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

This Article Abstract Full Text (PDF) All Versions of this Article: 292/2/C896    most recent 00649.2005v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Kuba, M. Articles by Kuba, K. Search for Related Content PubMed PubMed Citation Articles by Kuba, M. Articles by Kuba, K.

RECEPTORS AND SIGNAL TRANSDUCTION

Bidirectional Ca²⁺ coupling of mitochondria with the endoplasmic reticulum and regulation of multimodal Ca²⁺ entries in rat brown adipocytes

Laboratory of Anatomy and Physiology, School of Nutritional Sciences, Nagoya University of Arts and Sciences, Nissin, Aichi, Japan

Submitted 20 December 2005 ; accepted in final form 17 September 2006

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
How the endoplasmic reticulum (ER) and mitochondria communicate with each other and how they regulate plasmalemmal Ca²⁺ entry were studied in cultured rat brown adipocytes. Cytoplasmic Ca²⁺ or Mg²⁺ and mitochondrial membrane potential were measured by fluorometry. The sustained component of rises in cytoplasmic Ca²⁺ concentration ([Ca²⁺]_i) produced by thapsigargin was abolished by removing extracellular Ca²⁺, depressed by depleting extracellular Na⁺, and enhanced by raising extracellular pH. FCCP, dinitrophenol, and rotenone caused bi- or triphasic rises in [Ca²⁺]_i, in which the first phase was accompanied by mitochondrial depolarization. The FCCP-induced first phase was partially inhibited by oligomycin but not by ruthenium red, cyclosporine A, U-73122, a Ca²⁺-free EGTA solution, and an Na⁺-free solution. The FCCP-induced second phase paralleling mitochondrial repolarization was partially blocked by removing extracellular Ca²⁺ and fully blocked by oligomycin but not by thapsigargin or an Na⁺-deficient solution, was accompanied by a rise in cytoplasmic Mg²⁺ concentration, and was summated with a high pH-induced rise in [Ca²⁺]_i, whereas the extracellular Ca²⁺-independent component was blocked by U-73122 and cyclopiazonic acid. The FCCP-induced third phase was blocked by removing Ca²⁺ but not by thapsigargin, depressed by decreasing Na⁺, and enhanced by raising pH. Cyclopiazonic acid-evoked rises in [Ca²⁺]_i in a Ca²⁺-free solution were depressed after FCCP actions. Thus mitochondrial uncoupling causes Ca²⁺ release, activating Ca²⁺ release from the ER and store-operated Ca²⁺ entry, and directly elicits a novel plasmalemmal Ca²⁺ entry, whereas Ca²⁺ release from the ER activates Ca²⁺ accumulation in, or release from, mitochondria, indicating bidirectional mitochondria-ER couplings in rat brown adipocytes.

plasmalemmal calcium entry; calcium release; mitochondrial depolarization; FCCP

The role of mitochondria as a Ca²⁺ sink has been known in both excitable and nonexcitable cells (3, 14). At high [Ca²⁺]_i, mitochondria take up and accumulate Ca²⁺, whereas they release Ca²⁺ once the [Ca²⁺]_i level is decreased by other Ca²⁺ clearance mechanisms, thus buffering a transient rise in [Ca²⁺]_i (6–8, 14, 38, 45, 49). Moreover, the active role of mitochondria as a Ca²⁺ sink has recently been suggested in the mechanism of regulation of SOC. Despite the low affinity of the Ca²⁺ uniporter at the mitochondrial membrane (3), the localization of mitochondrial Ca²⁺ uptake sites close to the Ca²⁺-release sites of the ER enables the mitochondrial uniporter to take up Ca²⁺ from the high [Ca²⁺]_i produced in the microdomain between these organelles in several types of cell (35, 43–45, 51). This prevents refilling of the ER with Ca²⁺ by sarco(endo)plasmic Ca²⁺ pumps (SERCA) so that the activation of SOC is facilitated or maintained (11, 15, 36, 42). It is of great interest to study how these couplings of mitochondria and the ER operate in cells, where both organelles play important physiological roles.

Brown adipocytes possess numerous mitochondria and utilize them for thermogenesis, their major function (48). In response to cold exposure, heat is produced by uncoupling of respiration from ATP synthesis by the activation of uncoupling proteins under the enhanced hydrolysis of neutral lipids by the -action of norepinephrine (28, 32, 41). In mitochondrial respiration, Ca²⁺ is needed for the activation of Ca²⁺-dependent production of NADH and FADH₂, which is the absolute requirement of thermogenesis. In brown adipocytes, [Ca²⁺]_i is largely increased by the release of Ca²⁺ from the ER through IP₃ receptors (25, 30, 47, 54) and the subsequent activation of SOC (34) in response to the -action of norepinephrine. Thus one mode of norepinephrine action increasing [Ca²⁺]_i would be complementary to another mode of action generating heat in brown adipocytes. It is then possible that these functions of mitochondria and the ER interact with each other via Ca²⁺ dynamics. We have studied this issue in brown adipocytes and how such a coupling, if it exists, regulates Ca²⁺ entry at the plasma membrane. These problems were investigated by pharmacologically activating Ca²⁺ release from mitochondria by uncouplers and from the ER by inhibitors of the SERCA. The results demonstrate that Ca²⁺ release from mitochondria activates a new mode of plasmalemmal Ca²⁺ entry and causes Ca²⁺ depletion in the ER that activates SOC in a fraction of cells, whereas Ca²⁺ release from the ER results in depletion of Ca²⁺ in mitochondria as well as activation of SOC. Thus there is the two-way coupling of mitochondria and the ER, whose direction and tightness of the coupling differ in individual cells, thereby regulating Ca²⁺ entry in rat brown adipocytes. The results were reported in abstract form (23).

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Rats (Wistar ST or WKAH supplied from SLC Japan; 3–6 wk old) were used. Rats were kept at 4°C with free access to food and water for 5–7 h to deplete lipids stored in brown adipocytes, anesthetized with ether, and then killed. All procedures were performed in accordance with the Animal Experimentation Guidelines of Nagoya University. Interscapular brown adipose tissues were isolated and minced. Adipocytes were differentiated by treatment with collagenese type 2 and DNase I for 20 min, triturated three to four times every 5 min, and cultured for 2–7 days (34). Culture medium was composed of DMEM, 10% FBS, penicillin, and streptomycin. Adipocytes were loaded with fura 2-AM (5 µM), magfura 2-AM (5 µM), or Oregon green 488 BAPTA-1 (5 µM) for 30–50 min at 37.0°C for [Ca²⁺]_i measurement. In most experiments, rhodamine 123 (10 µM) was loaded for 10–15 min at 37.0°C after loading fura 2-AM or magfura 2-AM to record changes in the membrane potential of mitochondria (21).

Composition of normal Krebs-Ringer solution was (in mM) 112.8 NaCl, 5.6 KCl, 3.0 CaC1₂, 1.1 MgSO₄, 4.8 NaHCO₃, 2.3 NaH₂PO₄, 20.0 HEPES, and 5.4 glucose (pH 7.4 adjusted with NaOH). A nominally Ca²⁺-free solution was prepared by replacing total CaC1₂ with 4.5 mM NaCl in Krebs-Ringer. A Ca²⁺-free EGTA solution was made by adding 1 mM EGTA to the nominally Ca²⁺-free solution. A Ca²⁺-free, Mg²⁺-free EGTA solution was prepared by replacing total MgCl₂ with equimolar amounts of NaCl in a Ca²⁺-free EGTA solution. An Na⁺-deficient solution was made by replacing total NaCl with N-methyl-D-glucamine in equimolar amount and titrated with 1.0 N HCl. An Na⁺-free solution was prepared by replacing NaCl, NaHCO₃, HEPES, and NaH₂PO₄ with Tris·Cl (pH 7.4 adjusted with HCl). High-pH solutions were prepared by adding Tris·Cl in an appropriate amount. Various kinds of drugs were applied by changing a perfusing solution to the solution containing a drug(s). Time for complete exchange of solution in the bath was <1 min.

