Measurement of the intracellular free calcium concentration in salamander rods

Measurement of the free calcium concentration within a photo-receptor outer segment has been considered an important aim since the proposal by Hagins and Yoshikami that the primary event in phototransduction is a release of Ca (2+) inside the cell. More recent evidence has cast doubt on the calcium hypothesis, and the observations of Yau and Nakatani and Matthews et al. suggest that the internal Ca (2+) concentration ([Ca (2+)]i), may decrease after a flash of light. In the present study we have measured [Ca (2+)]i directly by using a new method for incorporating the Ca-sensitive photoprotein aequorin into an isolated rod. We report that the light response is accompanied by a decrease in [Ca (2+)]i, caused by the closure of light-sensitive channels which are the main route for Ca (2+) entry into the outer segment. Of the Ca (2+) entering through light-sensitive channels, about 95% is sequestered by a rapid and reversible buffering mechanism. Calcium is removed from the cell by an electrogenic pump in which 3 Na (+) ions are exchanged for each Ca (2+); the pump is highly active and the free Ca (2+) in the cell declines with a time constant of ~0.5 s after a flash of light.     

Light-suppressible, cyclic GMP-sensitive conductance in the plasma membrane of a truncated rod outer segment

Recent experiments by Fesenko et al and ourselves have shown that excised membrane patches from retinal rod outer segments contain a cyclic GMP-sensitive conductance which has electrical properties similar to those of the light-sensitive conductance. This finding supports the notion that cGMP mediates phototransduction (see ref. 3) by directly modulating the light-sensitive conductance. However, some uncertainty remained about whether the patch experiments had discriminated completely between plasma and intracellular disk membranes; thus the cGMP response in an excised membrane could have resulted from contaminating disk membrane fragments, which are known to contain a cGMP-regulated conductance. Furthermore, the patch conductance has not yet been shown to be light-suppressible, an ultimate criterion for identity with the light-sensitive conductance. We now report experiments on a truncated rod outer segment preparation which resolved these issues. The results demonstrated that the cGMP-sensitive conductance was present in the plasma membrane of the outer segment, and that in the presence of GTP the conductance could be suppressed by a light flash. With added ATP, the effectiveness of the light flash was reduced and the suppression was more transient. The effects of both GTP and ATP were consistent with the known biochemistry. From the maximum current inducible by cGMP, we estimate that approximately 1% of the light-sensitive conductance is normally open in the dark; this would give an effective free cGMP concentration of a few micromolar in the intact outer segment in the dark.     

Electrogenic Na-Ca exchange in retinal rod outer segment

Previous work has suggested that a Na-Ca exchanger may have a key role in visual transduction in retinal rods. This exchanger is thought to maintain a low internal free Ca2+ concentration in darkness and to contribute to the rod's recovery after light by removing any internally released Ca2+. Little else is known about this transport mechanism in rods. We describe here an inward membrane current recorded from single isolated rods which appears to be associated with such external Na+-dependent Ca2+ efflux activity. External Na+, but not Li+, could generate this current; high external K+ inhibited it while small amounts of La3+ (10 microM) completely abolished it. The exchanger can also transport Sr2+, but not Ba2+ or other divalent cations. The exchange ratio was estimated to be 3Na+:1Ca2+. As well as demonstrating clearly the Na-Ca exchanger in the rod outer segment, our experiments also cast serious doubt on the commonly held view that light simply releases internal Ca2+ to bind to and block the light-sensitive conductance.     

Cyclic GMP can increase rod outer-segment light-sensitive current 10-fold without delay of excitation

To test the hypothesis that cyclic GMP is the internal messenger coupling rhodopsin activation to membrane excitation in vertebrate rod photoreceptors, we used a novel technique combining measurement of membrane currents of isolated salamander rods with a suction electrode and the introduction of cyclic GMP through a whole-cell recording patch pipette. Rupture of an attached patch was followed by a rapid (approximately 10 s), approximately 10-fold increase in outer-segment membrane current, all of which was light-sensitive. There was little change in the rising phase of the response to a saturating flash, but the duration of the saturated phase of the response increased approximately 10-fold. The effects reversed completely within 3-4 min after withdrawal of the cyclic GMP-containing patch pipette. A formal kinetic analysis shows that the first two observations are inconsistent with the postulate that cyclic GMP opens the light-sensitive conductance by simple binding to channels, unless free cyclic GMP in the outer segment is assumed to be much lower than published estimates, and most of the outer-segment cyclic GMP is bound and inexchangeable on the timescale of 200 ms. Furthermore, our results suggest that rod cyclic GMP is not involved solely in keeping the light-sensitive conductance open, but may also affect the activity of the phosphodiesterase that mediates cyclic GMP hydrolysis.     

