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

J. Biol. Chem., Vol. 276, Issue 31, 29430-29439, August 3, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/31/29430    most recent M103274200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Arnaudeau, S. Articles by Demaurex, N. Search for Related Content PubMed PubMed Citation Articles by Arnaudeau, S. Articles by Demaurex, N. Social Bookmarking                      What's this?

Mitochondria Recycle Ca²⁺ to the Endoplasmic Reticulum and Prevent the Depletion of Neighboring Endoplasmic Reticulum Regions^*,

Serge Arnaudeau, William L. Kelley^§, John V. Walsh Jr.^¶, and Nicolas Demaurex

From the  Department of Physiology, University of Geneva, 1211 Geneva 4, Switzerland, the ^§ Division of Infectious Diseases, Geneva University Hospitals, 1211 Geneva 4, Switzerland, and the ^¶ Department of Physiology, University of Massachusetts Medical School, Worcester, Massachusetts 01605

Received for publication, April 12, 2001, and in revised form, May 14, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

To study Ca²⁺ fluxes between mitochondria and the endoplasmic reticulum (ER), we used "cameleon" indicators targeted to the cytosol, the ER lumen, and the mitochondrial matrix. High affinity mitochondrial probes saturated in ~20% of mitochondria during histamine stimulation of HeLa cells, whereas a low affinity probe reported averaged peak values of 106 ± 5 µM, indicating that Ca²⁺ transients reach high levels in a fraction of mitochondria. In concurrent ER measurements, [Ca²⁺]_ER averaged 371 ± 21 µM at rest and decreased to 133 ± 14 µM and 59 ± 5 µM upon stimulation with histamine and thapsigargin, respectively, indicating that substantial ER refilling occur during agonist stimulation. A larger ER depletion was observed when mitochondrial Ca²⁺ uptake was prevented by oligomycin and rotenone or when Ca²⁺ efflux from mitochondria was blocked by CGP 37157, indicating that some of the Ca²⁺ taken up by mitochondria is re-used for ER refilling. Accordingly, ER regions close to mitochondria released less Ca²⁺ than ER regions lacking mitochondria. The ER heterogeneity was abolished by thapsigargin, oligomycin/rotenone, or CGP 37157, indicating that mitochondrial Ca²⁺ uptake locally modulate ER refilling. These observations indicate that some mitochondria are very close to the sites of Ca²⁺ release and recycle a substantial portion of the captured Ca²⁺ back to vicinal ER domains. The distance between the two organelles thus determines both the amplitude of mitochondrial Ca²⁺ signals and the filling state of neighboring ER regions.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The calcium ion is a ubiquitous intracellular messenger that controls processes ranging from fertilization and cellular differentiation to muscle contraction and synaptic transmission (1, 2). The finely regulated spatial and temporal encoding of the calcium signal ensures that these various, and sometimes opposite, calcium-dependent processes are activated at the appropriate time and place within cells (3-5). Whereas the influx of calcium through voltage-dependent membrane channels triggers rapid secretion at synapses (6), the release of Ca²⁺ from stores in response to inositol 1,4,5-trisphosphate (IP₃)¹ can generate sustained [Ca²⁺]_cyt oscillations in both excitable and nonexcitable cells (7). These calcium oscillations can be decoded in the cytosol by frequency-sensitive effector proteins such as calmodulin-dependent kinase II (8), and have been shown to optimize both secretion (9) and gene expression (10, 11). Calcium oscillations can also be decoded by mitochondria (12), several dehydrogenases being activated as the free [Ca²⁺] increases within the mitochondrial matrix, thereby increasing the level of NAD(P)H and the production of ATP to meet the cell energy demand (12, 13).

In addition to being able to decode Ca²⁺ oscillations, mitochondria also participate actively in calcium signaling (reviewed in Refs. 14-16). Mitochondria take up calcium very efficiently and contribute to the local nature of the calcium signal by acting as a buffer barrier between cellular regions (17). Importantly, mitochondria are often in close contact with Ca²⁺ release sites in the endoplasmic reticulum (18, 19) or with Ca²⁺ influx channels at the plasma membrane (20). By acting as a Ca²⁺ buffering system at these strategic locations, mitochondria can modulate the rate of Ca²⁺ release by IP₃ receptors (21, 22) or the rate of capacitative Ca²⁺ entry through CRAC channels (20). Through this intimate connection with the calcium sources, mitochondria strongly shape calcium signals and, depending on the cellular context, can either potentiate or inhibit Ca²⁺ oscillations (23-26).

Our understanding of the calcium homeostasis of intracellular compartments is still incomplete, because the highly dynamic Ca²⁺ signals occurring within organelles are difficult to measure. Trapped fluorescent dyes such as Mag-fura and Mag-indo-1 have been used to measure calcium within the ER and mitochondria (27-31). However, these dyes are not specifically targeted and their selectivity for calcium over magnesium is poor. The cationic probe rhod2 has also been used to measure calcium within mitochondria (12, 20, 21, 24) but its specificity relies on the negative membrane potential of this organelle. The calcium-sensitive photoprotein aequorin, on the other hand, can be specifically targeted (32, 33) and has been used extensively to measure calcium dynamics within the mitochondria (13, 34-36), the ER (37-41), and the Golgi complex (42). However, the weak luminescence of the photoprotein and its irreversible consumption upon Ca²⁺ binding severely limits the use of this approach for calcium imaging. The "cameleon" indicators based on green fluorescent proteins and calmodulin developed in the group of R. Y. Tsien (43, 44) appear better suited for calcium measurements in organelles. The bright fluorescence of the green fluorescent protein mutants, combined with selective targeting sequences, allows one to visualize the calcium signals in organelles by fluorescence ratio imaging (45-47). Furthermore, the calcium affinity of calmodulin can be adjusted by molecular engineering, enabling one to match the calcium concentration within the organelle of interest (43). Despite these advantages, the cameleons have not yet found widespread applications, probably because their limited dynamic range and pH dependence requires careful in situ calibration to achieve quantitative Ca²⁺ measurements.