Changes in [Ca²⁺]_i were measured with two conventional Ca²⁺-imaging systems (charge-coupled device camera, C4742–12R; Hamamatsu Photonics) set on an upright microscope (Olympus, BX51W1 with an objective, x40 water, numerical aperture 0.8) or an inverted microscope (Nikon, TMD300 with an objective, x40 water, numerical aperture 1.15). In most experiments, fluorescence of fura 2 excited alternatively at 340 nm (D340, Chroma Technology, Rockingham, VT; width at half maximum = 16 nm) and 380 nm (D380, Chroma Technology; width at half maximum = 20 nm); results were recorded through a band-pass filter (HQ480/40; Chroma Technology) using a filter changer (EFC-1; Nikon, Tokyo, Japan) for the upright microscope or another changer (C8214; Hamamatsu Photonics) for the inverted microscope. One sequence of fluorescence data was obtained every 5–15 s. Fluorescence intensity was averaged over the region within the contour of each cell in an image plane. The ratio of fluorescence excited at 340 nm to that at 380 nm, (F₃₄₀/F₃₈₀) was analyzed by software (Aquacosmos, Hamamatsu Photonics). F₃₄₀/F₃₈₀ was converted to a [Ca²⁺]_i value by the equation reported in Ref. 13 using the dissociation constant of 145 nM, the ratio of the maximum F₃₄₀/F₃₈₀ to the minimum (14.1), and the fluorescence ratio of the free to the Ca²⁺-bound form (9.59). The ratio of magfura 2 fluorescence excited at 340 and 380 nm was taken but was not converted to a cytoplasmic Mg²⁺ concentration ([Mg²⁺]_i) value because it would have predominantly represented [Ca²⁺]_i, [Mg²⁺]_i, or both, depending on the condition of the cell (see text). Changes in mitochondrial membrane potential were recorded by measuring fluorescence of rhodamine 123 excited at 480 nm alternatively with fura 2 fluorescence at 340 or 380 nm, using either of the filter changers. In this experiment, fluorescence was separated by a dichroic mirror at >505 nm and recorded through a band-pass filter (HQ535/50; Chroma Technology).

Statistical analyses were made by Student's t-test, when necessary.

Collagenese type 2 (class 2) was obtained from Worthington Biochemical (Lakewood, NJ), and DNase I was from Roche Diagonistic (Indianapolis, IN). DMEM (low glucose type, 11885-084), penicillin, and streptomycin were from Invitrogen. Thapsigargin, cyclopiazonic acid (CPA), and rhodamine 123 were from Sigma-Aldrich (St. Louis, MO). FBS was from Thermo Electron (Melbourne, Australia). FCCP, oligomycin, rotenone, and ruthenium red were from Sigma-Aldrich. Dinitrophenol (DNP) was from Wako Pure Chemical. Cyclosporine A and U-73122 were from Merck. Fura 2-AM was from Molecular Probes (Eugene, OR) or Dojin (Kumamoto, Japan). Magfura 2-AM and Oregon green 488 BAPTA-1-AM were from Molecular Probes.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Blocking Ca²⁺ pump of the ER membrane liberates Ca²⁺ from the ER and activates SOC. The cells that showed basal levels of [Ca²⁺]_i ranging from 23.5 to 185.8 nM (n = 417) were assumed to be sound. The cells that showed [Ca²⁺]_i values >200 nM were eliminated from analyses because they often failed to show a phasic rise in [Ca²⁺]_i in response to norepinephrine, a natural agonist (unpublished observations). With this criterion, the basal [Ca²⁺]_i level of rat brown adipocytes in culture was judged to be 78.5 ± 2.4 (SE) nM (n = 417). Thapsigargin (1 µM) applied for 4–6 min caused a rise in [Ca²⁺]_i, which occurred in a mono- or biphasic manner (Fig. 1A). The increased [Ca²⁺]_i reached the peak (192.5 ± 20.9 nM, n = 82) within 10 min and very slowly declined over more than several tens of minutes to a plateau (152.3 ± 18.6 nM, n = 30). In the cells (n = 12) in which the increase in [Ca²⁺]_i was biphasic, the initial and second phases were 64.7 ± 12.5 and 168.1 ± 13.1 nM, respectively. The initial quick rise is likely the result of Ca²⁺ leak from the ER (27) by stopping the Ca²⁺ uptake by thapsigargin. The sustained high level of [Ca²⁺]_i during the plateau could be caused by either an increase in Ca²⁺ entry at the cell membrane or a decrease in Ca²⁺ extrusion.

View larger version (14K): [in this window] [in a new window]   Fig. 1. Thapsigargin-induced rises in cytosolic Ca2+ concentration ([Ca2+]i) and effects of Ca2+-free EGTA, Na+-deficient, and high-pH solutions. A: effects of Ca2+-free EGTA solution. A solution containing thapsigargin (1 µM) or Ca2+-free EGTA solution was superfused to the bath during a period indicated by black and gray horizontal bars, respectively. B: effects of Na+-deficient solution. A solution containing thapsigargin (1 µM) or Na+-deficient solution was superfused to the bath during a period indicated by black or gray horizontal bars, respectively. C: effects of high-pH (pH 8.3) solution. A solution containing thapsigargin (1 µM) or high-pH solution was superfused to the bath during a period indicated by black and gray horizontal bars, respectively.

We next examined a possible involvement of the inhibition of the plasma membrane Ca²⁺ pump in the plateau phase. Increasing external pH to 8.3 for 2 min, a condition that inhibits plasmalemmal Ca²⁺ pump (4), caused no or a small rise in [Ca²⁺]_i (to 13.0 ± 8.0 nM, n = 8) under the control condition. On the other hand, raising the external pH during the plateau phase under the effect of thapsigargin produced a large rise in [Ca²⁺]_i (by 36.2 ± 6.4 nM, n = 8; Fig. 1C). This implies that the activity of Ca²⁺ pumps at the cell membrane is enhanced, but not suppressed, during the plateau, counterbalancing the increased Ca²⁺ entry. Alternatively, the increase in [Ca²⁺]_i by raising pH during the plateau, but not at rest, could be explained by the effect of high pH to enhance SOC (24). In any cases, it is clear that the high level of [Ca²⁺]_i during the plateau is due to the activation of SOC as a result of Ca²⁺ depletion in the ER by the action of thapsigargin, as has been suggested in this type of cell (34) as well as many other cells (37). It is to be noted that thapsigargin did not affect mitochondrial membrane potential (see Fig. 13).

View larger version (13K): [in this window] [in a new window]   Fig. 2. Rises in [Ca2+]i and mitochondrial membrane depolarization induced by FCCP and effects of a Ca2+-free EGTA solution. A: triphasic rises in [Ca2+]i with mitochondrial membrane depolarization and effects of a Ca2+-free EGTA solution on the third phase of rises. Bottom traces: time course of changes in fluorescence of rhodamine 123 at 480 nm (F480), reflecting change in mitochondrial membrane potential. An increase in F480 implies a decrease in membrane potential. A solution containing FCCP (2 µM) or a Ca2+-free EGTA solution was superfused to the bath during periods indicated by black and gray horizontal bars, respectively. B: diphasic rises in [Ca2+]i with membrane depolarization and effects of a Ca2+-free EGTA solution. Explanations are similar to those in A.

