Abstract
In addition to vesicle release at synaptic ribbons, rod photoreceptors are capable of substantial slow release at non-ribbon release sites triggered by Ca2+-induced Ca2+ release (CICR) from intracellular stores. To maintain CICR as rods remain depolarized in darkness, we hypothesized that Ca2+ released into the cytoplasm from terminal endoplasmic reticulum (ER) can be replenished continuously by ions diffusing within the ER from the soma. We measured [Ca2+] changes in cytoplasm and ER of rods from Ambystoma tigrinum retina using various dyes. ER [Ca2+] changes were measured by loading ER with fluo-5N and then washing dye from the cytoplasm with a dye-free patch pipette solution. Small dye molecules diffused within ER between soma and terminal showing a single continuous ER compartment. Depolarization of rods to −40 mV depleted Ca2+ from terminal ER, followed by a decline in somatic ER [Ca2+]. Local activation of ryanodine receptors in terminals with a spatially confined puff of ryanodine caused a decline in terminal ER [Ca2+], followed by a secondary decrease in somatic ER. Localized photolytic uncaging of Ca2+ from o-nitrophenyl-EGTA in somatic ER caused an abrupt Ca2+ increase in somatic ER, followed by a slower Ca2+ increase in terminal ER. These data suggest that, during maintained depolarization, a soma-to-terminal [Ca2+] gradient develops within the ER that promotes diffusion of Ca2+ ions to resupply intraterminal ER Ca2+ stores and thus sustain CICR-mediated synaptic release. The ability of Ca2+ to move freely through the ER may also promote bidirectional communication of Ca2+ changes between soma and terminal.
SIGNIFICANCE STATEMENT Vertebrate rod and cone photoreceptors both release vesicles at synaptic ribbons, but rods also exhibit substantial slow release at non-ribbon sites triggered by Ca2+-induced Ca2+ release (CICR). Blocking CICR inhibits >50% of release from rods in darkness. How do rods maintain sufficiently high [Ca2+] in terminal endoplasmic reticulum (ER) to support sustained CICR-driven synaptic transmission? We show that maintained depolarization creates a [Ca2+] gradient within the rod ER lumen that promotes soma-to-terminal diffusion of Ca2+ to replenish intraterminal ER stores. This mechanism allows CICR-triggered synaptic release to be sustained indefinitely while rods remain depolarized in darkness. Free diffusion of Ca2+ within the ER may also communicate synaptic Ca2+ changes back to the soma to influence other critical cell processes.
- calcium imaging
- calcium-induced calcium release
- endoplasmic reticulum
- retina
- rod photoreceptors
- synaptic terminal
Introduction
Vertebrate photoreceptors transmit light-evoked voltage changes to second-order retinal neurons by changing glutamate release rates. Rods and cones both exhibit fast ribbon-mediated release of vesicles, but slow sustained release from rods also involves significant release at non-ribbon sites (Snellman et al., 2011; Chen et al., 2013, 2014). This slow non-ribbon release contributes to slower release kinetics in rods versus cones, paralleling the slower light response kinetics of rods (Schnapf and Copenhagen, 1982; Cadetti et al., 2005; Rabl et al., 2005). In amphibian and mammalian retinas, slow, non-ribbon release from rods is triggered by Ca2+-induced Ca2+ release (CICR) into cytoplasm from endoplasmic reticulum (ER) stores (Krizaj et al., 1999, 2003; Cadetti et al., 2006; Suryanarayanan and Slaughter, 2006; Babai et al., 2010; Chen et al., 2014). Blocking CICR in mouse and salamander retina inhibits rod-driven light responses of second-order neurons by 50–90%, indicating that CICR is a major mechanism for maintaining high cytoplasmic [Ca2+] needed to sustain vesicle release from rods in darkness (Cadetti et al., 2006; Suryanarayanan and Slaughter, 2006; Babai et al., 2010). Ca2+ exits ER during CICR in rods through ryanodine receptors (RyRs); immunohistochemical studies show the presence of RyRs but not inositol 1,4,5-trisphosphate receptors in rod terminals (Krizaj et al., 2003, 2004). The principal RyR subtype in rods is a RyR2 splice variant (Shoshan-Barmatz et al., 2005, 2007).