In this study, we used yellow cameleons Ca²⁺ indicators to measure Ca²⁺ signals in the cytosol, [Ca²⁺]_cyt, the endoplasmic reticulum, [Ca²⁺]_ER, and the mitochondria, [Ca²⁺]_mit in HeLa cells. Probes of different Ca²⁺ affinities and pH dependence were used (YC2, YC4_ER, YC2_mit, YC3.1_mit, and YC4.1_mit) and calibrated in situ, providing accurate estimates of the free Ca²⁺ concentration within the ER lumen and the mitochondrial matrix. Using this approach, we show that [Ca²⁺]_mit transients reach >100 µM in about 25% of mitochondria, and that part of the captured Ca²⁺ is returned back to the ER. The local cycle of Ca²⁺ between these two organelles prevents the depletion of ER regions bearing mitochondria, thereby generating two functionally distinct Ca²⁺ stores within the ER network.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Dulbecco's modified Eagle's culture medium, fetal calf serum, penicillin, streptomycin, geneticin were obtained from Life Technologies, Inc. (Paisley, Scotland). Histamine, epinephrine, thapsigargin, nigericin, monensin, ATP, and HEPES were purchased from Sigma. Ionomycin, EGTA, and HEEDTA were obtained from Fluka (Buchs, Switzerland). CGP 37157 was from Tocris (Bristol, United Kingdom). Transfast transfection reagent was purchased from Promega (Catalys AG, Switzerland). All other chemicals were of analytic grade and were obtained from Fluka or Sigma. The "Ca²⁺ medium" contained 140 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, 10 mM glucose, and 20 mM Hepes, pH 7.4. The "Ca²⁺-free medium" contained no CaCl₂ and 0.5 mM EGTA. Drugs were dissolved in dimethyl sulfoxide or ethanol and diluted in the recording medium on the day of use, at a final solvent concentration <0.1%.

Constructs-- Plasmids YC2, YC2.1, and YC4_ER were kindly provided by Dr. R. Y. Tsien. Plasmid YC2_mit was generated by ligation of the NotI YC2 insert into pCMV/myc/Mito (Invitrogen). Plasmid YC2.1_mit was then generated by exchanging the EYFP fragments of YC2.1 and YC2_mit using SacI and BstX1. Plasmids YC3_mit, YC3.1_mit, YC4_mit, and YC4.1_mit were prepared by the Quick Change method using Pfu polymerase (Stratagene) and YC2_mit or YC2.1_mit as templates. Complementary primer pairs: E31Q-A: 5'-gacggcaccatcaccacaaagcagctgggcaccgttatgaggtcgc-3'; E31Q-B: 5'-gcgacctcataacggtgcccagctgctttgtggtgatggtgccgtc-3'; E104Q-A: 5'-aacggctacatcagcgctgctcagctgcgtcacgtcatgacaaacc-3'; E104Q-B: 5'-ggtttgtcatgacgtgacgcagctgagcagcgctgatgtagccgtt-3', were designed for each mutant E31Q or E104Q in the calmodulin module. For convenience, a PvuII restriction site was introduced (underlined) with each mutant by simultaneously changing the adjacent leucine codon L32 or L105 to an alternative codon with high relative abundance in mammalian usage tables. Following 12 PCR cycles with annealing at 60 °C and extension at 68 °C, products were digested with DpnI, then electroporated into strain JM109. Transformants were first screened by NotI and PvuII digestion. The entire cameleon coding sequence for each mutant was then verified using dideoxy dye termination sequencing and appropriate primers. Plasmid DNAs were purified using a Qiagen column by the Maxi-prep purification protocol recommended by the manufacturer.

Cell Culture and Transfection-- HeLa and HEK-293 cells (purchased from American type Culture Collection, Rockville, MD) were grown in Dulbecco's modified Eagle's medium containing 10% heat-inactivated fetal calf serum, 50 units/ml penicillin, 50 µg/ml streptomycin, and were maintained in a humidified incubator at 37 °C in the presence of 5% CO₂, 95% air. Cells (~200,000) were plated on 25-mm glass coverslips. After they reached 60% of confluence, cells were transiently transfected with cDNAs encoding the yellow cameleons probes, using the calcium phosphate protocol for HEK-293 cells or Transfast reagent for HeLa cells. Cells were imaged 3 to 5 days after transfection. Stable HEK-293 transfectants were grown in the presence of geneticin (100 µg/ml) for 3 weeks and ~20 clones were expanded for each condition and tested for expression of the probes.

[Ca²⁺] Measurements-- Cells plated on 25-mm coverslips were superfused at 37 °C in a thermostatic chamber (Harvard Apparatus, Holliston, MA) equipped with gravity feed inlets and vacuum outlet for solution changes. The method for dual-emission ratio imaging of [Ca²⁺] with the use of cameleons was derived from Ref. 43. Cameleon fluorescence from cells was imaged on a Axiovert S100 TV using a 100X, 1.3 NA oil-immersion objective (Carl Zeiss AG, Feldbach, Switzerland). Cells were excited by the 430 ± 10 nm line from a monochromator (DeltaRam, Photon Technology International Inc., Monmouth Junction, NJ) through a 455DRLP dichroic mirror. Fluorescence emission from the cameleons was imaged using a cooled 12-bit CCD interlined camera (Visicam, Visitron System, Germany) at two emission wavelengths, using a filterwheel (Ludl Electronic Products, Hawthorn, NY) to alternatively change the two emission filters (475DF15 and 535DF25, Omega Optical, Brattleboro, VT). Image acquisition and analysis was performed with the Metamorph/Metafluor 3.5 software (Universal Imaging, West Chester, PA). Images were stored on optical discs for later analysis and archiving. Changes in fluorescence ratio R = (fluorescence intensity at 535 nm  background intensity at 535 nm)/(fluorescence intensity at 475 nm  background intensity at 475 nm) were calibrated in [Ca²⁺] using the equation,

(Eq. 1)

where R_max and R_min are the ratios obtained, respectively, in the absence of Ca²⁺ and at saturating Ca²⁺. K'_d is the apparent dissociation constant and n is the Hill coefficient of the Ca²⁺ calibrations curves obtained in situ for each cameleon. For optimal representation, wide-field fluorescence image stacks were deconvoluted after acquisition by treatment on a Silicon Graphics Octane work station using the Huygens 2 software (Bitplane AG, Zurich, Switzerland). For spectroscopic measurements, YC2 was extracted from stable transfectants by sonication. Membranes were removed by centrifugation (10,000 rpm, 10 min) and the supernatant diluted 1:8 before measuring the spectral response on a LS-5 fluorimeter (PerkinElmer Life Sciences).