View larger version (14K): [in this window] [in a new window]   Fig. 3. Effects of high-pH and Na+-deficient solutions on FCCP-induced rises in [Ca2+]i and effects of the latter on mitochondrial membrane depolarization. A: effects of high-pH solution on the second, spiky component of FCCP-induced rises in [Ca2+]i. FCCP (2 µM) was applied during period indicated by black bar. A high-pH (pH 8.2) solution was superfused to the bath during period indicated by gray bar. B: effects of Na+-deficient solution on FCCP-induced rises in [Ca2+]i and accompanied mitochondrial membrane depolarization. Bottom trace: time course of changes in F480, reflecting change in mitochondrial membrane potential. FCCP (2 µM) was applied during period indicated by black bar. Na+-deficient solution was superfused to the bath during period indicated by gray bar.

View larger version (13K): [in this window] [in a new window]   Fig. 4. Effects of high-pH and Na+-deficient solutions on the third phase of FCCP-induced rises in [Ca2+]i. A: effects of high-pH solution on the third phase of FCCP-induced rises in [Ca2+]i. FCCP (2 µM) was applied during period indicated by black bar. High-pH (pH 9.0) solution was superfused to the bath during period indicated by dark gray bar. B: effects of Na+-deficient solution on the third phase of FCCP-induced rises in [Ca2+]i. FCCP (2 µM) was applied during period indicated by black bar. Na+-deficient solution was superfused to the bath during period indicated by dark gray bar. Ca2+-free EGTA solution was superfused to the bath during period indicated by light gray bar.

View larger version (13K): [in this window] [in a new window]   Fig. 5. Effects of Ca2+-free EGTA solution on rises in [Ca2+]i induced by dinitrophenol (DNP) and rotenone and those on mitochondrial membrane potential changes evoked by rotenone. A and B: effects of Ca2+-free EGTA (A) and nominally Ca2+-free (B) solutions on changes in Oregon green 488 BAPTA-1 fluorescence induced by DNP. DNP (50 µM) was applied for 6 min (black bar), whereas Ca2+-free EGTA or a nominally Ca2+-free solution was superfused for the period indicated by gray bars. In A, cell exhibited diphasic rises in [Ca2+]i in response to DNP action in Krebs solution. In B, cell produced triphasic rises in [Ca2+]i. C: effects of Ca2+-free EGTA solution on rises in [Ca2+]i and mitochondrial membrane potential changes evoked by rotenone. Rotenone (1 µM) was applied for 6 min (black bar), whereas Ca2+-free EGTA solution was superfused for the period indicated by gray bar.

View larger version (15K): [in this window] [in a new window]   Fig. 6. Effects of Na+-free solution, ruthenium red, and cyclosporine A on the initial and second phases of FCCP-induced rises in [Ca2+]i and the associated mitochondrial membrane potential changes. A: effects of Na+-free solution on FCCP-induced rises in [Ca2+]i and the associated membrane potential changes. Bottom trace shows relative changes in F480 to the initial fluorescence, reflecting those in mitochondrial membrane potential. Increase in fluorescence ratio indicates depolarization of membrane potential. FCCP (2 µM) was applied for 3 min as shown by black bar, whereas Na+-free solution was perfused for the period indicated by gray bar. B: effects of ruthenium red on FCCP-induced rises in [Ca2+]i and the associated membrane potential changes. Explanations are the same as those in A except that ruthenium red (5 and 10 µM) was applied. C: effects of cyclosporine A on FCCP-induced rises in [Ca2+]i and the associated membrane potential changes. Explanations are the same as those in A except that cyclosporine A (1 µM) was applied.

View larger version (21K): [in this window] [in a new window]   Fig. 7. Comparison of the time courses of the Ca2+ entry-dependent component of the second spiky rise in [Ca2+]i and mitochondrial depolarization induced by FCCP. Ca2+ entry-dependent component of the second spiky rise (light gray trace) was isolated by subtracting the FCCP-induced response in a Ca2+-free EGTA solution (dark gray trace) from that in Krebs solution (black trace). Changes in mitochondrial membrane potential (Vmito) are shown in the reversed direction (interrupted trace) and superimposed so that a repolarizing phase can be compared with the time course of the Ca2+ entry-dependent component of the second spiky rise. FCCP (2 µM) was applied for 3 min, as shown by black horizontal bar. A and B are the analyses from 2 different cells. [Ca], change in Ca2+ concentration.

View larger version (13K): [in this window] [in a new window]   Fig. 8. Effects of a Ca2+-free, Mg2+-free EGTA solution on FCCP-induced changes in magfura 2 fluorescence ratio and effects of oligomycin on FCCP-induced rises in [Ca2+]i and mitochondrial membrane depolarization. A: effects of a Ca2+-free, Mg2+-free EGTA solution on the FCCP-induced changes in the ratio of magfura 2 fluorescence. Ratio of fluorescence at 340 and 380 nm was taken. FCCP (2 µM) was applied for 3 min as shown by black bar, whereas Ca2+-free, Mg2+-free EGTA solution was applied for the period indicated by gray bar. B: effects of oligomycin on FCCP-induced rises in [Ca2+]i and mitochondrial membrane depolarization. FCCP (2 µM) was applied for 3 min as shown by black bar, whereas oligomycin solution (5 µg/ml) was applied for the period indicated by gray bar. Bottom traces (F480) represent changes in mitochondrial membrane potential.

View larger version (13K): [in this window] [in a new window]   Fig. 9. Effects of the prior application of cyclopiazonic acid (CPA) on FCCP-induced rises in [Ca2+]i in Ca2+-free EGTA solution and effects of prior application of FCCP on CPA-induced rises in [Ca2+]i in Ca2+-free EGTA solution. A: effects of the preceding application of CPA on FCCP-induced rises in [Ca2+]i in Ca2+-free EGTA solution. FCCP (2 µM) was applied for 3 min as shown by black bar. Ca2+-free EGTA solution was applied for the period indicated by light gray bar, whereas CPA (10 µM) was applied for 8 min as indicated by dark gray bar. Bottom traces show relative changes in F480 to the initial fluorescence, reflecting changes in mitochondrial membrane potential. B: effects of the preceding application of FCCP on CPA-induced rises in [Ca2+]i in Ca2+-free EGTA solution. Explanations are similar to those in A.

View larger version (15K): [in this window] [in a new window]   Fig. 10. Functional Ca2+ couplings between mitochondria and the ER. A: Ca2+ coupling of mitochondria to the ER. FCCP produces triphasic rises in [Ca2+]i. Thapsigargin had no effects on [Ca2+]i during the third phase of rises in [Ca2+]i, and the third phase was completely abolished by removal of external Ca2+. FCCP (2 µM) and thapsigargin (1 µM) were applied during the period indicated by black and dark gray bars, respectively. Ca2+-free EGTA solution was superfused to the bath during period indicated by light gray bar. B: bidirectional Ca2+ couplings between mitochondria and the ER. FCCP produces triphasic rises in [Ca2+]i. Thapsigargin had again no effects on [Ca2+]i during the third phase of FCCP-induced rises in [Ca2+]i, and the third phase was completely abolished by removal of the external Ca2+ and transiently by the second application of FCCP. Other explanations are the same as those in A.

View larger version (22K): [in this window] [in a new window]   Fig. 11. Effects of a blocker of PLC on FCCP-induced rises in [Ca2+]i and associated changes in mitochondrial membrane potential in Krebs and Ca2+-free EGTA solutions. A: effects of U-73122 (2 µM) on FCCP-induced rises in [Ca2+]i and mitochondrial membrane potential changes in a Ca2+-free EGTA solution. FCCP (2 µM) was applied for 3 min as shown by black horizontal bar. A solution containing U-73122 (2 µM) was applied for the period indicated by dark gray bar, and a Ca2+-free EGTA solution was applied for the period indicated by light gray bar. Ba: superposed traces of FCCP-induced rises in [Ca2+]i in the absence and presence of U-73122 (2 µM) and the U-73122-sensitive component in a Ca2+-free solution shown in A. Bb: averaged U-73122-sensitive component (mean ± SE) obtained from 10 cells. C: effects of U-73122 (2 µM) on FCCP-induced rises in [Ca2+]i and mitochondrial membrane potential changes in Krebs solution. Explanations are the same as those in A except that U-73122 was applied in Krebs solution. Da: superposed traces of FCCP-induced rises in [Ca2+]i in the absence and presence of U-73122 (2 µM) and the U-73122-sensitive component of FCCP-induced response in Krebs solution shown in C. Db: superposed traces of control responses and those in the presence of U-73122 (2 µM) and the U-73122-sensitive component in Krebs solution shown in C. A part of the trace in the presence of U-73122 indicated by a light gray bar in Da was deleted to eliminate the delay in the onset of the second component of FCCP-induced rises. Dc: averaged U-73122-sensitive component (mean ± SE) obtained from 13 cells, in which a part of the trace in the presence of U-73122 was deleted.