How can CICR contribute indefinitely to elevation of Ca2+ in rod terminals without exhausting intraterminal ER Ca2+ stores? ER appears to be present in somas, axons, and terminals of rods from most, if not all, vertebrates (De Robertis and Franchi, 1956; De Robertis, 1956; Ladman, 1958; Ungar et al., 1981; Mercurio and Holtzman, 1982; Freihöfer et al., 1990; Johnson et al., 2007; Babai et al., 2010; Chen et al., 2014). Immunohistochemical labeling for sarco/ER Ca2+ ATPase type 2 (SERCA2) partially colocalizes with labeling for the ribbon protein Ribeye in mouse retina, supporting ultrastructural evidence that ER approaches close to ribbons (Babai et al., 2010). Fluorescence loss in photobleaching (FLIP) and other approaches indicate that the ER forms a continuous structure in many cells (Dayel et al., 1999; Park et al., 2000; Estrada de Martin et al., 2005; Verkhratsky, 2005). Similarly, fluorescence recovery after photobleaching (FRAP) experiments with ER-tracker dye suggested that the ER in salamander rods extends continuously from soma to terminal (Chen et al., 2014). Ca2+ accumulates within ER of rods, attaining especially high levels in somatic ER (Ungar et al., 1981; Somlyo and Walz, 1985). In other cell types, Ca2+ ions have been shown to diffuse within the ER from one region of a cell to another (Mogami et al., 1997; Park et al., 2000; Choi et al., 2006; Wu and Bers, 2006; Petersen and Verkhratsky, 2007; Swietach et al., 2008; Picht et al., 2011; Bers and Shannon, 2013). Experiments described in the present study showed that lengthy depolarization of salamander rods causes a sustained decline in somatic ER [Ca2+] without causing a substantial increase in cytoplasmic [Ca2+] of the soma. We hypothesized that Ca2+ depleted from rod terminal ER stores during sustained depolarization can be replenished by diffusion of Ca2+ through the ER lumen from reservoirs in the axon and soma. To test this idea, we combined use of ER and cytoplasmic Ca2+ dyes with various techniques, including FRAP, FLIP, voltage clamp, flash photolysis of caged Ca2+, and localized activation of RyRs. The results showed that activation of CICR during sustained depolarization of rods generates a soma-to-terminal [Ca2+] gradient within the ER, promoting diffusion of Ca2+ through the ER from the perikaryon to resupply Ca2+ stores in the synaptic terminal. This nearly inexhaustible mechanism for sustaining CICR helps to maintain high Ca2+ levels necessary for sustaining synaptic release from rods as they remain depolarized in darkness. Free Ca2+ diffusion through the ER may also communicate synaptic Ca2+ changes back to the soma, which can influence mitochondrial function, stress responses, or other processes.
Materials and Methods
Animal care and use.
Both sexes of aquatic tiger salamanders (Ambystoma tigrinum, 18–25 cm in length; Charles Sullivan) were used for experiments. They were maintained on a 12 h light/dark cycle and killed 1–2 h after the beginning of the dark cycle. Salamanders were anesthetized by bathing with MS-222 (0.25 g/L) for 30 min before being decapitated with heavy shears. The head was hemisected and the spinal cord pithed. Protocols were approved by the University of Nebraska Medical Center Institutional Animal Care and Use Committee.
Photoreceptor isolation.
Details of the photoreceptor isolation procedures have been described previously (Chen et al., 2013). Briefly, retinas were digested by incubation with papain (30 U/ml; Worthington) plus cysteine (0.2 mg/ml) in Ca2+-free amphibian saline solution containing the following (in mm): 116 NaCl, 2.5 KCl, 5 MgCl2, 10 HEPES, and 5 glucose, pH 7.4 (for 35 min at room temperature). The tissue was then washed in ice-cold, Ca2+-free amphibian saline containing 1% bovine serum albumin and deoxyribonuclease I (1 mg/ml; Worthington), followed by two additional washes in ice-cold, Ca2+-free saline. A piece of retina was triturated with a fire-polished Pasteur pipette, and the cell suspension was transferred onto glass slides or 1.78 refractive index coverslips (Olympus) coated with Cell-Tak (3.5 μg/cm2; BD Biosciences). After letting cells settle and adhere for 30 min at 4°C, they were superfused with oxygenated amphibian saline solution containing the following (in mm): 116 NaCl, 2.5 KCl, 1.8 CaCl2, 0.5 MgCl2, 10 HEPES, and 5 glucose, pH 7.8 (at room temperature). Ca2+ dyes were obtained from Invitrogen (Life Technologies). Unless otherwise specified, other reagents were obtained from Sigma-Aldrich. Rods were identified by their characteristic morphology including an axon and round or tear drop-shaped terminals. Intact terminals in retinal slices typically have a diameter of ∼5 μm. The terminals of isolated rods tended to flatten out against the coverslip to a thickness of 1–2 μm. Light-sensitive outer segments of rods were typically lost during trituration.
Imaging.
For FRAP, FLIP, and confined ryanodine puff experiments, isolated rods were studied on an inverted microscope (Olympus IX71) by either total internal reflection fluorescence (TIRF) or epifluorescence. For TIRF measurements of Ca2+ indicator dyes, a 488 nm solid-state laser (Melles Griot) was focused off-axis onto the back focal plane of a 1.65 numerical aperture (NA) objective (Apo 100× oil; Olympus). After leaving the objective, light traveled through a high refractive index (1.78) immersion fluid (Cargille Laboratories) and entered the coverslip, undergoing total internal reflection at the interface between the glass and lower refractive index of the cell membrane or overlying aqueous medium. The evanescent wave propagated at this interface had a length constant of 57 nm (Chen et al., 2013). Fluorescence emission was filtered by a 525 nm (45-nm-wide) bandpass filter (Semrock). For epifluorescence measurements, we used a 60×, 1.45 NA oil-immersion objective, illuminated isolated rods with 467–498 nm excitation light from a 120 W mercury lamp (XCite 120Q; Olympus), and collected the emitted fluorescence at 513–556 nm. Imaging data for both TIRF and epifluorescence experiments were acquired through an EMCCD camera (Hamamatsu ImageEM) at 31 ms/frame using MetaMorph software (Molecular Devices) and analyzed with MetaMorph or NIH ImageJ 1.46.
For monitoring Ca2+ changes during lengthy depolarizing steps applied to voltage-clamped rods or flash photolysis of caged Ca2+ compounds, fluorescence from Ca2+ dyes was collected through a 60×, 1.0 NA, water-immersion objective on an upright fixed-stage microscope (Nikon E600FN) equipped with a spinning-disk confocal scan head (Ultraview LC; PerkinElmer Life and Analytical Sciences). Excitation light at 488 or 568 nm was delivered from an argon/krypton laser, and emission was collected at 525 or 600 nm, respectively, by a cooled CCD camera (Hamamatsu OrcaER). Images were acquired and analyzed using PerkinElmer Imaging Suite version 5.5.
