THE DYNAMIC RANGE AND DOMAIN-SPECIFIC SIGNALS OF INTRACELLULAR CALCIUM IN PHOTORECEPTORS

Abstract

Vertebrate photoreceptors consist of strictly delimited subcellular domains: the outer segment, ellipsoid, cell body and synaptic terminal, each hosting crucial cellular functions, including phototransduction, oxidative metabolism, gene expression and transmitter release. We used optical imaging to explore the spatiotemporal dynamics of Ca²⁺ signaling in non-outer segment regions of rods and cones. Sustained depolarization, designed to emulate photoreceptor activation in the darkness, evoked a standing Ca²⁺ gradient in tiger salamander photoreceptors with spatially-averaged intracellular Ca²⁺ concentration within synaptic terminals of ∼2 μM and lower (∼750 nM) intracellular calcium concentration in the ellipsoid. Measurements from axotomized cell bodies and isolated ellipsoids showed that Ca²⁺ enters the two compartments via both local L-type Ca²⁺ channels and diffusion. The results from optical imaging studies were supported by immunostaining analysis. L-type voltage-operated Ca²⁺ channels and plasma membrane Ca²⁺ ATPases were highly expressed in synaptic terminals with progressively lower expression levels in the cell body and ellipsoid. These results show photoreceptor Ca²⁺ homeostasis is controlled in a region-specific manner by direct Ca²⁺ entry and diffusion as well as Ca²⁺ extrusion. Moreover, quantitative measurement of intracellular calcium concentration levels in different photoreceptor compartments indicates that the dynamic range of Ca²⁺ signaling in photoreceptors is approximately 40-fold, from ∼50 nM in the light to ∼2 μM in darkness.

The calcium ion is the most versatile known intracellular messenger, regulating almost all known cellular functions and processes (Berridge et al., 2003; Petersen et al., 2005). Changes in the intracellular calcium concentration, [Ca²⁺]i, regulate cellular processes by activating diverse signaling pathways and organelles and by coordinating the amplitude and timing of Ca²⁺-dependent processes across the cell. The distribution and abundance of Ca²⁺ channels, transporters and effector proteins create local and/or global changes in [Ca²⁺]i (Berridge et al., 2003; Delmas et al., 2004). Such domain-specific action of Ca²⁺ is especially evident in primary sensory neurons, highly polarized cells formed by several anatomical domains with distinct functions (Zufall et al., 2000; Dumont et al., 2001; Akopian and Witkovsky, 2002; Križaj and Copenhagen, 2002).

Rod and cone photoreceptors are highly compartmentalized neurons. For example, phototransduction is limited to the outer segment (OS), aerobic metabolism mainly to the ellipsoid, gene expression and protein synthesis to the cell body, and exocytosis to the synaptic terminal. Apoptosis, a Ca²⁺-dependent process that underlies most forms of photoreceptor degeneration, is initiated and develops in both the ellipsoid and the cell body (Doonan et al., 2003; Chiarini et al., 2003; Rohrer et al., 2004). All of these processes are regulated by light-evoked fluctuations in [Ca²⁺]i (Fain et al., 2001; Križaj and Copenhagen, 2002). However, it is not understood how [Ca²⁺]i changes in different regions of the photoreceptor during stimulation, nor is it known whether and how different regions communicate via local Ca²⁺ fluxes. Moreover, there are significant inconsistencies in the literature regarding absolute [Ca²⁺]i levels, the dynamic range and the role for Ca²⁺ diffusion in rods and cones.

A key question in photoreceptor signaling is related to understanding the mechanisms that compress the 4–5 log unit dynamic range of phototransduction in the OS to the approximately 25-fold range for exocytosis in the photoreceptor terminal (Choi et al., 2005). Presynaptic [Ca²⁺]i represents the final common path for this signal compression. A recent study, focusing on Ca²⁺ regulation in synaptic terminals of salamander rods and cones, suggested that average [Ca²⁺]i in the terminal ranges from ∼350 nM in hyperpolarized light-adapted cells to more than 39 μM in cells depolarized to dark potentials (Steele et al., 2005). Other studies, however, proposed that this dynamic range of average presynaptic [Ca²⁺]i in photoreceptors extends from tens of nanomoles to submicromolar or low micromolar (Rieke and Schwartz, 1996; Thoreson et al., 2004). This discrepancy has significant implications for our understanding of regulation of exocytosis at ribbon synapses in darkness and light (e.g. Rieke and Schwartz, 1996; Thoreson et al., 2004 but see Kreft et al., 2003; Heidelberger et al., 1994).

Studies also diverge on the mechanisms of Ca²⁺ entry into photoreceptor cell body and ellipsoid regions. It has been suggested that at physiological ‘dark’ membrane potentials [Ca²⁺]i in cell bodies and ellipsoids of rods is close to baseline values characteristic of light-adapted cells (Steele et al., 2005). Consistent with this, immunolocalization studies report that Kv1.4 Ca²⁺ channels are expressed exclusively in presynaptic terminals (Morgans et al., 1998; Morgans, 2001). On the other hand, other studies found that cell bodies of photoreceptors are also immunoreactive for subunits of voltage-activated Ca²⁺ channels (Nachman-Clewner et al., 1999; Morgans, 2001; Steele et al., 2005), suggestive of direct Ca²⁺ entry into these compartments.