Cyclic GMP-sensitive conductance in outer segment membrane of catfish cones

A cyclic GMP-sensitive conductance has recently been observed with patch-clamp recording in excised inside-out patches of plasma membrane from frog and toad rod outer segments. This conductance has properties suggesting that it is probably the light-sensitive conductance involved in visual transduction. We now report a similar conductance in the outer segment membrane of catfish cones. Cyclic GMP showed positive cooperativity in opening this conductance, with a Hill coefficient of 1.6-3.0 and a half-saturating cGMP concentration of 35-70 microM. Cyclic AMP at 1 mM, or changing Ca concentration (in the presence of Mg), had little effect on the conductance. In physiological solutions the cGMP-induced current had a reversal potential near +10 mV; the current amplitude increased roughly exponentially with membrane potential in both depolarizing and hyperpolarizing directions. Our results suggest that cGMP is also the internal transmitter for phototransduction in cones.     

Induction by cyclic GMP of cationic conductance in plasma membrane of retinal rod outer segment

Vertebrate rod photoreceptors hyperpolarize when illuminated, due to the closing of cation-selective channels in the plasma membrane. The mechanism controlling the opening and closing of these channels is still unclear, however. Both 3',5'-cyclic GMP and Ca2+ ions have been proposed as intracellular messengers for coupling the light activation of the photopigment rhodopsin to channel activity and thus modulating light-sensitive conductance. We have now studied the effects of possible conductance modulators on excised 'inside-out' patches from the plasma membrane of the rod outer segment (ROS), and have found that cyclic GMP acting from the inner side of the membrane markedly increases the cationic conductance of such patches (EC50 30 microM cyclic GMP) in a reversible manner, while Ca2+ is ineffective. The cyclic GMP-induced conductance increase occurs in the absence of nucleoside triphosphates and, hence, is not mediated by protein phosphorylation, but seems rather to result from a direct action of cyclic GMP on the membrane. The effect of cyclic GMP is highly specific; cyclic AMP and 2',3'-cyclic GMP are completely ineffective when applied in millimolar concentrations. We were unable to recognize discrete current steps that might represent single-channel openings and closings modulated by cyclic GMP. Analysis of membrane current noise shows the elementary event to be 3 fA with 110 mM Na+ on both sides of the membrane at a membrane potential of -30 mV. If the initial event is assumed to be the closure of a single cyclic GMP-sensitive channel, this value corresponds to a single-channel conductance of 100 fS. It seems probable that the cyclic GMP-sensitive conductance is responsible for the generation of the rod photoresponse in vivo.     

Control of Ca2+ in rod outer segment disks by light and cyclic GMP

Photons absorbed in vertebrate rods and cones probably cause electrochemical changes at the photoreceptor plasma membrane by changing the cytoplasmic concentration of a diffusible transmitter substance, reducing the Na+ current flowing into the outer segment of the cell in the dark, to produce the observed membrane hyperpolarization that is the initial excitatory response. Cyclic GMP has been proposed as the transmitter because a light-activated cyclic GMP phosphodiesterase (PDE) has been found in rod disk membranes and because intracellularly injected cyclic GMP reduces rod membrane potentials. Free Ca2+ has also been proposed because increasing external [Ca2+] quickly and reversibly reduces the dark current and divalent cationophores increase the Ca2+ sensitivity. Ca2+ efflux from rod outer segments (ROS) of intact retinas occurs simultaneously with light responses. Vesicles prepared from ROS disk membranes become more permeable on illumination, releasing trapped ions or molecules, but intact outer segment disks have not previously been found to store sufficient Ca2+ in darkness and to release enough in light to meet the theoretical requirements for control of the dark current by varying cytoplasmic Ca2+ (refs 14-18). We now report experiments that show the required Ca2+ storage and release from rod disk membranes suspended in media containing high-energy phosphate esters and electrolytes approximating the cytoplasmic composition of live rod cells. Cyclic GMP stimulates Ca2+ uptake by ROS disks in such media.     