Image Analysis-- The percentage of saturated pixels (Fig. 3) was determined image by image using a Metamorph routine, using a minimal size criteria of 5 contiguous pixels. A low-intensity threshold was used to define the cell-associated fluorescence signal (total pixel area) and a high-intensity threshold, corresponding to the average intensity of the total pixel area at saturating Ca²⁺ concentration, was applied to extract the saturated regions. To define ER regions close to mitochondria (Figs. 6 and 7), mitotracker red images were acquired just before and after agonist application. The location of mitochondria did not change significantly during this 2-min period. The red image was aligned to the yellow cameleon image using the nucleus boundaries, and used as a mask to define the mitochondria-associated ER region, which averaged 62% of the total ER area. To avoid the shading artifacts in the ratio images caused by the higher motility of the organelle edges, [Ca²⁺]_ER was measured as F535 fluorescence. The F535 data were spatially averaged within the two regions and normalized to the fluorescence at the start of the experiment (F/F₀). To compare ER regions located at similar distances from the cell border, the cell periphery was defined on the ER image and eroded by 5 and 10 pixels (1.3-2.6 µM). The nuclear area was excluded and the mitotracker image used to define mitochondria-rich ER regions.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Targeting of Probes to Organelles-- To measure Ca²⁺ variations in the cytoplasm and endoplasmic reticulum, we used two of the original yellow cameleons Ca²⁺ indicators: YC2, devoid of retention signal, and YC4_ER, a low affinity indicator with a KDEL sequence and a calreticulin signal peptide (43). A series of new cameleons bearing the cytochrome oxidase complex VIII targeting sequence were used to measure Ca²⁺ in the mitochondria: YC2_mit, a high affinity probe, as well as YC3.1_mit and YC4.1_mit, two mitochondrial probes of intermediate and low Ca²⁺ affinity, respectively. These two probes had a reduced pH dependence as they were based on the "improved" cameleons containing the V68L and Q69K mutations in the EYFP module (43, 44). Fig. 1 illustrates the subcellular distribution of the YC2, YC4_ER, and YC2_mit probes in transiently transfected HeLa cells. A three-dimensional rendering of wide-field images processed with a deconvolution algorithm is shown. As expected, the YC2 probe was uniformly distributed throughout the cytosol, with "holes" corresponding to the volume displaced by organelles (Fig. 1a). In contrast, the YC4_ER probe produced a delicate reticular staining pattern characteristic of the ER (Fig. 1b). The YC2_mit probe labeled discrete structures aligned along a largely interconnected network (Fig. 1c), consistent with the typical pattern of mitochondria in intact HeLa cells (19). These staining patterns were observed in most cells transfected at low efficiency (1-5%), but higher transfection efficiency was avoided as it caused significant mistargeting. To confirm that these staining patterns corresponded to the ER and mitochondria, the localization of the probes was verified by co-localization with specific markers. The YC4_ER probe largely co-localized with the ER marker protein calreticulin (Fig. 1b), while YC3.1_mit co-localized with the vital dye mitotracker red, as observed by confocal microscopy (Fig. 1c). This confirmed that the two probes specifically labeled the ER and the mitochondria, respectively.

View larger version (63K): [in this window] [in a new window]   Fig. 1.   Staining patterns and subcellular localization of the cytosolic, ER, and mitochondrial probes. a, typical fluorescence (430 ± 10 nm excitation, 535 ± 12.5 nm emission) of intact HeLa cells transfected with the YC2 probe. Images are shadow projections of 21 adjacent, 250 nm wide z sections deconvoluted with the iterative constrained Tikhonov-Miller restoration algorithm. b, left: HeLa cells transiently transfected with ER-targeted YC4ER probe. Right, the cells were fixed, permeabilized, and stained with an anti-calreticulin antibody; the YC4ER fluorescence co-localized with the calreticulin immunostaining. c, left: HeLa cells expressing YC2mit. Right, cells expressing the YC3.1mit probe were co-labeled with the vital dye mitotracker red to assess the specificity of the mitochondrial probe. Size bar: 5 µm.

