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

J. Biol. Chem., Vol. 276, Issue 28, 26411-26420, July 13, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/28/26411    most recent M101101200v1 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 Collins, T. J. Articles by Bootman, M. D. Search for Related Content PubMed PubMed Citation Articles by Collins, T. J. Articles by Bootman, M. D. Social Bookmarking                      What's this?

Mitochondrial Ca²⁺ Uptake Depends on the Spatial and Temporal Profile of Cytosolic Ca²⁺ Signals^*

Tony J. Collins, Peter Lipp^§, Michael J. Berridge^¶, and Martin D. Bootman^¶

From the  Laboratory of Molecular Signalling, The Babraham Institute, Babraham, Cambridge CB2 4AT, United Kingdom and ^¶ Department of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ, United Kingdom

Received for publication, February 5, 2001, and in revised form, April 18, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Using confocal imaging of Rhod-2-loaded HeLa cells, we examined the ability of mitochondria to sequester Ca²⁺ signals arising from different sources. Mitochondrial Ca²⁺ (Ca²⁺_mit) uptake was stimulated by inositol 1,4,5-trisphosphate (InsP₃)-evoked Ca²⁺ release, capacitative Ca²⁺ entry, and Ca²⁺ leaking from the endoplasmic reticulum. For each Ca²⁺ source, the relationship between cytosolic Ca²⁺ (Ca²⁺_cyt) concentration and Ca²⁺_mit was complex. With Ca²⁺_cyt < 300 nM, a slow and persistent Ca²⁺_mit uptake was observed. If Ca²⁺_cyt increased above ~400 nM, Ca²⁺_mit uptake accelerated sharply. For equivalent Ca²⁺_cyt increases, the rate of Ca²⁺_mit rise was greater with InsP₃-evoked Ca²⁺ signals than any other source. Spatial variation of the Ca²⁺_mit response was observed within individual cells. Both the fraction of responsive mitochondria and the amplitude of the Ca²⁺_mit response were graded in direct proportion to stimulus concentration. Trains of repetitive Ca²⁺ oscillations did not maintain elevated Ca²⁺_mit levels. Only low frequency Ca²⁺ transients (<1/15 min) evoked repetitive Ca²⁺_mit signals. Our data indicate that there is a lag between Ca²⁺_cyt and Ca²⁺_mit increases but that mitochondria will accumulate calcium when it is elevated over basal levels regardless of its source. Furthermore, in addition to the characteristics of Ca²⁺ signals, Ca²⁺ uniporter desensitization and proximity of mitochondria to InsP₃ receptors modulate mitochondrial Ca²⁺ responses.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Changes in the concentration of free calcium ions in the cytosol (Ca²⁺_cyt)¹ of cells have been implicated in the regulation of a wide variety of physiological events (1, 2). However, the toxic nature of Ca²⁺ requires any Ca²⁺_cyt load to be cleared quickly. Ca²⁺_cyt signals are either sequestered into endoplasmic/sarcoplasmic reticulum (ER/SR) stores via Ca²⁺-ATPases or removed from the cell across the plasma membrane by Ca²⁺-ATPases or Na⁺/Ca²⁺ exchangers. Supporting these mechanisms, mitochondria act to rapidly clear cytosolic Ca²⁺ loads before slowly releasing the Ca²⁺ back to the cytosol where it is dealt with by the aforementioned pumps and exchangers. Clearance of Ca²⁺_cyt loads by mitochondria has been demonstrated in a wide range of cell types (3-7). In addition to acting as a clearance mechanism, Ca²⁺_mit accumulation is a physiological response of cells to Ca²⁺_cyt-increasing stimuli, for example stimulating citric acid cycle enzymes and the respiratory chain (e.g. Refs 8-10; reviewed in Refs. 11 and 12).

Mitochondria sequester Ca²⁺ via a low-affinity, high-speed uniporter powered by the mitochondrial membrane potential (_mit). The in vitro affinity of the uniporter for Ca²⁺ (K_0.5 = 10-20 µM) (11, 13) is low in comparison with the global Ca²⁺ concentrations observed in vivo during physiological responses (100 nM-2 µM). However, mitochondria have been shown to sense and modulate both capacitative Ca²⁺ entry (14) and Ca²⁺ release from inositol 1,4,5-trisphosphate (InsP₃)-sensitive stores (15-20). To explain the paradox that mitochondria accumulate Ca²⁺ during physiological responses despite the insensitivity of the Ca²⁺ uniporter, it has been suggested that mitochondria are located in close proximity to sites of either Ca²⁺ release or Ca²⁺ entry (21-23). In these locations, mitochondria can respond to Ca²⁺ changes within the microdomains around the mouth of Ca²⁺ channels. The Ca²⁺ concentrations attainable in such microdomains are thought to be much higher than the observed global Ca²⁺ changes and closer to the in vitro affinity of the uniporter.

If mitochondria have privileged access to Ca²⁺ signals emanating from open channels, it would follow that portions of the cellular mitochondrial population would preferentially access certain types of Ca²⁺ signal, i.e. entry or release (e.g. Ref. 24), depending upon their relative cellular locations. Intuitively, perinuclear mitochondria would sense ER-derived Ca²⁺ signals in preference to peripheral mitochondria, and, conversely, peripheral mitochondria would more likely respond to Ca²⁺ entry signals.

The amplitude of Ca²⁺_cyt signals is also a likely determinant of Ca²⁺_mit accumulation. Previous studies have indicated a steep dependence of Ca²⁺_mit uptake on Ca²⁺_cyt, with a Hill coefficient  2 (for example see Ref. 25; reviewed in Refs. 26 and 27). The steep dependence of Ca²⁺_mit on Ca²⁺_cyt could indicate the co-operative binding and activation of the uniporter by Ca²⁺. However, it has also been suggested that mitochondria express an "activatory" Ca²⁺ binding site that allosterically stimulates the uniporter (reviewed in Refs. 26 and 27).

Aside from the source and amplitude of the Ca²⁺ signal, the kinetics of Ca²⁺_cyt transients have been shown to affect mitochondrial Ca²⁺ uptake (8, 28, 29). In hepatocytes, Ca²⁺ oscillations were more efficient at activating Ca²⁺-sensitive mitochondrial dehydrogenases than sustained Ca²⁺ elevations of similar magnitude (8). However, isolated rat heart mitochondria were found to sequester Ca²⁺ independently of the frequency of Ca²⁺ oscillations but in proportion to the mean extramitochondrial Ca²⁺ concentration (28).

