmol 1⁻¹ 1–2 s after the flash (Fig. 5B). Then [K⁺]_o rapidly declined, presumably because K⁺ enters glial cells (Coles & Tsacopoulos, 1979, 1981). It then undershot the baseline. This undershoot is abolished by strophanthidin (M. Tsacopoulos & J. A. Coles, unpublished) and so is probably due, in part at least, to the pumping of K⁺ into the photoreceptors (see also Heinemann & Lux, 1975). Anoxia was verified by measurements of the local (not shown) and occurred just before the response labelled 3. It was manifested by a slow increase of [K⁺]_o in the dark, presumably because of inhibition of the Na⁺ pump. The flash labelled 3 elicited a A[K⁺]_o response of almost normal amplitude but with a time course completely distorted when compared to that recorded in normoxia: the recovery to the baseline is incomplete, [K⁺]_o remained at a higher concentration in the dark than before the flash and even continued to increase slowly.

The simplest explanation of this phenomenon is that the Na⁺ pump is inhibited by the anoxia (due to lack of ATP) and thus the K⁺ released by the photoreceptor after the flash accumulates in the extracellular space. This is despite some influx into the glial cells due to the spatial buffering mechanism, which is not dependent on metabolic energy (Gardner-Medwin, Coles & Tsacopoulos, 1981; Coles & Orkand, 1983). Interestingly, after about 2 min, the [K⁺]_o increase reached a plateau, at about 30 mmol l⁻¹. At this level, a flash elicited only a small Δ [K⁺]_O, presumably because there was only a tiny light-induced Na⁺ influx into the photoreceptors. Tissue reoxygenation led to a rapid decrease of [K⁺]_o in the dark, towards the normal baseline existing before the exposure to anoxia. It then undershot the baseline since, following anoxia, the photoreceptors are loaded with Na⁺, which in the renewed presence of ATP greatly activates electrogenic sodium pumping (Thomas, 1972). A similar pattern for [K⁺]_o has been found to occur after anoxia in the mammalian cortex (Morris, 1974; Ullrich, Steinberg, Baierl & ten Bruggencate, 1982). Apparently, during the recovery from anoxia, the physiological redistribution of ions in the different tissue compartments is completed by the end of the undershoot of [K⁺]_o. Indeed, at this point, a flash elicited a Δ [K⁺]₀ response very similar to that recorded before the anoxia. (Fig. 5A). Inhibition of the sodium pump by anoxia and the related ionic perturbations recovered on average within 5 min (five experiments) after the onset of reoxygenation, i.e. faster than the resistance to anoxia. Similar results were obtained when the oxidative metabolism was inhibited by amytal (see below).

Conceivably anoxia could cause an increase of the intracellular ionized calcium (Lo, Wong & Pak, 1980) which does not necessarily recover with the same time course as the activity of the Na⁺ pump. This is important because it is known that an increase of [Ca²⁺]_i causes a decrease in sensitivity of invertebrate photoreceptors to light (Lisman & Brown, 1972; Bader et al. 1976). To examine whether there was a rise in [Ca²⁺]_i, the effects of anoxia upon the kinetics of the receptor potentials were examined (Figs 6, 7). Fig. 6A shows that during anoxia, the latency and the time-to-peak of the photoresponse progressively increased as the amplitude decreased. The decrease of amplitude could be due to a rise in [Ca²⁺]_i, but the first two effects are opposite to those expected for a rise in [Ca²⁺]_i, and are similar to those found in another invertebrate photoreceptor (Lantz & Mauro, 1978). Following reoxygenation, recovery of the latency and the time-to-peak occurred rapidly (Fig. 6B). Then, the fast repolarizing phase became progressively shorter by an amount comparable to that recorded by Bader et al. (1976) after intracellular Ca-EGTA injection, despite a return of the amplitude to the normal pre-anoxia level (Fig. 7, traces 9 and 11). The effects of Ca injection were fully reversible in 6 min (Bader et al. 1976, their Fig. 8); in our experiments the reversibility of the shortening was variable: in some cells there was a partial recovery in about 5 min after the complete recovery of the amplitude from anoxia, but in many other cells the shortening persisted for more than 60 min in normoxia. Careful comparison of the persistence of the shortening of the response with vulnerability to anoxia showed no correlation between the two.