Retinal slice preparation and electrophysiology.
We used retinal slice preparations for some experiments. Details of slice preparation and electrophysiological recordings have been described previously (Van Hook and Thoreson, 2013). To monitor [Ca2+] changes in the cytoplasm and ER simultaneously during sustained depolarization, we loaded the ER with a low-affinity Ca2+ indicator dye, fluo-5N AM (Kd of 90 μm) and monitored cytoplasmic changes with a higher-affinity dye, rhod-2 (Kd of 570 nm), introduced through a whole-cell patch pipette. First, retinal slices (125 μm thick) were incubated with fluo-5N AM (10 μm) for 2.5 h at 4°C to load dye into both cytoplasm and ER. Retinal slices were then placed on the upright fixed-stage microscope and superfused with oxygenated amphibian saline. Rods were voltage clamped using an Axopatch 200B (Molecular Devices) patch-clamp amplifier. Currents were acquired and analyzed using pClamp 9.2 software with Digidata 1322 interface (Molecular Devices). Cells with holding currents >300 pA were excluded from analysis.
Recording pipettes were pulled on a PP-830 vertical puller (Narishige International) from borosilicate glass pipettes (1.2 mm outer diameter, 0.9 mm inner diameter, with internal filament; World Precision Instruments). Pipette resistance was 12–18 MΩ. Rod pipettes were filled with the following (in mm): 40 cesium glutamate, 50 cesium gluconate, 9.4 tetraethylammonium-Cl, 3.5 NaCl, 1 MgCl2, 9.4 MgATP, 0.5 GTP, 5 EGTA, 10 HEPES, and 0.1 rhod-2 tripotassium salt, pH 7.2. After obtaining the whole-cell recording configuration, fluo-5N was washed out of the cytoplasm but remained in the ER (Solovyova and Verkhratsky, 2002). At the same time, the higher-affinity rhod-2 was introduced into the cytoplasm through the patch pipette. We waited at least 5 min to wash fluo-5N completely out of the cytoplasm. In experiments in which we measured ER Ca2+ changes without measuring cytoplasmic Ca2+, we omitted rhod-2 from the pipette solution. Rods were depolarized for 17 s from −70 to −40 mV, similar to the photoreceptor membrane potential in darkness. rhod-2 and fluo-5N were alternately illuminated by 568 and 488 nm light, respectively. With 488 nm excitation/525 nm emission, rhod-2 fluorescence was 10.8% of the intensity observed with 568 nm excitation/600 nm emission. Therefore, we corrected fluo-5N fluorescence collected at 600 nm for bleed-through from rhod-2 fluorescence by subtracting a 10.8% scaled version of the rhod-2 fluorescence. There was no detectable fluo-5N fluorescence observed in the rhod-2 channel.
FRAP and FLIP experiments.
To examine Ca2+ diffusion within ER, we loaded the ER of isolated rods with fluo-5N as described above for retinal slices. Isolated rods were incubated with fluo-5N AM (10 μm) for 1 h at 4°C to load dye into the cytoplasm and ER. After loading, we obtained whole-cell patch-clamp recordings using a Ca2+- and dye-free pipette solution to wash dye out of the cytoplasm. To examine FRAP with cytoplasmic dye, we loaded rods with a high-affinity Ca2+ indicator by incubating cells with fluo-4 AM for 30 min at 4°C. In FRAP and FLIP experiments, terminals or somas of rods were photobleached for 1–4 s by illumination with a small spot (5–8 μm diameter) from a 30 mW, 488 nm laser. Loss and recovery of fluo-5N or fluo-4 fluorescence were monitored by epifluorescence.
The epifluorescent measurement light used to monitor fluorescence recovery sometimes produced additional photobleaching during the recovery phase. To compensate for bleaching by the epifluorescent measurement light, we measured fluorescence declines in somas of neighboring cells that were not photobleached by the laser spot and fit fluorescence declines in those cells with an exponential function. In rods loaded with fluo-5N that were subject to laser photobleaching, fitting data from only the 2 s period before laser photobleach yielded a fluorescence decline attributable to the epifluorescent measurement light of 15.5% in 12.5 s, with a time constant (τ) of 16 s. The decline in soma fluorescence measured throughout the entire trial in neighboring rods that were not photobleached by the laser exhibited a similar rate of bleaching by the epifluorescent measurement light (15% in 12.5 s; τ = 18 s). During FLIP experiments, we used weaker epifluorescent illumination that did not produce noticeable bleaching.
Depletion of Ca2+ in terminal ER by confined puff application of ryanodine.
To activate RyRs and thus stimulate CICR in the terminal of isolated rods, we used a spatially confined pressure ejection (Toohey) of ryanodine (30–100 μm) applied to terminals for 200 to 500 ms. Ryanodine promotes opening of RyRs at low micromolar concentrations (1–10 μm) but blocks RyRs at high micromolar concentrations (50–100 μm). However, the short puff duration reduced its effective concentration so that ryanodine puffs acted as RyR agonists. The tip of the puffer pipette was positioned ∼3 μm away from the terminal. At the terminal, the diameter of the puff expanded to 6–7.2 μm as measured with fluorescein puffs. To monitor [Ca2+] changes in the cytoplasm, we loaded isolated rods with fluo-5F AM (10 μm, Kd of 2.3 μm) at 4°C for 30 min. To monitor [Ca2+] changes in ER, we loaded isolated rods with fluo-5N AM for 1 h and obtained whole-cell recordings with a dye-free pipette solution to wash away cytoplasmic dye as described above.
