HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Vellani, V. Articles by McNaughton, P. A. Search for Related Content PubMed PubMed Citation Articles by Vellani, V. Articles by McNaughton, P. A.

Modulation of the synaptic Ca²⁺ current in salamander photoreceptors by polyunsaturated fatty acids and retinoids

Vittorio Vellani, A. Martyn Reynolds and Peter A. McNaughton

MS 0874 Received 23 March 2000; accepted after revision 9 August 2000.

	ABSTRACT

Top Abstract Introduction Methods Results Discussion References

1. Synaptic transmission between retinal photoreceptors and second-order neurones is controlled by an L-type Ca²⁺ conductance (g_Ca) in the photoreceptor inner segment. Modulation of this conductance therefore influences the flow of visual information to higher centres.

2. Possible modulation of g_Ca by retinal factors was investigated using patch clamp and Ca²⁺ imaging. No significant modulation of g_Ca by retinal neurotransmitters nor by intracellular signalling pathways was found.

3. g_Ca was inhibited by retinoids (all-trans retinal) and by polyunsaturated fatty acids (PUFAs) such as arachidonic acid and docosahexaenoic acid, which are known to be released in the retina by exposure to light.

4. Some PUFAs tested are physiological substrates for the cyclo-oxygenase, lipoxygenase and epoxygenase pathways, but specific inhibitors of these pathways had no effect on the inhibition of g_Ca.

5. Treatments designed to activate or inhibit G-protein-coupled pathways or protein kinases A and C similarly had no effect on the inhibition by PUFAs nor on g_Ca itself. Inhibitors of phosphatases 1 and 2A were also largely ineffective.

6. The inhibition by PUFAs is, however, dependent on membrane potential, suggesting that it arises from a direct interaction of fatty acids with the Ca²⁺ channel. The effect was not use or frequency dependent, suggesting that the effect does not depend on channel gating state.

7. Control by retinoids and by PUFAs may be an important mechanism by which the Ca²⁺ conductance, and consequently the transmission of the visual signal, is modulated at the first retinal synapse.

	INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

The photoreceptor membrane potential is depolarized to -35 mV in darkness by a steady inward current flowing through the membrane of the photoreceptor outer segment. Light suppresses this current, thereby causing a relatively slow and maintained hyperpolarization. The hyperpolarization in turn suppresses a voltage-sensitive Ca²⁺ conductance (g_Ca) in the photoreceptor synaptic terminal, and consequently inhibits exocytosis of neurotransmitter onto horizontal and bipolar cells, the second-order visual neurones (Barnes et al. 1993). The characteristics of g_Ca at the photoreceptor synapse therefore need to be different from those at most synapses, where action potentials cause large but brief depolarizations. The principal Ca²⁺ channel in the photoreceptor inner segment and synaptic terminal, in keeping with the requirement for a tonic effect on synaptic transmission, is non-inactivating and suppressed by nifedipine, and is therefore L-type (Lasater & Witkovsky, 1991). The objective of the present study was to investigate factors that may be important in modulating synaptic g_Ca, and therefore in modulating information flow through this first retinal synapse.

Previous work by others has shown that the inner segment Ca²⁺ conductance is activated by nitric oxide and by increased extracellular pH (Barnes et al. 1993; Kurenny et al. 1994). Other possible factors which might modulate g_Ca include neurotransmitters such as dopamine and glutamate released from outer retinal neurones (Pourcho, 1996) and light-modulated paracrine factors such as melatonin (Witkovsky & Dearry, 1992). We investigated a possible modulatory action of these substances on g_Ca in the present study but found no effect. A profound effect was, however, found on application of polyunsaturated fatty acids (PUFAs) such as arachidonic acid (AA) and docosahexaenoic acid (DHA), and of retinoids such as all-trans retinal (ATR), all of which are known to be released in the retina by light. Strong light is thought to activate phospholipase A₂ (PLA₂) directly via an action involving the subunit of transducin (Jelsema & Axelrod, 1987). A light-invoked release of AA has been detected in rat retina, and its release was found to be antagonized by inhibitors of PLA₂ (Jung & Remé, 1994). DHA is also released by the activation of PLA₂ by light (Reinboth et al. 1996). Both of these PUFAs are therefore plausible regulators of g_Ca in vivo. The bleached chromophore all-trans retinal (ATR) is also produced in the photoreceptor outer segment, as a direct result of the action of light on rhodopsin, and is subsequently reduced to all-trans retinol (Fain et al. 1996). In the present study we show that g_Ca is rapidly and potently suppressed by PUFAs, by ATR and by other retinoid compounds, and we investigate the underlying mechanism of this effect.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Preparation and patch-clamp recording

Larval salamanders (Ambystoma tigrinum) were dark adapted for 2-5 h, killed by concussion, rapidly decapitated, double-pithed and the eyes enucleated. Rods were isolated by mechanical dissociation under infra-red illumination (Lagnado et al. 1992). The outer segment of an intact, isolated rod was drawn into a suction pipette, leaving the inner segment in front of a single-barrel perfusion tube, in which the flowing solution could be changed using a 6-way tap in less than 1 s. Whole-cell patch-clamp recordings were made from the inner segment using fire-polished borosilicate pipettes of bubble number 5·4-5·8. The patch-clamp amplifier used to record current was a List EPC-7. Command signals were generated and data acquired using the pCLAMP interface and software (Axon Instruments Inc.). Initial experiments were recorded with original software written in Visual Basic by A. M. Reynolds. The standard ramp protocol used to study the I-V relation of g_Ca is shown in the upper left inset of Fig. 1. The I-V relation was leak corrected using the value of leak conductance obtained from the initial step voltage change from the holding potential of -50 mV to the pre-ramp potential of -100 mV. Recordings in which the input resistance was less than 1 G were rejected.

All procedures apart from the initial isolation of rods were carried out in continuous bright light, which completely suppressed the light-sensitive conductance of the rod outer segment. As this is the only significant conductance in the outer segment (Lagnado & McNaughton, 1991) all of the current in these experiments was from the membrane of the inner segment. Some experiments were carried out on rods from which the outer segment had been removed during the isolation process, and the form of the I-V relation was identical to that observed in intact rods.

Immediately after beginning the recording the Ca²⁺ current amplitude increased to a stable level, within 10-30 s. The reason for this increase has not been established, but may be related to the reduction in [Ca²⁺]_i as the EGTA-buffered pipette solution diffused into the cell - the calculated time course of EGTA inflow is consistent with this hypothesis (Oliva et al. 1988). Once a steady current amplitude had been reached the current remained stable for 30-60 s before declining progressively, with a time constant of 3-5 min, a phenomenon that has been observed in many cell types (Horn & Korn, 1992). We took advantage of the period of 30-60 s when the Ca²⁺ current amplitude was reliably stable to test the effects of substances modulating the current amplitude. When no test substance was applied, the current at the end of a 30 s test period was unchanged (change = -1·2 ± 2·4 %, n = 6), and therefore little error is likely to have been introduced by current rundown using this protocol. Enzyme inhibitors/activators were pre-applied 5 min before the beginning of the recording when appropriate (see text).