In the following section of this paper, we present evidence suggesting that the resistance of the receptor potential to anoxia is related to the recovery of the oxidative metabolism.

As mentioned above, a flash of light induces a rapid increase of the oxygen consumption which in turn causes a transient drop of the local (Tsacopoulos & Poitry, 1982).

In the experiment shown in Fig. 8, a photoreceptor located about 150 μm from the exposed surface of a retinal slice was impaled and the membrane potential was recorded. The was also recorded at two sites: one at the exposed surface and another located deep in the tissue at about the same depth as the impaled photoreceptor. The replacement of O₂ by N₂ caused a drop of the local ⁠. It reached a value close to zero faster deep in the tissue than at the surface because in addition to the reduction of the O₂ flux, the superficial layers of the tissue consumed O₂. The last light-induced is indicated by an arrow on record D. The tissue anoxia was followed by depolarization of the photoreceptor, and progressive diminution of the amplitude of the receptor potential until it was completely blocked. This record clearly shows the tight coupling between oxidative metabolism and photoreceptor function. When the in the Ringer was reintroduced, increased rapidly at the surface and more slowly inside the tissue. This slowness in the recovery of deep in the tissue was due to a higher in the dark following the anoxia. This became clearer when the oxidative metabolism was transitorily inhibited by Na-amobarbital (amytal) because this drug rapidly blocked the light-induced extra O₂ uptake, but for short times affected but little the resting (small steady increase).

Amytal blocks the mitochondrial respiratory chain (see e.g. Nishiki, Erecinska & Wilson, 1979) by inhibiting the transfer of electrons at the level of flavoproteins (Chance & Hollunger, 1960), whereas anoxia blocks the transfer of electrons from cytochrome a to O₂. Fig. 9 shows that, like anoxia, 3 mmol l⁻¹ amytal inhibited the light-induced and Δ [K⁺]_O and suppressed the receptor potentials. As with anoxia, exposure to amytal caused [K⁺]_o to increase to a plateau of 30 mmol l⁻¹ (Fig. 9B). The small increase of the local in the dark (Fig. 9C,D) indicates that basal was also partly inhibited by amytal (see Tsacopoulos et al. 1981). After removal of amytal, returned to the baseline in the dark, and transitorily undershot it. This undershoot clearly indicates that in the dark was transitorily higher than before inhibition of the oxidative metabolism.

Inspection of Figs 8 and 9 reveals that the recovery of the amplitude of is much slower than the recovery of the receptor potential or the activity of the Na-pump as expressed by the time course of the [K⁺]_o undershoot. This indicates that the light-induced recovers, after a transient inhibition of the oxidative metabolism, more slowly than most of the electrophysiological parameters, including the return of the post-anoxic transient undershoot of [K⁺]_o to the baseline. The return of to normal after a brief inhibition of the oxidative metabolism by anoxia or amytal was studied quantitatively in the experiments illustrated in Fig. 10. In Fig. 10A, was recorded at a depth of 160 μm in a 300μm slice, and at the exposed surface. The retina was stimulated with strong light flashes, one every 2 min. During oxygenation, an intense flash induced a which reached a peak of 45glg^−,min⁻¹ above the baseline in the dark (control: trace C). The retina was then exposed to anoxia for a period sufficient to block the receptor potential (not shown). When O₂ was reintroduced to the Ringer, rose but, as in the other experiments, the return towards the baseline was much slower than the drop during the exposure to N2. The response recorded 4 min after the onset of the increase of the local had an amplitude about 50% of the control. Note that at this time both the receptor potential and the activity of the pump had recovered almost completely to the pre-anoxia values (Fig. 8A,B). It took 12 min more before the amplitude of had recovered to about 90% of the amplitude of the control response. Similar results were obtained in six slices and also, as shown in Fig. 10B, when the retina was exposed to 3 mmol 1⁻¹ amytal: 6min after the removal of amytal, the amplitude and the time course (Fig. 9A) of the receptor potential as well as of [K⁺]_o had completely recovered, but it took about 23 min for to recover.