To resolve these conflicting reports on Ca²⁺ dynamics in photoreceptor regions, we studied depolarization-evoked Ca²⁺ signals in post-OS domains of rods and cones. Our results show that voltage-gated Ca²⁺ influx at membrane potentials that mimic physiological depolarizations elevates [Ca²⁺]i from ∼50–100 nM to ∼2 μM in terminals. Our results also suggest that Ca²⁺ enters the ellipsoid and the cell body regions directly through local Ca²⁺ channels, with an additional potential component from diffusion. Finally, we find that the spatiotemporal [Ca²⁺]i gradient is correlated with the distribution of voltage-operated Ca²⁺ channels and plasma membrane calcium ATPases (PMCAs), which control the dynamic range of cytosolic [Ca²⁺]i in a region-specific manner. Our data suggest that interactions between plasma membrane Ca²⁺ entry and clearance mechanisms compress the four-five-fold range of photon-evoked responses in the OS into a 20-fold dynamic range of presynaptic [Ca²⁺]i.

EXPERIMENTAL PROCEDURES

Preparation of isolated cells

Care was taken to minimize the number of animals used and their suffering. Larval stage tiger salamanders (Ambystoma tigrinum) were decapitated and pithed using procedures approved by the UCSF Committee for Animal Care and in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Dissociated photoreceptors were prepared following protocols described previously (Križaj and Copenhagen, 1998). Briefly, retinas were dissected from enucleated eyes, and dissociated in 0 Ca²⁺/papain (10 U/ml; Worthington, Freehold, NJ, USA) saline for 20 min at room temperature (20–22 °C). The retina was cut into small pieces using a razorblade; each piece was dissociated separately immediately prior to the experiment. Cells were plated onto coverslips coated with 0.2 mg/ml concanavalin A (Sigma, St. Louis, MO, USA). Although the salamander retina contains at least four classes of cones (Mariani, 1986), all cones studied here responded to pharmacological manipulations in an identical manner; the data obtained from different cone classes have therefore been tabulated together. The viability of isolated cone ellipsoids was tested by 30 s incubation with 0.01% Toluidine Blue. The dye was excluded from isolated ellipsoids, suggesting that these subregions reseal and form viable compartments following dissociation. The recording chamber was superfused via two electronically controlled multi-inlet manifolds (MP-8, Warner Instruments, Hamden, CT, USA). The control saline solution contained, in mM, 97 NaCl, 2 KCl, 2 CaCl₂, 2 MgCl₂, 10 HEPES, 2 lactic acid, 0.3 ascorbic acid and 1 taurine at 240 mOsm. The pH was adjusted to 7.6 with NaOH. To stimulate glycolysis, the glucose concentration in all saline solutions was 10 mM (e.g. Winkler, 1983). All chemicals were obtained from Sigma.

[Ca²⁺]i measurements

These methods are fully described elsewhere (Križaj and Copenhagen, 1998). Briefly, photoreceptors were loaded with 2–5 μM fura-2 AM (fura 2-acetoxymethylester; Molecular Probes, Eugene, OR, USA) for 10 min and subsequently washed for 20 min. The fluorescence signals were acquired on an inverted microscope (Nikon Eclipse 200) using a dry 40× objective (N.A.=0.8) or an oil 100× objective (N.A.=1.2). Image acquisition was generally binned at 3×3 pixels (except for Fig. 2 where the binning was 2×2) and was run at 0.5–1 Hz by a cooled 12 bit digital CCD camera (Cascade, Photometrics, Tucson, AZ, USA). The camera and the shutter (Lambda DG-4, Sutter Instruments, Novato, CA, USA) were controlled by commercial software (Metafluor 6.1; Universal Imaging, West Chester, PA, USA). Separate regions of interest (ROI) were defined that encompassed the synaptic terminal, cell body and ellipsoid regions. ROI size was ∼4 μm² for the terminal, 20 μm² for the cell body, and 10 μm² for the ellipsoid. The proximity of the cell body and terminal in cones generally precluded accurate estimation of cone terminal [Ca²⁺]i, therefore the average [Ca²⁺]i elevations were measured in rod terminals with axons >5 μm in length. Emission ratios between the 340 nm and 380 nm excitation wavelengths were calculated after subtraction of the background fluorescence.

Fig. 2.

Spatiotemporal [Ca²⁺]i differences in IS subregions in response to depolarization. (A) Rod photoreceptor superfused for 20 s with 30 mM KCl. Fast kinetics, high amplitude and spatially localized depolarization-evoked [Ca²⁺]_i are observed in the synaptic terminal. The free [Ca²⁺]i signal amplitude is sharply delimited between the terminal, the cell body, the subellipsoid and the ellipsoid regions. The gain of the synaptic terminal region in the image was increased relative to the rest of the cell (as denoted by the yellow line) to enhance weak fura-2 fluorescence within the terminal. (B) Detail of [Ca²⁺]_i elevation in a rod terminal triggered by transient depolarization. High-amplitude hotspot [Ca²⁺] increases are observed during depolarization (arrowheads).

Free [Ca²⁺] levels were calibrated in vivo with 10 μM ionomycin in 0 and 10 mM [Ca²⁺]o saline using the standard relationship (Grynkiewicz et al., 1985), and the K_d for Ca²⁺ binding to fura-2 was taken to be 224 nM (Almers and Neher, 1985; Neher, 1995). It is possible that the intracellular environment caused small changes in Ca²⁺ binding to fura-2, which would be expected to change errors in estimation of the K_d parameter (Grynkiewicz et al., 1985). To test for potential compartmentalization of fura-2 dye within the mitochondria and other intracellular organelles, fura-2-loaded cells were depolarized with 20 mM KCl together with 2 mM Mn²⁺. Greater than 95% of the ellipsoid fura-2 signal was quenched by exposure to Mn²⁺, indicating that most of the dye was cytoplasmic (data not shown). The fura-2 dye, with a K_d of 224 nM, is ∼90% saturated at ∼2 μM [Ca²⁺]i (Augustine and Neher, 1992; Zhou and Neher, 1993), possibly resulting in an underestimation of [Ca²⁺] at >2.5 μM. We were occasionally unable to calibrate [Ca²⁺]i levels; therefore, data for some cells are presented as 340/380 nm ratios. The decay time constants were calculated by fitting single exponentials to the decrease in [Ca²⁺]i following exposure to KCl (Križaj and Copenhagen, 1998). [Ca²⁺]i rise times, which equal the rate of entry minus the rate of clearance, were estimated as the slope of the linear rising phase from the onset to peak of the [Ca²⁺]i elevation (Igor Pro, Wave-metrics, Lake Oswego, OR, USA). All pooled data are presented as mean±S.E.M. Significance was determined using the t-test.