Calibration and pH Dependence-- The cameleons probes were calibrated in situ by incubating HEK-293 cells stably expressing YC2 or YC4_ER for several minutes in heavily Ca²⁺-buffered solutions containing 10 µM ionomycin. To ensure a good Ca²⁺ equilibration across the ER membrane, 5 µg/ml digitonin was included for YC4_ER calibration. As shown in Fig. 2a, the YC2 calibration curve was monophasic, with an apparent dissociation constant (K'_d) of 1.24 µM and a Hill coefficient (n) of 0.79. The small shoulder reflecting Ca²⁺ binding to the high affinity site of the calmodulin-M13 hybrid protein was not resolved in our in situ calibration curve, which was slightly shifted to the left compared with the curve obtained in vitro (43). In contrast, the YC4_ER curve was clearly biphasic, with K'_d of 39 nM and 292 µM, and n of 0.57 and 0.60, consistent with the large shift in the low-affinity component produced by the E31Q mutation. The YC2 probe had a similar Ca²⁺ dependence when calibrated at pH 8.0 (squares), indicating that the probe behaved adequately in the pH range 7.0-8.0.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Ca2+ calibration and pH dependence. a, in situ Ca2+ calibrations of HEK-293 cells stably expressing YC2 () and YC4ER. (). Cells were incubated for several minutes with 10 µM ionomycin at the indicated calcium concentration, pH 7.4. For YC4ER calibration, the solution also contained 5 µg/ml digitonin. To mimic the conditions in the mitochondrial matrix, YC2 was also calibrated at pH 8.0 in the presence of 10 µM nigericin and 10 µM monensin (). b, changes in EYFP emission fluorescence during Ca2+ and pH changes in HeLa cells expressing YC4ER () or YC2mit (). Cells were stimulated sequentially with 50 µM histamine (Hist), 1 µM thapsigargin (Tg), 1 µM ionomycin (iono), 1 µM of the protonophore carbonylcyanide m-chlorophenylhydrazone (CCCP), and 25 mM of the permeant weak base NH4Cl. The pH of the organelle was then equilibrated at the indicated pH values with 10 µM nigericin and 10 µM monensin. c, in situ Ca2+ calibrations of HeLa cells transiently expressing the mitochondrial probes YC3.1mit and YC4.1mit. Cells were equilibrated with 10 µM ionomycin, 10 µM nigericin, and 10 µM monensin at the indicated calcium concentration, pH 8.0.

To confirm that the pH dependence of the YC probes did not interfere with Ca²⁺ measurements, we monitored the pH of the ER and mitochondria. By directly exciting the pH-sensitive EYFP module (excitation/emission: 480/535 nm), Ca²⁺-dependent FRET is bypassed and the probe reports only pH changes. As shown in Fig. 2b, no pH changes were observed within the ER and mitochondria during histamine or thapsigargin stimulation. Ionomycin caused a slight decrease as Ca²⁺ was exchanged for H⁺ across the organelle membrane. In contrast, large changes in EYFP fluorescence were observed with the protonophore carbonylcyanide m-chlorophenylhydrazone or the permeant weak base NH₄Cl. A pH titration curve in the presence of nigericin and monensin confirmed that the probe adequately reported the alkaline pH of mitochondria and the near-neutral pH of the ER. Thus, consistent with previous studies (48, 49), the pH of the ER and mitochondria is stable during Ca²⁺ transients, implying that, despite their pH dependence, the YC probes can be used to measure Ca²⁺ changes within these organelles.

The pH independent probes targeted to the mitochondria were also calibrated in situ, using transiently transfected HeLa cells (Fig. 2c). The low-affinity mitochondrial probe YC4.1_mit retained its biphasic Ca²⁺ dependence, with K'_d of 105 nM and 104 µM, and n of 0.81 and 0.62. The YC3.1_mit probe had intermediate Ca²⁺ affinity, with a K'_d of 3.98 µM and n of 0.67, as expected from the E104Q mutation which completely eliminates the high-affinity Ca²⁺-binding site. Thus, the Ca²⁺ dependence of the probes was not affected by the environment of the mitochondrial matrix. Together with YC2_mit, whose behavior was similar to the YC2 probe (not shown), the new mitochondrial probes allowed [Ca²⁺]_mit measurements over a wide range of Ca²⁺ concentrations.

[Ca²⁺]_mit Transients Reach Submillimolar Levels in a Fraction of Mitochondria-- Fig. 3 illustrate the [Ca²⁺]_mit responses reported by the three mitochondrial probes. Individual organelles were difficult to resolve due to the high density and the motility of mitochondria, and the shrinkage of the organelle during equilibration at high [Ca²⁺] prevented a pixel-by-pixel calibration. In these conditions, only the spatially averaged signal could be adequately calibrated. Using the apparent dissociation constants obtained in situ, the resting [Ca²⁺]_mit values measured with the high affinity YC2_mit probe (K_d ~1.2 µM) averaged 188 ± 25 nM (n = 15, range 48-445) and increased to 3.34 ± 0.36 µM upon stimulation with histamine. This averaged value was clearly an underestimate, as ~25% of the signal was saturated during the peak histamine response (defined as objects larger than 5 contiguous pixels exceeding the averaged R_max, Fig. 3a, right panel). Because the YC2_mit probe saturates above 100 µM (Fig. 2), this suggested that the peak [Ca²⁺]_mit values were higher than 100 µM in a fraction of mitochondria. A similar behavior was observed with the YC3.1_mit probe (K_d ~4 µM), which saturates above 300 µM (Fig. 2). Although the spatially averaged values were substantially higher (Table I), a fraction of the signal was still saturated during the histamine response and a subsequent Ca²⁺ titration to 100 µM and 1 mM caused only a slight increase in ratio (Fig. 3b, middle panel). To determine unambiguously the peak [Ca²⁺]_mit levels in these mitochondria, cells were transfected with the low affinity probe YC4.1_mit (K_d ~105 µM), which responds adequately in the range 10 µM to 3 mM (Fig. 2c). The peak [Ca²⁺]_mit reported by the YC4.1_mit probe averaged 106 ± 5 µM, and a further increase was observed when cells were equilibrated with 1 mM free [Ca²⁺] (Fig. 3c and Table I). The peak [Ca²⁺]_mit value now comprised almost all responding mitochondria, as only ~3% of the YC4.1_mit signal saturated during the histamine response. Thus, the response of mitochondria was quite heterogenous, the [Ca²⁺]_mit transients reaching millimolar levels in some mitochondria, which are likely to be very close to the sites of Ca²⁺ release.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   [Ca2+]mit transients reported by the three mitochondrial probes. a, left: YC2mit fluorescence image illustrating the total measured area (white line) and the regions larger than 5 contiguous pixels that saturated during the histamine response (red). The saturation level was defined as the average intensity of the total pixel area at saturating [Ca2+], as described under "Experimental Procedures." Middle, changes in ratio fluorescence during stimulation with 50 µM histamine and subsequent equilibration at 10 mM and 1 nM [Ca2+]. Right, time course of the pixel saturation during histamine application and subsequent equilibration at saturating [Ca2+]. The YC2mit probe was saturated in ~25% of mitochondria during the peak [Ca2+]mit response. b, similar experiment illustrating the response obtained with the YC3.1mit probe, which saturated in ~18% of mitochondria. A titration calibration was performed at the end of the experiments to estimate the peak [Ca2+]mit values. c, the YC4.1mit probe saturated in only ~3% of mitochondria and reported peak [Ca2+]mit levels around 100 µM. Traces are representative of 15, 8, and 14 experiments. Size bar: 5 µm.