It appears, therefore, that the physiological resistance of the photoreceptors to anoxia is related to the recovery of the oxidative metabolism. This is clearly shown in Fig. 11. In this experiment, the receptor potential was recorded simultaneously with the local at the same depth (150 μm). The retina was exposed to anoxia until 90% of the receptor potential was suppressed. Then the retina was allowed to recover for 22 min, until and the light-induced had recovered exactly to the same level as before the first anoxia. Thus, the first and the second anoxias were applied under the same conditions of oxidative metabolism. It then took the same time for the second anoxia to suppress 90% of the receptor potential.

The results presented in this paper unequivocally confirm earlier findings (Baumann & Mauro, 1973; Wong et al. 1976; Lantz & Mauro, 1978; Payne, 1981) showing that anoxia depolarizes the membrane of arthropod photoreceptors and abolishes the receptor potential. The observed accumulation of K⁺ in the extracellular space during anoxia or exposure to amytal indicates that the depolarization produced by anoxia is probably due to inhibition of the Na⁺ pump. This accumulation occurred slowly in the dark, but stimulation with light accelerated it (Figs 5,9), presumably because light opens Na⁺ channels (Fulpius & Baumann, 1969; Coles & Orkand, 1982; Bacigalupo & Lisman, 1983) and much more Na⁺ enters the photoreceptor (exchanged with K⁺) than during the ‘leakage’ current in the dark. However, after about 2 min of anoxia or longer exposure to amytal, both the accumulation of K⁺ and the depolarization of the membrane greatly slowed down, presumably because less Na⁺ entered the photoreceptors, since anoxia blocks the opening of the Na⁺ channels by light (see Baumann & Mauro, 1973). The depolarization and the suppression of the receptor potential, as well as their recovery, did not occur in parallel. This is more evident in Limulus or barnacle photoreceptors (Fuortes, 1965; Lantz & Mauro, 1978).

Measurements of local Po, allowed a precise determination of the occurrence of complete anoxia and therefore of the arrest of the oxidative metabolism. Thus it was shown that depolarization starts after a certain delay following the anoxia and the abolition of the receptor potential also occurs progressively. In fresh preparations, it took about 3–4 min for anoxia to suppress completely the receptor potential (Fig. 2). We define this as a resistance of the photoreceptors to anoxia. As shown in a previous article (Coles & Tsacopoulos, 1985), this resistance is unlikely to be related to activation of anaerobic glycolysis leading to the accumulation of lactate in the tissue as is the case in the mammalian brain (McIlwain & Bachelard, 1971) or retina (Tsacopoulos, 1979). We hypothesize that the suppression of the receptor potential rather is related to the progressive consumption of a substance from an intracellular store which is not renewable during anoxia. This substance may be ATP or another high energy phosphate. Consequently, the resistance of the photoreceptor to anoxia would depend on the quantity of this energy substance and the rate of its use during complete inhibition of oxidative metabolism. This hypothesis is supported by the observations that the suppression of the receptor potential was progressive and that the time required for complete abolition was dependent on the intensity of the light stimulation (Fig. 3) and on the interval between successive exposures to anoxia (Fig. 4) (see below). In addition, recent findings have shown that exposure of retinal slices for 3–4 min to anoxia causes about a 50% decrease of the total retinal ATP concentration (Tsacopoulos & Coles, 1984). A single flash of light induces a transient increase of in the dark which reaches a peak of about 50 μlg⁻¹ min⁻¹ (Tsacopoulos & Poitry, 1982). This is accompanied by a transient increase in the ATP concentration (Coles, Tsacopoulos & Dunant, 1984). Also, a single flash induces a transient increase of the intracellular Na⁺ concentration (Δ [Na⁺]_i) of 3–4 mmol l⁻¹ (Coles & Orkand, 1982; Tsacopoulos et al. 1983). Most of the extra ATP produced after a flash appears to be used by the Na⁺ pump (Tsacopoulos et al. 1983). Consequently, during stimulation of the retina with light flashes there is an increase in both production and consumption of ATP by the photoreceptors. Tsacopoulos & Coles (1984) have found that repetitive stimulation of the retina at 0’5 Hz did not cause any significant change of the [ATP] in the steady state.