Elevation of Ca2+ by localized photolysis of o-nitrophenyl-EGTA.
To examine diffusion of Ca2+ in the cytoplasm and ER, we used a caged Ca2+ compound, ο-nitrophenyl (NP)-EGTA (Invitrogen). Kd of NP-EGTA for Ca2+ increases from 80 to >1 mm during UV illumination. Chelated Ca2+ was released by a 1.5 ms UV light flash from a xenon arc flash lamp (JML-C2 Flash Lamp System; Rapp OptoElectronic) with a diameter of 6 μm and centered on the rod soma.
To examine diffusion of Ca2+ in the cytoplasm of isolated rods, cells were incubated with NP-EGTA AM (10 μm) and fluo-4 AM (10 μm) for 45 min at 4°C. In one set of experiments, retinal slices were incubated with NP-EGTA AM (10 μm), and then rods were voltage clamped with a pipette solution containing Oregon Green BAPTA-6F (OGB-6F; 500 μm) and ryanodine (2 μm). In recordings from slices, we used OGB-6F because it exhibits higher basal fluorescence than fluo-5F, making it easier to see rod terminals under confocal microscopy. Ryanodine was included in the pipette to stimulate the opening of RyR channels.
To examine diffusion of Ca2+ within the ER lumen, isolated rods were incubated with NP-EGTA AM (10 μm) and fluo-5N AM (10 μm) for 2.5 h at 4°C. For measurements of ER Ca2+ changes in these experiments, NP-EGTA and fluo-5N in cytoplasm were washed out of the cytoplasm by introducing dye-free intracellular solution through patch pipettes. To confirm that NP-EGTA was washed away completely from cytoplasm, we conducted control experiments in which we loaded ER with NP-EGTA and then introduced membrane-impermeant fluo-5F pentapotassium salt (100 μm) through the patch pipette to monitor [Ca2+] changes in the cytoplasm. We also conducted similar control experiments in rods from retinal slices using the dye OGB-6F (500 μm) after loading the ER with NP-EGTA-AM.
Statistical analysis.
We performed statistical analysis using GraphPad Prism 4. Results are presented as mean ± SEM, and statistical significance was determined using Student's t test. When comparing experiments involving multiple measurements (see Fig. 2), we compared the average of five data points from each cell in both conditions. We chose p < 0.05 to be the criterion for statistical significance.
Results
ER lumen is continuous from soma to terminal
ER appears to be present in the cell bodies, axons, and terminals of all vertebrate rods that have been examined, including rat, mouse, rabbit, frog, salamander, and teleost fish (De Robertis and Franchi, 1956; De Robertis, 1956; Ladman, 1958; Ungar et al., 1981; Mercurio and Holtzman, 1982; Freihöfer et al., 1990; Johnson et al., 2007; Babai et al., 2010). FRAP experiments using ER-tracker green dye that labels K+ channels in the ER membrane showed that the ER membrane in rods is continuous from terminal to soma (Chen et al., 2014). To test whether small molecules can diffuse within the ER lumen from soma to terminal, we loaded ER in isolated rods with a low-affinity Ca2+ indicator fluo-5N (Kd of 90 μm). We selected this dye because it can report the high levels of Ca2+ that are present in the ER (60–1000 μm; Bygrave and Benedetti, 1996; Michalak et al., 2002; Solovyova and Verkhratsky, 2002). After loading cells with fluo-5N AM, we obtained whole-cell patch-clamp recordings using a Ca2+- and dye-free pipette solution to wash fluo-5N out of the cytoplasm, leaving it within the ER lumen (Solovyova and Verkhratsky, 2002; Chen et al., 2014). After waiting for 5 min after patch rupture, depolarizing voltage steps caused intraterminal fluo-5N fluorescence to decrease (Chen et al., 2014). This depolarization-evoked decline in fluo-5N fluorescence showed that Ca2+ leaves the ER during activation of CICR and confirmed that Ca2+ dye was washed out of the cytoplasm (Chen et al., 2014).
As shown in Figure 1A, isolated rods were incubated with fluo-5N AM for 1 h and then patched with a dye-free pipette solution to wash dye out of the cytoplasm but not the ER. Axonal ER is very thin, and so its fluorescence is very faint. We bleached the terminal using a spatially confined laser spot (488 nm, 30 mW, 8 μm in diameter; Fig. 1A, white circle) applied for 3 s while monitoring changes in fluo-5N fluorescence. Rod terminals that were photobleached showed almost complete recovery of fluorescence (Fig. 1B). Bleaching by the epifluorescent measurement light during the recovery phase was estimated and corrected with an exponential decay function (see Materials and Methods). After correction, the time constant of fluorescence recovery in the terminal averaged 6.3 ± 0.95 s (n = 8; Table 1).
We also conducted FRAP experiments to examine diffusion of dye through the cytoplasm. For these experiments, we used a higher-affinity dye (fluo-4 AM), loaded rods for only 30 min to preferentially load cytoplasm, and did not voltage clamp the cells. After laser photobleach of terminal fluorescence for 1–3 s, partial recovery of cytoplasmic fluorescence was observed (Fig. 1C) that averaged τ = 2.3 ± 0.54 s (n = 5; Table 1). This is approximately threefold faster than FRAP of fluo-5N in the ER. The rapid kinetics of cytoplasmic recovery suggests that dye diffused back into the terminal during the 1–3 s bleaching period. Rapid return of dye during the photobleach period would explain the modest bleaching produced by the laser. Bleaching of additional returning dye would also explain why there was only partial recovery of fluorescence after the bleaching period. The faster recovery of fluorescence in the cytoplasm compared with ER is presumably attributable to a lower degree of tortuosity.