Most experiments were carried out using the ruptured patch variant of the whole-cell voltage-clamp technique, but in some experiments where we wished to maintain intracellular signalling pathways in an undisturbed state (see Results), the recordings were performed with the perforated patch technique (Rae et al. 1991). Briefly, 150 mg ml^-1 of amphotericin B was added to the standard intracellular solution, and then back-filled into the recording pipette. Access (resistance 7-15 M) normally occurred within 2-3 min, and allowed stable recordings for up to 8-12 min. Series resistance compensation was used (approximately 70 % could be compensated without oscillation) but in view of the small currents recorded in these experiments, the error introduced by series resistance was negligible. Whole-cell capacitance was also compensated.

All experiments were carried out at room temperature (18-22°C).

Measuring [Ca²⁺]_i in the inner segment

Rods isolated as described above were allowed to attach to clean borosilicate coverslips and were loaded with the Ca²⁺-sensitive indicator fluo-3 (Molecular Probes) by incubation with the acetoxy-methyl ester (AM) form (10 µM) for 30-60 min at 4°C in the dark. The Ca²⁺-dependent fluorescence in the inner segment was imaged using an MRC-600 confocal microscope (BioRad) as previously described (Nadal et al. 1996). At the end of the experiment the Ca²⁺-dependent fluorescence was calibrated by first elevating [Ca²⁺]_i to a saturating level by permeabilizing the rod with ionomycin (10 µM) in the presence of 30 mM [Ca²⁺]_o, followed by lysis in digitonin (20-50 µM) in order to measure background fluorescence from loaded organelles and from instrumental sources (for details see Nadal et al. 1996).

Solutions

Rods were isolated and initial patch-clamp recordings were made in Ringer solution, of composition (mM): 110 NaCl, 2·5 KCl, 1·0 CaCl₂, 1·6 MgCl₂, 10 Hepes, 2 NaHCO₃, neutralized to pH 7·6 with NaOH. Glucose (5 mM) was added to the Ringer solution before use. To record the L-type current we used Ba²⁺ as Ca²⁺ surrogate, and replaced other external cations with tetraethylammonium (TEA⁺) and anions with methanesulphonate (CH₃SO₃^-) in a solution of composition (mM): 105 TEA-CH₃SO₃, 6 BaCl₂, 10 Hepes, neutralized to pH 7·6 with TEA-OH. In the pipette solution we used TEA and Cs⁺ to block K⁺ conductances in a solution of composition (mM): 25 TEA-OH, 95 CsCl, 10 EGTA, 10 Hepes, neutralized to pH 7·2 with CsOH. To maintain G-protein-coupled pathways functional 0·2 mM Li-GTP and 2·5 mM Mg-ATP were added to the intracellular solution. Liquid junction potentials between the intracellular and extracellular solutions were less than 2 mV, and appropriate correction for this was made.

To avoid oxidation, fatty acids were purchased in 10 mg amounts stored under xenon, and were dissolved in DMSO at 5 mM concentration. These stock aliquots were stored at -70°C and were kept for no more than 2-3 days. Fatty acids were added to Ba²⁺ extracellular solution at the desired concentration, while sonicating to aid dispersal. All chemicals were from Sigma unless otherwise stated.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Characteristics of the Ca²⁺ current in rod inner segments

The Ca²⁺ current was routinely measured using 6 mM Ba²⁺ as a Ca²⁺ surrogate, and with a ramp voltage-clamp protocol from a starting potential of -100 mV (see top left inset to Fig. 1). A typical current-voltage relation obtained using this protocol is shown in Fig. 1. Inactivation of the Ca²⁺ current is negligible (lower left inset in Fig. 1) and the current-voltage relation observed using voltage steps (circles) was identical within experimental error to that obtained with the ramp protocol (continuous curve). As the ramp protocol was much less time consuming it was therefore adopted for most subsequent studies. The peak Ca²⁺ current was monitored every 2 s as an index of Ca²⁺ conductance. Possible changes in Ca²⁺ channel gating as a function of membrane potential were monitored by measuring the position on the voltage axis of the peak current and the point of inflexion of current activation. The threshold for activation of g_Ca was -40 mV, and the current attained its peak at -5·5 ± 0·4 mV. The inflection point and zero crossing point of the I-V relation were -21·5 ± 0·4 and +35·0 ± 0·9 mV, respectively. Peak current amplitude was -148 ± 8 pA (n = 255; all values are means ± S.E.M.).

		View larger version [in this window] [in a new window]
Figure 1. Characteristics of Ca2+ current in salamander rod inner segments Curves show current-voltage relations in response to the ramp protocol shown in the upper inset on the left. Each trace is the average of 3 records, leak subtracted using the value of leak conductance obtained on stepping from the holding potential (-50 mV) to the pre-ramp potential (-100 mV). Points show current amplitude at the end of a 500 ms voltage step (see voltage protocol in upper inset on the right) from a prepulse potential of -100 mV (n = 7, error bars show ±S.D.). The small residual current in nifedipine (3 µM), which was observed only in some cells, was blocked by Cd2+ (50 µM). The Ca2+ current showed little inactivation during a prolonged depolarizing voltage pulse (lower inset on the left).

About 95-100 % of the inward current was blocked by the dihydropyridines nifedipine (3 µM, Fig. 1) and nitrendipine (3 µM, not shown). A remaining small inward current with similar I-V properties, completely suppressed by Cd²⁺ at 50 µM, was present in some cells (see e.g. Fig. 1). The high threshold of activation, the absence of inactivation, the block by dihydropyridines and the enhancement by Bay K 8644 (increase 620 ± 130 %, mean ± S.E.M., n = 12) are all characteristic properties of L-type Ca²⁺ channels. We searched for the possible presence of rapidly inactivating or low threshold Ca²⁺ currents by using a voltage step from -100 to -30 mV, in the presence of nifedipine (3 µM). The leak-compensated inward current in these conditions was only 1·7 ± 0·6 % (n = 20) of the maximum current observed in the absence of nifedipine. This small residual current showed no sign of inactivation, and therefore can almost certainly be attributed to incomplete block of L-type Ca²⁺ channels by 3 µM nifedipine, as has been observed in other preparations, rather than to the presence of other Ca²⁺ channel types (Taylor & Morgans, 1998). The g_Ca in this preparation of rod inner segments is therefore almost completely L-type, and there is no evidence for significant current components carried by other types of Ca²⁺ channels.

Possible modulation of Ca²⁺ current by neurotransmitters

The presynaptic Ca²⁺ current in a number of other cell types is modulated by neurotransmitters such as noradrenaline, dopamine and glutamate (Starke et al. 1989). In the outer retina glutamate is released by photoreceptors, GABA by horizontal cells and dopamine by interplexiform cells (Pourcho, 1996). We investigated whether modulation by these neurotransmitters might play a role in regulating g_Ca, and therefore synaptic gain, at the first retinal synapse.

The application of catecholamines at high concentrations had no significant effect on the peak Ca²⁺ current (Fig. 2A), nor on the position of the activation curve on the voltage axis (not shown). Similar results were obtained in other experiments with glutamate, serotonin and GABA, all at 100 µM. The possibility that melatonin, whose concentration in the retina shows a pronounced diurnal variation, might modulate g_Ca was also investigated and no significant effect was found (Fig. 2A).