In the present study, we examined the effect of different spatial and temporal Ca²⁺ signals on mitochondrial Ca²⁺ uptake in HeLa cells. Our data indicate that Ca²⁺_mit accumulation does not necessarily depend on microdomains of elevated Ca²⁺_cyt, because sources that do not generate such local Ca²⁺ gradients can be sequestered by mitochondria. The temporal response of mitochondria depends on the kinetics and amplitude of Ca²⁺_cyt increases. High amplitude Ca²⁺_cyt signals caused mitochondrial Ca²⁺ uptake to desensitize, whereas low amplitude Ca²⁺_cyt signals gave a slower but persistent Ca²⁺_mit accumulation. Mitochondria did not display an all-or-none response to Ca²⁺_cyt increases, but instead distinct mitochondria could be recruited on an individual basis, and the amplitude of their responses was graded in proportion to Ca²⁺_cyt.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell Culture, Chemicals, and Transfection

HeLa cell culture was performed as described previously (30). All subsequent experimental steps were carried out at room temperature (20-22 °C). The culture medium was replaced with an extracellular medium (EM) containing (in mM): NaCl, 121; KCl, 5.4; MgCl₂, 0.8; CaCl₂, 1.8; NaHCO₃, 6; D-glucose, 5.5; Hepes, 25, pH 7.3. Histamine, thapsigargin, and bromo-A23187 were obtained from Sigma. All fluorescent dyes were purchased from Molecular Probes. Transfection of HeLa cells was achieved with Effectene Transfection Reagent (Qiagen, Crawley, UK) using the manufacturer's suggested protocol. The pDsRed1-Mito construct was purchased from CLONTECH (Palo Alto, CA). The InsP₃ ester was kindly provided by Dr. Jan Thuring and Dr. Andrew Holmes (Dept. of Chemistry, University of Cambridge, UK).

Imaging Ca²⁺_mit and Ca²⁺_cyt

Fura-2 Imaging-- Cells were loaded with Fura-2 by incubation in 1 µM Fura-2 acetoxymethyl ester (AM) for 45 min followed by a 30 min de-esterification period. Coverslips were mounted on a Nikon Diaphot inverted microscope and imaged as described previously (30). Briefly, this involved excitation with 340 and 380 nm from twin monochromators, the excitation wavelengths being switched with a rotating mirror chopper (Glen Creston Instruments, Stanmore, UK). Emitted light was filtered at 510 nm and collected with an intensified charge-coupled device camera (Photonic Science, Tunbridge Wells, UK). The video signal was digitized, stored, and subsequently processed off-line with an Imagine image-processing unit (Synoptics Ltd., Cambridge, UK). Ratio images were acquired at 2-s intervals.

Rhod-2-- Imaging of the Ca²⁺-sensitive dye Rhod-2 was used to provide simultaneous measurements of both Ca²⁺_mit and Ca²⁺_cyt (see also Ref. 15). Cells were incubated in 1 µM Rhod-2 acetoxymethyl ester (AM) for 30 min followed by a 30 min de-esterification period. Imaging was performed with an Oz confocal microscope (Noran Instruments, Milton Keynes, UK). Rhod-2 was excited at 514 nm, and the Ca²⁺-sensitive emission >525 nm was selected using a long-pass filter. Median optical sections of cells loaded with Rhod-2 were acquired for 2 min at a rate of 1 image/0.533 s (i.e. 1.875 Hz). Ca²⁺_mit was determined from either individual mitochondrion or small groups of adjacent mitochondria. Although most of the Rhod-2 was localized within mitochondria, a small proportion remained within the cytosolic and nuclear compartment. By monitoring the nuclear Rhod-2 signal, the dynamics of the Ca²⁺_cyt signal were determined indirectly. Although the nucleoplasmic Rhod-2 fluorescence was low, averaging the signal over the entire nuclear optical section provided a sensitive measure of changes in Ca²⁺_cyt. Furthermore, because of the narrow confocal section (<1 µm), the nuclear Rhod-2 fluorescence was independent of Ca²⁺_mit changes.

The Ca²⁺ concentration ([Ca²⁺]) was calculated from background-corrected fluorescence values (F) using the equation, (Ca²⁺] = K_d × [(F  F_min)/(F_max  F)). Fluorescence at minimum [Ca²⁺] (F_min) was determined by incubating the cells with 10 µM bromo-A23187 in EM supplemented with 4 mM EGTA. Maximum fluorescence [Ca²⁺] (F_max) was obtained by treating the cells with 10 µM bromo-A23187 in normal EM (i.e. containing 1.8 mM CaCl₂). The dissociation constant (K_d) of Rhod-2 was determined by equilibrating bromo-A23187-treated cells (10 µM) with standard Mg²⁺-containing Ca²⁺ solutions (Molecular Probes). The K_d values of Rhod-2 were 1.0 µM for the nucleus and 1.3 µM for mitochondria (see "Results"). To aid the equilibration of intracellular and extracellular [Ca²⁺], the Rhod-2-loaded cells were pretreated with bromo-A23187 in EM containing no added CaCl₂, plus 4 mM EGTA, 2 µM antimycin, and 2 µM oligomycin and with the D-glucose substituted with 5.5 mM 2-deoxy-glucose. This pretreatment was necessary because HeLa cells are able to maintain resting cytosolic Ca²⁺ levels in the presence of A23187 despite large transmembrane Ca²⁺ gradients.

Simultaneous Rhod-2 and Fluo-3 Measurements-- Cells were loaded with Fluo-3/AM (2 µM for 40 min plus 30 min de-esterification) before being loaded with Rhod-2 as described above. Simultaneous dual channel confocal images were acquired at 7.5 Hz with a Noran Oz confocal microscope (514 nm excitation; green emission, 510-560 band pass; red emission, 590 long pass filter). Cross-talk from the green into the red channel was calculated from Fluo-3 only loaded cells and corrected for in subsequent experiments.

Cytosolic pH

Cells were loaded with 2 µM cSNARF1 diacetate for 40 min, followed by 30 min de-esterification period. The base form of cSNARF1 was imaged with a Bio-Rad MRC1024 LSCM (488 nm excitation; 680 emission with 30 nm band pass). Cytosolic pH was adjusted by the addition of pH 7.4 extracellular medium containing varying millimolar ratios of weak acid (butyric acid) and base (trimethylamine) to alter cytosolic pH from resting to 6.8 (butyric acid /trimethylamine of 16:1) to 7.4 (1:1) to 8.0 (1:16) (31). The change in cSNARF1 fluorescence during manipulation of cytosolic pH is expressed as F/F₀ where F₀ is the fluorescence resting before treatment with the acid/base mixture. The effect of these changes in pH on Rhod-2 was determined by imaging cells loaded with Rhod-2 preferentially loaded in the cytosol, i.e. cells were incubated with 10 µM Rhod-2 in the presence of (2 µM) antimycin and (2 µM) oligomycin.