It is of interest that hypoxia blocks synaptic transmission in hippocampal slices from guinea pig, largely as a result of a decline in [ATP], and that a resistance to this effect can be induced by raising phosphocreatine levels (Lipton & Whittingham, 1982). Such resistance is unlikely to be caused by creatine phosphate in the drone, since none can be detected in the retina (Tsacopoulos & Poitry, 1982), but might be induced by some other high energy phosphate.

The diminution of ATP during anoxia should inhibit the recently discovered light-induced phosphorylation reactions that probably lead to the opening of Na⁺ channels on stimulation of invertebrate photoreceptors (Fein et al. 1984; Brown et al. 1984). Also, an increase in intracellular free Ca²⁺ occurring during metabolic inhibition could reduce the sensitivity of the photoreceptors to light (see Brown & Lisman, 1975). Wong, Lantz & Mauro (1979) suggested such a mechanism since they observed that the intracellular injection of the Ca²⁺ chelator EGTA in the Limulus ventral photoreceptor transitorily restored some of the receptor potential which was abolished by the metabolic inhibitor DNP. Pressure injection of calcium into drone photoreceptors reduces the amplitude of the receptor potential and shortens the fast repolarizing phase of the response (Bader et al. 1976) (see Baumann, 1974, for terminology); in contrast, injection of EGTA makes the response much larger. Also, Brown & Lisman (1975) have shown that injection of calcium reduces the latency and time-to-peak of the photoresponse. The effects of anoxia on the time course of the receptor potential (Figs 6,7) suggest that if a Ca²⁺ increase in the drone photoreceptor is the cause of changes due to anoxia, this occurs mainly during the recovery phases.

Thus, it appears difficult to suggest a causality between an increase of [Ca²⁺]_i during anoxia and the progressive inhibition of the receptor potential. More probably, anoxia or amytal blocks one step in phototransduction in which the phosphorylation by a protein kinase is involved (Fein et al. 1984; Brown et al. 1984).