		View larger version [in this window] [in a new window]
Figure 2. Effect of neurotransmitters and modulators of signalling pathways on the peak amplitude of Ca2+ current The peak amplitude of Ca2+ current was measured by applying voltage ramps, as shown in Fig. 1, at 2 s intervals. A, fractional suppression of Ca2+ current in response to adrenaline, noradrenaline, dopamine, serotonin, melatonin, glutamate and GABA (all 100 µM). Error bars show ±S.E.M. Numbers of experiments are shown above each bar. None of the changes in Ca2+ current was statistically significant. B, fractional suppression of Ca2+ current in response to application of the PKC and PKA inhibitor staurosporine (1 µM), the selective PKC inhibitor RO 31-8220 (1 µM), the adenylate cyclase activator forskolin (20 µM), the specific PKC activator phorbol myristate acetate (PMA, 1 µM), and the phosphatase inhibitors calyculin A, okadaic acid and mycrocystin (all at 20 µM). For the experiments in B the perforated patch configuration was used (see Methods). The suppression of current by staurosporine, RO 31-8220, forskolin and PMA was significant at the 5 % level. Error bars show ±S.E.M.

We then investigated the possibility that g_Ca might be modulated by intracellular signalling pathways. In this series of experiments the perforated patch technique was used (see Methods) in order to prevent wash-out of small molecules from the cell. A small but significant suppression of current was observed with application of forskolin (20 µM), which activates adenylate cyclase, the protein kinase C (PKC) activator phorbol 12-myristate 13-acetate (PMA, 1 µM), the protein kinase A (PKA) and PKC inhibitor staurosporine (1 µM), and the specific PKC inhibitor RO 31-8220 (see Fig. 2B). These effects are likely to be non-specific, however, because other treatments designed to modulate G-protein-coupled pathways were ineffective (see below), and because PMA and the PKC inhibitors had similar, rather than opposite, effects. No significant effect was observed with calyculin A, microcystin and okadaic acid, inhibitors of phosphatases 1 and 2A (20 µM, see Fig. 2B).

In conventional whole-cell recordings the maximum value of Ca²⁺ current was not significantly different from control when 1·25 mM cAMP was included in the patch pipette (135 ± 18 pA, n = 13, compared with 148 ± 8 pA in control cells, see above). The rundown in Ca²⁺ current with time was also unaffected by cAMP. Inclusion of the pre-activated catalytic subunit of protein kinase A (62 units ml^-1; Sigma) in the patch pipette also had no effect on the maximal Ca²⁺ current (121 ± 17 pA, n = 4). A possible involvement of G-protein-coupled pathways was investigated by intracellular perfusion with GTP analogues. High concentrations of GTP analogues were used to allow appropriate perfusion of the cytoplasm, with typical measured access resistances of 3-5 M, within 20-40 s of the beginning of the whole-cell recording (Oliva et al. 1988). Maximal Ca²⁺ current was 152 ± 30 pA (n = 6) in the presence of 2 mM GTP, 210 ± 47 pA (n = 7) in the presence of 2 mM GDP--S and 151 ± 20 pA (n = 12) with the G-protein activator GTP--S. None of these values is significantly different from control values obtained with 0·2 mM GTP in the patch pipette.

The lack of effect of either cAMP or activated PKA shows that the synaptic g_Ca in rod photoreceptors, unlike g_Ca in cardiac muscle and elsewhere, is not subject to modulation by PKA. The lack of effect of GDP--S, which stabilizes the inactive form of G-proteins, or GTP--S, which stabilizes the active form, argues against an involvement of G-protein-coupled signalling pathways more generally, although it is possible that G-protein turnover in the absence of activation is too low to allow significant nucleotide exchange within the time course of the recordings.

Suppression of Ca²⁺ current by polyunsaturated fatty acids and retinoid compounds

Figure 3 shows that AA and ATR potently suppress the photoreceptor synaptic Ca²⁺ current, with no change in the voltage dependence of channel activation. The suppression was rapid but not instantaneous: the effect of nifedipine, which binds directly to the L-type Ca²⁺ channel at an external site (Nakayama et al. 1991), is considerably more rapid than that induced by either AA or ATR (Fig. 3B).

		View larger version [in this window] [in a new window]
Figure 3. Arachidonic acid and all-trans retinal inhibit Ca2+ current A, steady-state inhibition of peak Ca2+ current by arachidonic acid (AA, 10 µM) and by all-trans retinal (ATR, 10 µM). Note that the form of the current-voltage relation is unaffected. Each trace represents the average of 3 traces taken at the end of a 30 s application of the compound. B, the upper panel shows the time course of inhibition of peak Ca2+ current by AA and ATR, compared with suppression by nifedipine (N, 3 µM). Peak amplitudes was measured every 2 s by the ramp protocol shown in Fig. 1. Time zero is the moment of attaining the whole-cell configuration. Lower graph shows the position on the voltage axis of the peak Ca2+ current () and the point of inflection () as a function of time.

The potency of a number of fatty acids and related compounds in suppressing g_Ca is shown in Fig. 4. Complete concentration-response relations for docosahexaenoic acid (DHA), arachidonic acid (AA) and eicosapentaenoic acid (EPA) (see Table 1) were obtained by exposing the inner segment for a standard period of 30 s to a test concentration of fatty acid, and measuring the suppression of peak Ca²⁺ current at the end of this period. A relatively brief exposure was used because the recovery from long exposures was often incomplete, perhaps because PUFAs partition into internal membranes. An exposure of 30 s was sufficiently long to allow the suppression to reach a steady-state value, while largely avoiding the problem of incomplete recovery. Recovery after exposure to PUFAs was about 90 % complete within 30-90 s for PUFA concentrations 10 µM (see Fig. 3B). Ca²⁺ current was stable for a period of 30-60 s in control cells before current rundown began (see Methods), so no compensation needed to be made for rundown provided the application was made during this time window.

		View larger version [in this window] [in a new window]
Figure 4. Inhibition of Ca2+ current by long-chain fatty acids and related compounds A, concentration-response curves for inhibition of peak Ca2+ current by docosahexaenoic acid (DHA, see Table 1), arachidonic acid (AA) and eicosapentaenoic acid (EPA). The points are fitted by the Hill equation (continuous curves) with the values of 50 % inhibition (K½) and Hill coefficient (n) as given. Error bars show ±S.D. B, comparison of the effectiveness of various long-chain fatty acids and related compounds in inhibiting peak Ca2+ current. Concentration was 10 µM unless otherwise stated. From left: docosahexaenoic acid (DHA, see Table 1), eicosatrienoic acid (ETA), arachidonic acid (AA), eicosapentaenoic acid (EPA), linolenic acid (LN), linoleic acid (LA), oleic acid (OA, 20 µM), myristic acid (MA, 20 µM), eicosatetraynoic acid (ETYA, analogue of AA with double bonds replaced by triple bonds, 20 µM), all-trans retinal, all-trans retinoic acid, all-trans retinol. Error bars show ±S.E.M.