Cytosolic Mg²⁺

Cells were loaded with either 2 µM Magnesium Green/AM (40 min loading; 30 min de-esterification) or Rhod-2 preferentially in the cytosol (see above). Magnesium Green-loaded cells were imaged with a Bio-Rad MRC1024 LSCM (488 nm excitation; 515BP emission). Cytosolic Mg²⁺ was adjusted from near zero to 96 mM in the presence of 1.2 µM Ca²⁺ with the addition of 1 µM A23187, 3 mM EDTA, 14 mM EGTA, and 16.3 mM CaCl₂ mixture followed by the addition of 1 µM A23187, 3 mM EDTA, 3 mM EGTA, 100 mM MgCl₂, and 2.0 mM CaCl₂ (based on MaxChelator calculations).

DsRed1

pDsRed1-Mito transfected cells were stained with MitoTracker Green (0.2 µM, 25 min). Dual channel emission images (525DF30 and 590LP) were acquired by simultaneous excitation with 514 and 488 nm laser lines. Please note that we did not find any detectable green emission from the DsRed1-expressing cells.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Measurement of Ca²⁺_mit and Ca²⁺_cyt with Rhod-2-- In the present study, Ca²⁺_mit and Ca²⁺_cyt were simultaneously monitored using Rhod-2. The AM ester form of Rhod-2 is cationic, causing it to load relatively selectively into respiring mitochondria. To confirm the mitochondrial sequestration of Rhod-2, we compared the subcellular fluorescence distribution of cells loaded with Rhod-2, MitoTracker Green and pDsRed1-Mito, which encodes a fusion of red fluorescent protein, DsRed1, and the mitochondrial targeting sequence from subunit VIII of human cytochrome c oxidase. The co-localization of these reporters to mitochondria is illustrated in Fig. 1. Furthermore, the loading of Rhod-2 into organelles was inhibited by agents that depolarize _mit (data not shown). These data indicate that the only intracellular structures in which Rhod-2 accumulated in HeLa cells were mitochondria.

View larger version (62K): [in this window] [in a new window]   Fig. 1.   Co-localization of Rhod-2, MitoTracker Green, and pDsRed1-Mito. Panels A and B show examples of HeLa cells co-stained with MitoTracker Green/FM and either Rhod-2 or pDsRed1-Mito. The images of rhod-2-loaded cells were obtained during thapsigargin treatment when both cytosolic and mitochondrial calcium concentrations were high, to allow clear identification of Rhod-2 localization. To compare the localization of the fluorescent markers, the images in Aa and Ab or Ba and Bb were converted to binary black and white images. These 1-bit images were then converted back to 8-bit gray scale images at half-maximal intensity (i.e. 128 of 256 gray levels). Merging of the images then shows an overlap of the fluorescent pixels as white areas, whereas the few non-overlapping pixels remain gray (marked by arrows). Calculation of the pixel intensities in Ac and Bc indicated that <7% of the pixels were not white. Therefore the fluorescence images in Aa and Ab or Ba and Bb overlap by >93%.

Because of the positive charge of Rhod-2 AM, most of the dye was sequestered into mitochondria. In addition, a small amount of Rhod-2 was localized to the cytosol and nucleus (see also ref 15). Because Ca²⁺_cyt and nuclear Ca²⁺ are in equilibrium in HeLa cells at the time resolution used in the present study (32, 33), confocal monitoring of Rhod-2 fluorescence in the nucleus allowed measurement of Ca²⁺_cyt without contamination from Ca²⁺_mit. Furthermore, as the cytosolic concentration of Rhod-2 was substantially less than that in the mitochondria, the determination of Ca²⁺_mit was not significantly affected by background changes of Ca²⁺_cyt (see below). To allow accurate estimation of cytosolic and mitochondrial Ca²⁺ concentrations, we calibrated the Rhod-2 fluorescence in both compartments as illustrated in Fig. 2A. Rhod-2 displayed similar affinities for Ca²⁺ in both the nuclear and mitochondrial compartments but showed a much smaller dynamic range (maximum fluorescence relative to minimum fluorescence) in the nucleus. The lower dynamic range of Rhod-2 in the nucleus is a disadvantage in discriminating small changes in Ca²⁺ concentration. However, averaging the Rhod-2 fluorescence over the large area of the nucleus observed in the confocal sections (~100 µm²) gave signals with good signal to noise ratios (see below). Although a few mitochondria were quite motile, the majority did not move appreciably during the course of confocal recordings. Only non-motile mitochondria were analyzed in the present study.

View larger version (32K): [in this window] [in a new window]   Fig. 2.   Calibration of cytosolic and nuclear Rhod-2. For panel A, HeLa cells were equilibrated with solutions containing different Ca2+ concentrations as described under "Materials and Methods." The dynamic range (DR) indicates the ratio of maximal to minimal fluorescence. Each data point represents the mean data obtained from 10 cells. The lack of effect of pH and Mg2+ on the reporting of Ca2+ by Rhod-2 is shown in panels B and C, respectively. B, cells loaded with Rhod-2 in the cytoplasm were treated with butyric acid and trimethylamine buffers of varying ratio to adjust the cytosolic from rest to pH 6.8 to 7.4 to 8.0. Ca2+cyt was clamped to 1.2 µM (i.e. near the Kd of Rhod-2 for Ca2+) by the addition of 1 µM A23187 in a solution containing 1.2 µM Ca2+. The data indicate the mean ± S.E., n = 12. The inset histogram shows the verification of changes in intracellular pH for cells treated with the butyric acid and trimethylamine buffers. To detect the pH change, the cells were loaded with cSNARF-1. The bars indicate the mean ± S.E., n = 12. C, cells loaded with Rhod-2 in the cytoplasm were incubated with 1 µM A23187 plus 1.2 µM CaCl2. The cells were then subjected to near zero and 96 mM free Mg2+ (changes in Mg2+ were detected using Magnesium Green; see inset). Finally, 1.8 mM CaCl2 was added to provide a value for Fmax and to demonstrate that although Rhod-2 did not detect Mg2+ it was still sensitive to Ca2+. The bars are mean ± S.E., n = 10 cells.

The protocol used to calibrate the Rhod-2 fluorescence would have collapsed any existing pH or Mg²⁺ gradient across the inner mitochondrial membrane. To determine whether the alkaline pH (34) of the mitochondria was affecting the Rhod-2 fluorescence or shifting the K_d of the dye for calcium, the pH sensitivity of the indicator was determined. To do this, Rhod-2 was preferentially accumulated into the cytoplasm by incubating cells with 10 µM Rhod-2AM in the presence of antimycin (2 µM) and oligomycin (2 µM). This allowed us to more reliably manipulate the pH and Mg²⁺ concentration in the same compartment as the indicator, which would be difficult to achieve with the Rhod-2 localized to mitochondria. The cytosolic pH was adjusted from resting (~7.4) to 6.8 to 7.4 to 8.0 using varying ratios of butyric acid and trimethylamine (31). Changes in cytosolic pH were verified in parallel experiments using the pH-sensitive dye cSNARF1 (Fig. 2B).