Additional confirmation of continuity between soma and terminal ER was provided by FLIP experiments. As in Figure 1A–C, we loaded isolated rods with fluo-5N for 1 h and then patch clamped them to wash dye out of the cytoplasm. We bleached dye in a local region of somatic ER by applying a 488 nm laser spot (5 μm; Fig. 1D,E). In the example shown in Figure 1, fluorescence declined in the bleached area with a time constant of 2 s. Fluorescence also declined in the unbleached terminal, and this fluorescence loss was delayed with respect to that of the soma (Fig. 1D,E), consistent with dye diffusion from terminal to soma during the photobleach period. Similar delayed FLIP of terminal ER fluorescence was observed while photobleaching the somas of eight rods. In summary, FRAP and FLIP experiments showed that small molecules can move freely between the soma and terminal within the ER lumen.
Depletion of terminal ER Ca2+ during long depolarizing steps reduced somatic ER Ca2+
In darkness, rods have a resting membrane potential of approximately −40 mV, stimulating the opening of voltage-gated Ca2+ channels in the terminal. The influx of Ca2+ through open Ca2+ channels in turn activates CICR in the terminal. The ability of rods to release glutamate-filled vesicles continuously in darkness indicates that intraterminal [Ca2+] must remain elevated indefinitely. To determine whether this sustained elevation of intraterminal [Ca2+] involves sustained activation of CICR, we loaded rods with both fluo-5N in the ER and rhod-2 in the cytoplasm to examine Ca2+ changes in the two cell compartments simultaneously. The low-affinity Ca2+ indicator fluo-5N AM was loaded into ER and cytoplasm by incubating retinal slices with the dye. Rods were then patched with a pipette containing rhod-2 (100 μm) to wash fluo-5N out of the cytoplasm and replace it with rhod-2 (Fig. 2A). [Ca2+] changes in ER and cytoplasm were then monitored by alternately measuring fluo-5N fluorescence with a 488 nm laser and rhod-2 fluorescence with a 568 nm laser on a spinning-disk confocal microscope. Cytoplasmic rhod-2 fluorescence increased slowly in the terminal during a 17 s depolarizing step from −70 to −40 mV (p = 0.013, n = 6; Fig. 2B, red), attaining a plateau after ∼12 s. At the same time, fluo-5N fluorescence in terminal ER declined (p = 0.04, n = 6; Fig. 2C, red), showing a progressive depletion of Ca2+ stores that also appeared to plateau after 12 s. The parallel between the rise in cytoplasmic Ca2+ and decline in ER Ca2+ suggests that the rise in cytoplasmic Ca2+ was not only attributable to Ca2+ influx through membrane Ca2+ channels but also involved continued Ca2+ release from ER stores.
Similar to the terminal, cytoplasmic [Ca2+] also rose in the soma (p = 0.004, n = 6; Fig. 2B, black) whereas ER [Ca2+] declined (p = 0.0008, n = 6; Fig. 2C, black) during sustained depolarization. However, the depolarization-evoked increase in cytoplasmic [Ca2+] in the soma was only ∼3% of the cytoplasmic [Ca2+] increase observed in the terminal, as measured by ΔF/F with rhod-2 (p = 0.004, n = 6; Fig. 2B, black). In comparison, the decline in somatic ER [Ca2+] measured with fluo-5N (Fig. 2C, black) was a larger fraction (∼25%) of the decline in terminal ER [Ca2+] (n = 6). The decline in terminal ER [Ca2+] also appeared larger than the decline in somatic ER [Ca2+], although the difference did not attain statistical significance (p = 0.12). Although differing dye Ca2+ affinities and differences in the volumes of ER and cytoplasm make precise quantitative comparisons difficult, the 3% increase in cytoplasmic ΔF/F observed with rhod-2 (Kd of 570 nm) is likely to involve a smaller Ca2+ change than the 25% decrease in ER ΔF/F observed with the much lower-affinity dye fluo-5N (Kd of 90 μm). This suggests that not all of the Ca2+ ions that leave the somatic ER enter the surrounding cytoplasm, but instead some of them diffuse through ER to the terminal. In addition to release of Ca2+ from somatic ER, the rise of cytoplasmic [Ca2+] in the soma might also involve diffusion of Ca2+ through the cytoplasm from the terminal or entry of Ca2+ through membrane channels in the soma (e.g., L-type Ca2+ channels or store-operated channels). The evidence for smaller ER [Ca2+] declines in the soma than in the terminal suggests that, with maintained depolarization, a gradient of [Ca2+] develops within the ER, with higher levels in the soma and lower levels in the terminal. This concentration gradient would help to drive diffusion of Ca2+ through the ER from soma to terminal and thereby help to sustain CICR in the terminal during maintained depolarization.