Table 1. Summary of fatty acid structure

Docosahexaenoic acid (DHA)	22:6	cis-4,7,10,13,16,19	-3
Eicosapentaenoic acid (EPA)	20:5	cis-5,8,11,14,17	-3
Arachidonic acid (AA)	20:4	cis-5,8,11,14	-6
Eicosatrienoic acid (ETA)	20:3	cis-8,11,14	-3
Linolenic acid (LN, octadecatrienoic acid)	18:3	cis-9,12,15	-3
Linoleic acid (LA, octadecadienoic acid)	18:2	cis-9,12	-6
Oleic acid (OA, octadecenoic acid)	18:1	cis-9

The inhibitory effect of PUFAs was not affected by current rundown. When 7·5 µM AA was applied for 30 s within 0·5-1 min of reaching the steady state (see Methods), 82 ± 3 % of the current was suppressed (n = 11). If current was allowed to run down to a level equal to 20 ± 2 % of the maximum initial level (3-5 min after obtaining the whole-cell configuration) the same concentration of AA suppressed 75 ± 9 % of the current (n = 4). Using the perforated patch whole-cell configuration, in which the rate of rundown was greatly reduced, 7·5 µM AA suppressed 78 ± 3 % of the Ca²⁺ current (n = 8).

Concentration-response relations were well fitted by the Hill equation with a co-operativity coefficient close to 2 (see Fig. 4A). This may imply that two fatty acid molecules are required to occlude a single Ca²⁺ channel if the interaction is direct (see experiments below). However, some overestimation of the Hill value might be caused by the short time of exposure. The most potent block was obtained with DHA, with a K_½ of 3·0 µM. The 20-carbon chain AA and EPA were slightly less potent (K_½ of 4·7 µM and 7·0 µM, respectively). These inhibitory effects of PUFAs on the photoreceptor Ca²⁺ channel are more potent than most others reported to date (Meves, 1994).

A wider range of fatty acids was screened by observing the inhibition of g_Ca in response to a single concentration (usually 10 µM). When the chain length was reduced to 18 carbons, as in linolenic acid (LN, see Table 1) or linoleic acid (LA) the potency was sharply reduced (see Fig. 4B). Reducing the number of double bonds also substantially reduced the potency, with oleic acid (OA) and myristic acid (MA) producing no inhibition even at the higher concentration of 20-50 µM. The acetylenic compound eicosatetraynoic acid (ETYA), identical to AA except that the four double bonds of AA are changed to triple bonds, was completely ineffective in inhibiting g_Ca (Fig. 4B). The presence of double bonds is therefore essential. ETYA has the same effect on membrane fluidity as PUFAs (Meves 1994), so the absence of any effect of ETYA shows that non-specific changes on the physical properties of the lipid bilayer do not underlie the effect of PUFAs on the rod Ca²⁺ current.

ATR, retinoic acid and retinol contain a long chain of alternating single and double bonds, so perhaps it is not surprising that they also inhibit g_Ca, with a potency only slightly less than that of AA (Fig. 4B). The double bonds, unlike all the other PUFAs tested, are in the trans configuration. This difference does not seem to influence the ability to produce the effect, as linolelaidic acid (18:2, trans-^9,12octadecadienoic acid) also produced a 26 ± 12 % inhibition of g_Ca (n = 3, not shown), similar to its cis- equivalent, linoleic acid. The effect of the retinoid compounds differed from that of PUFAs in that it was irreversible, at least within the time course of the recording. The inhibitory effects of retinoid compounds are particularly interesting as these compounds are produced in the photoreceptor outer segment by a direct action of light on rhodopsin, and therefore may play a physiological role in reducing the gain of synaptic transmission in light.

Arachidonic acid inhibits the voltage-dependent increase in [Ca²⁺]_i

Changes in the photoreceptor membrane potential gate voltage-dependent Ca²⁺ channels, and consequently lead to changes in intracellular [Ca²⁺]. PUFAs such as AA should therefore reduce the rise in [Ca²⁺]_i caused by depolarization, as a consequence of their inhibitory action on Ca²⁺ channels. We tested whether AA has effects on [Ca²⁺]_i in order to investigate whether its effects are still seen in an intact cell bathed in a physiological external solution and with its normal physiological cytoplasmic environment not replaced by patch pipette solution.

Depolarizing the photoreceptor inner segment with a pulse of 20 mM KCl caused a reversible increase in Ca²⁺-dependent fluorescence (see Fig. 5A). The increase is entirely attributable to a Ca²⁺ influx from the external medium, because it is completely abolished in 0 Ca²⁺ (not shown). The cell was then exposed to AA for 60 s before repeating the KCl application, this time in the presence of the fatty acid, and the increase in fluorescence (F_AA) was measured again. Note that F varies almost linearly with [Ca²⁺]_i in these experiments, because F_CTRL was only 39·8 ± 7 % of the value of F_max, the saturating level of fluorescence measured at high [Ca²⁺]_i. The inhibition in [Ca²⁺]_i caused by PUFA application was therefore calculated as [1 - (F_AA/F_CTRL)] (see Fig. 5B).

		View larger version [in this window] [in a new window]
Figure 5. Inhibition by AA of [Ca2+]i increase caused by depolarization A, [Ca2+]i-dependent fluo-3 fluorescence recorded from inner segment during depolarization by 20 mM KCl (horizontal bar). Ordinate gives the change in fluorescence observed upon application of 20 KCl, normalized to peak change in absence of AA (CTRL). B, inhibition of fluorescence increase caused by 20 mM KCl, as a function of concentration of AA. Points fitted by the Hill equation with the parameters given. Error bars show ±S.E.M.

As expected from its action in inhibiting g_Ca, AA reduced the [Ca²⁺]_i increase caused by depolarization. The concentration-response relation for the inhibition was well fitted by a Hill equation with co-operativity coefficient close to 2 (Fig. 5B). The K_½ of 2·7 µM was also close to the value of 4·7 µM observed for inhibition of Ca²⁺ current (Fig. 4A). Block of the depolarization-induced [Ca²⁺]_i increase was also observed with DHA and other PUFAs (not shown). These experiments provide independent confirmation that AA and related compounds inhibit the voltage-gated Ca²⁺ influx in intact rods. They also show that changes in [Ca²⁺]_i, which are responsible for controlling synaptic exocytosis, are indeed modulated by AA.

Metabolic pathways are not involved in the suppression of Ca²⁺ current

AA is the substrate for a number of metabolic pathways, such as the cyclo-oxygenase, lipoxygenase and epoxygenase/cytochrome P450 pathways (Meves, 1994). We therefore tested the possibility that the inhibition of g_Ca is due to an action of one of the products of these pathways rather than to a direct action of AA itself. The possibility that a substrate of the cyclo-oxygenase pathway might inhibit the Ca²⁺ current is ruled out by the observation (see Fig. 6) that indomethacin, a specific cyclo-oxygenase inhibitor, has no effect on the inhibition of Ca²⁺ current produced by AA. Similarly, the lipoxygenase pathway is not involved because intracellular dialysis with the lipoxygenase inhibitor nordihydroguaiaretic acid (NDGA) does not affect the inhibition of Ca²⁺ current. NDGA had to be applied via the patch pipette as, when applied extracellularly, it completely blocked g_Ca, as has also been observed in pituitary cells (Korn & Horn, 1990) and in frog cardiac myocytes (Petit-Jaques & Hartzell, 1996). NDGA is also an inhibitor of non-specific lipid peroxidation (Huang et al. 1992), so this mechanism is unlikely to be involved. ETYA is a competitive inhibitor of cytochrome P450 as well as of 12- and 15-lipoxygenase (Capdevila et al. 1988). ETYA itself does not mimic any of the effect of PUFAs (Fig. 4) and is not able to prevent their effect when applied at the same time or pre-applied (Fig. 6). These experimental observations therefore argue against the possibility that the inhibition is due to a product of a metabolic pathway utilizing AA as substrate.