Single cyclic GMP-activated channel activity in excised patches of rod outer segment membrane

The plasma membrane of retinal rod outer segments contains a cyclic GMP-activated conductance which appears to be the light-sensitive conductance involved in phototransduction. Recently, it has been found that this conductance is partially blocked by Mg2+ and Ca2+ at physiological concentrations, thus possibly accounting for the absence of observable single-channel activity in excised membrane patches and for the unusually small apparent unit conductance deduced from noise measurements on intact cells. We now report that, as expected from this idea, single cGMP-activated channel activity can be detected from an excised rod membrane patch in the absence of divalent cations. The most prominent unitary current had a mean conductance of approximately 25 pS. Both individual channel openings (mean open time approximately 1 ms) and short bursts of openings (mean burst duration of about a few milliseconds) were observed. In addition, there were smaller events which probably represented other states of the conductance. The mean current increased with the third power of cGMP concentration, suggesting that there are at least three cGMP-binding sites on the channel molecule. With 0.2 mM Mg2+ in the cGMP-containing solution, a flickering block of the open channel was observed; the effect of Ca2+ was similar. The results resolve a puzzle about the light-sensitive conductance by demonstrating that it is an aqueous pore rather than a carrier.     

Effect of ions on the light-sensitive current in retinal rods

The effect of ions on the light-sensitive current of retinal rods was studied by sucking the inner segment into a tightly fitting capillary with the outer segment projecting into a flowing solution. This new method showed that the light-sensitive pathway, in which Na+ is the normal carrier of current, has an ionic selectivity different from that of other known sodium channels. Externàl calcium has a striking effect on the current, which increased about 20-fold when all calcium was removed. Reducing the sodium concentration gradient greatly prolonged the response to a flash of light, as would be expected if internal calcium blocks sodium channels and if light releases calcium which is subsequently extruded by a sodium-calcium exchange mechanism.     

Effects on the photoresponse of calcium buffers and cyclic GMP incorporated into the cytoplasm of retinal rods

It is generally accepted that the light response in retinal rods involves a reduction of ionic permeability (predominantly to Na+) in the plasma membrane of the outer segment and that this is mediated by an internal messenger which diffuses between the disk and plasma membranes. There is controversy, however, over the identity of the diffusible substance; two alternative schemes have received widespread support (for review see refs 1,2). According to the 'calcium hypothesis', light stimulates the release into the cytoplasm of calcium, leading to the blockage of channels which are normally open in darkness, whereas based on the 'cyclic nucleotide hypothesis', cyclic GMP causes the opening of channels in the dark, but is hydrolysed by a light-activated phosphodiesterase. We report here effects of introducing calcium buffers and cyclic GMP into the rod cytoplasm by means of a patch pipette, which seem to be inconsistent with the calcium hypothesis.     

Cyclic GMP-sensitive conductance of retinal rods consists of aqueous pores

The surface membrane of retinal rod and cone outer segments contains a cation-selective conductance which is activated by 3',5'-cyclic guanosine monophosphate (cGMP). Reduction of this conductance by a light-induced decrease in the cytoplasmic concentration of cGMP appears to generate the electrical response to light, but little is known about the molecular nature of the conductance. The estimated unitary conductance is so small that ion transport might occur via either a carrier or a pore mechanism. Here we report recordings of cGMP-activated single-channel currents from excised rod outer segment patches bathed in solutions low in divalent cations. Two elementary conductances, of approximately 24 and 8 pS, were observed. These conductances are too large to be accounted for by carrier transport, indicating that the cGMP-activated conductance consists of aqueous pores. The dependence of the channel activation on the concentration of cGMP suggests that opening of the pore is triggered by cooperative binding of at least three cGMP molecules.     

Extrusion of calcium from rod outer segments is driven by both sodium and potassium gradients

Calcium is transported across the surface membrane of both nerve and muscle by a Na+-dependent mechanism, usually termed the Na:Ca exchange. It is well established from experiments on rod outer segments that one net positive charge enters the cell for every Ca2+ ion extruded by the exchange, which is generally interpreted to imply an exchange stoichiometry of 3 Na+:1 Ca2+. We have measured the currents associated with the operation of the exchange in both forward and reversed modes in isolated rod outer segments and we find that the reversed mode, in which Ca2+ enters the cell in exchange for Na+, depends strongly on the presence of external K+. The ability of changes in external K+ concentration ([K+]o) to perturb the equilibrium level of [Ca2+]i indicates that K+ is co-transported with calcium. From an examination of the relative changes of [Ca2+]o, [Na+]o, [K+]o and membrane potential required to maintain the exchange at equilibrium, we conclude that the exchange stoichiometry is 4 Na+:1 Ca2+, 1 K+ and we propose that the exchange should be renamed the Na:Ca, K exchange. Harnessing the outward K+ gradient should allow the exchange to maintain a Ca2+ efflux down to levels of internal [Ca2+] that are considerably lower than would be possible with a 3 Na+:1 Ca2+ exchange.     