Decreasing terminal ER Ca2+ caused a secondary decrease in somatic ER Ca2+
To test further whether Ca2+ can diffuse from soma to terminal within the ER, we depleted Ca2+ locally from intraterminal ER by applying spatially confined puffs of ryanodine (30–100 μm, 6–7.5 μm diameter) to terminals of isolated rods. To test whether ryanodine puffs could activate terminal RyRs without also activating RyRs in the soma, we loaded isolated rods with a Ca2+ indicator fluo-5F (Kd of 2.3 μm) to monitor cytoplasmic [Ca2+] changes (Fig. 3A,B). We then puffed ryanodine (30–100 μm) onto the terminal (Fig. 3A) or soma (Fig. 3B) for 200–500 ms. Use of a short puff duration reduced its effective concentration so that ryanodine acted as an agonist at RyRs. In most cells, puffing ryanodine (30–100 μm) directly onto terminals stimulated an elevation of terminal cytoplasmic [Ca2+] (30 μm, 9 of 12; 100 μm, 17 of 23 cells; Fig. 3A, open red circles), consistent with release of Ca2+ from ER. Typically, there was no change in somatic [Ca2+] after a puff onto the terminal (Fig. 3B, filled black circles), although in some cells we observed a small cytoplasmic [Ca2+] increase in the soma (data not shown). The small increase in somatic [Ca2+] in these cells was probably secondary to intracellular diffusion of Ca2+ from the terminal because direct application of ryanodine puffs to the soma did not stimulate Ca2+ increases in the soma (Fig. 3B) regardless of whether puffs were applied before (100 μm, 0 of 14 cells) or after (30 μm, 0 of 3 cells; 100 μm, 0 of 9 cells) puffs to the terminal. Although puffing ryanodine onto the soma did not stimulate increases in somatic [Ca2+], on a few occasions, puffing ryanodine onto the soma stimulated intraterminal [Ca2+] increases (30 μm, 0 of 3 cells; 100 μm, 3 of 23 cells). These data indicate that Ca2+ release can be stimulated more easily in the terminal than soma and show that confined puffs to the terminal can directly activate RyRs in that compartment.
Next, we measured Ca2+ changes in the ER by incubating cells with fluo-5N AM (Kd of 90 μm) and then washing dye out of the cytoplasm by introducing a dye-free solution into the rod through a patch pipette (Fig. 3C). We found that ER [Ca2+] decreased within the terminal after a localized ryanodine puff to the terminal (normalized ΔF/F = −0.25 ± 0.046, p = 0.0006, n = 8), consistent with release of Ca2+ from ER to cytoplasm by activation of RyRs. Localized activation of RyR-mediated Ca2+ release in the terminal also caused a secondary decline in somatic ER [Ca2+] (normalized ΔF/F = −0.064 ± 0.018, p = 0.0033, n = 8). The decline in somatic [Ca2+] developed more slowly than the decline in intraterminal [Ca2+], consistent with the possibility that it resulted from diffusion of Ca2+ through the ER from soma to terminal. The decline in the soma was smaller than that in the terminal (p = 0.0058, n = 8), consistent with a larger Ca2+ store in the soma than the terminal. Figure 3C shows average data from eight cells visualized by TIRF microscopy. When we measured fluo-5N fluorescence changes using epifluorescent illumination, the ΔF/F changes were smaller (−0.036 ± 0.013, n = 6, p = 0.0018 compared with fluorescence changes monitored by TIRF), suggesting that spatially averaged changes are smaller than local submembrane changes in ER Ca2+. The finding that depletion of terminal ER Ca2+ is followed by a delayed decline in somatic ER Ca2+ supports the hypothesis that terminal and somatic ER form a single interconnected Ca2+ store in rod photoreceptors.
Increasing somatic ER Ca2+ caused a secondary rise in terminal ER Ca2+
We also tested interconnectedness of the ER lumen by determining whether increasing [Ca2+] in somatic ER causes a secondary increase in terminal ER. To do so, we loaded the ER with a caged Ca2+ compound, NP-EGTA, that releases Ca2+ during photolysis by UV light. We then flashed a small spot (6 μm diameter) of UV light onto the soma to uncage Ca2+ only in the soma. To load the ER with NP-EGTA, we incubated rods with NP-EGTA AM for 2.5 h and then washed NP-EGTA out of the cytoplasm by introducing a drug-free solution into the cell through a whole-cell patch pipette. To test whether NP-EGTA was successfully washed out of the cytoplasm, we performed control experiments in which we monitored cytoplasmic [Ca2+] changes by including membrane-impermeant fluo-5F or OGB-6F salts in the pipette solution. After waiting at least 3 min after patch rupture, flashing a confined spot of UV light (6 μm diameter) onto the soma did not evoke [Ca2+] increases in either somatic or terminal cytoplasm of isolated rods (n = 6; Fig. 4A) or rods from retinal slices (n = 6). The transients in the example in Figure 4A were artifacts of the brief uncaging flashes.
We next loaded the ER of rods from retinal slices with NP-EGTA and monitored cytoplasmic Ca2+ with OGB-6F salts introduced through the patch pipette. This is similar to the experimental configuration described in Figure 4A except that we also included 2 μm ryanodine in the patch pipette solution to activate RyRs and thus permit Ca2+ to exit from the ER into the cytoplasm. When ryanodine was included in the pipette solution, localized uncaging of Ca2+ in the somatic ER by application of a spatially confined UV light flash stimulated an abrupt increase in cytoplasmic Ca2+ of the soma (Fig. 4B, filled black circles), followed by a secondary slow rise in terminal cytoplasmic Ca2+ (Fig. 4B, open red circles). As shown by the control experiments in Figure 4A, this increase in cytoplasmic Ca2+ was not attributable to residual NP-EGTA in the cytoplasm but was instead a consequence of uncaging Ca2+ within the ER, followed by diffusion of Ca2+ ions out of the ER and into the cytoplasm through open RyR channels. The slow increase in terminal cytoplasmic Ca2+ was attributable to diffusion of Ca2+ down the axon. This could be a result of diffusion through the cytoplasm or diffusion through the ER lumen, followed by exit into the cytoplasm through open RyR channels in the terminal.