		View larger version [in this window] [in a new window]
Figure 6. Effect of inhibitors of metabolic pathways utilizing AA on the suppression of peak Ca2+ current The fractional suppression caused by 7·5 µM AA (column 1) is unchanged in the presence of the cyclo-oxygenase inhibitor indomethacin (INDO, 20 µM, pre-applied for 5 min, column 2), the lipoxygenase inhibitor nordihydroguaiaretic acid (NDGA, 20-40 µM, applied via the patch pipette, column 3) and the non-metabolizable AA analogue eicosatetraynoic acid (ETYA, 20 µM, column 5). NDGA and INDO applied at the same time were equally ineffective (column 4). Longer pre-application times (15-60 min) and higher concentrations (100 µM INDO, 50 µM ETYA) were also tested with a similar lack of effect (not shown). Error bars show ±S.E.M.

AA is known to modulate the activity of G-protein-coupled signalling cascades (reviewed in Meves, 1994). We did not uncover any evidence for modulation of the Ca²⁺ current by G-protein-coupled pathways (see Fig. 2), but we thought it worthwhile nonetheless to investigate the possibility that the inhibition of Ca²⁺ current by AA may be directly mediated by a G-protein. We measured the inhibition of Ca²⁺ current caused by a half-saturating and a near-saturating concentration of AA (5 µM and 10 µM), and compared the inhibitory effect with that observed in the presence of increased intracellular GTP (from 0·2 to 2 mM), with the G-protein inhibitor GDP--S (2 mM) and with the activator GTP--S (2 mM). None of these treatments had any influence on the effect of AA (Fig. 7) nor on the rate of recovery of current after AA removal (not shown). The inhibition caused by EPA and ATR was also not affected (not shown). These results suggest that G-proteins are not involved in the mechanism of the inhibitory effect of PUFAs on g_Ca.

		View larger version [in this window] [in a new window]
Figure 7. Modulators of G-protein activity do not affect suppression of Ca2+ current by AA Mean inhibition caused by 5 µM AA (columns to left) and 10 µM AA (columns to right) in the presence of 0·2 mM GTP (control), 2 mM GTP, 2 mM GDP--S with 0 GTP, and 2 mM GTP--S with 0 GTP. Error bars show ±S.E.M.

Possible involvement of kinases and phosphatases

AA is known to modulate the activity of protein kinases (Meves, 1994), and the experiments in Fig. 8 were therefore undertaken to investigate whether protein kinases are involved in the effect of AA on the Ca²⁺ current. Figure 8 shows that the inhibition caused by 7·5 µM AA was not significantly affected by either inhibition of PKC, using the specific inhibitor RO 31-8220, or activation of PKC, using the specific activator PMA. Inhibition of both PKC and PKA by the non-specific inhibitor staurosporine, or activation of PKA, either by applying the adenylate cyclase activator forskolin, or by inclusion of 1·25 mM cAMP in the patch pipette solution, were also without effect on the AA response. The lack of effect with any of these treatments does not favour the possibility that AA might act via modulation of either PKA or PKC.

		View larger version [in this window] [in a new window]
Figure 8. Inhibition of gCa by AA is not affected by activators and inhibitors of PKC and PKA The effect of AA (7·5 µM) on peak Ca2+ current was not significantly affected by a specific inhibitor of PKC (RO 31-8220, 1 µM), by the non-specific PKC/PKA inhibitor staurosporine (1 µM), by the specific PKC activator PMA (1 µM), nor by the PKA activator forskolin (20 µM). Adding cAMP (1·25 mM) to the pipette solution was without effect. Experiments were performed with the perforated patch technique, except for that involving intracellular perfusion with cAMP. Error bars show ±S.E.M.

Another way of testing the question of whether the effect of AA and other fatty acids is mediated by phosphorylation is to inhibit phosphatase activity using specific blockers. In the experiments summarized in Fig. 9 we tested the inhibition caused by a near-saturating level of AA (7·5 µM) under control conditions and in the presence of the phosphatase inhibitors calyculin A, microcystin LR and okadaic acid. All of these will rapidly produce a complete inhibition of protein phosphatases 1 (PP1) and 2A (PP2A) at the high concentrations used (20 µM). The effect of AA therefore should be strongly modulated by these treatments if it is attributable either to modulation of PP1 or PP2A activity, or alternatively to modulation of a kinase which in turn might phosphorylate a site susceptible to dephosphorylation by PP1 or PP2A. Figure 9 shows that a small but significant reduction of the AA inhibition of Ca²⁺ current was observed with calyculin A and okadaic acid, but not with microcystin. A similar inhibition was also observed with a lower concentration of calyculin A (1 µM). We believe that the effect is likely to be non-specific in origin, because the complete inhibition of phosphatase activity caused by the high concentrations used should cause a much more radical change in the AA effect if it depended on the modulation of phosphatase activity. A second reason for believing the effect to be non-specific is that microcystin LR, at a concentration which should also produce a strong inhibition of PP1 and PP2A, had no significant effect.

		View larger version [in this window] [in a new window]
Figure 9. Effect of inhibitors of phosphatase activity Okadaic acid (20 µM) and calyculin A (20 µM) had a small but significant effect (P < 0·01) on the inhibition caused by AA (7·5 µM), but microcystin (20 µM) was without effect. Error bars show ±S.E.M.

Inhibition by PUFAs is voltage dependent

The experiments described in the preceding two sections offer no support for the idea that the inhibition of g_Ca by AA might involve metabolic pathways such as the cyclo-oxygenase or lipoxygenase pathways, G-protein-coupled signalling cascades, or phosphorylation/dephosphorylation reactions mediated by kinases or phosphatases. It is difficult to rule out all such possibilities simply by a process of elimination, but the absence of any evidence for an action involving cellular pathways increases the likelihood that the inhibitory action of AA is due instead to a direct interaction with the channel.

Fatty acids are ionized at physiological pH and are negatively charged. One test for a direct interaction with Ca²⁺ channels is therefore to look for a voltage dependence of the inhibition. If AA and other PUFAs interact directly to block the pore of the Ca²⁺ channel, in such a way that the negatively charged C-terminal is within the transmembrane voltage gradient, then depolarizing the membrane potential will increase the apparent inhibition by attracting a higher concentration of fatty acid molecules into the pore region of the channel.

In the experiment shown in Fig. 10A the holding potential of the rod inner segment was maintained at -100 mV (left trace) and -50 mV (right trace). In this experiment the magnitude of the Ca²⁺ current was tested with a train of brief voltage pulses to 0 mV rather than with the ramps used in other experiments in order to reduce the time spent depolarized in testing the amplitude of g_Ca. A significantly greater inhibition of Ca²⁺ current by DHA is observed when the holding potential is -50 mV than when it is kept at -100 mV.