Because the effects of pH on the affinity of Rhod-2 for Ca²⁺ would be most easily detected around the K_d, the cells were treated with 1 µM A23187 in the presence of 1.2 µM external calcium during the manipulation of cytosolic pH. As shown in Fig. 2B, changes in pH from 6.8 to 8.0 had no significant effect on Rhod-2 fluorescence. Similar observations were made in cells not treated with A23187, which would have a resting Ca²⁺ concentration of ~100 nM (data not shown).

In some cell types, there is also a severalfold Mg²⁺ gradient across the inner mitochondrial membrane: [Mg²⁺]_cyt ~ 1-5 mM; [Mg²⁺]_mit ~ 20-40 mM (35). To determine whether high matrix Mg²⁺ was perturbing the calibration of Rhod-2, the sensitivity of Rhod-2 to Mg²⁺ was evaluated. Similar to the investigation of the effects of pH on Rhod-2 fluorescence described above, the indicator was loaded solely into the cytoplasm by depolarizing the mitochondria with antimycin + oligomycin. The fluorescence of Rhod-2 was then monitored after changing the cytoplasmic Mg²⁺ concentration from near zero to 100 mM while maintaining Ca²⁺_cyt at 1.2 µM. See "Materials and Methods" for the composition of the Mg²⁺ solutions. Although the change in cytoplasmic Mg²⁺ concentration was detectable using Magnesium Green, Rhod-2 fluorescence was unaffected (Fig. 2C).

These results show that Rhod-2 is neither pH- nor Mg²⁺-sensitive over a wide range. Therefore, even though the calibration protocol may have altered the pH and Mg²⁺ gradients within the cell, this would not have affected the subsequent conversion of Rhod-2 fluorescence into actual Ca²⁺ concentration.

Sensitivity of Mitochondrial Ca²⁺ Uptake to Different Ca²⁺ Sources-- To compare the sensitivity of mitochondria to cytosolic Ca²⁺ signals arising from distinct sources, HeLa cells were stimulated with a variety of treatments to elicit Ca²⁺ release from InsP₃-sensitive stores, Ca²⁺ leak from the ER, capacitative Ca²⁺ entry (CCE) and ionophore-induced Ca²⁺ equilibration. For InsP₃-evoked Ca²⁺ signals, HeLa cells were treated with 100 µM histamine in extracellular Ca²⁺ (Ca²⁺_cyt)-containing extracellular medium. The ER Ca²⁺ leak was unveiled by superfusion with the Ca²⁺ATPase inhibitor thapsigargin (1 µM) in Ca²⁺_o-free medium. Subsequent reapplication of Ca²⁺_o evoked a cytosolic Ca²⁺ rise because of CCE. Finally, the addition of ionophore (10 µM bromo-A23187) in Ca²⁺_o-containing medium was used to equilibrate Ca²⁺ throughout the cells. The typical cytosolic Ca²⁺ responses to these treatments are illustrated in Fig. 3A. The least signal, in terms of amplitude, was consistently the thapsigargin-induced ER Ca²⁺ leak. InsP₃-induced Ca²⁺ release, CCE, and ionophore-induced Ca²⁺ elevations were of similar amplitude. The kinetics of the cytosolic Ca²⁺ signals evoked by these treatments were different. Essentially, the order in which the Ca²⁺ signals arose was InsP₃-evoked Ca²⁺ release > ionophore > CCE > ER Ca²⁺ leak (Fig. 3B).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Ca2+cyt signals arising from different sources. Panel A shows the Ca2+cyt response of a single Fura-2-loaded HeLa cell to the treatments indicated by the upper bars. The concentrations of the agents used were: histamine, 100 µM; thapsigargin, 1 µM; Br-A23187, 10 µM. For comparison, the rising phases of the Ca2+cyt are shown on an expanded time scale in panel B.

The trace in Fig. 3A was obtained using video imaging of Fura-2-loaded cells. Because Fura-2 is a ratiometric indicator and bleaches only slowly during excitation with a xenon arc lamp on our system, we were able to demonstrate the response to each Ca²⁺-increasing treatment when consecutively applied to the same cells. For subsequent experiments monitoring both mitochondrial Ca²⁺ (Ca²⁺_mit) and cytosolic Ca²⁺ (Ca²⁺_cyt), laser scanning confocal imaging of Rhod-2 was employed. As Rhod-2 is not a ratiometric indicator and bleaches detectably during laser excitation, the effect of each Ca²⁺-increasing treatment was investigated separately using naïve cells. Essentially, the cells were treated/pretreated as depicted in Fig. 3A, but the Rhod-2 signal was only recorded just prior to and during the test period.

Relationship between Ca²⁺_cyt and Ca²⁺_mit-- Each of the Ca²⁺_cyt-increasing treatments illustrated in Fig. 3 also elevated Ca²⁺_mit, although there were distinct differences in the characteristics of some of the Ca²⁺_mit signals (Fig. 4, A-D). Histamine stimulation resulted in the largest rise of Ca²⁺_mit (Table I), with CCE and ionophore evoking significantly smaller responses (Table I). The thapsigargin-induced ER Ca²⁺ leak, which gave the least Ca²⁺_cyt increase, also evoked the lowest Ca²⁺_mit elevation (Table I).

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Comparison of Ca2+cyt and Ca2+mit. Panels A-D indicate Ca2+cyt (dotted line) and Ca2+mit (solid line) in response to the treatments depicted in Fig. 3A. In panels E-H, Ca2+cyt and Ca2+mit are plotted against each other. The times shown in panels E-G represent the approximate durations of the segments delineated by the dashed lines. The traces represent averaged responses from at least 10 cells (~10 individual mitochondria examined per cell).

The thapsigargin responses in HeLa cells were not due to activation of InsP₃ receptors by Ca²⁺-induced Ca²⁺ release, as has been demonstrated for mouse lacrimal cells (36), because there was no effect of 40 mM caffeine (an InsP₃ receptor antagonist) on the characteristics of thapsigargin-induced Ca²⁺_cyt signals monitored with Fura-2. In contrast 40 mM caffeine was sufficient to completely inhibit responses to supramaximal concentrations of histamine (data not shown). The inclusion of 40 mM caffeine while monitoring Ca²⁺_cyt and Ca²⁺_mit with Rhod-2 gave curves similar to those shown in Fig. 4B (data not shown). The CCE response also did not involve any component of Ca²⁺ release, because the prior application of thapsigargin would have completely discharged the InsP₃-sensitive Ca²⁺ pool.