We estimated the diffusion coefficient for Ca2+ movement down the axon by modeling this process as diffusion through a pipe using the following formula (Berg, 1983): F(t) = Fmax/2[1 − erf(x/4Dt)1/2)], where F(t) is the fluorescence at each point in time, Fmax is the plateau fluorescence value, x is the distance from the soma/axon boundary to the center of the terminal, and t is time. The error function (erf) was approximated numerically (Abramowitz and Stegun, 1972, their Eq. 7.1.27), and data were fit by nonlinear regression. This analysis yielded a diffusion coefficient (DCa2+Mix) of 30.6 ± 4.1 μm2/s (n = 5; Fig. 4B, blue line; Table 1).
We next studied [Ca2+] changes within the ER by loading the ER of isolated rods with both fluo-5N AM and NP-EGTA AM for 2.5 h and then washing both compounds out of the cytoplasm by introducing a dye- and drug-free patch pipette solution. Photolytic uncaging of NP-EGTA in the soma with a spatially confined UV light flash stimulated an abrupt, large Ca2+ increase in somatic ER (Fig. 4C, black). This direct elevation of Ca2+ in somatic ER caused by localized photolytic activation of NP-EGTA in somatic ER was followed by a slower secondary increase in terminal ER Ca2+ (Fig. 4C, red). The slow rise in terminal ER Ca2+ that followed Ca2+ uncaging in somatic ER was fit with equation above, yielding DCa2+ER of 23.1 ± 2.9 μm2/s (n = 14; Fig. 4C; Table 1). Uncaging flashes applied directly to the terminal evoked only small intraterminal Ca2+ increases that were not sufficient to stimulate a detectable secondary Ca2+ increase in the soma (n = 3 cells).
For comparison, we also measured diffusion of Ca2+ through the cytoplasm by loading isolated rods with both fluo-4 AM and NP-EGTA-AM for 1 h (Fig. 4D). The shorter incubation time was used to minimize loading into the ER, and the higher Ca2+ affinity fluo-4 was chosen to limit measurements to cytoplasmic Ca2+ changes because high levels of Ca2+ in the ER (Bygrave and Benedetti, 1996) would be expected to saturate fluo-4. Uncaging Ca2+ from cytoplasmic NP-EGTA in the soma stimulated an abrupt increase in cytoplasmic [Ca2+] in the soma (Fig. 4D, black). This abrupt rise in somatic [Ca2+] was followed by a delayed increase in terminal cytoplasmic [Ca2+] (Fig. 4D, red). The slower increase in terminal cytoplasmic [Ca2+] caused by uncaging Ca2+ in soma cytoplasm was presumably attributable to diffusion of Ca2+ through axonal cytoplasm. Fitting the intraterminal increase in cytoplasmic [Ca2+] with equation above (Fig. 4D, blue line) yielded a diffusion coefficient (DCa2+Cyto) of 32.6 ± 7.3 μm2/s (n = 8; Table 1) for Ca2+ movement from soma to terminal through the cytoplasm.
The diffusion coefficient values found in the three uncaging experiments were similar to one another (p = 0.26, ANOVA; Table 1), and so differences in diffusion coefficient could not be used to distinguish whether the secondary increase in terminal cytoplasmic Ca2+ in the presence of open RyRs (Fig. 4B) was attributable to diffusion of Ca2+ through axonal cytoplasm, axonal ER, or both. The finding that DCa2+Cyto and DCa2+ER were similar to one another (p = 0.17, t test) suggests that Ca2+ ions can move with relative freedom through the ER from soma to terminal.
Discussion
ER forms a single continuous Ca2+ store throughout rods
FRAP and FLIP experiments with fluo-5N confirmed that the ER forms a continuous structure in rods that allows for the diffusion of small molecules from soma to terminal (Mercurio and Holtzman, 1982; Ungar et al., 1984; Chen et al., 2014). The time constant for recovery of fluo-5N fluorescence within the ER after localized photobleach was threefold slower than recovery of cytoplasmic fluo-4 fluorescence, presumably because of the greater tortuosity of the ER lumen compared with the cytoplasm. Similarly, diffusion of fluo-5N through the sarcoplasmic reticulum (SR) of cardiac myocytes is 3- to fourfold slower than cytoplasmic diffusion (Wu and Bers, 2006). Although dye diffusion coefficients differed between ER and cytoplasm, the Ca2+ diffusion coefficients DCa2+ER (23 μm2/s) and DCa2+Cyto (33 μm2/s) were not significantly different. DCa2+Cyto was similar to a previous estimate derived from the spread of Ca2+ waves through rod terminals (34–40 μm2/s; Cadetti et al., 2006). In cardiac myocytes, Swietach et al. (2008) calculated DCa2+ in SR of 8–9 μm2/s, slightly less than the cytoplasmic DCa2+ of 14 μm2/s (Wu and Bers, 2006). However, other studies in cardiac myocytes have found a much higher value for DCa2+ in the SR (60 μm2/s; Wu and Bers, 2006; Picht et al., 2011). Studies on pancreatic acinar cells also found that Ca2+ diffuses through the ER more freely than through the cytoplasm (Park et al., 2000). This is thought to be attributable to weaker Ca2+ binding in the ER (Mogami et al., 1999; Wu and Bers, 2006; Picht et al., 2011). The finding that DCa2+ER and DCa2+Cyto of rods are similar to one another suggests that the low affinity of Ca2+ binding sites within the ER may compensate for effects of spatial tortuosity and allow relatively free movement of Ca2+ through the ER.