Metarhodopsin I/metarhodopsin II transition triggers light-induced change in calcium binding at rod disk membranes

The hypothesis of Yoshikami and Hagins that calcium ions act as diffusible transmitter molecules between the photochemistry of rhodopsin and the subsequent electrical events at the outer plasma membrane of rods initiated many investigations on light-stimulated calcium release in vertebrate photoreceptor cells (see refs 2, 3). Although it not seems firmly established that light has some effect on the redistribution of calcium in various disk preparations, reconstituted systems and intact rod outer segments, the physiological significance remained unclear. We previously reported a rapid, light-triggered calcium release from binding sites at the disk membrane in the presence of calcium ionophore A23187 (refs 3, 8). However, there is no evidence for rapid calcium release into the cytosol in the absence of ionophore. On fragmentation of intact rod outer segments, calcium release due to a light-requlated change of calcium binding appeared almost completely abolished. We describe here experiments with sonicated rod outer segments in which the previously observed loss of the calcium release capacity has been prevented. Calcium release in sonicated disks in the presence of A23187 kinetically follows the metarhodopsin I/metarhodopsin II transition (tau 1/2 = 10 ms, activation energy EA = 34 kcal mol-1), suggesting that calcium release is triggered by this photochemical transition.     

Highly cooperative feedback control of retinal rod guanylate cyclase by calcium ions

Visual excitation in retinal rod cells is mediated by a cascade that leads to the amplified hydrolysis of cyclic GMP (cGMP) and the consequent closure of cGMP-activated cation-specific channels in the plasma membrane. Recovery of the dark state requires the resynthesis of cGMP, which is catalysed by guanylate cyclase, an axoneme-associated enzyme. The lowering of the cytosolic calcium concentration (Cai) following illumination is thought to be important in stimulating cyclase activity. This hypothesis is supported by the finding that the cGMP content of rod outer segments increases several-fold when Cai is lowered to less than 10 nM. It is evident that cGMP and Cai levels are reciprocally controlled by negative feedback. Guanylate cyclase from toad ROS is strongly stimulated when the calcium level is lowered from 10 microM to 10 nM, but only if they are excited by light. We show here that the guanylate cyclase activity of unilluminated bovine rod outer segments increases markedly (5 to 20-fold) when the calcium level is lowered from 200 nM to 50 nM. This steep dependence of guanylate cyclase activity on the calcium level in the physiological range has a Hill coefficient of 3.9. Stimulation at low calcium levels is mediated by a protein that can be released from the outer segment membranes by washing with a low salt buffer. Calcium sensitivity is partially restored by adding the soluble extract back to the washed membranes. The highly cooperative activation of guanylate cyclase by the light-induced lowering of Cai is likely to be a key event in restoring the dark current after excitation.     

Is hyperosmotic neurosecretion from motor nerve endings a calcium-dependent process?

Spontaneous liberation of neurotransmitter quanta is strongly affected by the osmotic pressure of the extracellular fluid. Elevation of the osmolarity by 20-30% increases the rate of release from motor nerve endings by more than one order of magnitude. In this respect the neuromuscular junction resembles some other secretory systems. The mechanism of this hyperosmotic neurosecretion is not yet understood; extracellular calcium ions are not directly responsible, since this effect can be produced in their absence. Recently, it has been suggested that the liberation of neurotransmitter is regulated by the intracellular concentration of free calcium ions. We have therefore examined the hypothesis that hyperosmotic neurosecretion originates from an increase in internal calcium concentration ([Ca]in). At the frog neuromuscular synapse however, it is impossible at present to estimate directly free [Ca]in; hence we used an indirect technique, which is based on two assumptions; first, the frequency of the miniature endplate potentials (m.e.p.p.s.) reflects free [Ca]in. Second, the movement of calcium ions across the presynaptic membrane is governed by the electrochemical gradient, and by the calcium conductance (g(Ca)). If hyperosmotic neurosecretion is caused by an increase in [Ca]in, then increasing g(Ca), under reversed electrochemical gradient for the calcium should cause a reduction in the effect of hyperosmotic stress on transmitter release. We report that hyperosmotic neurosecretion is dependent on [Ca]in.     