For each Ca²⁺_cyt-increasing treatment except ionophore, Ca²⁺_cyt rose significantly earlier than Ca²⁺_mit. This was true both for the time taken for Ca²⁺_cyt and Ca²⁺_mit to reach half-maximal amplitude and for the absolute latency for Ca²⁺_cyt and Ca²⁺_mit to show an increase over basal levels (Table II). The discrepancy in the kinetics of Ca²⁺_mit and Ca²⁺_mit increases is clearly apparent when these two parameters are plotted against each other (Fig. 4, E-H). For histamine, ER Ca²⁺ leak, and CCE, there was an initial increase in Ca²⁺_cyt of a few hundred nanomolar with only a modest change in Ca²⁺_mit. This was followed by a period when Ca²⁺_cyt did not substantially change, although Ca²⁺_mit rose markedly. Finally, Ca²⁺_mit reached a plateau as Ca²⁺_cyt started to decline (Fig. 4, E-G). As expected from its ability to shuttle Ca²⁺ across membranes, the response to ionophore treatment gave a linear relationship between Ca²⁺_cyt and Ca²⁺_mit concentrations (Fig. 4H).

The data presented in Fig. 4 indicate that net Ca²⁺_mit accumulation began as soon as Ca²⁺_cyt was increased above the 100 nM resting level. More significantly, however, was the rapid increase in Ca²⁺_mit uptake when Ca²⁺_cyt reached 400-600 nM (Fig. 4, E-G). The inflection points in the curves in Fig. 4, E-G, indicates that for only a small increase in Ca²⁺_cyt, there was a dramatic enhancement of Ca²⁺ uniporter activity. Before and after the inflection points, Ca²⁺_mit accumulation varied approximately linearly with Ca²⁺_cyt. The rates of Ca²⁺_mit uptake for each of the Ca²⁺ sources are listed in Table III. These data indicate that before Ca²⁺_cyt reached the 400-600 nM threshold, a similarly low rate of Ca²⁺_mit uptake was observed irrespective of the source of Ca²⁺_cyt. However, after the Ca²⁺_cyt threshold had been attained, there was a significant increase in the rate of Ca²⁺_mit uptake. It should be noted that the 400-600 nM threshold represents the value obtained by averaging several mitochondria. Plots of Ca²⁺_cyt versus Ca²⁺_mit for single mitochondria showed similarly shaped curves to those shown in Fig. 4, but the threshold values for Ca²⁺_mit uptake varied between mitochondria from 200-1000 nM.

The rapid acceleration of Ca²⁺_mit uptake observed when Ca²⁺_cyt reached an average of 400 to 600 nM is consistent with the notion of a cytosolic activatory Ca²⁺ binding site that allosterically regulates the uniporter (see Refs. 11 and 26). To demonstrate that reaching a threshold intramitochondrial matrix Ca²⁺ concentration did not cause the acceleration of Ca²⁺_mit uptake, we examined the kinetics of Ca²⁺_mit accumulation during a prolonged Ca²⁺_cyt signal of less than 300 nM. To evoke such a sustained low-amplitude Ca²⁺_cyt response in HeLa cells, we activated CCE by depleting ER Ca²⁺ stores using thapsigargin in Ca²⁺_o-free medium and subsequently applied a solution containing a low (250 µM) Ca²⁺_o concentration. The weak influx of Ca²⁺ via CCE under these conditions caused a slowly elevating Ca²⁺_cyt signal that reached a plateau after ~30 min (Fig. 5A). The kinetics of the Ca²⁺_mit response mirrored the Ca²⁺_cyt signal, although the amplitude of the mitochondrial response was significantly higher than that in the cytoplasm (Fig. 5A). Furthermore, a comparison of cytosolic and mitochondrial signals indicated that there was no acceleration of Ca²⁺_mit uptake (Fig. 5B). Instead, Ca²⁺_mit varied linearly with Ca²⁺_cyt. The peak rate of Ca²⁺_mit uptake under these conditions was 0.5 ± 0.1 nM/s.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Prolonged Ca2+mit accumulation in response to a low-amplitude sustained Ca2+cyt signal. HeLa cells were pretreated with 2 µM thapsigargin in nominally Ca2+o-free medium to deplete intracellular Ca2+ stores and activate CCE. At time = 0 in panel A, Ca2+o was increased to 250 µM. The resultant increases in Ca2+cyt and Ca2+mit were recorded using confocal imaging of Rhod-2. The thick line in panel B depicts the relationship between Ca2+cyt and Ca2+mit for the responses shown in panel A. For comparison, the thin line in panel B shows the response of cells to the addition of 1.8 mM Ca2+o.

These data indicate that the slow rate of Ca²⁺_mit uptake observed when Ca²⁺_cyt was below 300 nM was able to cause significant net accumulation of Ca²⁺ into the mitochondria over a prolonged period of time. It is therefore unlikely that the acceleration of Ca²⁺_mit uptake observed in Fig. 4, EG, was due to a threshold intramitochondrial matrix Ca²⁺ concentration or the time-dependent binding of Ca²⁺ to the uniporter.

Mitochondria "Tune-out" of High Frequency Ca²⁺ Oscillations-- HeLa cell mitochondria showed only a transient response to high amplitude repetitive Ca²⁺ oscillations. Even though the peaks of the Ca²⁺_cyt spikes were above the ~400-600 nM threshold for rapid uptake, Ca²⁺_mit relaxed back to basal values within several minutes. A similar transient Ca²⁺_mit accumulation was observed in cells displaying Ca²⁺ oscillations in response to either continuous perfusion with histamine (Fig. 6A) or pulsatile histamine applications (Fig. 6B).

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Ca2+mit responses desensitize during prolonged Ca2+ signals. Panels A and B illustrate the typical transient Ca2+mit response to oscillatory Ca2+ increases. In panel A, the Ca2+ oscillations were evoked by constant superfusion of a cell with histamine as shown by the filled bar. Because histamine-evoked Ca2+ responses diminish with time, a pulsatile application of histamine was used to evoke a non-decremental train of Ca2+cyt oscillations (panel B). The lower parts of panels A and B illustrate that either treatment evoked only a transient Ca2+mit increase. Panel C shows a comparison of Ca2+mit signals in response to a single Ca2+cyt transient (single), a series of Ca2+cyt oscillations (osc.), multiple rapid Ca2+cyt signals caused by pulsatile histamine applications (multi.), and Ca2+ oscillations in the presence of CGP-37157.