ER Refilling and Heterogeneity during Stimulation with IP₃-generating Agonists-- Similar to mitochondria, the ER was also quite motile and exhibited some "ruffling" that prevented a pixel-by-pixel calibration, but in this case the stability of the overall structure allowed to compare distinct ER regions. Fig. 4 illustrates the [Ca²⁺]_ER responses elicited by histamine (Fig. 4a, linked video: YC4ER-hist.mov) and by the SERCA ATPase inhibitor thapsigargin (Fig. 4c, linked video: YC4ER-TG.mov). The resting [Ca²⁺]_ER levels averaged 371 ± 21 µM in the absence of external Ca²⁺ (n = 35, range 187-585) and decreased rapidly to 133 ± 14 µM upon stimulation with histamine (Fig. 4a, linked video: YC4ER-hist.mov). This decrease sometimes displayed an oscillatory pattern (not shown), suggesting that the ER refilled during stimulation. Accordingly, [Ca²⁺]_ER rapidly returned to prestimulatory levels upon Ca²⁺ readdition. In contrast, TG induced a slower but significantly larger [Ca²⁺]_ER decrease than histamine (Fig. 4c, linked video: YC4ER-TG.mov) the final levels averaging 59 ± 5 µM (Fig. 4d, n = 8). Subsequent readdition of Ca²⁺ had no effects, confirming that SERCAs had been fully inhibited and that this value reflected the fully depleted [Ca²⁺]_ER levels. Thus, during stimulation with IP₃-generating agonists, cells maintain [Ca²⁺]_ER well above depleted levels, indicating that substantial ER refilling occur even in the absence of external Ca²⁺.

View larger version (41K): [in this window] [in a new window]   Fig. 4.   Heterogeneity of [Ca2+]ER responses to IP3 generating agonists. a, intensity modulated YC4ER ratio image (linked video: YC4ER-hist.mov) illustrating the total cell area (white lines) and the measured regions of interest. b, time course of the [Ca2+]ER response to histamine in Ca2+-free medium, followed by wash of the agonist and Ca2+ readdition. Regions far from the nucleus (dotted lines, R2, R5) released more Ca2+ than regions close to nucleus (dashed lines, R1, R3, R4). The black line is the mean response of the entire cell. c, left: intensity modulated YC4ER ratio image (linked video: YC4ER-TG.mov) of cells stimulated with the SERCA ATPases inhibitor thapsigargin (Tg). d, time course of the Tg response. Tg induced a slower and more homogenous decrease in [Ca2+]ER.

Subcellular analysis of the ER responses revealed that distinct ER regions had similar resting [Ca²⁺]_ER values, and that thapsigargin released Ca²⁺ with similar kinetics throughout the ER network (Fig. 4d). This suggests that Ca²⁺ equilibrates freely within the ER lumen, both at rest and during the slow depletion induced by TG. In contrast, stimulation with histamine produced a more heterogeneous [Ca²⁺]_ER response, the ER depleting faster in regions closer to the plasma membrane (Fig. 4b). This heterogeneity was not due to a higher motility of the organelle during the histamine response, as verified by time-lapse recordings (see movies). Rather, it might reflect the increased activity of plasma-membrane ATPases upon stimulation with IP₃ generating agonists. Alternatively, the heterogeneity might reflect the distinct mitochondria density of the central and perinuclear ER regions, which could differentially modulate the kinetics of IP₃-induced Ca²⁺ release.

Ca²⁺ Cycles between Mitochondria and the ER-- To assess the functional role of the mitochondria/ER interactions, we studied the effect of mitochondrial inhibitors on the Ca²⁺ signals measured in the different cell compartments (Fig. 5). Histamine induced oscillations in [Ca²⁺]_cyt that were readily detected by mitochondria (Fig. 5, a and b, left panels). These oscillations originated from the ER as they occurred in the absence of external Ca²⁺, and could be reproduced repeatedly when cells were allowed to refill between consecutive histamine applications (Fig. 5a). This approach allowed us to study the contribution of mitochondria without compounding effects due to store-operated Ca²⁺ influx, which has been shown to be strongly inhibited by mitochondrial inhibitors (50). To collapse the mitochondrial membrane potential while minimizing ATP depletion, we used a combination of rotenone, an inhibitor of mitochondrial respiratory complex I, and oligomycin, to prevent ATP consumption by the reverse function of the mitochondrial H⁺-ATPase (51, 52). As expected, oligomycin/rotenone fully inhibited Ca²⁺ uptake by mitochondria, the peak [Ca²⁺]_mit becoming almost undetectable even with the high-affinity YC2_mit probe (Fig. 5b, peak [Ca²⁺]_mit = 2.92 ± 0.43 µM versus 0.48 ± 0.11 µM, n = 7). More Ca²⁺ was released into the cytosol when mitochondrial Ca²⁺ uptake was blocked, the averaged [Ca²⁺]_cyt levels measured with the cytosolic YC2 probe increasing by 20 ± 8% (n = 6, p < 0.05). Interestingly, the peak response was not significantly affected, but [Ca²⁺]_cyt failed to return to basal levels between oscillations (Fig. 5a, right panel), suggesting that Ca²⁺ clearance from the cytosol was affected by the lack of functional mitochondria.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Effects of mitochondria inhibitors on Ca2+ signals. Cells were stimulated with histamine in the absence of external calcium and incubated for 10 min with the mitochondrial inhibitors in calcium-containing medium between histamine applications. a-c, effect of 5 µg/ml oligomycin and 25 µM rotenone on the changes in [Ca2+]cyt, [Ca2+]mit, and [Ca2+]ER induced by histamine. The same cell was stimulated twice, except for [Ca2+]ER measurements. In c, 1 mM Ca2+ was added at the end of the histamine stimulation to illustrate the ER refilling kinetics. The mitochondrial inhibitors largely abolished the cytosolic Ca2+ oscillations and completely blocked mitochondrial Ca2+ uptake, but potentiated the decrease in [Ca2+]ER. d and e, effect of 10 µM CGP 37157 on the histamine-induced changes in [Ca2+]mit and [Ca2+]ER. The block of the mitochondrial Na+/Ca2+ exchanger prolonged the [Ca2+]mit signal and increased the ER depletion. Traces are representative of 6-14 experiments.