View larger version (14K): [in this window] [in a new window]   Fig. 12. Effects of FCCP on [Ca2+]i during the sustained rise in [Ca2+]i induced by thapsigargin. A: reduction of [Ca2+]i by FCCP (2 µM) during the thapsigargin-induced rise in [Ca2+]i. The second spiky component of rises remains during the FCCP-induced reduction of [Ca2+]i but is blocked in Ca2+-free EGTA solution. FCCP and thapsigargin were applied during the period indicated by black and dark gray bars, respectively. A Ca2+-free EGTA solution was superfused to the bath during the period indicated by light gray bar. B: diphasic rises induced by FCCP during the sustained rise in [Ca2+]i caused by thapsigargin. The reduction of [Ca2+]i after the initial transient rise underlies the second spiky rise. FCCP and thapsigargin were applied during the period indicated by black and dark gray bar, respectively. CPA (10 µM) was applied during the period indicated by light gray bar to examine the potency of thapsigargin to deplete Ca2+ in the ER.

View larger version (13K): [in this window] [in a new window]   Fig. 13. Top: FCCP-induced rises in [Ca2+]i and mitochondrial membrane depolarization during the thapsigargin-induced rise in [Ca2+]i. Bottom: time course of changes in F480, reflecting a change in mitochondrial membrane potential. FCCP (2 µM) was applied during the period indicated by black bar. Thapsigargin (1 µM) was applied to the bath during the period indicated by gray bar.

FCCP (2 µM) applied for 2–5 min produced rises in [Ca²⁺]_i in a biphasic or triphasic manner in rat brown adipocytes. The initial transient was 60.6 ± 5.1 nM (n = 283), and the subsequent spiky rise was 28.9 ± 2.9 nM (measured from the end of the first phase, n = 279; Figs. 2–4, 6–9, and 11–13). In a fraction of cells, these phases of rises were followed by a late slow rise, the amplitude of which was largely varied from 6.6 to 430 nM among cells (68.5 ± 7.7 nM, n = 109; Figs. 2A, 4, 10, and 12A). The mechanisms of these three phases of FCCP-induced rises were examined below.

The initial transient phase was consistently accompanied by a transient increase in rhodamine 123 fluorescence (by 50.5 ± 4.0% of the basal intensity, n = 84) in all of the cells where it was measured (Figs. 2, 3B, 6–9, 11, and 13), reflecting a decrease in mitochondrial membrane potential. The mitochondrial depolarization induced by FCCP (2 µM) applied for 3 min lasted for 15–20 min, indicating that FCCP remains in the cell during this period. The initial transient phase was generated in a nominally Ca²⁺-free solution (65.1 ± 8.1 nM, n = 36) or a Ca²⁺-free EGTA solution (60.0 ± 5.1 nM, n = 36; Figs. 2B, 7, 9, and 11) and an Na⁺-deficient solution applied for a short period (100.4 ± 7.5% of the control, n = 22; Fig. 3B) or in an Na⁺-free solution superfused for a long period (see below). Thus the initial phase is apparently caused by Ca²⁺ release from mitochondria.

The second spiky rise in [Ca²⁺]_i was partially abolished in a nominally Ca²⁺-free or Ca²⁺-free EGTA solution, leaving a small external Ca²⁺-independent component in two-thirds of the cells (n = 36). The "apparent" amplitude of the remaining component was 5.6 ± 0.9 nM (n = 23) and 8.0 ± 1.4 nM (n = 23), respectively, when measured from the level just before the remaining component (Figs. 2B and 7). (The real magnitude revealed by a blocker of PLC, however, was much greater than these values; see Fig. 11, B and D.)

There are two possible mechanisms for the external Ca²⁺-dependent component. First, blockade of mitochondrial energetics by FCCP would cause an increase in the cytosolic H⁺ concentration (5) via ATP hydrolysis, which enhances Na⁺ entry via the Na⁺/H⁺ exchanger at the plasma membrane and subsequently inhibits the plasmalemmal Na⁺/Ca²⁺ exchanger, thereby increasing [Ca²⁺]_i. The second spiky component, however, was not affected by an Na⁺-deficient solution applied for a short period (123.0 ± 8.6% of the control, n = 22; Fig. 3B), thus ruling out the role of the plasmalemmal Na⁺/H⁺ exchanger and Na⁺/Ca²⁺ exchanger in the mechanism. The possible involvement of the inhibition of Ca²⁺ pump at the cell membrane was tested by raising the external pH that inhibits its activity. The spiky rise, however, was enhanced in a high-pH (pH 8.2) solution to 82.6 ± 10.1 nM from 31.3 ± 3.7 nM (n = 44, P < 0.0001; Fig. 3A). Because an increase in [Ca²⁺]_i by a high-pH solution alone was 47.6 ± 3.9 nM (n = 47), the enhancement of the second spiky rise appeared to result from the simple summation of the rise caused by the high-pH-induced inhibition of Ca²⁺ pumps to the FCCP-induced spiky rise. The results thus indicate that mitochondrial depolarization or accompanying Ca²⁺ release activates a transient Ca²⁺ entry at the cell membrane and a Ca²⁺ release presumably from other Ca²⁺ stores. These mechanisms are described in more detail later.

The late slow rise was abolished by the removal of external Ca²⁺ (Figs. 2A and 10). The high level of [Ca²⁺]_i during the late slow phase was quickly reduced to a level (23.4 ± 1.6 nM, n = 60) lower than the initial basal [Ca²⁺]_i in a Ca²⁺-free EGTA solution. Na⁺-deficient solution reduced the high level of [Ca²⁺]_i during the late slow phase (by 54.0 ± 12.0 nM, n = 28; Fig. 4B) with no effect under the control condition (3.0 ± 0.7 nM, n = 28). This effect of reducing the external Na⁺ is similar to that on the sustained rise in [Ca²⁺]_i by thapsigargin, suggesting the role of Na⁺ in the mechanism of high level of [Ca²⁺]_i, not the inhibition of the Na⁺/Ca²⁺ exchanger at the plasma membrane (see below and DISCUSSION). The depressant effect of an Na⁺-deficient solution on the late slow rise produced by FCCP might be caused by the mechanism involving the Na⁺/H⁺ exchanger at the plasma membrane as discussed above as a mechanism for the second spiky rise. This mechanism, however, is unlikely because mitochondrial membrane potential and ATP synthesis are well restored during the third phase (see DISCUSSION). Raising the external pH during the late slow rise in [Ca²⁺]_i caused a sharp, transient, large rise in [Ca²⁺]_i (26.5 ± 6.6 times the control, n = 16, P < 0.001; Fig. 4A). Thus the late slow rise in [Ca²⁺]_i by FCCP is not caused by the inhibition of Ca²⁺ extrusion via the plasmalemmal Ca²⁺ pump but by the activation of Ca²⁺ entry.