For live staining, cells were labeled for 15 min in 0.5–1 μM solution of MitoTracker Red or 200 nM MitoTracker Green (Molecular Probes). MitoTracker Green contains a thiol-reactive chloromethyl moiety that results in stable peptide conjugates after accumulation in the mitochondria. The uptake of this probe is independent of the mitochondrial membrane potential. TMRM (tetramethylrhodamine), a cationic rhodamine-based dye that reversibly accumulates in the highly negatively charged mitochondrial matrix and fluoresces at negative membrane potentials, was used at 100 nM; dye loading was 30 min. Stock solutions of MitoTracker and TMRM dyes were made in DMSO. TMRM fluorescence was analyzed using the linescan function of the confocal microscope at the 543 nm line (1%; LSM5 Pa, Zeiss, Tarrytown, NY, USA).

Immunostaining

Immunostaining procedures were performed as described before (Križaj et al., 2002). Retinal sections on slides were washed in PB for 15 min before permeabilization and blocking in 0.5% Triton X-100 and 10% goat serum. The antibodies used were raised against the α1 subunit of voltage-gated Ca²⁺ channels (Alomone ACC-004), Chemicon (MAB427); for PMCA labeling, Affinity BioReagents PMCA1 antibody (PA-914) was used at 1:100 dilution. Goat anti-mouse or goat anti-rabbit IgG (H+L) conjugated to fluorophores (Alexa 488 and Alexa 594 conjugates (Invitrogen), diluted 1:500 or 1:1000, were used as secondary antibodies. After incubation, sections on slides were washed in PB and mounted with Vectashield (Vector, Burlingame, CA, USA). Negative controls for non-specific staining of secondary antibodies were performed for every set of experiments by omitting the primary antibodies. Immunofluorescent and differential interference contrast (DIC) images were acquired at depths of 12 bits on a confocal microscope using 488 nm Ar and 594 nm He/Ne lines for fluorophore excitation, suitable band-pass or long-pass filters for emission detection, and a 63×/0.9 NA water objective.

RESULTS

Photoreceptor regions proximal to the OS—the ellipsoid, cell body and synaptic terminal—are uniquely defined by their morphology (Fig. 1; Mercurio and Holtzman, 1982; Nilsson, 1985; Townes-Anderson et al., 1985). The aim of this report is to determine the dynamic range of [Ca²⁺]i in the different regions of hyperpolarized and depolarized photoreceptors to obtain a basis for understanding how Ca²⁺ regulates diverse signaling pathways in light and darkness, respectively. Moreover, we find that Ca²⁺ influx into cell bodies and ellipsoids occurs via indigenous local voltage-operated Ca²⁺ channels.

Fig. 1.

(A) Schematized generic photoreceptor. In addition to the OS which hosts the phototransduction apparatus, the cell consists of three distinct subregions: the ellipsoid contains most of the cell's mitochondria; the cell body is filled by the nucleus and ER cisternae and the synaptic terminal is packed with vesicles and a few cisternae of smooth ER. (B) Dissociated salamander rod and (C) salamander cone photoreceptor. Abbreviations: ER, endoplasmic reticulum; VGCC, voltage-gated channel. Ca²⁺ sequestration and release from the mitochondria occurs via Ca²⁺ uniporter channels and Na⁺/Ca²⁺ transporters, respectively. Scale bars=5 μm.

Depolarization-evoked Ca²⁺ signals have a distinct spatiotemporal pattern

The spatial and temporal dynamics of depolarization-evoked [Ca²⁺]i responses in salamander photoreceptor non-OS regions were studied with functional Ca²⁺ imaging. Rods and cones were freshly isolated and loaded with the Ca²⁺ indicator dye fura-2 AM. The baseline [Ca²⁺]i in all non-OS regions of rod photoreceptors was under 100 nM (synaptic terminal, 98±24 nM; cell body, 57±7 nM; ellipsoid, 56±5 nM). Cells were depolarized by superfusion with elevated concentrations of KCl (high K⁺). Depolarization-evoked [Ca²⁺]i exhibited a highly reproducible standing [Ca²⁺]i gradient across the terminal, cell body, and ellipsoid (Fig. 2). A representative experiment showing the effects of depolarization on photoreceptor [Ca²⁺]i domains is illustrated in Fig. 2.

A statistical analysis summarizing the amplitudes of [Ca²⁺]_i elevations evoked by 30 mM KCl in the terminal, cell body and ellipsoid is shown in Fig. 3. The amplitude of depolarization-evoked [Ca²⁺]_i increases in rod terminals was 2160±196 nM (n=5). [Ca²⁺]i elevations occurred at discrete puncta within the terminal, raising average [Ca²⁺]i within the terminal above 2 μM (Fig. 2B and C). The increases were significantly lower in the cell body (1168±87 nM; P<0.005; n=19) and were sharply delimited from the subellipsoid, an endoplasmic reticulum-rich space that separates the cell body from the ellipsoid (Fig. 2B). By far the lowest concentration and the slowest kinetics of evoked responses were observed in the ellipsoid (754±55 nM, n=19) (Figs. 2A and 3).

Fig. 3.

Summary of the high K⁺ evoked [Ca²⁺]_i elevations in regions of rod and cone photoreceptors downstream from the OS.