cGMP-dependent cation channel of retinal rod outer segments

Light-modulated cytoplasmic cGMP simultaneously controls plasma membrane Na+ conductance in visual excitation and Ca2+ entry into rods by direct interaction with the cation channel. Cytoplasmic Ca2+ in turn may set operating points and contribute to the dynamics of several enzymes that regulate cGMP levels in the dark, recovery from excitation and receptor adaptation or down regulation. Similar channels may couple electrical activity to internal nucleotide metabolism in other tissues. We here report the identification, partial purification and behaviour after reconstitution of a protein of relative molecular mass 39,000 (Mr 39K) present in both disk and plasma membranes from bovine rod outer segments that mediates these cGMP-dependent cation fluxes. Its cGMP agonist specificity, kinetic cooperativity, ionic selectivity, membrane density and other features closely match the properties of the visual cGMP-dependent conductance inferred from electrophysiological measurements.     

Action potentials must admit calcium to evoke transmitter release.

There are two hypotheses to explain how neurons release transmitter. The calcium hypothesis proposes that membrane depolarization is necessary only for opening calcium channels and increasing internal calcium concentration ([Ca2+]i) near membrane transmitter-release sites. These calcium ions trigger a transient release of neurotransmitter. The calcium-voltage hypothesis postulates that voltage induces a conformational change in a membrane protein rendering it sensitive to calcium such that, in the presence of high [Ca2+]i, depolarization directly triggers transmitter release. Here we report that when calcium influx is blocked by cobalt or manganese ions in a calcium-free Ringer, as measured with Fura-2, and [Ca2+]i is elevated by liberation from a caged calcium compound, transmitter release at the crayfish neuromuscular junction is unaffected by presynaptic action potentials. These results support the calcium hypothesis.     

Membrane depolarization evokes neurotransmitter release in the absence of calcium entry

The discovery that Ca2+ is necessary for the release of neurotransmitter, the primary means by which nerve cells communicate, led to the calcium hypothesis of neutransmitter release, in which release is initiated after an action potential only by an increase in intracellular Ca2+ concentration near the release sites and is terminated (1-2 ms) by the rapid removal of Ca2+. Since then, the calcium-voltage hypothesis has been proposed, in which the depolarization of the presynaptic terminals has two functions. First, in common with the calcium hypothesis, the Ca2+ conductance is increased, thereby permitting Ca2+ entry. Second, a conformational change is induced in a membrane molecule that renders it sensitive to Ca2+, and then binding of Ca2+ to this active form triggers release of neurotransmitter. When the membrane is repolarized, the molecule is inactivated and release is terminated, regardless of the local Ca2+ concentration at that moment. This hypothesis, in contrast to the calcium hypothesis, accounts for the insensitivity of the time course of release to experimental manipulations of intracellular Ca2+ concentration. Furthermore, it explains rapid termination of release after depolarization, even though Ca2+ concentration may still be high. Here we describe experiments that distinguish between these two hypotheses and find that our results support the calcium voltage hypothesis.     

Circadian rhythm and light regulate opsin mRNA in rod photoreceptors

Disk membranes in the outer segment of rod photoreceptors are continuously renewed, being assembled at the outer segment base, displaced outward by new disks and eventually shed at the tip. In lower vertebrates, disk assembly occurs with a diurnal rhythm with 2-4% of the outer segment length produced daily. We have discovered that in toad and fish retinas the level of mRNA for opsin, the most abundant protein in rod disks, fluctuates with a daily rhythm and is regulated both by light and by a circadian oscillator. The mRNA level rises before light onset, remains high during the light phase of a diurnal cycle and decreases four to tenfold during the dark phase. In constant darkness, mRNA elevation occurs during subjective daytime. At night, rod opsin mRNA can be elevated by exposure to light.     

Cyclic GMP is involved in the excitation of invertebrate photoreceptors

The hyperpolarizing receptor potential in vertebrate rod photoreceptors appears to be mediated by the second messenger, cyclic GMP. Injection of cGMP into rods or application of cGMP to inside-out membrane patches activates a conductance resembling that produced by light. Light produces a rapid reduction of cGMP in living rods, leading to closure of sodium channels and membrane hyperpolarization. In most invertebrate photoreceptors the response to light is depolarizing. We have investigated whether cGMP is involved in controlling the increase in sodium conductance that underlies this depolarization. We show here that injection of cGMP into Limulus photoreceptors produces a depolarization that mimics the receptor potential. We also show that the cGMP concentration of the squid retina increases rapidly during exposure to light. These results support the hypothesis that cGMP mediates the light-induced depolarization in invertebrate photoreceptors and suggests that vertebrate and invertebrate phototransduction may be more similar than previously thought.