To confirm whether these effects of FCCP result from the actions on mitochondria, we then applied another protonophore, DNP, and a blocker of electron transport, rotenone. DNP (50 µM) produced di- or triphasic rises in [Ca²⁺]_i similar to those by FCCP. The amplitude of the initial phase was 26.6 ± 1.7% (n = 45) of the basal fluorescence intensity, whereas that of the second phase was 18.7 ± 1.3% (measured from the initial base; Fig. 5, A and B). The second phase was largely abolished by a Ca²⁺-free EGTA solution with or without a small remaining transient rise. In separate experiments, DNP (50 µM) depolarized mitochondrial membrane potential (not shown). Likewise, rotenone caused multiphasic rises in [Ca²⁺]_i (Fig. 5C), which were accompanied by mitochondrial depolarization. A Ca²⁺-free solution abolished the later phase of [Ca²⁺]_rise and suppressed that of depolarizaton (see below for possible reasons). The initial phase was 70.4 ± 7.0 nM (n = 30), whereas the second phase was 89.0 ± 20.0 nM (n = 30; when measured from the base) or 13.1 ± 5.2 nM (n = 5; when measured from the level after the end of the first phase). Thus these actions of DNP and rotenone are similar to those of FCCP except for depression of rotenone-induced mitochondrial depolarization in a Ca²⁺-free solution. Consequently, the actions of FCCP on [Ca²⁺]_i indeed result from their actions on mitochondria. {The rotenone-induced depolarization, particularly in the later phase, was strongly depressed in a Ca²⁺-free solution. Reductions in the basal [Ca²⁺]_i and/or mitochondrial Ca²⁺ uptake under this condition would have prevented the opening of permeability transition pore (PTP; Ref. 50) that may have underlain the later phase. Alternatively, the decreased mitochondrial Ca²⁺ uptake should have depressed TCA cycle and reduced the activity of electron transport and so membrane potential. Under this condition, FCCP-induced depolarization would be smaller.} The mechanisms of the three phases of FCCP-induced rises were further analyzed below.

Mitochondrial Ca²⁺ release occurs through a Ca²⁺ pathway other than Ca²⁺ uniporter, Na⁺/Ca²⁺ exchanger, or PTP. The initial transient phase was not affected by long-term treatment with an Na⁺-free solution (>30 min: to 95.5 ± 8.0% of the control, n = 41; Fig. 6A), which would have reduced intracellular Na⁺ concentration. We next examined effects of ruthenium red, which blocks Ca²⁺ uniporters at the mitochondrial membrane with a high affinity of 9 nM (22). Although ruthenium red is known to be membrane impermeable, it does permeate the plasma membrane when applied extracellularly at a high (µM) concentration (9). The permeation of ruthenium red across the cell membrane can be assessed by the action to reduce rhodamine 123 fluorescence via its direct action and/or the mitochondrial hyperpolarization as a result of blockade of Ca²⁺ uniporter (9). Ruthenium red applied at concentrations of 5, 10, and 50 µM indeed decreased rhodamine 123 fluorescence to 88.7 ± 1.3% (n = 50), 80.9 ± 1.5% (n = 11), and 21.9 ± 0.3% (n = 7), respectively, of that before application (see Fig. 6B for the data at 5 and 10 µM). Furthermore, it increased (or did not affect) FCCP-induced mitochondrial depolarization to 116.3 ± 8.6% (n = 62), 106.3 ± 8.9% (n = 11), and 323.6 ± 29.7% (n = 7), respectively, of the control (see Fig. 6B for the data at 5 and 10 µM) presumably because of mitochondrial hyperpolarization as a result of blockade of Ca²⁺ uptake through Ca²⁺ uniporters. Under this condition, the amplitude of the first phase of FCCP-induced rises in [Ca²⁺]_i was slightly reduced to 93.5 ± 4.1% (n = 62), 85.9 ± 8.4% (n = 11), and 95.6 ± 11.3% (n = 7), respectively, of the control (see Fig. 6B for the data at 5 and 10 µM).

We then tested effects of cyclosporine A, a blocker of PTP. Cyclosporine A (1 µM) did not affect the initial transient phase of FCCP-induced rises (87.6 ± 4.8%, n = 54; Fig. 6C). Thus depolarization of the inner mitochondrial membrane by shunting effects of FCCP causes Ca²⁺ release through a Ca²⁺ pathway other than Ca²⁺ uniporter, Na⁺/Ca²⁺ exchanger, and PTP. By exclusion, an Na⁺-independent Ca²⁺ efflux pathway (1, 7, 14) would play a major role in the generation of the initial transient rise in [Ca²⁺]_i.

Ca²⁺ entry-dependent component of the second spiky rise is related to mitochondrial repolarization. When the time course of the second spiky rise in [Ca²⁺]_i was compared with that of FCCP-induced mitochondrial depolarization, it closely paralleled the repolarizing phase of the depolarization, as shown in Fig. 7, in which the amplitude of depolarization was shown upside down. This indicates that the mechanism is closely associated with the recovery of ATP synthesis that would have accompanied membrane repolarization.

This possibility was examined by recording [Mg²⁺]_i by magfura 2 fluorescence. FCCP increased the ratio of magfura 2 fluorescence at 340 and 380 nm (F340/F380) in a Ca²⁺-free, Mg²⁺-free solution (Fig. 8A). The time course of FCCP-induced changes in the ratio of magfura 2 fluorescence in Krebs solution is more or less similar to that of [Ca²⁺]_i. The amplitudes of the first and second components were 23.0 ± 1.1% and 26.4 ± 1.5% of the initial F₃₄₀/F₃₈₀ value (n = 42). In a Ca²⁺-free, Mg²⁺-free solution, the first component was depressed to 65.4 ± 2.5% of that in Krebs solution (n = 42). The distinct difference between the responses measured with magfura 2 and fura 2 was that the second component of the former was not much depressed in the absence of the external Ca²⁺. The extent of the decrease in the second component of magfura 2 response in a Ca²⁺-free, Mg²⁺-free solution (65.1 ± 8.1%, n = 42) was similar to that of the first component. Furthermore, the FCCP-induced rise in magfura 2 fluorescence ratio was sustained for a certain period even after the end of repolarization. The duration of the FCCP-induced magfura 2 response in a Ca²⁺-free, Mg²⁺-free solution was 146.9 ± 11.8% (n = 38, P < 0.0004; measured at the half maximum of the amplitude of the second peak) of that in Krebs solution. The initial quick rise could be caused by a rise in [Ca²⁺]_i, which was sensed by magfura 2, whereas the later phase must reflect a rise in [Mg²⁺]_i, since the [Ca²⁺]_i level was much reduced during this phase (see Figs. 2B and 7). Because ATP hydrolysis accompanies liberation of Mg²⁺ from ATP molecules, the rise in [Mg²⁺]_i would indicate hydrolysis of ATP during the course of mitochondrial depolarization, whereas the restoration of the initial [Mg²⁺]_i level must reflect the recovery of ATP synthesis. The prolongation of [Mg²⁺]_i rise in a Ca²⁺-free, Mg²⁺-free solution could be due to the sustained blockade of ATP synthesis as the result of a decrease in [Ca²⁺]_i, which should have decreased Ca²⁺ concentration in the mitochondrial matrix. Thus the spiky Ca²⁺ entry is likely produced by a mechanism that depends on the recovery of ATP synthesis.

This idea was further supported by the action of oligomycin, a blocker of H⁺-ATPase. In the presence of oligomycin (2.5–5 µg/ml), the second spiky component of FCCP-induced rises was not generated (n = 44; Fig. 8B). This effect remained even 50 min after its removal. Oligomycin per se produced a small transient rise in [Ca²⁺]_i (17.1 ± 3.2 nM, n = 44), whereas the first phase of FCCP-induced rises was depressed (to 7.5 ± 2.7% of the control, n = 44) in the presence of oligomycin with little change in mitochondrial depolarization (108.3 ± 6.9%, n = 38).