The gradient of depolarization-evoked [Ca²⁺]i across the cell was stable for the duration of the experiment. The region-specific amplitude and kinetics of depolarization-evoked [Ca²⁺]i pattern illustrated in Figs. 2 and 3 were observed in all intact cells and were highly conserved across rods and all subtypes of cones. The differences between cell regions were statistically significant (P<0.001 to P<0.005). In cells lacking synaptic terminals, Ca²⁺ signals were first observed in the cell body domain closest to the axon, possibly as a result of higher density of Ca²⁺ channel expression in this region or due to retraction of the Ca²⁺ channel-rich terminal into the soma (data not shown; Townes-Anderson et al., 1985, but see Nachman-Clewner et al., 1999; Steele et al., 2005).

These results show that depolarization evokes micro-molar increases in global [Ca²⁺]i in rod terminals and submicromolar changes in perikaryal and ellipsoid [Ca²⁺]i. Below we analyze the mechanisms responsible for controlling the amplitude and kinetics of depolarization-evoked responses.

Depolarization triggers direct Ca²⁺ influx into the photoreceptor cell body

A recent report using optical imaging suggested that, following stimulation with 30 mM KCl, there is at least a 250-fold difference between the amplitude of depolarization-evoked [Ca²⁺]i responses in the soma and the terminal of rods (Steele et al., 2005). This study suggested that under physiological conditions [Ca²⁺]i elevations in the cell body would occur via diffusion from the synaptic terminal. The same study, however, found the apparent intensity of immunostaining for α_1D subunit-containing L-type Ca²⁺ channels to be only ∼three-fold higher within the terminal compared with the soma (Steele et al., 2005). To address this discrepancy, and to determine the fraction of Ca²⁺ entering the cell body by diffusion, we recorded Ca²⁺ responses from rods with and without terminals. If most perikaryal [Ca²⁺]i is due to Ca²⁺ diffusion from the terminal, then depolarization should evoke little response in rods in which synaptic terminals had been cut off experimentally. In contrast, if cell bodies express functional voltage-operated Ca²⁺ channels, depolarization-evoked responses should be detected in axotomized cells. The terminals were cut in 0 Ca²⁺ saline with a sharp microelectrode (Fig. 4A, B). Following a resealing period of 15–30 min, cells were returned to control saline and [Ca²⁺]i responses measured in 30 mM KCl. As illustrated in Fig. 4D, cell bodies in control and axotomized cells had comparable [Ca²⁺]i baselines of 71±20 nM for controls and 79±29 for axotomized cells. High K⁺ raised [Ca²⁺]i by 988±83 nM in four of five axotomized rods, not significantly different from 1168±255 nM measured in intact cells (n=19; P=0.0656; Figs. 3 and 4C, D). In cells possessing intact synaptic terminals, perikaryal [Ca²⁺]i responses were 1250±253 (n=5), not significantly different from axotomized rods (P=0.1086). A small (45 nM) [Ca²⁺]i elevation was detected in the cell body of one of five axotomized cells; this cell body had a relatively high basal [Ca²⁺]i of 100 nM. These data are strongly suggestive of endogenous voltage-operated Ca²⁺ channels in cell bodies of photoreceptors.

Fig. 4.

The amplitude of depolarization-evoked [Ca²⁺]i increases in cell bodies of axotomized rods is not significantly different from [Ca²⁺]i increases in intact cells. (A) Transmission image of an intact rod with a synaptic terminal. (B) Transmission image following terminal axotomy by a sharp microelectrode. (C) Summary data for baseline [Ca²⁺]i (2 mM KCl) and depolarization-evoked (30 mM KCl) [Ca²⁺]i levels in cell bodies from intact cells possessing synaptic terminals. (D) Summary data for baseline [Ca²⁺]i (2 mM KCl) and depolarization-evoked (30 mM KCl) [Ca²⁺]i levels in cell bodies of cells with severed terminals.

[Ca²⁺]i regulation in the photoreceptor ellipsoid

To determine whether the ellipsoid [Ca²⁺]i elevation exemplified in Fig. 2 could result from Ca²⁺ influx through local plasma membrane channels, depolarization-evoked [Ca²⁺]i responses were analyzed in isolated ellipsoids. These experiments were performed in cones, because it was possible to separate cone, but not rod, ellipsoids mechanically from perikarya. Three types of control experiment were performed to determine the viability and integrity of isolated cone ellipsoids. First, exposure to the vital dye Toluidine Blue showed that isolated ellipsoids exclude the dye, suggesting that ellipsoid membrane maintains its integrity following dissociation (Fig. 5A). Second, cells were labeled simultaneously with TMRM (100 nM), a cationic rhodamine-based dye that reversibly accumulates in the highly negatively charged mitochondrial matrix (Scaduto and Grotyohann, 1999), and MitoTracker Green (200 nM), a mitochondria-specific indicator that is independent of the membrane potential (Pendergrass et al., 2004). We reasoned that if mitochondrial function in the isolated ellipsoid is normal, the MitoTracker Green signal should occupy the same area as TMRM. Indeed, complete superposition between TMRM and Mitotracker Green signals was observed in intact ellipsoids (Fig. 5Bi, Bii). The intensity of the TMRM signal was identical in isolated and intact ellipsoids (Fig. 5Bi; n=9 pairs). We additionally tested the mitochondrial membrane potential by exposing cells to the mitochondrial uncoupler antimycin. Following application of 8 μM antimycin, the intensity of the TMRM signal in isolated ellipsoids decreased with a time constant of ∼6.6 min (Fig. 5C, D). These results suggest that the mitochondrial membrane potential in isolated ellipsoids is normal. The baseline [Ca²⁺]i in isolated ellipsoids was slightly elevated (58.18±8.92 nM) with respect to controls (45±4.83 nM); however, this difference was not significant (P=0.2775). Taken together, these four control experiments indicate that following dissociation the isolated ellipsoid remains a structurally and functionally intact compartment.