A single Ca²⁺ oscillation evoked by perfusion of cells with 100 µM histamine for 30 s induced a Ca²⁺_mit transient that decayed with a half-time of 3.2 min (Fig. 6C), whereas the Ca²⁺_cyt increase recovered with a half-time of ~30 s. Generation of a sustained train of Ca²⁺_cyt oscillations by perfusing histamine in a pulsatile manner (100 µM histamine; 30-s application followed by 30-s recovery period) caused a Ca²⁺_mit increase that decayed back to prestimulated levels with a half-time of 5.6 min (Fig. 6C). The peak amplitudes of the Ca²⁺_mit increases with single or multiple oscillations were similar (data not shown). These data indicate that there was no additivity in the amplitude of the Ca²⁺_mit response during multiple Ca²⁺_cyt oscillations but that the oscillations prolonged the mitochondrial Ca²⁺ accumulation for a couple of minutes. However, once Ca²⁺_mit had recovered back to basal levels, the ongoing Ca²⁺_cyt oscillations were not sequestered. Incubation of cells with the mitochondrial Na⁺/Ca²⁺ exchange inhibitor CGP-37157 increased the half-time for recovery of Ca²⁺_mit during a train of Ca²⁺_cyt oscillations to 10.2 min. Despite the presence of CGP-35157 there was no additivity of the Ca²⁺_mit response, and it eventually declined back to resting levels (Fig. 6C).

Using a paired-pulse protocol with progressively longer periods of quiescence between histamine applications, we observed that mitochondrial Ca²⁺ accumulation recovered with a half-time of ~10 min (Fig. 7). With a sufficient rest interval, Ca²⁺_mit uptake was indistinguishable between two consecutive histamine applications (Fig. 7). These data indicate that the apparent failure of Ca²⁺_mit accumulation was not due to irreversible damage of the mitochondria or Rhod-2 bleaching. Furthermore, the lack of Ca²⁺ uptake was not due to activation of the mitochondrial permeability transition pore, because cyclosporin A (10 µM) did not affect the kinetics of Ca²⁺_mit recovery. In addition, monitoring of _mit with TMRE during histamine-evoked Ca²⁺_cyt pulses indicated that the mitochondria were not depolarized (data not shown).

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Recovery of Ca2+mit uptake following desensitization. Pulses of histamine (100 µM, applied for 30 s) were given at increasing intervals, and the corresponding Ca2+cyt and Ca2+mit responses were recorded. The bars indicate the amplitude of the second histamine response relative to the first. The open bars illustrate that the Ca2+cyt response had a reproducible amplitude for stimulation intervals of 2-30 min. In contrast, 2 min after the original histamine application, the Ca2+mit response was inhibited by 75% but subsequently recovered to 100% after 20 min (filled bars). The data indicate the mean ± S.E. of 5 cells.

Mitochondria Show a Spatially Graded Pattern of Recruitment-- Stimulation of HeLa cells with increasing histamine concentrations evokes progressively higher amplitude global Ca²⁺_cyt increases (30, 37), most likely because of the summation of Ca²⁺ release from an escalating number of activated InsP₃ receptors. Similarly, HeLa cell mitochondria also displayed a graded response to increasing histamine stimulation (Fig. 8). The threshold histamine concentration for evoking a Ca²⁺_mit increase within a single cell differed between individual mitochondria (Fig. 8, B and C). In addition, the amplitudes of the Ca²⁺ rises were markedly different between individual mitochondria (Fig. 8B). Because of desensitization of Ca²⁺_mit uptake, as described above, it was not possible to construct full concentration-response curves for histamine-evoked Ca²⁺_mit signals. However, our empirical observations suggested that increasing histamine concentrations evoked a Ca²⁺_mit response in a larger fraction of the mitochondrial population and gave progressively higher amplitude Ca²⁺_mit signals.

View larger version (41K): [in this window] [in a new window]   Fig. 8.   Different thresholds for Ca2+mit uptake in a single HeLa cell. Panel A depicts the outline of a single HeLa cell (the nucleus is shown as a thin dashed line). The Ca2+cyt and Ca2+mit responses to the application of increasing histamine concentrations are shown in panel B. The relative positions of the analyzed mitochondria are depicted by the correspondingly numbered circles in panel A. The spatial inhomogeneity of the Ca2+mit responses in this cell is also illustrated by the surface plots in panel C, where the amplitude of the Ca2+mit signal is encoded by the height and color of the image. The arrows indicate the typical localized responses observed at regions 1 and 7.

The inability of some mitochondria to sequester Ca²⁺ during signals such as those illustrated in Fig. 8B was surprising, because with each histamine concentration the global Ca²⁺_cyt level increased above the ~400-600 nM threshold required for rapid Ca²⁺_mit uptake. A plausible explanation for these data is that the mitochondria were not equally exposed to the microdomains of elevated Ca²⁺_cyt around the mouth of the open InsP₃ receptors. Furthermore, the Ca²⁺_cyt signals were too brief for the mitochondria that did not see Ca²⁺ microdomains to accumulate significant amounts of calcium. Outside such Ca²⁺ microdomains the mitochondria will presumably accumulate Ca²⁺ with a similar latency (~20 s; Table II) and rate to that seen with the ER Ca²⁺ leak. This scheme suggests that there will be temporal delays between individual mitochondrial responses, reflecting their proximity to activated InsP₃ receptors.

We further investigated whether the duration of a Ca²⁺ signal determined the ability of mitochondria to respond. Essentially, HeLa cells were stimulated with a maximal histamine concentration (100 µM) for varying durations. With stimulation periods  3 s, all mitochondria responded. However, when the duration of histamine application was reduced to 0.5 s, the Ca²⁺_mit response failed (Fig. 9A). The 0.5-s histamine application evoked a global response that was greater than the 400-600 nM threshold. These data therefore indicate that mitochondrial Ca²⁺ sequestration lags temporally behind Ca²⁺_cyt changes and that high amplitude Ca²⁺_cyt signals of short duration will not be sequestered by mitochondria.

View larger version (26K): [in this window] [in a new window]   Fig. 9.   The duration of cell stimulation determines Ca2+mit uptake. Panel A illustrates the failure of mitochondria to accumulate Ca2+ from rapid global Ca2+cyt signals. Panel Aa depicts the outline of a single HeLa cell. The Ca2+cyt and Ca2+mit responses to application of 100 µM histamine for either 0.5 or 30 s are shown in panel Ab. The relative positions of the analyzed mitochondria are depicted by the correspondingly numbered circles in panel Aa. For both 0.5 and 30 s of histamine stimulation the Ca2+cyt signal was a global response. Panel B depicts the lack of Ca2+mit response to a cytosolic Ca2+ puff. The outline of a single HeLa cell is shown in Ba. The Ca2+cyt and Ca2+mit responses triggered by application of a membrane-permeable InsP3 ester are shown in Panel Bb. Ca2+cyt was monitored with Fluo-3, and Ca2+mit was monitored with Rhod-2.