		View larger version [in this window] [in a new window]
Figure 10. The inhibition of Ca2+ current by PUFAs is potentiated by membrane depolarization A, a subsaturating concentration of DHA (5 µM) suppressed 65 % of peak Ca2+ current from a holding potential of -100 mV (left trace), but 81 % from a holding potential of -50 mV (right trace). Current amplitude was tested using steps (see pulse protocols in insets) to avoid potential-dependent changes in fatty acid binding during the course of a test ramp. B, summary of results from 6 experiments similar to that shown in A. Difference was significant (P = 0·03). Error bars show ±S.E.M.

Figure 10B shows the mean results of six similar experiments (3 performed in the order shown, 3 in reverse order). The mean fraction of Ca²⁺ current suppressed when the fatty acid was applied at -50 mV was 0·81, while from -100 mV it was 0·65, a difference which was significant at the 5 % level (P = 0·03). A similar experiment performed using AA gave a similar significant result (P = 0·04, n = 5). We conclude that the inhibition of g_Ca depends on membrane potential, consistent with a direct interaction of negatively charged fatty acid molecules with Ca²⁺ channels.

We investigated whether the channels need to be open in order to be blocked by PUFAs or, in other words, if the mechanism is use dependent (Hille, 1992). A cell was kept at -100 mV for 2 min, the current amplitude checked once with the step protocol, then AA was applied for 30 s. Finally the Ca²⁺ current suppression was tested again with a single 200 ms step to 0 mV. In a similar set of experiments, Ca²⁺ current amplitude was checked with the same step applied every 1·2 s. No significant difference was found (P = 0·33, n = 5), and therefore it can be concluded that PUFAs can produce their effect even when the channel is totally closed.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

In the present study we investigated the possibility that the L-type synaptic Ca²⁺ conductance of the rod inner segment may be modulated by endogenous retinal factors. Modulation of g_Ca in this way would provide a powerful means of regulating the flow of visual information through the first retinal synapse. A number of neurotransmitters released from neurones intimately associated with the first synapse were investigated as possible modulators of g_Ca but, perhaps surprisingly, none was found to produce detectable changes. Melatonin was also found to be inactive. The possibility that neurotransmitters or extrinsic factors other than those tested could modulate intracellular signalling pathways affecting g_Ca was also investigated by the use of a variety of treatments designed to activate or inhibit G-proteins, PKA, PKC and phosphatases. A small but consistent suppressive effect of forskolin and other PKA activators has been reported before (Stella et al. 1997), and was confirmed in the present study (see Fig. 2B), but we believe this effect to be non-specific in view of its small size and the lack of confirmation from other experiments (e.g. the absence of a significant effect of modulators of G-protein function, of directly infused cAMP, or of activated PKA). Small effects of modulators of PKC activity were also observed (Fig. 2B), but are unlikely to be specific as activators and inhibitors of PKC had similar effects. Phosphatase inhibitors were without effect. Our experiments therefore do not provide support for a modulation of g_Ca by conventional intracellular signalling pathways.

AA and other long-chain, polyunsaturated fatty acids do, however, have a potent effect in suppressing the photoreceptor synaptic g_Ca. The most potent effect was observed with the 22-carbon chain DHA, which inhibits g_Ca with a half-effective concentration of 3·0 µM, though the 20-carbon chain unsaturated acids, such as AA, were only slightly less effective (Fig. 4). Reducing the chain length to 18 carbons or below and reducing the number of double bonds diminished or abolished the inhibitory effect. The compound ETYA, identical to AA but with triple bonds, was without effect. Therefore, as observed for other L-type Ca²⁺ channels (Meves, 1994; Petit-Jaques & Hartzell, 1996; Xiao et al. 1997), the structural requirements necessary to produce an effect seem to be a long-chain hydrocarbon, two or more double bonds, and a charged group at one end. A similar strong inhibition was observed with retinal, retinol and retinoic acid, all of which share these structural features and are produced within the retina by the action of light upon rhodopsin or by subsequent reactions.

Mechanism of inhibition of g_Ca

AA and other related fatty acids are involved in a large number of cellular metabolic and signalling pathways. Apart from acting as the substrate for the cyclo-oxygenase, lipoxygenase and other metabolic pathways, AA is also known to activate PKC directly (Meves, 1994). We investigated whether these or other pathways might underlie the effect on g_Ca, but found no convincing effect of blockers of the cyclo-oxygenase and lipoxygenase pathways, of manipulations activating or inhibiting G-protein-coupled pathways, of modulation of the activity of PKA and PKC, nor of inhibition of the activity of phosphatases 1 and 2A. The fact that PUFAs which are not precursors of active metabolites produce the same inhibitory effect also argues against a role for an intracellular metabolic pathway which uses AA as substrate. One possibility which has not been excluded by the present experiments is that AA and other PUFAs might directly activate a phosphatase such as the AA-dependent phosphatase 5, which was recently isolated by Rossie's group (Skinner et al. 1997). However, we note that oleic acid, which stimulates phosphatase 5, does not inhibit g_Ca (see Fig. 4).

An alternative explanation for the inhibitory action of PUFAs on g_Ca is a direct block of the channel. If the binding site is within the transmembrane electric field, for example if the site is within the pore itself, then the block may be voltage dependent, because the carboxyl terminals of PUFAs are deprotonated at physiological pH and the PUFAs are therefore negatively charged. When the cell membrane potential was depolarized by 50 mV a small but significant increase in the inhibition was indeed observed (Fig. 10), consistent with the idea that depolarization attracts the negatively charged fatty acid into the channel, thereby increasing the effectiveness of the block. The voltage dependence of the block is weak, however, implying that the blocking site is located near the external face of the membrane, and consequently senses only a small fraction of the transmembrane potential. While this model provides a straightforward interpretation of the results, the evidence for it is indirect. Other explanations, such as a direct interaction of PUFAs with the intracellular side, or within the hydrophobic, membrane-embedded part of the channel molecule still remain possible.

Effects of PUFAs on g_Ca in other systems

AA and related PUFAs have been reported to have inhibitory or activatory effects on ion channels in a wide variety of cell types (reviewed by Ordway et al. 1991; Meves, 1994). The mechanisms proposed embrace non-specific actions (e.g. change in membrane fluidity, detergent effects or surface charge screening at high concentrations, etc.), actions on signalling pathways modulating PKA or PKC, actions via oxygen radicals, or direct interactions with the ion channel. AA has been found in many studies to block Ca²⁺ influx, and more rarely to activate it. In most cells investigated more than one type of g_Ca may be present, opening up the possibility that different Ca²⁺ channel types may be modulated by different and possibly conflicting mechanisms. Keyser & Alger (1990) found an effect of AA on L-type Ca²⁺ channels from hippocampal pyramidal neurones which is mediated by PKC and oxygen radicals, but it is worth noting that in their paper the half-maximal AA concentration necessary to produce the effect is 10-fold higher than in the photoreceptor g_Ca, and the effects take much longer times to appear. In frog cardiac myocytes, Petit-Jacques & Hartzell (1996) found, when the L-type Ca²⁺ channel had been phosphorylated by PKA activation, that the inhibition caused by AA was partially relieved by phosphatase inhibitors, and they proposed therefore that part of the inhibition caused by AA was due to enhancement of phosphatase activity. Such a mechanism is not involved in modulating the photoreceptor synaptic Ca²⁺ current, for two reasons: the Ca²⁺ current is not modulated either by phosphorylation or dephosphorylation (Fig. 2), and the inhibition caused by AA was only marginally affected by saturating levels of two phosphatase inhibitors (calyculin A and okadaic acid) and was unaffected by a third inhibitor (microcystin, see Fig. 9). In rat ventricular myocytes Xiao et al. (1997) report a similar voltage dependence of the inhibition by PUFAs to that found in the present study, which they also attribute to a direct interaction of PUFAs with Ca²⁺ channels.