Another characteristic of the spiky Ca²⁺ entry was that it occurs independent of Ca²⁺ release from mitochondria and the ER. The spiky component was produced even after the quick decrease (instead of a rise) in [Ca²⁺]_i by FCCP applied under the effect of thapsigargin (see Fig. 12A). In addition, the spiky Ca²⁺ entry was abolished during the late slow rise produced by the preceding application of FCCP (8.6 ± 3.1 nM, n = 40; see Fig. 10). Thus it is likely that its activation mechanism could be related to the process that activates the subsequent slow phase.

Ca²⁺ release-dependent component of the second phase. There was a residual rise in [Ca²⁺]_i after the first phase of FCCP-induced rises in a nominally Ca²⁺-free or Ca²⁺-free EGTA solution, indicating Ca²⁺ release from Ca²⁺ stores other than mitochondria, namely, the ER. If this is the case, it is expected that, once the ER is depleted of Ca²⁺ by inhibiting Ca²⁺ uptake, the Ca²⁺ release-dependent component of the second phase would disappear. After the application of CPA (10 µM), a blocker of SERCA, FCCP caused only a small monotonic rise in [Ca²⁺]_i in a Ca²⁺-free EGTA solution (5.0 ± 1.4% of the control, n = 7; Fig. 9A), indicating the depression of the first phase of FCCP-induced responses and the disappearance of Ca²⁺ release involved in the second phase. Thus it is likely that Ca²⁺ release from the ER is involved in the second phase. Furthermore, the results indicate a decrease in free Ca²⁺ concentration in mitochondria after Ca²⁺ depletion of the ER (see below).

Mitochondrial uncoupling depletes Ca²⁺ in the ER and activates SOC. The foregoing results suggest that mitochondrial uncoupling causes Ca²⁺ release from the ER. If so, the ER would be depleted of Ca²⁺ under this condition. The amount of Ca²⁺ stored in the ER was estimated by means of the action of CPA to deplete Ca²⁺ in the ER in a Ca²⁺-free EGTA solution. This must be reflected in the magnitude of CPA-induced rise in [Ca²⁺]_i in a Ca²⁺-free EGTA solution. The CPA-induced rise in [Ca²⁺]_i in a Ca²⁺-free solution was markedly decreased to 32.6 ± 5.2% of the control (n = 22) after the application of FCCP, which produced a transient rise (Fig. 9B). This was further assessed by the action of thapsigargin on the late phase of FCCP-induced rises in [Ca²⁺]_i. Application of thapsigargin during the late slow phase affected little the high level of [Ca²⁺]_i (1.73 ± 1.77 nM, n = 32; Fig. 10), indicating the depletion of Ca²⁺ in the ER. Thus it is possible that the Ca²⁺ released from mitochondria activates Ca²⁺ release from the ER presumably via a Ca²⁺-induced Ca²⁺ release (CICR) mechanism through IP₃ receptors or ryanodine receptors, if not the former. This would deplete Ca²⁺ in the ER and activate SOC. Alternatively, the mitochondrial matrix may exchange Ca²⁺ with the inside of the ER via a directly coupled Ca²⁺ pathway. Interestingly, it was also found that Ca²⁺ depletion in the ER caused by mitochondrial Ca²⁺ release maintained the state of Ca²⁺ depletion in mitochondria in a fraction of cells (Fig. 10B). The high level of [Ca²⁺]_i during the late slow rise in [Ca²⁺]_i produced by the first application of FCCP was quickly lowered by the second application of FCCP in a fraction of cells (by 34.6 ± 6.0 nM, n = 27) among 57 cells, in which FCCP was applied (Fig. 10B).

Involvement of PLC in the mitochondrial coupling-induced Ca²⁺ release from the ER. To examine a possible involvement of PLC in the activation of Ca²⁺ release from the ER by mitochondrial uncoupling or Ca²⁺ release, we observed effects of a blocker of PLC, U-73122, on the residual second component of FCCP-induced rises in a Ca²⁺-free solution. This component of the rise in a nominally Ca²⁺-free solution was abolished in the presence of 2 µM U-73122 (Fig. 11, A and B; n = 17). Similar effects of U-73122 were obtained in a Ca²⁺-free EGTA solution (n = 5; not shown). Under this condition, there was no change in the amplitude of mitochondrial depolarization (106.3 ± 8.4%, n = 17), although the decay phase was somewhat facilitated (Fig. 11A). The U-73122-sensitive component of FCCP-induced rises in a Ca²⁺-free or Ca²⁺-free EGTA solution, which would represent a PLC-dependent component of rise, was isolated by subtracting the response in the presence of U-73122 from that in its absence (see Fig. 11B for nominally Ca²⁺-free solution data).

The depressant action of U-73122 on intracellular Ca²⁺ release can also be inferred from the action of U-73122 on FCCP-induced rises in [Ca²⁺]_i in Krebs solution. U-73122 (2 µM) decreased the "valley" of [Ca²⁺]_i between the first and second phases (to 78.5 ± 3.7% of the control, n = 30; Fig. 11, C and D) and delayed the onset of the second phase. The latter action resulted in a complex shape of the U-73122-sensitive component of FCCP-induced rises in a Krebs solution (Fig. 11Da). When the action to prolong FCCP-induced rise was removed by deleting several data points (gray bar in Fig. 11Da) during the midphase of the response in the presence of U-73122, the U-73122-sensitive component was similar to that in a Ca²⁺-free solution in amplitude and more or less similar in shape, albeit not identical (Fig. 11, Db and Dc). The results indicate that mitochondrial membrane depolarization, protonation, and/or Ca²⁺ release activates a transient Ca²⁺ release presumably from the ER under the activation of PLC.

Activation of Ca²⁺ release from, or uptake at, mitochondria by Ca²⁺ release from the ER. The results above suggested activation of Ca²⁺ release from the ER by mitochondrial Ca²⁺ release. Conversely, it is possible that Ca²⁺ release from the ER activates mitochondrial Ca²⁺ release because CICR has been reported to occur through several Ca²⁺ pathways at the mitochondrial inner membrane (20, 29). This possibility can be tested by applying FCCP during the sustained rise in [Ca²⁺]_i by thapsigargin (Fig. 12). If mitochondria is depleted of Ca²⁺ under this condition, FCCP would fail to increase [Ca²⁺]_i.

FCCP (2 µM) applied during the sustained rise in [Ca²⁺]_i by thapsigargin produced two types of actions depending on the cell. In one type of cell, it caused a sharp rise in [Ca²⁺]_i (by 36.4 ± 7.9 nM, n = 24), which was followed by a decrease in [Ca²⁺]_i to a level (54.1 ± 4.6 nM, n = 40) lower than that before application of FCCP (Fig. 12B). The magnitude of rise was varied among cells; it was greater or smaller than that before application of thapsigargin. The greater rise implies that Ca²⁺ is taken up further into mitochondria from the high level of [Ca²⁺]_i during the plateau produced by thapsigargin, whereas the smaller rise indicates release of a fraction of Ca²⁺ stored in mitochondria. In another type of cells, FCCP did not elevate [Ca²⁺]_i further but quickly reduced the high level of [Ca²⁺]_i to a level lower than that before application of FCCP (by 34.6 ± 6.0 nM, n = 27; Fig. 12A). Thus depletion of Ca²⁺ in the ER by the action of thapsigargin causes strong, if not complete, depletion of Ca²⁺ in mitochondria in this type of cells. Even under this condition, FCCP depolarized mitochondrial membrane (Fig. 13), although the extent of depolarization was smaller (77.9 ± 5.8% of the control, n = 29), no matter whether it increased or decreased [Ca²⁺]_i during the sustained rise in [Ca²⁺]_i.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
The present study demonstrates the bidirectional interaction of mitochondria and the ER with respect to Ca²⁺ dynamics and their regulation of plasmalemmal Ca²⁺ entry in various modes. First, depletion of Ca²⁺ in the ER by thapsigargin activates SOC, as suggested in previous studies (see Introduction for references). Second, mitochondrial uncoupling activates Ca²⁺ release from mitochondria, causing a transient rise in [Ca²⁺]_i, the subsequent spiky Ca²⁺ entry, and Ca²⁺ release from the ER in most cells. In a fraction of cells, the Ca²⁺ release from the ER activated by mitochondrial Ca²⁺ release causes Ca²⁺ depletion in the ER that activates SOC for several tens of minutes. Thus mitochondrial membrane depolarization, protonation, and/or the resultant Ca²⁺ release activates two types of Ca²⁺ entry at the cell membrane: a spiky Ca²⁺ entry and long-lasting SOC. Finally, Ca²⁺ release from the ER causes mitochondrial Ca²⁺ depletion in a fraction of cells. How mitochondria and the ER communicate with each other via Ca²⁺ flux, how such communications regulate plasmalemmal Ca²⁺ entry, and how the variability in mitochondrial ER couplings occurs among cells are discussed below.