Fig. 5.

(A) Isolated ellipsoid and a putative spiking neuron incubated for 30 s in 0.01% Toluidine Blue. The ellipsoid was not stained by the vital dye. Scale bar=10 μm. (Bi–iii) A large single cone and an isolated ellipsoid double labeled with TMRM (100 nM) and MitoTracker Green (200 nM). Both indicator dyes selectively label the ellipsoid region in intact cells and following ellipsoid dissociation from the cell body. (Ci–iii) Antimycin (10 μM) reduces the TMRM, but not MitoTracker Green, signal within the ellipsoid. (D) Confocal linescan of TMRM-positive region in isolated cone ellipsoid during exposure to the mitochondrial uncoupler antimycin. Antimycin depolarizes the membrane potential across the inner mitochondrial membrane, seen as a decrease in ellipsoid TMRM fluorescence.

L-type Ca²⁺ channel blockers suppress depolarization-evoked [Ca²⁺]i responses in both cell body and ellipsoid of the intact cell (Fig. 6A). To test whether Ca²⁺ entry occurs directly into the ellipsoid itself, evoked responses were analyzed in isolated ellipsoids. Stimulation with high K⁺ induced significant [Ca²⁺]i elevations in isolated ellipsoids (Fig. 3). To amplify depolarization-evoked responses from isolated ellipsoids, KCl concentration in a subset of experiments was raised to 60 mM. Superfusion with 60 mM KCl increased isolated ellipsoid [Ca²⁺]_i from 33±7 nM to 586±131 nM (n=4; Fig. 6B). In the presence of nifedipine, depolarization elevated [Ca²⁺]_ellipsoid by only ∼33 nM (from 48±9 to 81±13 nM; Fig. 6B), suggesting ellipsoids express functional L-type Ca²⁺ channels. In control experiments, 60 mM KCl elevated [Ca²⁺]i in ellipsoids and cell bodies of intact cones to 1407±320 nM and 3117±349nM, respectively (Fig. 3). These data show that responses from isolated ellipsoids are roughly about half the elevations observed in intact ellipsoids and five-fold lower than increases in the cell body.

Fig. 6.

L-type Ca²⁺ channels are expressed in cone ellipsoids. (A) Intact cone transiently depolarized by puffs of 90 mM KCl; 2 μM nifedipine reversibly reduced depolarization-evoked [Ca²⁺]i transients in the cell body (solid trace) and the ellipsoid (dotted trace). (B) Isolated cone ellipsoids express indigenous voltage-operated Ca²⁺ channels; 60 mM KCl-induced [Ca²⁺]_i elevation is strongly suppressed by 2 μM nifedipine.

If diffusion provides a component of the depolarization-evoked [Ca²⁺]i signal within the ellipsoid, the rate of Ca²⁺ increase (d[Ca²⁺]i/dt) within isolated ellipsoids should be slower compared with ellipsoids in intact cells. Fig. 7A shows the time course of depolarization-evoked [Ca²⁺]i elevations recorded from an intact cell body and its attached ellipsoid, and an isolated cone ellipsoid. The rise-time and decay constant of depolarization-evoked Ca²⁺ transients in isolated ellipsoids was almost three-fold slower than in ellipsoids from intact cones (Fig. 7).

Fig. 7.

Kinetics of depolarization-evoked [Ca²⁺]i responses in photoreceptor compartments. (A) Simultaneous recording of [Ca²⁺]_i elevations in the cone cell body (black), ellipsoid (blue) and an isolated cone ellipsoid (red trace). The position of the regions of interest is indicated in the inset at the upper right panel. The fastest and largest [Ca²⁺]i change occurred in the cell body while a moderate elevation was recorded in the ellipsoid. The amplitude and kinetics of the high K⁺-evoked [Ca²⁺]_i response were further reduced in the isolated ellipsoid. (B) High resolution detail from A emphasizes the differences in kinetics between cell regions by enlarging the first segment of the high K⁺-induced [Ca²⁺]_i response. (C, E) Kinetics of [Ca²⁺]_i rise times and (D, F) [Ca²⁺]_i decays in cone cell bodies, ellipsoids and isolated ellipsoids in response to 30 mM and 60 mM KCl stimulation, respectively.

Assuming intracellular [K⁺] in amphibian excitable cells to be ∼124 mM (Aidley, 1989), 30 mM and 60 mM KCl depolarized ellipsoids to −36 mV and −18 mV, respectively. Doubling [K⁺]o increased [Ca²⁺]i in intact ellipsoids by 211±23%, compared with 164±22% in isolated ellipsoids. Interestingly, while doubling the concentration of extracellular K⁺ caused an approximately six-fold increase in the rate of [Ca²⁺]_i rise in the cell body, only a two-fold increase was observed in the intact ellipsoid (Fig. 7C, E).

Unlike the cell body, which responded to the increase in extracellular K⁺ with an increase in the rate of Ca²⁺ clearance (the time constant decreased from 20±4 to 10±5 s), an equivalent increase in [Ca²⁺]_i clearance was not observed in the ellipsoid of intact cone photoreceptors (24±4 vs. 19±3 s) nor in the isolated ellipsoid (32±4 vs. 35±11 s) (Fig. 5C).

These data suggest that Ca²⁺ enters the ellipsoid via local Ca²⁺ channels, with a possible additional diffusion component from the cell body. The kinetics of onset and offset of depolarization-evoked [Ca²⁺]i signals in the ellipsoid were greatly reduced vis à vis the cell body, consistent with low expression levels of plasma membrane Ca²⁺ channels and pumps (see below) and Ca²⁺ sequestration into intracellular compartments such as the mitochondria (Križaj et al., 2003).