Release of PUFAs and retinoid componds in the retina by light

DHA and AA are the major PUFAs present in photoreceptor membrane phospholipids, at 43·3 mol% and 5·3 mol%, respectively (Dratz & Deese, 1986), and therefore could rapidly be produced following photoreceptor PLA₂ activation. As noted in the Introduction, light activates PLA₂ (Jelsema & Axelrod, 1987), and both AA and DHA have been shown to be released in the retina by light (Jung & Remé, 1994; Reinboth et al. 1996). ATR and its derivatives retinol and retinoic acid are also likely to play an important physiological role. These compounds proved to be potent and rapid inhibitors of g_Ca, with half-maximal effects in the micromolar range (see Fig. 4). ATR, the most potent of the three, is released from rhodopsin after bleaching by light. The rhodopsin concentration in rod outer segments is 3-6 mM (Liebman et al. 1987), and as the outer segments are densely packed and ATR is readily membrane permeable, uniform bright illumination causing a 0·1 % bleach could produce an elevation of its concentration in the regions adjacent to the photoreceptor layer in the micromolar range. A significant light-induced increase of the ATR metabolite, retinoic acid, has been demonstrated recently in the retina (McCaffery et al. 1996). Recent studies have shown that retinoic acid at sub-micromolar concentrations mimics the characteristic effects of light adaptation on horizontal cell synaptic morphology and gap junction permeability, suggesting that retinoic acid plays a physiological role in light adaptation (Weiler et al. 1998, 1999).

Possible role of PUFAs and retinoids in light adaptation

It is well known that light causes adaptation of the phototransduction machinery in the outer segment (see e.g. McNaughton, 1990), but other possible sites of adaptation in the retina are seldom considered. The present study suggests that the gain of synaptic transmission may be reduced in light by the liberation of PUFAs and retinoids, and that the first retinal synapse may therefore also be a site of light adaptation. The gain of synaptic transmission, or the number of millivolts of postsynaptic voltage change per millivolt presynaptic change, depends on the slope of the Ca²⁺ current vs. membrane potential relation (see e.g. Fig. 1). The gain is therefore largest at the photoreceptor resting potential, of around -30 mV, and declines as the rod hyperpolarizes in light (Attwell et al. 1987). Liberation of PUFAs and retinoids by light will cause a further reduction in synaptic gain by partially suppressing the Ca²⁺ current, and the incremental response of second-order neurones to changes in light intensity will therefore be reduced, as predicted by the Weber-Fechner law. In contrast with the adaptation processes occurring within the machinery of phototransduction, which are localized to a particular outer segment, the diffusible nature of PUFAs and retinoids means that synaptic adaptation could spread from the cell in which photon capture occurs. One possible consequence is that cone synaptic transmission could be influenced by PUFAs and retinoids released from the much larger and more numerous rod outer segments, even if the light intensity is in the photopic range and the rods themselves are saturated.