How does mitochondrial uncoupling deplete Ca²⁺ in the ER or Ca²⁺ liberation from the ER activates mitochondrial Ca²⁺ release? Existence of numerous mitochondria and occurrence of a large Ca²⁺ release from the ER through IP₃ receptors suggest the intimate coupling of mitochondria and the ER in terms of Ca²⁺ dynamics in brown adipocytes. The close functional coupling between mitochondria and the ER could also be indicated by the synthesis of lipid in the ER (39) and its consumption in mitochondria in brown adipocytes (see Introduction for references). Although the definite morphological evidence has not been well described (33), it is possible that mitochondria and the ER are juxtaposed to each other. The high [Ca²⁺]_i produced in the microdomain between these organelles by Ca²⁺ release from either of the organelles would enable Ca²⁺ uptake into or release from the other. Although mitochondrial Ca²⁺ uptake and its role in Ca²⁺ buffering have been well documented (see Introduction for references), evidence for the activation of Ca²⁺ release from these organelles by the Ca²⁺ released from another is now found in the present study.

The third slow phase of the rise in [Ca²⁺]_i by FCCP appears to be produced by Ca²⁺ entry. This Ca²⁺ entry seems to be caused by activation of SOC because thapsigargin had little effect on [Ca²⁺]_i and CPA was not able to release Ca²⁺ from the ER after the action of FCCP. Thus the ER is depleted of Ca²⁺ under this condition. This implies that mitochondrial Ca²⁺ release, protonation, and/or depolarization itself activates Ca²⁺ release from the ER. In support of this, the second component of the FCCP-induced rise in [Ca²⁺]_i (as well as the first component) remaining in a Ca²⁺-free EGTA solution was abolished after Ca²⁺ depletion in the ER by CPA (see below for the significance of this abolition). The activation of Ca²⁺ release from the ER by mitochondrial depolarization itself is unlikely because there has been so far no evidence for the direct physical coupling of a mitochondrial membrane molecule to a Ca²⁺ release channel at the ER membrane. The Ca²⁺ release from the ER by mitochondrial Ca²⁺ release could then be caused by a CICR mechanism through IP₃ receptors or ryanodine receptors. Although the ryanodine receptor is a possible candidate in rat brown adipocytes (unpublished observations), the IP₃ receptor could also be a likely candidate because it causes a powerful release of Ca²⁺ by the -action of epinephrine (34). IP₃ may be produced by the activation of PLC resident on the membranes of mitochondria or the ER (53) by a local rise in [Ca²⁺]_i as the result of mitochondrial Ca²⁺ release. In fact, the second component of FCCP-induced rise in [Ca²⁺]_i remaining in a Ca²⁺-free EGTA solution was abolished by a blocker of PLC.

The present study demonstrates that the depletion of Ca²⁺ in the ER causes Ca²⁺ release from mitochondria. FCCP was not able to increase [Ca²⁺]_i; instead, it reduced the high level of [Ca²⁺]_i during the activation of SOC by thapsigargin in a fraction of the cells. Similar absence of rises in [Ca²⁺]_i by FCCP was seen during the activation of SOC by the preceding application of FCCP. Aside from the depressant action on the high level of [Ca²⁺]_i, the lack of effects of FCCP to raise [Ca²⁺]_i implies depletion of Ca²⁺ in mitochondria as a result of Ca²⁺ depletion in the ER. One possible mechanism would be that the high [Ca²⁺]_i in the microdomain caused by Ca²⁺ leak from the ER could activate Ca²⁺ release from mitochondria in brown adipocytes. In line with this, it was suggested that CICR occurs in mitochondria through PTP (20) or Ca²⁺ uniporters (29). The possible involvement of PTP or Ca²⁺ uniporters in the mechanism of Ca²⁺ release from mitochondria, however, is unlikely because a blocker of either of these did not affect the first and second component of FCCP-induced rises in [Ca²⁺]_i. Alternatively, the increased Ca²⁺ in the mitochondrial matrix via enhanced Ca²⁺ uptake from the high-Ca²⁺ microdomain may activate phospholipase A₂ in the matrix (10). This would result in the activation of PTP and Ca²⁺ release from mitochondria. This mechanism is also unlikely because of absence of effects of a PTP blocker.

Another possible mechanism for depletion of mitochondrial Ca²⁺ after Ca²⁺ depletion in the ER may be that Ca²⁺ stored in mitochondria is supplied in part by the Ca²⁺ stored in the ER through a direct pathway connecting the insides of both organelles or indirectly via the local [Ca²⁺]_i domain. This mechanism may be supported by the observation that FCCP-induced mitochondrial Ca²⁺ release disappeared in a Ca²⁺-free EGTA solution, when the ER is depleted of Ca²⁺ by CPA. This mechanism requires further investigation. Ca²⁺ release from mitochondria by the action of 10 µM thapsigargin was also reported in Trypomastigotes, although the mechanism was ascribed to mitochondrial membrane depolarization (52). In brown adipocytes, however, thapsigargin did not depolarize the mitochondrial membrane.

The couplings between mitochondria and the ER are thus bidirectional. Accordingly, Ca²⁺ movement occurs from mitochondria to the ER or vice versa. Which direction of coupling and how tight the coupling occurs, however, differ depending on the cell. Consequently, four types of cells were seen: tight or loose coupling in both directions and tight coupling in one direction and loose in the other (see Figs. 10 and 12). This variability may result from differences in the tightness of juxtaposition of these organelles or the effective localization or amount of molecules involved in Ca²⁺ flux in the membrane of these organelles.

Regulation of plasmalemmal Ca²⁺ entry by mitochondrial uncoupling or Ca²⁺ release from the ER. In this study, we demonstrated various modes of activation of Ca²⁺ entry at the cell membrane by mitochondrial uncoupling and Ca²⁺ release from the ER. First, activation of SOC by depletion of Ca²⁺ in the ER by the action of thapsigargin was demonstrated as reported in a previous study (34). One notable characteristic of SOC activation in brown adipocytes is that it is depressed to some extent by removal of most of the external Na⁺. This suggests that SOC contains a component whose activation depends on external Na⁺. One possible mode of dependence of Ca²⁺ entry on Na⁺ is caused by the operation of Na⁺/Ca²⁺ exchange in the reversed mode. Rosker et al. (46) reported this reversed mode of operation of Na⁺/Ca²⁺ exchange as a result of Na⁺ entry through a type of transient receptor potential (TRP) protein, TRPC3, transfected in HEK-293 cells. They showed that this operation was enabled by the physical coupling of the TRPC3 with the Na⁺/Ca²⁺ exchanger, in which the exchanger senses the high-Na⁺ concentration in the microdomain close to the orifice of the TRPC3 channel. Another possible role of Na⁺ in Ca²⁺ entry would be that the Ca²⁺ channel involved in SOC in brown adipocytes is permeable to both Ca²⁺ and Na⁺ and voltage dependent. The magnitude of depolarization caused by the activation of this channel must be reduced in an Na⁺-deficient solution and may lead to a decrease in the number of opened channels and thus Ca²⁺ entry. The second mode of SOC activation found in this study is the activation by mitochondrial Ca²⁺ release from Ca²⁺ depletion in the ER through their couplings as already discussed.