Expression of Ca²⁺ influx and extrusion mechanisms in the IS is consistent with spatiotemporal dynamics of depolarization-evoked Ca²⁺ signals

We next tested whether spatiotemporal variation in depolarization-evoked signals in photoreceptors could be attributed to differential distribution of Ca²⁺ channels and PMCA pumps within the non-OS region. Since the identity of Ca²⁺ channels in salamander photoreceptors has not yet been unequivocally established, we immunostained salamander retinal sections with antibodies raised against the consensus sequence common to all L-type α₁ Ca²⁺ channel subunits. Prominent α₁ subunit immunofluorescence was observed in synaptic terminals in the outer plexiform layer (OPL) (Fig. 8B, arrows) and the inner plexiform layer (not shown). Müller cell processes enveloping photoreceptors and neurons throughout the retina were α₁-immunoreactive. At higher confocal gains, the pan-α₁ signal was observed in the cell body of rods and cones, consistent with sparser Ca²⁺ channel expression. In contrast, little Ca²⁺ channel expression in the ellipsoid area was seen using this antibody.

Fig. 8.

Compartmentalized expression of voltage-operated Ca²⁺ channels and PMCAs in salamander photoreceptors. (A) DIC image of the vertical sections of the outer retina of the tiger salamander and (B) fluorescent micrographs of the retina stained with antibodies raised against α₁ subunits of L-type Ca²⁺ channels (pan α₁). Puncta in the OPL, corresponding to photoreceptor terminals, are strongly stained (arrows), with moderate labeling of the cell body (arrowheads). The putative processes of Müller glia are labeled throughout. (C) DIC image of the salamander outer retina. (D) Fluorescent micrographs of the retina labeled for the PMCA transporter isoform 1 (PMCA1). Synaptic terminals are strongly immunoreactive (arrows), whereas a weaker PMCA1 signal is observed in the cell body (arrowheads). Scale bars=10 μm.

We also attempted to correlate the kinetics of Ca²⁺ decay following depolarization with expression of the dominant plasma membrane Ca²⁺ clearance mechanism in photoreceptors. Ca²⁺ extrusion in non-OS regions of salamander photoreceptors is supported exclusively by PMCAs (Križaj and Copenhagen, 1998). A recent analysis of PMCA expression in the salamander retina showed that salamander rods and cones express PMCA isoforms 1, 2 and 4(Križaj et al., 2004); however, a region-specific analysis was not performed in that study. The salamander retina was immunostained with the antibody PMCA1, the dominant PMCA isoform in vertebrate photoreceptors (Križaj et al., 2002). As illustrated in Fig. 8D, the antibody strongly labeled photoreceptor terminals, with a moderate expression in the cell body. These results suggest that high rate of Ca²⁺ clearance from the synaptic terminal observed physiologically (Križaj and Copenhagen, 1998) reflects the high PMCA density in synaptic terminals of rods and cones.

DISCUSSION

In darkness, photoreceptors are tonically depolarized, resulting in sustained entry of Ca²⁺ and elevation of [Ca²⁺]i. The aim of this study was to characterize the spatiotemporal pattern of elevated free [Ca²⁺]i in the three non-phototransducing regions of depolarized rods and cones and to determine the dynamic range of physiological Ca²⁺ concentrations across photoreceptor regions. Using optical imaging we show that (1) depolarization is associated with a standing [Ca²⁺]i gradient across non-OS regions of rod and cone photoreceptors. This gradient is produced in part by polarized expression of Ca²⁺ channels and pumps (2) Ca²⁺ entry into the cell body and ellipsoid occurs via indigenous Ca²⁺ channels with a minor contribution from Ca²⁺ diffusion. (3) Physiological depolarization evokes micromolar increases in terminal [Ca²⁺]i with 20–40-fold dynamic range.

Spatiotemporal Ca²⁺ gradient within the photoreceptor inner segment

The amplitude of depolarization-evoked [Ca²⁺]i signals in rod photoreceptor synaptic terminal is ∼double those in the cell body and ∼2.5-fold higher than in the ellipsoid (Figs. 2 and 3) whereas the rates of Ca²⁺ entry and extrusion were 1.5-three-fold higher for the cell body compared with the ellipsoid. The gradient of depolarization-evoked [Ca²⁺]i across the cell was remarkably stable for the duration of the experiments (>1 h) and could have originated either from Ca²⁺ influx from the extracellular space or Ca²⁺ release from intracellular Ca²⁺ stores. The dihydropyridine blocker nifedipine suppressed depolarization-evoked Ca²⁺ signals in all non-OS regions, suggesting most voltage-activated Ca²⁺ influx occurs via the L-type class of Ca²⁺ channel. Consistent with this, the [Ca²⁺]i gradient correlated with the expression levels of L-type Ca²⁺ channels and PMCA pumps (Fig. 8; Križaj et al., 2004; see also Nachman-Clewner et al., 1999; Morgans, 2001). The pattern of depolarization-evoked [Ca²⁺]i signals in cones was similar to the responses observed in rods. The additional advantage of the cone preparation was that it allowed us to directly study Ca²⁺ entry pathways in the ellipsoid region (see below).

The data presented suggest that at physiological depolarizations the average [Ca²⁺]i within photoreceptor terminals of rods spans the operating range from 50 to 100 nM in the light to ∼2 μM in darkness. Assuming ∼60 channels are open in a salamander rod terminal at the dark resting potential (Heidelberger et al., 2005) and an average number of 7.3 Ca²⁺ channel clusters that produce [Ca²⁺]i hotspots per terminal (Nachman-Clewner et al., 1999), each hotspot at dark membrane potentials would have ∼eight open channels and contribute ∼287 nM free Ca²⁺ within an area of ∼0.76 um². While we did detect local elevations in terminal [Ca²⁺]i (Fig. 2B), the relationship between exocytosis and localized increases in terminal [Ca²⁺]i remains to be studied. However, our data show that photoreceptors compress the 4–5 log unit dynamic range obtained by the phototransduction cascade in the OS (Fain et al., 2001) into a 20-fold operating range of [Ca²⁺]i within the presynaptic terminal. This resembles the 28-fold range for exocytosis in synaptic terminals of cones (Choi et al., 2005) and suggests that photoreceptor output is controlled by Ca²⁺ homeostatic mechanisms within the terminal. Thus presynaptic gain control by Ca²⁺ extrusion, release from stores (Križaj et al., 1999, 2003), neuromodulation (Stella et al., 2002) and feedback from horizontal cells (Verweij et al., 1996; Vessey et al., 2005) may mediate presynaptic adaptation via regulation of terminal [Ca²⁺].