Parallel [Ca²⁺]_ER measurements revealed that the lack of mitochondrial Ca²⁺ uptake was associated with an increased depletion of ER Ca²⁺ stores (Fig. 5c). Due to the small dynamic range of YC4_ER, reproducible [Ca²⁺]_ER responses could not be obtained by repetitive stimulation of the same cell. However, in independent experiments [Ca²⁺]_ER levels decreased from 304 ± 46 to 126 ± 24 µM in control and from 299 ± 43 to 69 ± 15 µM in oligomycin/rotenone-treated cells, corresponding to a [Ca²⁺]_ER decrease of 59 ± 3 versus 75 ± 5%, respectively (mean ± S.E., n = 6, p = 0.03, unpaired t test). Interestingly, the initial drop in [Ca²⁺]_ER induced by the agonist was preserved, but [Ca²⁺]_ER continued to decrease in the presence of the mitochondrial inhibitors during the stimulation (Fig. 5c, right). This did not reflect a lack of Ca²⁺ pumping ability of the ER, as upon Ca²⁺ readdition [Ca²⁺]_ER increased with similar kinetics regardless of the presence of the inhibitors (Fig. 5c, _1/2 = 29 ± 7 s versus 38 ± 10 s for oligomycin/rotenone). This indicates that the SERCA ATPases were still fully functional and that the increased ER depletion was not due to a decrease in ATP levels. To test whether the larger ER depletion reflected the lack of Ca²⁺ returning from mitochondria, we used CGP 37157, a blocker of the mitochondrial Na⁺/Ca²⁺ exchanger. CGP 37157 caused a marked prolongation of the [Ca²⁺]_mit signal as Ca²⁺ remained trapped in the mitochondria (Fig. 5d). Although this effect is expected to increase the production of oxidative ATP, the decrease in [Ca²⁺]_ER was larger and more sustained in the presence of CGP 37157 (Fig. 5e). In nine independent experiments, [Ca²⁺]_ER decreased from 381 ± 25 to 175 ± 19 µM in control and from 314 ± 26 to 90 ± 10 µM in CGP-treated cells, corresponding to a 53 ± 5 versus 71 ± 2% decrease in [Ca²⁺]_ER, respectively (p = 0.013, unpaired t test). The increased ER depletion observed in the presence of CGP 37157 thus suggests that part of the Ca²⁺ captured by mitochondria is normally re-used for ER refilling.

Mitochondria Define Two Functional Ca²⁺ Stores in the ER-- To determine whether mitochondria indeed sustain the refilling of the ER, the kinetics of Ca²⁺ release were determined in ER regions containing or lacking mitochondria (Fig. 6). Cells were co-labeled with mitotracker red to locate mitochondria (Fig. 6b), and [Ca²⁺]_ER measured within overlapping or non-overlapping ER regions, which comprised 61.6 ± 1.6 and 38.4 ± 1.6% of the total YC4_ER staining, respectively. To avoid artifacts caused by the organelle motility, [Ca²⁺]_ER was measured as the ratio of F535 fluorescence normalized to the fluorescence at the start of the experiment (F/F₀). Consistent with the results of Fig. 4 upon stimulation with histamine [Ca²⁺]_ER decreased more slowly in the central ER regions that contained mitochondria (Fig. 6c, gray dots). Statistical analysis confirmed that the [Ca²⁺]_ER levels measured at the end of the histamine stimulation were significantly lower in ER regions lacking mitochondria (Table II). The difference persisted when ER regions located between 1.3 and 2.6 µM from the cell border were compared (0.982 ± 0.005 versus 0.956 ± 0.014, n = 7, p < 0.05), indicating that the faster depletion of the mitochondria-free compartment did not reflect its closer proximity to the plasma membrane. Consistent with a role of mitochondria in ER refilling, the differences disappeared in the presence of thapsigargin, the ER depletion being even slightly faster in ER regions containing mitochondria (Fig. 6, d-f, and Table II). Ca²⁺ release occurred with identical kinetics throughout the ER network in the presence of oligomycin/rotenone (Fig. 7, a-c, and Table II), confirming that the local refilling of neighboring ER regions required functional mitochondria. No differences were observed between the two ER regions in the presence of CGP 37157 (Fig. 7, d-f, and Table II), confirming that localized ER refilling depended on the supply of mitochondrial Ca²⁺, rather than on the local ATP levels. Taken together, these experiments indicate that mitochondria prevent the depletion of vicinal ER regions by returning Ca²⁺ back into the ER. The presence of active mitochondria thus defines two functionally distinct Ca²⁺ stores within the ER lumen.