	REFERENCES

Top Abstract Introduction Methods Results Discussion References

Attwell, D., Borges, S., Wu, S. M. & Wilson, M. (1987). Signal clipping by the rod output synapse. Nature 328, 522-524.	[Medline]
Barnes, S., Merchant, V. & Mahmud, F. (1993). Modulation of transmission gain by protons at the photoreceptor output synapse. Proceedings of the National Academy of Sciences of the USA 90, 10081-10085.	[Abstract]
Capdevila, J., Gil, L., Orellana, M., Marnett, L. J., Mason, J. I., Yadagiri, P. & Falck, J. R. (1988). Inhibitors of cytochrome p-450-dependent arachidonic-acid metabolism. Archives of Biochemistry and Biophysics 261, 257-263.	[Medline]
Dratz, E. A. & Deese, A. J. (1986). The role of docosahexaenoic acid (22:6 n-3) in biological membranes: examples from photoreceptors and model membrane bilayers. In Health Effects of Polyunsaturated Fatty Acids in Seafoods, ed. Simopoulos, A. P., Kifer, R. R. & Martin, R. E., pp. 319-351. Academic Press, New York.
Fain, G. L., Matthews, H. R. & Cornwall, M. C. (1996). Dark adaptation in vertebrate photoreceptors. Trends in Neurosciences 19, 502-507.	[Medline]
Hille, B. (1992). Ionic Channels of Excitable Membranes. Sinauer, Sunderland, MA, USA.
Horn, R. & Korn, S. J. (1992). Prevention of rundown in electrophysiological recording. Methods in Enzymology 207, 149-155.	[Medline]
Huang, J. M. C., Xian, H. & Bacaner, M. (1992). Long-chain fatty acids activate calcium channels in ventricular myocytes. Proceedings of the National Academy of Sciences of the USA 89, 6452-6456.	[Abstract]
Jelsema, C. L. & Axelrod, J. (1987). Stimulation of phospholipase A2 activity in bovine rod outer segments by the --subunits of transducin and its inhibition by the -subunit. Journal of Biological Chemistry 262, 163-168.	[Abstract]
Jung, H. H. & Remé, C. E. (1994). Light-evoked arachidonic acid release in the retina: Illuminance/duration dependence and the effects of quinacrine, mellitin and lithium. Graefe's Archives of Clinical and Experimental Ophthalmology 232, 167-175.
Keyser, D. O. & Alger, B. E. (1990). Arachidonic acid modulates hippocampal calcium current via protein kinase C and oxygen radicals. Neuron 5, 545-553.	[Medline]
Korn, S. J. & Horn, R. (1990). Nordihydroguaiaretic acid inhibits voltage-activated Ca2+ currents independently of lipoxygenase inhibition. Molecular Pharmacology 38, 524-530.	[Abstract]
Kurenny, D. E., Thurlow, G. A., Turner, R. W., Moroz, L. L., Sharkey, K. A. & Barnes, S. (1994). Modulation of ion channels in rod photoreceptors by nitric oxide. Neuron 13, 315-324.	[Medline]
Lagnado, L., Cervetto, L. & McNaughton, P. A. (1992). Calcium homeostasis in the outer segments of retinal rods from the tiger salamander. The Journal of Physiology 455, 111-142.	[Abstract]
Lagnado, L. & McNaughton, P. A. (1991). Net charge transport during sodium-dependent calcium extrusion in isolated salamander rod outer segments. Journal of General Physiology 98, 479-495.	[Abstract]
Lasater, E. M. & Witkovsky, P. (1991). The calcium current of turtle cone photoreceptor axon terminals. Neuroscience Research 12, 165-173.
Liebman, P. A., Parker, K. R. & Dratz, E. A. (1987). The molecular mechanism of visual excitation and its relation to the structure and composition of the rod outer segment. Annual Review of Physiology 49, 765-791.	[Medline]
McCaffery, P., Mey, J. & Drager, U. C. (1996). Light-mediated retinoic acid production. Proceedings of the National Academy of Sciences of the USA 93, 12570-12574.	[Abstract/Full Text]
McNaughton, P. A. (1990). Light response of vertebrate photoreceptors. Physiological Reviews 70, 847-883.	[Medline]
Meves, H. (1994). Modulation of ion channels by arachidonic acid. Progress in Neurobiology 43, 175-186.	[Medline]
Nadal, A., Fuentes, E. & McNaughton, P. A. (1996). Albumin stimulates uptake of calcium into subcellular stores in rat cortical astrocytes. The Journal of Physiology 492, 737-750.	[Abstract]
Nakayama, H., Taki, M., Striessnig, J., Glossmann, H., Catterall, W. A. & Kanaoka, Y. (1991). Identification of 1,4-dihydropyridine binding regions within the -1 subunit of skeletal-muscle Ca2+ channels by photoaffinity-labeling with diazipine. Proceedings of the National Academy of Sciences of the USA 88, 9203-9207.	[Abstract]
Oliva, C., Cohen, I. S. & Mathias, R. T. (1988). Calculation of time constants for intracellular diffusion in whole cell patch clamp configuration. Biophysical Journal 54, 791-799.	[Abstract]
Ordway, R. W., Singer, J. J. & Walsh, J. V. (1991). Direct regulation of ion channels by fatty acids. Trends in Neurosciences 14, 96-100.	[Medline]
Petit-Jaques, J. & Hartzell, H. C. (1996). Effect of arachidonic acid on the L-type calcium current in frog cardiac myocytes. The Journal of Physiology 493, 67-81.	[Abstract]
Pourcho, R. G. (1996). Neurotransmitters in the retina. Current Eye Research 15, 797-803.	[Medline]
Rae, J., Cooper, K., Gates, P. & Watsky, M. (1991). Low access resistance perforated patch recordings using amphotericin-B. Journal of Neuroscience Methods 37, 15-26.	[Medline]
Reinboth, J.-J., Clausen, M. & Remé, C. E. (1996). Light elicits the release of docosahexaenoic acid from membrane phospholipids in the rat retina in vitro. Experimental Eye Research 63, 277-284.	[Medline]
Skinner, J., Sinclair, C., Romeo, C., Armstrong, D., Charbonneau, H. & Rossie, S. (1997). Purification of a fatty acid-stimulated protein-serine/threonine phosphatase from bovine brain and its identification as a homolog of protein phosphatase 5. Journal of Biological Chemistry 272, 22464-22471.	[Abstract/Full Text]
Starke, K., Gothert, M. & Kilbinger, H. (1989). Modulation of neurotransmitter release by presynaptic autoreceptors. Physiological Reviews 69, 864-989.	[Medline]
Stella, S. L. Jr, Toews, M. L. & Thoreson, W. B. (1997). Forskolin suppresses L-type calcium currents in vertebrate rods. Society for Neuroscience Abstracts 23, 1817.
Taylor, W. R. & Morgans, C. (1998). Localization and properties of voltage-gated calcium channels in cone photoreceptors of Tupaia belangeri. Visual Neuroscience 15, 541-552.
Weiler, R., He, S. & Vaney, D. I. (1999). Retinoic acid modulates gap junction permeability between horizontal cells of the mammalian retina. European Journal of Neuroscience 11, 3346-3350.	[Medline]
Weiler, R., Schultz, K., Pottek, M., Tieding, S. & JanssenBienhold, U. (1998). Retinoic acid has light-adaptive effects on horizontal cells in the retina. Proceedings of the National Academy of Sciences of the USA 95, 7139-7144.	[Abstract/Full Text]
Witkovsky, P. & Dearry, A. (1992). Functional roles of dopamine in the vertebrate retina. In Progress in Retinal Research, ed. Osborne, N. N. & Chader, G. J., pp. 247-292. Pergamon, Oxford.
Xiao, Y. F., Gomez, A. M., Morgan, J. P., Lederer, W. J. & Leaf, A. (1997). Suppression of voltage-gated L-type Ca2+ currents by polyunsaturated fatty acids in adult and neonatal rat ventricular myocytes. Proceedings of the National Academy of Sciences of the USA 94, 4182-4187.	[Abstract/Full Text]

Acknowledgements

We thank Andrea Moriondo, Paolo Cesare, Jay-Yoon Sul, Alessandro Sardini and Enrico Nasi for helpful discussions, and Sarah L. Fergusson for manuscript editing. V. Vellani was supported by the University of Modena (Italy), by EMBO (Fellowship ASTF 8563) and by the European Commission (TMR Fellowship ERBFMBICT972456).

Corresponding author's present address

P. A. McNaughton: Department of Pharmacology, University of Cambridge, Tennis Court Road, Cambridge CB2 1QJ, UK.

Email: pam42{at}cam.ac.uk

Author's present address

V. Vellani: Department of Pharmacology, University of Cambridge, Tennis Court Road, Cambridge CB2 1QJ, UK.

This article has been cited by other articles:

				A. M. Cameron, O. A. R. Mahroo, and T. D. Lamb Dark adaptation of human rod bipolar cells measured from the b-wave of the scotopic electroretinogram J. Physiol., September 1, 2006; 575(2): 507 - 526. [Abstract] [Full Text] [PDF]

				A. Kasus-Jacobi, J. Ou, D. G. Birch, K. G. Locke, J. M. Shelton, J. A. Richardson, A. J. Murphy, D. M. Valenzuela, G. D. Yancopoulos, and A. O. Edwards Functional Characterization of Mouse RDH11 as a Retinol Dehydrogenase Involved in Dark Adaptation in Vivo J. Biol. Chem., May 27, 2005; 280(21): 20413 - 20420. [Abstract] [Full Text] [PDF]

				S. L. McCabe, D. M. Pelosi, M. Tetreault, A. Miri, W. Nguitragool, P. Kovithvathanaphong, R. Mahajan, and A. L. Zimmerman All-trans-retinal Is a Closed-state Inhibitor of Rod Cyclic Nucleotide-gated Ion Channels J. Gen. Physiol., April 26, 2004; 123(5): 521 - 531. [Abstract] [Full Text] [PDF]

				E. A. Schwartz Transport-Mediated Synapses in the Retina Physiol Rev, October 1, 2002; 82(4): 875 - 891. [Abstract] [Full Text] [PDF]

				D. M. Dean, W. Nguitragool, A. Miri, S. L. McCabe, and A. L. Zimmerman All-trans-retinal shuts down rod cyclic nucleotide-gated ion channels: A novel role for photoreceptor retinoids in the response to bright light? PNAS, June 11, 2002; 99(12): 8372 - 8377. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Vellani, V. Articles by McNaughton, P. A. Search for Related Content PubMed PubMed Citation Articles by Vellani, V. Articles by McNaughton, P. A.