Our results differ from the recently reported data that suggested [Ca²⁺]i baseline in synaptic terminals of light adapted rods is ∼350 nM, exceeding 39 μM during depolarization with 30 mM KCl (Steele et al., 2005). The conclusions reached by Steele et al. are difficult to interpret given that the fura-2 dye used in that study saturates at 2–3 μM [Ca²⁺]i (Neher 1995). Furthermore, the relative Ca²⁺ concentrations in different cell regions in Steele et al. (2005) are difficult to extrapolate because cell regions were not calibrated in situ.

Immunocytochemical studies of Ca²⁺ channel expression in photoreceptors have been equivocal. In rat photo-receptors, L-type Ca²⁺ immunoreactivity with antibodies against the retina-specific α_1F subunit was prominent within the synaptic terminal whereas the signal in the cell body and the ellipsoid was undetectable (Morgans et al., 1998; Morgans, 2001). However, other antibodies raised against the α_1F subunit showed more prominent immunostaining of the ONL (Morgans, 2001) suggesting the possibility that different Ca²⁺ channel isoforms are expressed within different non-OS compartments. Immunostaining of dissociated salamander photoreceptors with α_1C subunit antibodies showed similar immunostaining in the cell body and the synaptic terminals of isolated rods and cones whereas little α_1C, signal was detected in the ellipsoid (Nachman-Clewner et al., 1999). However, an α_1D antibody appeared to stain terminals of dissociated salamander rods about three-fold more strongly than cell bodies (Steele et al., 2005). Given the difficulty of quantifying the absolute densities of antigens in immunostained sections and the questions about the α₁ subunits expressed in salamander photoreceptors (Nachman-Clewner et al., 1999; Steele et al., 2005) we performed qualitative immunostaining of retinal sections with pan Ca²⁺ channel antibodies recognizing all α₁ isoforms. Our results suggest that (i) voltage-operated Ca²⁺ channels are expressed at low densities in somas of rods and cones, consistent with direct Ca²⁺ entry into the soma, and (ii) the ellipsoid membrane contains Ca²⁺ channels at a density that is below the resolution of our antibody staining.

Consistent with low L-type Ca²⁺ channel immunoreactivity in the ellipsoid (Fig. 8; see also Morgans et al., 1998; Nachman-Clewner et al., 1999), a recent report suggested that depolarization with 30 mM KCl evokes little or no response from cell bodies and ellipsoids of rods (Steele et al., 2005). We took the advantage of the high sensitivity of functional Ca²⁺ imaging to show directly that depolarization increases [Ca²⁺]i in both isolated cell bodies and ellipsoids. In our experiments, depolarization-evoked [Ca²⁺]i increases were observed in axotomized cell bodies and in dissociated ellipsoids, suggesting that both regions express indigenous voltage-operated Ca²⁺ channels activated at dark potentials (∼—35 mV in 30 mM KCl). Taken together, more than 95% of rod cell bodies, with and without visible axons, and including axotomized cells, were responsive to 20 mM and 30 mM KCl (n≫150, including data from Križaj and Copenhagen, 1998, 1999; Križaj et al., 2003).

There is increasing evidence that somatic [Ca²⁺] is not only controlled by plasma membrane Ca²⁺-permeable channels but also by intracellular stores (Gerasimenko and Gerasimenko, 2004; Križaj and Szikra unpublished observations). Calcium-induced Ca²⁺ release, which may contribute up to 40% to the depolarization-evoked Ca²⁺ signal within the cell body (Križaj et al., 2003), could also contribute to the discrepancy between the relatively weak expression of L-type Ca²⁺ channels observed with immunostaining and prominent [Ca²⁺]i elevations observed in physiological experiments. The mean amplitude of depolarization-evoked [Ca²⁺]i increases in the cell body of axotomized cells was ∼80% of those observed in intact cells (Figs. 2-4). It is possible that Ca²⁺ diffusion from the terminal into the cell body plays a larger role at higher depolarization levels (Steele et al., 2005).

A potential explanation for the low amplitude and slow time course of Ca²⁺ signals within the ellipsoid might lie in compartmentalization of fura-2 dye within the mitochondria. However, >95% of the ellipsoid fura-2 signal was quenched by exposure to Mn²⁺, indicating that most of the dye was cytoplasmic (data not shown). We propose that the ellipsoid plasma membrane contains Ca²⁺ channels at a density that is below the resolution of our antibody staining. Interestingly, the mean amplitudes and kinetics of depolarization-evoked Ca²⁺ elevation and decay in isolated ellipsoids were consistently ∼50% of those in intact ellipsoids. Although this could be attributed to damage during the isolation procedure, we found that isolated ellipsoids excluded vital dyes such as Toluidine Blue and appeared to have inner mitochondrial membrane potentials indistinguishable from those in intact cells. Moreover, baseline [Ca²⁺]i levels in isolated ellipsoids were not statistically different from [Ca²⁺]i levels in intact ellipsoids, arguing against increased Ca²⁺-mediated inactivation of Ca²⁺ entry in the former. Finally, the doubling of extracellular KCl resulted in supralinear increase in ellipsoid [Ca²⁺]i. Although at the moment there is no direct evidence for intra-compartmental Ca²⁺ diffusion or its potential modulation by intracellular stores and PMCAs in rods and cones, the (a) [Ca²⁺]-dependence of the Ca²⁺ diffusion coefficient (Allbritton et al., 1992), (b) gradient of expression of voltage-operated Ca²⁺ channels and PMCAs in the inner segment and (c) a large potential Ca²⁺ sink represented by the ellipsoid mitochondria (Fig. 5), lead us to hypothesize that Ca²⁺ diffusion from the cell body toward an ellipsoid sink might additionally contribute to the inter-regional [Ca²⁺]i gradient in depolarized cells.