After the abolition of the receptor potential, when O₂ was reintroduced in the Ringer’s solution or amytal was washed out, oxidative metabolism resumed. It is reasonable to expect that the immediate consequence of this should be the progressive restoration of all ATP-dependent functions of the photoreceptor which were suppressed or diminished during the inhibition of the oxidative metabolism. Apparently, the first function that recovered was Na⁺ pumping, since the photoreceptors quickly repolarized and even hyperpolarized, presumably because this pump is electrogenic (Thomas, 1972). Consistent with this finding was the parallel decrease of [K⁺]_o, which, after a large undershoot, returned to the baseline in about 5 min (Fig. 5). The amplitude and the kinetics of the receptor potential and the light-induced membrane conductance (Baumann & Mauro, 1973) recovered after the repolarization. It has been observed that after a brief exposure of the retina to 100 mmol l⁻¹ K⁺ the repolarization of the membrane and the recovery of the amplitude and the kinetics of the receptor potential are simultaneous (Dimitracos, 1983; see also Fulpius & Baumann, 1969). Consequently, the recovery from the effect of anoxia is not that of a simple ionic perturbation, but more complex, determined essentially by the time course of the recovery of metabolic energy-dependent processes that were inhibited during the anoxia. Among these processes the Na pump recovers first, maybe because it can use the first new molecules of ATP phosphorylated by the mitochondria immediately following the onset of oxidative metabolism, and this simply for anatomical reasons. In the drone photoreceptors, the mitochondria are located at the periphery of the cell very close to the cytoplasmic membrane facing the glial cells (Fig. 1), with which the photoreceptors display ionic exchanges (Coles & Tsacopoulos, 1981). Thus, it appears plausible that the pumping sites are located on the cytoplasmic membrane close to the mitochondria. When oxidative metabolism is resumed the new molecules of ATP may well be preferentially used for Na⁺ pumping. As Na⁺ is extruded from the cell (as seen in the recovery of the [K⁺]_o undershoot, Figs 5,9), more and more molecules of ATP produced by the mitochondria may be able to diffuse towards more distant sites of the cells, namely the subrhabdomeric cistemae (Perrelet, 1970) and the microvilli of the rhabdom, and allow ATP-dependent processes of the phototransduction process to recover from anoxia. Thus, cellular compartmentation and intracellular diffusion may explain the sequence of recovery of the various functions of the photoreceptors.

The effect of the duration of oxygenation upon the resistance of the photoreceptor to subsequent anoxia can be explained in terms of an ATP store. The effect could be produced if the duration of oxygenation was insufficient to allow replenishment of the store. This might be due to an increase in the basal use of ATP, as observed experimentally. The light-induced which normally produces a net increase of ATP (Coles et al. 1984) also does not recover completely during this period (Fig. 10).

During recovery from anoxia or exposure to amytal, the resting of the drone retina was transitorily higher than before the inhibition of the oxidative metabolism, and of smaller amplitude (Figs 8, 9, 10, 11). Similarly, the oxidative metabolism of mammalian nervous tissue is transitorily higher after brief anoxia, and it has been suggested that this is due to an accumulation of ADP and phosphate during anoxia (Kreisman, Sick, Lamanna & Rosenthal, 1981; Chance & Schoener, 1962). An alternative explanation, at least in the drone retina, could be the following: during anoxia or exposure to amytal, the photoreceptor becomes loaded with both Na⁺ and metabolic substrate (a product of glycogen from the glial cells; Evêquoz, Stadelmann & Tsacopoulos, 1983) which is not consumed because of the lack of O₂ or the blockage of the respiratory chain by amytal. When O₂ is reintroduced, or amytal removed, is strongly activated because the photoreceptors are loaded with this substrate and because of the large increase in activity of the sodium pump. However, it is not yet understood why the light-induced recovers so slowly after anoxia or amytal, because the mechanism for the regulation of is still not explained (see Tsacopoulos et al. 1983).

In conclusion, inhibition of the oxidative metabolism by anoxia or amytal depolarizes the photoreceptor and suppresses the receptor potential. The photoreceptor exhibits a certain resistance to anoxia which is probably due to an ATP store which is progressively consumed during the anoxia. The faster this store (or part of it) is consumed, the faster occurs the suppression of the receptor potential. A second exposure to anoxia, if applied shortly after the first, suppresses the receptor potential faster, even though most of the electrophysiological parameters are restored. A longer time interval between periods of anoxia is necessary for full recovery and this time interval is similar to that necessary for the recovery of light-activated oxidative metabolism. The reason for this link between full recovery from anoxia and recovery of light-induced remains to be elucidated.

We thank Drs J. A. Coles and G. J. Jones and Mr S Poitry for helpful discussions and comments on the manuscript. All computer programmes used in this study were written by Mr S. Poitry. Messrs. J. -L. Munoz and Ph. Perrottet made many of the microelectrodes, D. Pollet constructed much of the special apparatus. Supported by Swiss NSF Grants 3.399.0-78 and 3.316-0.82 and the Georges Kernen Foundation.