View larger version (54K): [in this window] [in a new window]   Fig. 6.   Mitochondria define two functional Ca2+ stores in the ER. a-c, YC4ER cells were co-labeled with mitotracker red to locate mitochondria, and [Ca2+]ER was measured in regions containing (white line) and lacking mitochondria (dashed gray line). The two regions had similar ratio values prior to the histamine stimulation (2.52 versus 2.62), and [Ca2+]ER was measured by following the decrease in F535 fluorescence (F/F0) to avoid artifacts in the ratio image caused by the motility of the organelle edges. The ER region containing mitochondria (open circles) had higher [Ca2+]ER levels during the histamine stimulation than the ER region lacking mitochondria (dashed gray circles). d-f, the same experiment was performed in the presence of thapsigargin to inhibit the refilling of the ER. The differences between ER regions disappeared when SERCA ATPases were inhibited. Data are representative of seven (a-c) and five (d-f) experiments. Size bar, 5 µm.

View this table: [in this window] [in a new window]   Table II YC4ER responses in ER regions containing or lacking mitochondria The decrease in [Ca2+]ER (F/F0) was measured in YC4ER-labeled regions that co-localized or not with the mitotracker red staining (% mito), as described under "Experimental Procedures." Data are mean ± S.E. p = paired Student's t test.

View larger version (68K): [in this window] [in a new window]   Fig. 7.   Mitochondria recycle Ca2+ back to vicinal ER domains. Experiments were performed as described in the legend to Fig. 6, but in the presence of oligomycin/rotenone (a-c) or CGP 37157 (d-f). The differences between the two ER regions disappeared in the presence of the mitochondrial inhibitors. Data are representative of four (a-c) and seven (d-f) experiments. Size bar: 5 µm.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In this study, we used green fluorescent protein-based cameleon probes to measure Ca²⁺ changes within the cytosol, ER, and mitochondria. These genetically encoded Ca²⁺ indicators offer several advantages over other approaches used to measure Ca²⁺ changes in organelles. Their bright fluorescence and molecular targeting allowed time-resolved imaging of Ca²⁺ signals in defined intracellular compartments. The tunable Ca²⁺ affinity of the ratiometric probes ensured quantitative measurements within a wide range of Ca²⁺ concentrations. Furthermore, the probes could be used in conjunction with red shifted dyes to study the interactions between organelles (Figs. 6 and 7). Among the disadvantages, we observed that the targeting efficiency depended on cameleon expression levels, and that the small dynamic range of the probes limited the precision of the measurements. A more significant drawback was the pH dependence of the first generation of cameleons, which precluded the use of protonophores and required independent determination of the pH of the organelle. The stable pH of the ER and the alkaline pH of the mitochondria, however, allowed us to use the pH-sensitive cameleons for Ca²⁺ measurements in these organelles.

The resting [Ca²⁺]_ER levels averaged 371 µM in HeLa cells, values that agree well with previous reports using aequorin or cameleons (39, 41, 43, 47, 53). Substantial ER refilling was observed during agonist stimulation even in the absence of extracellular Ca²⁺, consistent with a recent study using targeted cameleon (47). No Ca²⁺ gradients were observed within the ER at rest or during stimulation with thapsigargin (Fig. 4), suggesting that the ER behaves as a single continuous compartment when IP₃ levels are low. In contrast, two functionally distinct ER compartments were observed during stimulation with IP₃-generating agonists. During histamine stimulation, regions rich in mitochondria, located deep in the cytosol, had higher [Ca²⁺]_ER levels that regions poor in mitochondria, located at the periphery of the cell (Fig. 6). The difference persisted in ER regions located at similar distance from the cell border, indicating that it did not reflect Ca²⁺ extrusion by the plasma-membrane ATPases. Instead, the two functional ER subdomains reflected the differential activity of SERCA ATPases, as the ER inhomogeneity disappeared in the presence of thapsigargin (Fig. 6b). A large part of the repumped Ca²⁺ originated from mitochondria, as the depletion of ER Ca²⁺ stores increased by 18% when mitochondrial Ca²⁺ efflux was blocked with CGP 37157 (Fig. 5).

Previous studies have shown that mitochondria are very close to the ER in HeLa cells (19), and that Ca²⁺ signal transmission between these organelles is quasisynaptic (54). Consistent with these observations, we observed that [Ca²⁺]_mit transients reached submillimolar values in a fraction of mitochondria (Fig. 3), which must be very close to the sites of Ca²⁺ release. These high levels were not detected by the high affinity YC2_mit probe, which was near saturation within this range of Ca²⁺ concentrations, but were readily detected by the low-affinity YC4.1_mit probe, which reported an average peak [Ca²⁺]_mit around 100 µM (Fig. 3). This finding is consistent with recent imaging data obtained in HeLa cells with a permutated green fluorescent protein engineered to sense Ca²⁺ (55), suggesting that earlier reports using high-affinity fluorescent dyes or aequorin might have underestimated the peak [Ca²⁺]_mit response. Higher, millimolar values were reported in chromaffin cells using a mutated aequorin of reduced Ca²⁺ affinity (36), possibly reflecting the close proximity of these mitochondria to both the plasma membrane and Ca²⁺ release channels (36). In HeLa cells, mitochondria are located far from the plasma membrane. The high [Ca²⁺]_mit levels observed in HeLa cells thus indicate that mitochondria take up a significant portion of the Ca²⁺ released by the ER. Some of this Ca²⁺ is then re-used for ER refilling, suggesting that Ca²⁺ cycles back and forth between the ER and mitochondria during Ca²⁺ oscillations.

Mitochondria might increase ER refilling by providing a local source of either Ca²⁺ or ATP, thereby enhancing the activity of SERCA ATPases. The Ca²⁺ effect appear predominant, as block of mitochondrial Ca²⁺ efflux by CGP 37157, which should increase the production of ATP by mitochondria, enhanced the depletion of the ER and led to the disappearance of the mitochondria-associated compartment. Furthermore, the ER refilled with similar kinetics regardless of the presence of mitochondria, indicating that the mitochondrial ATP did not contribute significantly to the activity of the SERCA ATPases. Accordingly, the ER refilling kinetics were not affected by oligomycin/rotenone or CGP 37157, consistent with earlier observations indicating that ATP generation is mainly glycolytic in HeLa cells (56). Thus, the predominant effect of mitochondria is to provide the local source of Ca²⁺ for refilling the ER. The close proximity of mitochondria from the sites of Ca²⁺ release ensures that a large part of the escaped Ca²⁺ is returned back into the ER.