Functional implications of the spatiotemporal [Ca²⁺]i gradient

The separation of distinct Ca²⁺ signaling mechanisms in spatial domains of a photoreceptor is well suited for photoreceptor signaling. [Ca²⁺]i levels in the OS determine the rate and extent of adaptation to light (Fain et al., 2001). To preserve the fidelity of the light response, it is imperative to protect the gain control mechanisms in the OS from interference by Ca²⁺ fluxes in the inner segment. This may be especially important for cones, as baseline [Ca²⁺]i in saturated light-adapted cone OSs is several-fold lower (∼5 nM; Sampath et al., 1999) compared with inner segment [Ca²⁺]i (∼20–40 nM). The low density of voltage-operated Ca²⁺ channels and Ca²⁺ sequestration by its resident mitochondria (Nilsson, 1985; Townes-Anderson et al., 1985; Hoang et al., 2002; Križaj et al., 2003) ensures that the ellipsoid serves as a buffer for Ca²⁺ fluxes between the OS and the rest of the cell. Consistent with a mitochondrial Ca²⁺ sink within the ellipsoid (Szikra and Križaj, 2005), the kinetics of Ca²⁺ signals in the ellipsoid was slow and independent of doubling of [KCl]o (Fig. 5). Why are voltage-operated Ca²⁺ channels expressed in the ellipsoid at all? Photoreceptors are among the biggest energy consumers in the brain (Stone et al., 1999). Mitochondria undergo large changes in matrix [Ca²⁺]i following Ca²⁺ influx via plasma membrane Ca²⁺ channels (Rutter et al., 1993; Babcock et al., 1997). In turn, the amplitude of the Ca²⁺ rise inside the mitochondria determines the increases in cytosolic and mitochondrial ATP concentrations (Fein and Tsacopoulos, 1988; Jouaville et al., 1999). We propose that Ca²⁺ diffusion from the cell body does not provide adequate stimulation of mitochondrial dehydrogenases needed to sustain the metabolic needs of photoreceptors.

The large depolarization-evoked Ca²⁺ signals in the cell body may play a role in nuclear gene activation (Dolmetsch et al., 2001). The c-fos promoter, which appears to be able to selectively respond to changes in cytoplasmic and nuclear [Ca²⁺]i would be a prime candidate here (Hardingham et al., 1997; Grimm et al., 2000). There is increasing evidence that somatic [Ca²⁺] is controlled by both plasma membrane Ca²⁺-permeable channels and intracellular stores (Hardingham et al., 1997; Gerasimenko and Gerasimenko, 2004). Ca²⁺ release from perikaryal stores is likely to additionally amplify the depolarization-evoked signal amplitude and thus influence gene expression (Mao and Wang, 2003; Carrasco et al., 2004; see also Marius et al., 2006). Ryanodine, inositol triphosphate receptors and sarcoplasmic-endoplasmic reticulum calcium ATPase (SERCA) transporters are expressed in the cell body and terminal but not the ellipsoid regions (Križaj et al., 2004), consistent with the observations of caffeine-evoked [Ca²⁺]i signals in the cell body and the terminal (Križaj et al., 1999). Finally, high density of L-type channels in the terminal is well suited for controlling rapid and reliable exocytosis. The high surface-to-volume ratio of the synaptic terminal and high density of PMCAs are likely to limit fast high amplitude [Ca²⁺]i signals to the terminal, ensuring that [Ca²⁺]i levels and the exocytotic rate drop rapidly during the light onset. Parenthetically, a compartmentalized pattern of Ca²⁺ channels, Ca²⁺ store transporters and PMCA expression was also reported for retinal bipolar cells and cochlear/vestibular hair cells (Satoh et al., 1998; Dumont et al., 2001; Križaj et al., 2002). In bipolar cells, depolarization triggers compartmentalized [Ca²⁺]i elevations (Heidelberger and Matthews, 1992) that mirror those observed in photoreceptors. For example, slow low-amplitude depolarization-evoked signals in bipolar cell bodies were observed following severance of the synaptic terminal, indicating expression of indigenous voltage-operated Ca²⁺ channels (Heidelberger and Matthews, 1992).

CONCLUSION

In conclusion, we determined the amplitude and kinetics of [Ca²⁺]i signals in different regions of the photoreceptor cell and ascertained the contribution of endogenous voltage-operated Ca²⁺ channels in the cell body and ellipsoid regions. Our results underscore the importance of careful analysis of Ca²⁺ dynamics in photoreceptor compartments when using ratiometric Ca²⁺ indicators. We hypothesize that [Ca²⁺]i can be independently controlled between the outer and inner segment, and between different non-OS domains of rods and cones, allowing for intricate local modulation by separate and distinct Ca²⁺-dependent signaling pathways.

Abbreviations

[Ca²⁺]i

intracellular calcium concentration

DIC

differential interference contrast

fura-2 AM

fura 2-acetoxymethylestre

OPL

outer plexiform layer

OS

outer segment

PMCA

plasma membrane calcium ATPase

ROI

region of interest

SERCA

sarcoplasmicendoplasmic reticulum calcium ATPase

TMRM

tetramethylrhodamien