Consistent with this finding, we also observed that elementary Ca²⁺ signals, i.e. Ca²⁺ puffs (32, 37), were not sensed by mitochondria (Fig. 9B), unlike the situation in myocytes (38) and myotubes (39). To simultaneously monitor Ca²⁺ puffs and Ca²⁺_mit, the HeLa cells were co-loaded with Fluo-3 and Rhod-2 (see "Materials and Methods"). Using this combination of indicators we could image with sufficient speed and sensitivity to resolve the cytosolic Ca²⁺ puff events. The InsP₃ receptors were activated using a membrane-permeable InsP₃ ester (10 µM) (16). In the 20 cells examined, we did not observed any change of Ca²⁺_mit in the mitochondria within the vicinity of Ca²⁺ puffs.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The aim of the present study was to investigate the ability of mitochondria to accumulate Ca²⁺ in response to Ca²⁺_cyt signals with differing spatiotemporal characteristics. By treating HeLa cells with agents that mobilize InsP₃-sensitive Ca²⁺ stores, reveal the ER Ca²⁺ leak, or activate CCE, we were able to generate Ca²⁺_cyt signals with varying amplitudes and kinetics (Fig. 3). On first inspection, the data presented in Fig. 4, AD, suggested that for each Ca²⁺ source Ca²⁺_mit mirrored Ca²⁺_cyt; the kinetics of Ca²⁺_mit uptake appeared to be relative to rate of Ca²⁺_cyt increase, and the amplitude of Ca²⁺_mit increased in direct proportion to the magnitude of Ca²⁺_cyt. However, a detailed comparison of Ca²⁺_mit and Ca²⁺_cyt suggested a more complex relationship. With histamine, the ER Ca²⁺ leak, and CCE, mitochondria began to accumulate Ca²⁺ soon after Ca²⁺_cyt was elevated above the 100 nM resting level, but Ca²⁺_mit uptake greatly accelerated when Ca²⁺_cyt reached 400-600 nM (Fig. 4, E-G). In addition, for each Ca²⁺ source there were distinct differences in the maximal rates of Ca²⁺_mit uptake upon reaching the 400-600 nM threshold (Fig. 4, E-G; Table III). Furthermore, Ca²⁺_mit reached a plateau long after Ca²⁺_cyt had stopped increasing and declined substantially more slowly than Ca²⁺_cyt (Fig. 4, A-C). Ionophore treatment served as a useful control to demonstrate that the rate and extent of mitochondrial Ca²⁺ accumulation is directly proportional to Ca²⁺_cyt without any threshold (Fig. 4H), providing the inner mitochondrial membrane is sufficiently permeable.

The substantial increase in the rate of Ca²⁺_mit uptake upon reaching a Ca²⁺_cyt threshold of 400-600 nM is consistent with the binding of Ca²⁺ to an activatory site that stimulates the Ca²⁺ uniporter (see, for examples, Refs. 40 and 41). Because the ER Ca²⁺ leak responses were unaffected by caffeine (see above), and the CCE response is caused by the prior depletion of Ca²⁺ stores with thapsigargin, these signals are unlikely to have given rise to microdomains of high Ca²⁺_cyt concentration in the vicinity of mitochondria. We therefore suggest that the 400-600 nM threshold represents the in situ affinity of the activatory Ca²⁺ binding site. The fact that histamine also caused an acceleration of Ca²⁺_mit uptake at a similar 400-600 nM Ca²⁺_cyt threshold may be coincidental. At the time when Ca²⁺_mit started to accelerate (a few seconds after histamine addition; Fig. 4E) the histamine response would have comprised both microdomains of Ca²⁺ around the mouth of activated InsP₃ receptors and Ca²⁺ that had diffused away from those microdomains. With an InsP₃-generating agonist, it is therefore difficult to estimate affinities, because the Ca²⁺_cyt signal is spatially heterogeneous. However, the Ca²⁺_cyt responses to CCE and ER Ca²⁺ leak are slowly rising and do not involve microdomains (see above), meaning that the recorded Ca²⁺_cyt signal is a reasonable global average.

Although rapid Ca²⁺_mit uptake has a threshold of ~400-600 nM, the different rates of accumulation observed with histamine, the ER Ca²⁺ leak and CCE (Table III) indicate that other factors impinge on the kinetics of Ca²⁺ sequestration. Most likely, the mitochondrial Ca²⁺ uniporter is sensitive not only to the absolute Ca²⁺_cyt level but also to the gradient of Ca²⁺ across the inner mitochondrial membrane (26). In this case, InsP₃-induced Ca²⁺ release can be considered to be the most effective way of inducing Ca²⁺_mit accumulation. InsP₃ rapidly elevated Ca²⁺_cyt to the 400-600 nM threshold (Table II). In addition, the microdomains of Ca²⁺ that form around the mouth of the open InsP₃ receptors provide a concentration gradient that drives rapid Ca²⁺_mit uptake. Because the ER Ca²⁺ leak and CCE signals do not generate such microdomains, they do not stimulate the mitochondrial uniporter to the same degree.

A spatial distinction between mitochondria became apparent when the histamine concentration was varied (Fig. 8). The threshold histamine concentration required to evoke a Ca²⁺_mit response differed between mitochondria. Furthermore, the amplitudes of the Ca²⁺_mit elevations were extremely variable. A similar heterogeneous mitochondrial Ca²⁺ uptake was observed in single rat aortic myocytes responding to two different Ca²⁺-mobilizing agonists (42). The weaker agonist, which would approximate to stimulation of HeLa cells with low histamine concentrations, gave only a modest Ca²⁺_mit response in a subpopulation of mitochondria. In contrast, application of an agonist that evoked a more substantial Ca²⁺_cyt increase evoked a larger Ca²⁺_mit response and recruited more of the mitochondrial population. An obvious explanation for such heterogeneous Ca²⁺_mit uptake is that the Ca²⁺_cyt signals are localized within the cell. However, in the experiment illustrated in Fig. 8, the Ca²⁺_cyt responses were all "global" signals that propagated throughout the cell. This was evident from the small background response of cytoplasmic Rhod-2 in non-mitochondrial areas (data not shown). Therefore the spatial recruitment of mitochondria as depicted in Fig. 8 is not simply due to subcellular localization of the Ca²⁺_cyt increase. A more likely explanation is that only the mitochondria in the vicinity of the activated InsP₃ receptors are able to rapidly respond. As the stimulus is increased, more InsP₃ receptors are activated, and thus a greater fraction of mitochondria respond. This scheme applies only to transient Ca²⁺_cyt increases, such as those depicted in Fig. 8, because any prolonged elevation of Ca²⁺_cyt above the basal Ca²⁺_cyt level will eventually lead to significant Ca²⁺_mit uptake (Fig. 4). Therefore, Ca²⁺_mit responses are dependent on the proximity of mitochondria to InsP₃ receptors. Furthermore, the duration of the Ca²⁺_cyt signal can modulate Ca²⁺_mit accumulation, because mitochondrial Ca²⁺ uptake lags significantly behind Ca²⁺_cyt (Table II and Fig. 9) such that rapid Ca²⁺_cyt transients are not sequestered by HeLa cell mitochondria.