In previous studies using trapped fluorescent dyes or BAPTA-loaded cells, mitochondria have been shown to increase, rather than prevent, the depletion of the ER (50, 53). This might reflect the effect of Ca²⁺ buffers on the ER/mitochondria interactions, as trapped fluorescent dyes have been shown to accumulate not only into the ER lumen, but also in the mitochondrial matrix (57). Because the peak [Ca²⁺]_mit are much larger than previously thought (Fig. 3), low-affinity dyes such as Mag-fura can be significantly affected by the [Ca²⁺]_mit changes. Accordingly, both increase and decreases in [Ca²⁺] have been reported during agonist stimulation using Mag-fura, depending on the relative contribution of mitochondria to the Mag-fura signal (57). The decreased [Ca²⁺] release observed with Mag-fura in the presence of mitochondrial inhibitors thus likely reflects the decreased [Ca²⁺]_mit signal, rather than the true [Ca²⁺]_ER response (50). In BAPTA-loaded cells, on the other hand, accumulation of the Ca²⁺ chelator might trap Ca²⁺ into the mitochondrial matrix, increasing the ability of mitochondria to buffer [Ca²⁺] but minimizing their contribution in the ER refilling process. Consistent with the ability of mitochondria to increase local Ca²⁺ buffering, we observed that mitochondria slightly increased the depletion of vicinal ER regions when refilling was blocked by thapsigargin (Fig. 6b).

The local cycle of Ca²⁺ between a portion of the ER and its neighboring mitochondria generates two functionally distinct Ca²⁺ stores within the ER network. The two stores are not structurally distinct ER pools with different Ca²⁺ transport characteristics, because they behaved similarly in the presence of thapsigargin or when mitochondrial Ca²⁺ cycling was inhibited (Figs. 6 and 7). Instead, they reflected the imbalance between the local Ca²⁺ refilling in ER regions bearing mitochondria and the Ca²⁺ drag occurring in remaining ER regions. Recent studies indicate that the ER is a lumenally continuous organelle, and that Ca²⁺ changes induced by local uncaging propagate over several micrometers (58). Accordingly, we found that [Ca²⁺]_ER changes were homogenous within the ER lumen in the absence of functional mitochondria. In contrast, the presence of functional mitochondria allowed neighboring ER regions to maintain higher [Ca²⁺]_ER levels. Although Ca²⁺ diffusion within the ER lumen would tend to dissipate these Ca²⁺ gradients, the high [Ca²⁺]_mit levels detected in some mitochondria indicate that mitochondria might provide sufficient amounts of Ca²⁺ to achieve a local control of [Ca²⁺]_ER levels in neighboring ER regions_. The strategic location of mitochondria thus appears to be a key determinant of their ability to modulate Ca²⁺ signals. By preventing the depletion of a defined ER region, mitochondria can restrict Ca²⁺ signals to specific regions of the cell. Depletion of the peripheral ER region could be required for the activation of store-operated Ca²⁺ influx, while refilling of the central ER region allow generating Ca²⁺ oscillations near the nucleus. The depletion of the central ER region was indeed associated with an altered cytosolic Ca²⁺ signal as, in the presence of mitochondria inhibitors, [Ca²⁺]_cyt failed to return to basal levels between oscillations (Fig. 4). The cycling of calcium between the ER and mitochondria thus not only controls the filling state of the ER, but also the spatio-temporal pattern of the cytosolic Ca²⁺ signal.

	ACKNOWLEDGEMENTS

We thank Cyril Castelbou for dedicated technical assistance, Drs. R. Y. Tsien and A. Miyawaki for providing the cameleon constructs, Dr. Uta Schmidt for expert advice, and Drs L. Bernheim, M. Mühlethaler, and W. Graier for critical reading of the manuscript.

	FOOTNOTES

^* This work was supported by operating Grants 31-46859.96 and 31-55344.98/1 from the Swiss National Science Foundation and in part by Grant HL61297 (to J. W.) from the National Institutes of Health.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The on-line version of this article (available at http://www.jbc.org) contains videos (Fig. 4).

Fellow from the Prof. Dr. Max Cloëtta Foundation. To whom correspondence should be addressed: Dept. of Physiology, University of Geneva Medical Center, 1, Michel-Servet, CH-1211 Geneva 4, Switzerland. Tel.: 4122-702-5399; Fax: 4122-702-5402; E-mail, Nicolas.Demaurex@medecine.unige.ch.

Published, JBC Papers in Press, May 17, 2001, DOI 10.1074/jbc.M103274200

	ABBREVIATIONS

The abbreviations used are: IP₃, inositol 1,4,5-trisphosphate; [Ca²⁺]_cyt, cytosolic free Ca²⁺ concentration; [Ca²⁺]_ER, endoplasmic reticulum free Ca²⁺ concentration; [Ca²⁺]_mit, mitochondrial matrix free Ca²⁺ concentration; CGP 37157, 7-chloro-5-(2-chlorophenyl)-1,5-dihydro-4,1-benzothiazepin-2(³H)-one; TG, thapsigargin; SERCA, sarco/endoplasmic reticulum Ca²⁺-ATPase; YC, yellow cameleon; EYFP, enhanced yellow fluorescent protein.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				J. E. Geiger and N. S. Magoski Ca2+-Induced Ca2+ Release in Aplysia Bag Cell Neurons Requires Interaction Between Mitochondrial and Endoplasmic Reticulum Stores J Neurophysiol, July 1, 2008; 100(1): 24 - 37. [Abstract] [Full Text] [PDF]

K. H. Weber, E. K. Lee, U. Basavanna, S. Lin