Ca2+ in somatic ER helps replenish depleted Ca2+ stores in terminals
We found that localized depletion of intraterminal ER Ca2+ (stimulated by local ryanodine puffs or modest membrane depolarization) was followed by a secondary reduction in somatic ER Ca2+. The decrease in ER Ca2+ in the soma appeared larger than the corresponding increase in cytoplasmic Ca2+ in the soma. This could reflect dilution of Ca2+ into the cytoplasmic volume but could also be explained by diffusion of Ca2+ through the ER from soma to terminal. With maintained depolarization, there was a greater Ca2+ decline in terminal ER than somatic ER, indicating that a gradient of Ca2+ developed between the soma and terminal ER. This concentration gradient would drive Ca2+ through the ER from soma to terminal. Because Ca2+ ions diffuse through the ER from the soma to replenish ions depleted from terminal ER during CICR, Ca2+ may be simultaneously restored to the ER in the soma and other parts of the cell by store-operated Ca2+ entry (SOCE) across the plasma membrane (Szikra et al., 2008; García-Sancho, 2014) and uptake of Ca2+ into the ER via SERCA2 (Krizaj, 2005; Szikra and Krizaj, 2007). Although we did not investigate this possibility directly, the small slow Ca2+ increase in soma cytoplasm observed after activation of CICR and depletion of Ca2+ stores in the terminal may involve SOCE. Contributions of SOCE to maintaining ER Ca2+ levels and the diffusion of Ca2+ from soma to terminal may explain why blocking SOCE channels, like blocking CICR, inhibited slower components of glutamate release from rods but had little effect on fast, transient release evoked by brief depolarizing steps (Szikra et al., 2008).
In rods, Ca2+ released from terminal ER stores can trigger vesicle fusion at both ribbon and non-ribbon sites (Suryanarayanan and Slaughter, 2006; Chen et al., 2014), and at least 50% of the sustained glutamate release from both mammalian and amphibian rods in darkness appears to be driven by CICR (Cadetti et al., 2006; Suryanarayanan and Slaughter, 2006; Babai et al., 2010). Release from cones occurs only at ribbon sites and does not involve CICR (Cadetti et al., 2006; Snellman et al., 2011). The continued influx of Ca2+ through L-type Ca2+ channels located deep within invaginating rod synapses causes a decline in synaptic cleft Ca2+ levels during sustained depolarization that is sufficient to cause a large reduction in ICa amplitude (Rabl and Thoreson, 2002). The remaining small influx of Ca2+ during sustained depolarization can be amplified by CICR (Krizaj et al., 1999, 2003), which in turn amplifies release (Cadetti et al., 2006; Suryanarayanan and Slaughter, 2006; Babai et al., 2010). The soma-to-terminal Ca2+ gradient that develops during sustained depolarization and the ability of Ca2+ to diffuse freely through the ER promotes the continuous refilling of intraterminal Ca2+ stores required to sustain CICR in rod terminals indefinitely during long periods of darkness. This mechanism of Ca2+ tunneling from soma to terminal through the ER appears to be essential for maintaining synaptic release from rods in darkness. Choi et al. (2006) showed that the ER extends from the soma into the dendrites of neurons. Our results showed that the ER network also extends from the soma into presynaptic terminals. This is consistent with findings from a number of other neurons showing submembrane cisterns of ER in both presynaptic and postsynaptic processes (Bouchard et al., 2003; Fuchs et al., 2014; Segal and Korkotian, 2014). CICR triggers release directly in rods (Suryanarayanan and Slaughter, 2006; Chen et al., 2014), in part because of the submicromolar affinity of the exocytotic Ca2+ sensor in photoreceptors (Thoreson et al., 2004). It is unclear whether CICR is capable of triggering release directly in other neurons, but it can provide a source of Ca2+ to enhance release (Verkhratsky, 2005; for review, see Castellano-Muñoz and Ricci, 2014). CICR in presynaptic terminals also contributes to synaptic plasticity and synaptic dysfunction in neurodegenerative diseases (for review, see Stutzmann and Mattson, 2011). In addition to providing a source of Ca2+, the ER can also serve as a Ca2+ sink under certain conditions, limiting the influence of Ca2+ entering through ion channels (Castonguay and Robitaille, 2001; Im et al., 2014).
Our study focused on the ability of ER Ca2+ movements to sustain synaptic release, but free movement of Ca2+ through the ER can also communicate synaptic Ca2+ changes back to the soma to influence a diverse array of processes, including mitochondrial function, gene expression, and protein folding (Verkhratsky, 2005; Araki and Nagata, 2011; Kaufman and Malhotra, 2014). Bidirectional communication of Ca2+ within the ER between soma and terminal can impart beneficial adaptability to neurons but also contribute to damaging apoptotic and ER stress responses (Verkhratsky, 2005).
Footnotes
This research was supported by National Institutes of Health Grants R01EY10542 (W.B.T.) and F32EY023864 (M.J.V.H.), a Senior Scientific Investigator Award from Research to Prevent Blindness (W.B.T.), and a University of Nebraska Medical Center Graduate Fellowship (M.C.).
The authors declare no competing financial interests.
- Correspondence should be addressed to Dr. Wallace B. Thoreson, Department of Ophthalmology and Visual Sciences, University of Nebraska Medical Center, 4050 Durham Research Center I, Omaha, NE 68198-5840. wbthores{at}unmc.edu