We found an entirely new type of activation of Ca²⁺ entry by mitochondrial uncoupling in the present study. FCCP caused a spiky rise in [Ca²⁺]_i, which was partially abolished by the removal of external Ca²⁺ but remained after the application of thapsigargin. There are three important characteristics of this Ca²⁺ entry. First, its generation is tightly coupled with the repolarizing phase of the mitochondrial depolarization. Second, it is abolished by a blocker of ATP synthesis. Third, there was a rise in [Mg²⁺]_i during the course of spiky Ca²⁺ entry, indicating ATP hydrolysis. Thus a process involved in the restoration of ATP synthesis following the membrane repolarization after disappearance of FCCP may activate a mechanism of transient Ca²⁺ entry at the plasma membrane. It is totally unknown, however, what molecule is involved in the mechanism. Another characteristic of the spiky Ca²⁺ entry is that it is abolished during the sustained rise in [Ca²⁺]_i produced by the preceding application of FCCP. A messenger that activates a spiky Ca²⁺ entry may disappear, once the mechanism for Ca²⁺ depletion in the ER is activated by Ca²⁺ release from mitochondria. The Ca²⁺ pathway responsible for this mode of Ca²⁺ entry would differ from that involved in activation of SOC, since the former was not affected by lowering the external Na⁺ concentration, whereas the latter was depressed. It is suggested that some types of TRP channels, TRPC6 and TRPC3, are activated by diacylglycerol (18). The involvement of diacylglycerol possibly produced by the activation of PLC in the activation of the spiky Ca²⁺ entry, however, is unlikely, because it was not blocked by a PLC blocker (Fig. 11C).

In this study, we have also shown negative modes of regulation of Ca²⁺ entry by the processes that activate Ca²⁺ entry, once plasmalemmal Ca²⁺ entry is activated by any means. FCCP or thapsigargin, when applied during the sustained rise in [Ca²⁺]_i caused by either of these drugs, produced a quick reduction of [Ca²⁺]_i, which was slowly restored to a level higher than the initial basal level under the control condition. In rat basophilic leukemia cells, a similar depression of SOC by mitochondrial depolarization was reported (12). The mechanism of these inhibitory effects of activators of Ca²⁺ release or Ca²⁺ entry on the high level of [Ca²⁺]_i was not investigated further and left for future studies. This mechanism of negative regulation must be a physiological event, since similar restoring actions were elicited by norepinephrine once plasmalemmal Ca²⁺ entry was activated by Ca²⁺ depletion in the ER or mitochondria (unpublished observations).

Physiological significance of mitochondrial ER coupling and plasmalemmal Ca²⁺ entries. Under the physiological conditions, the actions of FCCP observed in the present study could be reproduced by the activation of uncoupling proteins by the -action of norepinephrine released from sympathetic nerves in response to cold exposure (see Introduction for references). On the other hand, the action of thapsigargin could be reproduced by the -action of norepinephrine (see Introduction for references), although the former is irreversible, whereas the latter is reversible. Thus the two modes of the actions of norepinephrine can produce those induced by FCCP and thapsigargin. Indeed, we have observed effects of -agonist similar to those of FCCP actions on mitochondria. Isoprenaline caused Ca²⁺ release from mitochondria and subsequent Ca²⁺ entry, resulting in complex modes of rises in [Ca²⁺]_i (16). In contrast to the pharmacological system utilizing an uncoupler, another mode of -actions enhances hydrolysis of neutral lipids, which elevates mitochondrial membrane potential via the enhancements of TCA cycle and the subsequent electron transport chain. This effect on energetics must be accelerated by the increase in [Ca²⁺]_i produced by both the - and -actions via Ca²⁺ release from the ER and two modes of Ca²⁺ entry. However, it is not yet known how each mode of coupling between mitochondria and the ER and the two modes of Ca²⁺ entry contribute to the regulation of thermogenesis.

	GRANTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
This work was supported by Grant-in-Aid 18590211 from the Japanese Ministry of Education, Science, and Culture (to K. Kuba).

	ACKNOWLEDGMENTS

 
We thank Dr. M. Omatsu-Kanbe for useful advice on culturing methods of rat brown adipocytes.

	FOOTNOTES

 

Address for reprint requests and other correspondence: K. Kuba, Laboratory of Anatomy and Physiology, School of Nutritional Sciences, Nagoya Univ. of Arts and Sciences, 57 Takenoyama, Iwasaki-cho, Nissin, Aichi 470-0196, Japan (e-mail: kubak{at}nuas.ac.jp)

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.

	REFERENCES

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
1. Bernardi P. Mitochondrial transport of cations: channels, exchangers and permeability transition. Physiol Rev 79: 1127–1155, 1999.[Abstract/Free Full Text]

2. Berridge MJ. Inositol trisphophate and calcium signalling. Nature 361: 315–325, 1993.[CrossRef][Medline]

3. Carafoli E. Intracellular calcium homeostasis. Annu Rev Biochem 56: 395–433, 1987.[CrossRef][ISI][Medline]

4. Carafoli E. Biogenesis, plasma membrane calcium ATPase: 15 years of work on the purified enzyme. FASEB J 99: 993–1002, 1994.

5. Cavallini L, Coassin M, Borean A, Alexandre A. Arachidonic acid activates a proton conductance pathway and the Na⁺/H⁺ exchanger in platelets. Biochem J 319: 567–574, 1996.

6. Colegrove SL, Albrecht MA, Friel DD. Dissection of mitochondrial Ca²⁺ uptake and release fluxes in situ after depolarization-evoked [Ca²⁺]_i elevations in sympathetic neurons. J Gen Physiol 115: 351–369, 2000.[Abstract/Free Full Text]

7. Crompton M. The regulation of mitochondrial calcium transport in heart. Curr Top Membr Transp 25: 231–276, 1985.

8. David G. Mitochondrial clearance of cytosolic Ca²⁺ in stimulated lizard motor nerve terminals proceeds without progressive elevation of mitochondrial-matrix [Ca²⁺]. J Neurosci 19: 7495–7506, 1999.[Abstract/Free Full Text]

9. Duchen MR. Ca²⁺-dependent changes in the mitochondrial energetics in single dissociated mouse sensory neurons. Biochem J 283: 41–50, 1992.

10. Gadd ME, Broekemeier KM, Crouser ED, Kumar J, Graff G, Pfeiffer DR. Mitochondrial iPLA₂ activity modulates the release of cytochrome c from mitochondria and influences the permeability transition. J Biol Chem 281: 6931–6939, 2006.[Abstract/Free Full Text]

11. Gilabert JA, Parekh AB. Respiring mitochondria determine the pattern of activation and inactivation of the store-operated Ca²⁺ current I_CRAC. EMBO J 19: 6401–6407, 2000.[CrossRef][ISI][Medline]

12. Glitsch MD, Bakowski D, Parekh AB. Store-operated Ca²⁺ entry on mitochondrial Ca²⁺ uptake. EMBO J 21: 6744–6754, 2002.[CrossRef][ISI][Medline]

13. Grynkiewicz G, Poenie M, Tsien RY. A new generation of Ca²⁺ indicators with greatly improved fluorescence properties. J Biol Chem 260: 3440–3450, 1985.[Abstract/Free Full Text]

14. Gunter TE, Pfeiffer DR. Mechanisms by which mitochondria transport calcium. Am J Physiol Cell Physiol 258: C755–C786, 1990.