In addition to a dependence on the spatial properties of Ca²⁺_cyt increases, it has been suggested that Ca²⁺_mit accumulation is sensitive to the frequency of Ca²⁺ signals. For example, hepatocyte mitochondria accumulate Ca²⁺ only transiently during a tonic Ca²⁺_cyt increase but can respond with a sustained Ca²⁺_mit signal to rapid Ca²⁺_cyt oscillations (8). Unlike the responses in hepatocytes, sustained Ca²⁺_mit accumulation was not observed during rapid Ca²⁺ oscillations (Fig. 6). Instead, Ca²⁺_mit increased only during the initial Ca²⁺_cyt oscillation and subsequently decayed back to resting levels. A train of oscillations approximately doubled the relaxation time for Ca²⁺_mit, suggesting that there was a limited sequestration of Ca²⁺ accumulation occurring after the first Ca²⁺_cyt spike. If the cells were allowed a period of quiescence for ~30 min, then Ca²⁺_mit uptake appeared to recover to the degree observed in naïve cells (Fig. 7).

The mechanism underlying desensitization of Ca²⁺_mit uptake during oscillatory Ca²⁺_cyt increases is unclear. It has been suggested that mitochondria accumulate sequential pulses of Ca²⁺ before reaching a threshold matrix concentration when the permeability transition pore opens, resulting in a rapid efflux of Ca²⁺ in a process known as mitochondrial Ca²⁺-induced Ca²⁺ release (43, 44). In the results presented here, Ca²⁺ accumulation was not additive, despite the matrix Ca²⁺ failing to return to rest between oscillations. Therefore, unless the first Ca²⁺_cyt spike is sufficient to elevate the matrix Ca²⁺ concentration above the threshold for activation of mitochondrial Ca²⁺-induced Ca²⁺ release, it is unlikely that this mechanism accounts for the diminution of Ca²⁺_mit in the face of a sustained or oscillatory Ca²⁺_cyt signal. Furthermore, activation of permeability transition pores would be associated with rapid mitochondrial depolarization, which was not observed during responses to histamine.

A similar loss of Ca²⁺_mit uptake was observed in pancreatic  cells (45), where pulsatile application of various Ca²⁺-increasing stimuli at <30-min intervals gave only a single mitochondrial response. The electron transport chain and _mit were unaffected by such stimulation, and the mitochondria were still able to produce ATP. It was therefore suggested that desensitization of the Ca²⁺ uniporter was responsible for the inhibition of mitochondrial Ca²⁺ uptake (45). Because Ca²⁺_mit uptake is a positive feedback system, whereby elevated matrix Ca²⁺ activates dehydrogenases that serve to hyperpolarize the inner mitochondrial membrane and further increase the driving force for Ca²⁺ influx, it is reasonable that Ca²⁺ accumulation would be limited by a mechanism independent of other crucial mitochondrial functions. Although we have no direct evidence for inhibition of the uniporter in HeLa cells, our data are consistent with such a scheme; elevation of Ca²⁺_cyt or Ca²⁺_mit did not depolarize _mit or activate permeability transition pores, yet the Ca²⁺ accumulation was transient. In addition, slowing the mitochondrial Na⁺/Ca²⁺ exchange using the inhibitor CGP-37157 did not prevent the decline in Ca²⁺_mit levels, suggesting that Ca²⁺ uptake was limiting. It seems unlikely that Ca²⁺ itself can directly mediate desensitization of the Ca²⁺ uniporter, because the inhibition of Ca²⁺ uptake persists when Ca²⁺_cyt and Ca²⁺_mit have fully recovered. Incubation of the cells with the general kinase inhibitors H7 (100 µM) and KN93 (1 µM) did not prevent the desensitization of Ca²⁺_mit uptake (data not shown), arguing against a long-term effect of uniporter phosphorylation.

The desensitization of Ca²⁺_mit uptake observed in the present study is clearly not a feature of mitochondria in all cell types. In many cases, Ca²⁺_mit mirrors Ca²⁺_cyt increases for prolonged periods (8). Furthermore, rapid pulsatile Ca²⁺_cyt increases can cause additive stepwise Ca²⁺_mit increases in some cells (46, 47).

The rapid desensitization of Ca²⁺_mit uptake by repetitive Ca²⁺ oscillations (Fig. 6) contrasts with the prolonged Ca²⁺ accumulation observed with CCE in the presence of 250 µM extracellular Ca²⁺ (Fig. 5). In the latter experiment, Ca²⁺_mit uptake persisted for the 40-min experimental duration, gradually causing a Ca²⁺_mit increase in excess of that achieved with CCE in the presence of 1.8 mM extracellular Ca²⁺ (Fig. 5B). These apparently paradoxical observations could be reconciled if the amplitude of the Ca²⁺_cyt rise determined the desensitization of Ca²⁺_mit uptake. We suggest that Ca²⁺_cyt signals below the 400-600 nM threshold for rapid Ca²⁺_mit accumulation do not cause desensitization. In contrast, larger Ca²⁺_cyt increases cause desensitization to limit Ca²⁺_mit uptake.

In summary, our data indicate that multiple factors determine the ability of mitochondria to respond to Ca²⁺_cyt signals. The characteristics of the Ca²⁺_cyt signal are clearly important in determining the ability of mitochondria to sequester Ca²⁺. In addition, the lag between Ca²⁺_cyt rises and Ca²⁺_mit increases, uniporter desensitization, and the relative positions of InsP₃ receptors and mitochondria all control the extent of the mitochondrial response. Furthermore, we would suggest that substantial Ca²⁺_mit accumulation can occur when Ca²⁺_cyt is modestly elevated above basal levels and in situations in which Ca²⁺_cyt microdomains are not present.

	FOOTNOTES

^* This work was supported by the Biotechnology and Biological Sciences Research Council.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.

Supported by a Royal Society University Research Fellowship.

^§ To whom correspondence should be addressed. Tel.: 01223-496515; Fax: 01223-496033; E-mail: peter.lipp@bbsrc.ac.uk.

Published, JBC Papers in Press, May 1, 2001, DOI 10.1074/jbc.M101101200

	ABBREVIATIONS

The abbreviations used are: Ca²⁺_cyt, cytosolic Ca²⁺; Ca²⁺_mit, mitochondrial Ca²⁺; InsP₃, inositol 1,4,5-trisphosphate; ER, endoplasmic reticulum; EM, extracellular medium; CCE, capacitative Ca²⁺ entry; AM, acetoxymethyl ester..

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES