Mitochondrial Free [Ca²⁺] Increases during ATP/ADP Antiport and ADP Phosphorylation: Exploration of Mechanisms

Abstract

ADP influx and ADP phosphorylation may alter mitochondrial free [Ca²⁺] ([Ca²⁺]_m) and consequently mitochondrial bioenergetics by several postulated mechanisms. We tested how [Ca²⁺]_m is affected by H₂PO₄⁻ (P_i), Mg²⁺, calcium uniporter activity, matrix volume changes, and the bioenergetic state. We measured [Ca²⁺]_m, membrane potential, redox state, matrix volume, pH_m, and O₂ consumption in guinea pig heart mitochondria with or without ruthenium red, carboxyatractyloside, or oligomycin, and at several levels of Mg²⁺ and P_i. Energized mitochondria showed a dose-dependent increase in [Ca²⁺]_m after adding CaCl₂ equivalent to 20, 114, and 485 nM extramatrix free [Ca²⁺] ([Ca²⁺]_e); this uptake was attenuated at higher buffer Mg²⁺. Adding ADP transiently increased [Ca²⁺]_m up to twofold. The ADP effect on increasing [Ca²⁺]_m could be partially attributed to matrix contraction, but was little affected by ruthenium red or changes in Mg²⁺ or P_i. Oligomycin largely reduced the increase in [Ca²⁺]_m by ADP compared to control, and [Ca²⁺]_m did not return to baseline. Carboxyatractyloside prevented the ADP-induced [Ca²⁺]_m increase. Adding CaCl₂ had no effect on bioenergetics, except for a small increase in state 2 and state 4 respiration at 485 nM [Ca²⁺]_e. These data suggest that matrix ADP influx and subsequent phosphorylation increase [Ca²⁺]_m largely due to the interaction of matrix Ca²⁺ with ATP, ADP, P_i, and cation buffering proteins in the matrix.

Introduction

Matrix free Ca²⁺ ([Ca²⁺]_m) may play a major role in regulating mitochondrial function. Studies have shown a correlation between increased bioenergetics and increased [Ca²⁺]_m (1–6). However, excess [Ca²⁺]_m predisposes the mitochondria to form and open the permeability transition pore (mPTP) (7–11), a key factor in cell apoptosis; inhibition of mPTP formation reduces ischemia-reperfusion injury (12–17). The importance of [Ca²⁺]_m in both physiological and pathological conditions implies a necessity to tightly regulate [Ca²⁺]_m.

[Ca²⁺]_m is regulated in part by voltage-dependent cation fluxes via a series of poorly identified cation channels and exchangers on the inner mitochondrial membrane (IMM) (7,8,10,18). The primary route for matrix Ca²⁺ uptake is via the ruthenium-red (RR)-sensitive Ca²⁺ uniporter (CU), whereas the principal Ca²⁺ efflux pathway is the Na⁺/Ca²⁺ exchanger (NCE). There may also be Ca²⁺ efflux through a Na⁺-independent Ca²⁺ exchanger (NICE), putatively a Ca²⁺/H⁺ exchanger (CHE) (19–22). Transport through the CU and NCE are dependent on the mitochondrial membrane potential (ΔΨ_m), whereas the CHE is thought to be concentration-, and not ΔΨ_m-dependent (21,23,24). Several independent studies have shown a correlation between a larger ΔΨ_m and a higher [Ca²⁺]_m (10,22).

In measuring [Ca²⁺]_m, we observed an increase in [Ca²⁺]_m with no added buffer CaCl₂ during state 3 respiration, a state in which ΔΨ_m is decreased. This paradoxical increase in [Ca²⁺]_m implies a mechanism other than ΔΨ_m in modulating [Ca²⁺]_m. We tested five mechanisms that might play a role in this phenomenon:

The first mechanism that might explain an increase in [Ca²⁺]_m is allosteric activation of the CU by ADP to increase [Ca²⁺]_m (25). As ADP becomes phosphorylated, allosteric activation would decrease as [ADP] falls and [Ca²⁺]_m would return to the preexisting level. A second mechanism is that ADP phosphorylation might also induce a large matrix volume contraction, thereby raising [Ca²⁺]_m.

A third possible mechanism is modulation of [Ca²⁺]_m by precipitation of free Ca²⁺ with P_i, such as Ca₃(PO₄)₂ (26). The complexing of Ca²⁺ with P_i is dependent on the product of concentrations of different species of P_i and Ca²⁺. Thus, a decrease in matrix [P_i] would facilitate an increase in [Ca²⁺]_m. In this situation, matrix [P_i] would be lower during state 3 respiration, as P_i becomes phosphorylated to ADP. After phosphorylation of ADP, matrix [P_i] would increase to near the level before phosphorylation, as Ca²⁺ would again form a complex with P_i.

A fourth possible mechanism is that the increased [Ca²⁺]_m results from basic physicochemical differences, i.e., different binding affinities (K_d) of ADP and ATP for Ca²⁺ (27). These differences in K_d predict an increase in [Ca²⁺]_m whenever the mitochondrial ADP/ATP ratio increases, and vice versa. Since ATP, ADP, and P_i bind variably to other cations (e.g., Mg²⁺ and H⁺) as well, a change in concentration of these ions may also result in a significant change of free ATP, ADP, and P_i (ATP⁴⁻, ADP³⁻, and H₂PO₄⁻) and thus alter the buffering capacity for Ca²⁺.

A fifth possible mechanism is that the altered bioenergetic state during state 3 respiration leads to release of matrix Ca²⁺ stores by an unknown mechanism.

To explore which of these possible mechanisms might underlie the large change in [Ca²⁺]_m during transition to and from state 3 respiration, we measured matrix [Ca²⁺]_m as a function of extramatrix free [Ca²⁺] ([Ca²⁺]_e) by increasing buffer [CaCl₂]. We also examined changes in [Ca²⁺]_m at a higher and lower buffer [P_i] and [Mg²⁺], and assessed the effect of Δ[ADP] on matrix volume. Experiments were done with or without the CU blocker RR, the F₁F₀-ATPase blocker oligomycin (OMN), and the ADP/ATP carrier (AAC) blocker carboxy-atractyloside (CATR). Na⁺/Ca²⁺ exchange (NCE) was inactive and was eliminated as a factor because there was no Na⁺ present in the experimental buffer.

Materials and Methods

Fluorescence measurements

Fluorescence spectrophotometry was used to measure matrix free Ca²⁺, NADH, pH, and ΔΨ_m (Qm-8, Photon Technology, Birmingham, NJ) (28–30). Isolated mitochondria (5 mg/ml) were incubated for 20 min at room temperature (25°C) with 5 μM indo-1 AM to measure [Ca²⁺]_m or with 5 μM BCECF AM to measure pH_m (Invitrogen, Carlsbad, CA) followed by suspension in 25 ml isolation buffer and repeated centrifugation at 8000 × g. The AM form of the dyes is taken up into the matrix where it is deesterified and retained. The dye-loaded pellet was resuspended in 0.5 ml isolation buffer, and protein concentration was measured again and diluted to 12.5 mg mitochondrial protein/ml. In other mitochondria, background autofluorescence (AF), which at 456 nm represents NADH (redox state), was measured and ΔΨ_m was determined using rhodamine (Rh) 123. Mitochondria were kept on ice for the duration of the studies. All studies were conducted at room temperature. Please refer to the Supporting Material for detailed information on methods to assess matrix and extramatrix [Ca²⁺]_m and Mg²⁺, redox state, matrix pH, ΔΨ_m, matrix volume, and respiration.

Experimental groups and protocol

Guinea pig heart mitochondria were isolated (see Supporting Material) as described previously (31,32) and diluted (33). Isolated mitochondria were divided into seven treatment groups: control (CON, 5 mM P_i), high P_i (HP, 10 mM P_i), low P_i (LP, 1 mM P_i), RR at start (RRS), RR later, after CaCl₂ addition (RRL), and OMN or CATR, each given at t = −120 s (Fig. 1). The CON mitochondria were suspended (0.5 mg/ml) in experimental buffer containing (in mM) 130 KCl (EMD Biosciences, San Diego, CA), 5 K₂HPO₄, 20 MOPS, 0.016 bovine serum albumin, and 0.04 EGTA, pH 7.15 (adjusted with KOH). To adjust for osmolarity, the HP and LP experimental buffers were adjusted to contain 123 mM KCl and 10 mM K₂HPO₄ (HP) or 136 mM KCl and 1 mM K₂HPO₄ (LP). MOPS, BSA, and EGTA concentrations and pH were the same for all groups. In the RRS group, 25 μM RR was added to the CON experimental buffer before energizing mitochondria. In the RRL group, 25 μM RR was added after the addition of CaCl₂. In the OMN and CATR groups, 100 μM OMN or 1.3 μM CATR was added before energizing mitochondria. In selected studies, 1 mM MgCl₂ was present while the CaCl₂ and ADP were added (see Supporting Material).

Figure 1.

Time line for adding substances to experimental buffer. PA, pyruvic acid; RRS and RRL, ruthenium red at start or later; ADP, adenosine diphosphate; CATR, carboxyatractyloside; OMN, oligomycin; CCCP, carbonylcyanide m-chlorophenylhydrazone.

Experiments were initiated at t = −120 s, when OMN, CATR, and RR were added to comprise the OMN, CATR, and RRS groups; at t = −90 s, mitochondria were added (Fig. 1). At t = 0 s, pyruvic acid (PA, 0.5 mM) was added, followed by either of two concentrations of CaCl₂ (10 or 25 μM in deionized H₂O) or vehicle (0 CaCl₂) at t = 120 s. At t = 180 s, RR was added to the RRL group. At t = 240 s, ADP (250 μM) was added, followed by 4 μM of the protonophore carbonylcyanide m-chlorophenylhydrazone (CCCP) at t = 400 s to maximally depolarize the IMM. All buffers and reagents, including substrates, were Na⁺-free to prevent NCE activation. Inactivity of the NCE was verified by comparing data from experiments with and without added CGP-37157, a specific mitochondrial NCE inhibitor (data not shown). When no drug or CaCl₂ was added, the appropriate vehicle was added to the mitochondrial suspension. All chemicals were obtained from Sigma-Aldrich (St. Louis, MO) unless noted otherwise.

Statistical analyses

All data are presented as the mean ± SE. Repeated-measures ANOVA followed by a post hoc analysis using Student-Newman-Keuls' test was performed to determine statistically significant differences between and within groups. Data for analysis were collected at the times noted above. A P value of <0.05 (two-tailed) was considered significant.

Results

Extramatrix and matrix free [Ca²⁺] and inhibition of CU

Adding 0, 10, or 25 μM CaCl₂ to the mitochondrial suspension caused a rapid, concentration-dependent increase in [Ca²⁺]_e to initial values of 20 ± 3, 114 ± 13, and 485 ± 40 nM (Fig. 2) due to the presence of 40 μM of EGTA. This was followed by a slower, steady decline in [Ca²⁺]_e as Ca²⁺ was transported into mitochondria via the CU; this was most prominent after adding 25 μM CaCl₂. Transport of Ca²⁺ through the CU was confirmed in the RR groups because RR blocked the subsequent decrease in [Ca²⁺]_e. Adding ADP did not significantly affect [Ca²⁺]_e in the absence of RR. In the presence of RR there was a significant decrease in [Ca²⁺]_e when ADP was added.

Figure 2.

Extramatrix free [Ca²⁺] ([Ca²⁺]_e) over time measured using indo-1 (in the presence of 40 μM EGTA). Adding CaCl₂ increased [Ca²⁺]_e in a concentration-dependent manner. Adding 25 μM CaCl₂ increased [Ca²⁺]_e up to 485 nM, which then slowly declined as Ca²⁺ was taken up into the matrix. Adding 10 μM CaCl₂ increased [Ca²⁺]_e to 114 nM with a lesser and slower subsequent Ca²⁺ uptake. RR blocked matrix uptake of Ca²⁺ through the CU so that [Ca²⁺]_e remained constant. ADP caused a small decrease in [Ca²⁺]_e only in the presence of RR.

Adding CaCl₂ caused a concentration-dependent increase in matrix [Ca²⁺]_m (Fig. 3), which took place via the CU, as verified again by the nearly complete block in response to RR (Fig. 3 C). After adding 10 μM CaCl₂ (initially equivalent to 114 nM [Ca²⁺]_e), [Ca²⁺]_m increased from 80 ± 5 nM to 183 ± 7 nM. After adding 25 μM CaCl₂ (initially equivalent to 485 nM [Ca²⁺]_e), [Ca²⁺]_m increased from 80 ± 5 to 518 ± 44 nM. Note that [Ca²⁺]_m (Fig. 3) remained at a steady-state level, whereas [Ca²⁺]_e (Fig. 2) continued to decrease in the CON group from its initial value of 485 nM [Ca²⁺] (after adding 25 μM CaCl₂).

Figure 3.

Effects of ADP and RR on [Ca²⁺]_m over time, measured using indo-1 AM. Dynamic, time-dependent changes (left column) and mean changes (right column) taken at the vertical dotted lines in the left column are displayed. (A) Adding 10 and 25 μM CaCl₂ in energized mitochondria (after PA) caused abrupt, graded increases in [Ca²⁺]_m to steady-state values. Adding ADP caused an abrupt but transient rise in [Ca²⁺]_m above the values established with added CaCl₂. (B) Adding RR after 25 μM CaCl₂ (RRL) did not blunt the rise in [Ca²⁺]_m induced by adding ADP. (C) Adding RR before CaCl₂ (RRS) blocked Ca²⁺ uptake, but the ADP-induced increase in [Ca²⁺]_m remained evident, although it was much smaller (note scale) after adding 25 μM CaCl₂. In the right column, for P < 0.05, # indicates state 3 versus states 2 and 4 respiration; ^∗ indicates RRL and RRS versus the CON group at the same respiration state and CaCl₂.

Adding ADP caused a proportional increase in [Ca²⁺]_m in the CON, RRS, and RRL groups (Fig. 3), i.e., the magnitude of increase was dependent on the [Ca²⁺]_m before ADP was added. The effect of ADP was absent in the RR group with 10 μM added CaCl₂, as [Ca²⁺]_m was low (80 ± 5 nM). ADP addition increased [Ca²⁺]_m only a little (Fig. 3 C), probably because adding CaCl₂ after RR did not increase [Ca²⁺]_m. The effect of ADP on increasing [Ca²⁺]_m was similar in the RRL and CON groups and smaller in the RRS group after adding CaCl₂. There was an attenuated increase of [Ca²⁺]_m (25 μM Ca²⁺; 485 nM [Ca²⁺]_e), though not a significant one, in the RRL versus the CON group. Matrix uptake of Ca²⁺ continued in the CON group. Adding RR after 180 s stopped this slow uptake, resulting in a smaller increase in [Ca²⁺]_m after adding 10 μM (114 nM [Ca²⁺]_e). However, extending the time of Ca²⁺ uptake before adding RR in the 10-μM group resulted in an increase in [Ca²⁺]_m beyond 180 s similar to that of the CON group. Adding ADP after RR then increased [Ca²⁺]_m as much as in the CON group (data not shown).

[Ca²⁺]_m increased as a function of [Ca²⁺]_e, and this was attenuated during state 3 respiration (Fig. 4). These data show that [Ca²⁺]_m reached a value twofold higher than that of [Ca²⁺]_e at t = 180 s (state 2 respiration). [Ca²⁺]_m was 80 ± 5, 212 ± 10, and 518 ± 44 nM, whereas [Ca²⁺]_e was 26 ± 3, 112 ± 12, and 266 ± 12 nM after adding 0, 10, and 25 μM CaCl₂. This plot also shows that [Ca²⁺]_m increased 1.75- and twofold during state 3 versus state 2 respiration (from 518 ± 44 nM to 1039 ± 83 nM), whereas [Ca²⁺]_e decreased slightly. These data support the proposal that as more Ca²⁺ becomes buffered in the matrix after CaCl₂ addition, some is released during ADP/ATP antiport and ADP phosphorylation, possibly due to a decrease in matrix Ca²⁺ buffering sites (e.g., ADP, proteins).

Figure 4.

[Ca²⁺]_m as a function of [Ca²⁺]_e. [Ca²⁺]_e and [Ca²⁺]_m were measured at t = 100 s (before added CaCl₂), at t =180 s (after added CaCl₂), and at t = 250 s (in the presence of ADP). ADP caused a more than additive increase in [Ca²⁺]_m as a function of [Ca²⁺]_e, and the increase was dependent on the existing [Ca²⁺]_m level. Data were taken from Figs. 2 and 3A. For P < 0.05, # indicates state 3 versus states 2 and 4 respiration.

Effect of buffer P_i and Mg²⁺ on matrix free [Ca²⁺]

The effect on [Ca²⁺]_m of altering buffer [P_i] was evaluated at different [CaCl₂]_e and during ADP addition (Fig. 5). Buffer [KCl] was adjusted against [K₂HPO₄] to maintain the same osmolarity. With no added CaCl₂, there were no detectable changes in [Ca²⁺]_m among the three [P_i] groups. A change in buffer [P_i] from 5 to 1, and from 5 to 10 mM, did not significantly affect [Ca²⁺]_m after adding CaCl₂; however, at 25 μM CaCl₂ and 1 mM [P_i] (LP), the initial (fast) uptake of Ca²⁺ was followed by a slow decrease in [Ca²⁺]_m. The increase in [Ca²⁺]_m after adding ADP did not differ among the P_i groups.

Figure 5.

Effects of ADP and altered buffer P_i on [Ca²⁺]_m. (A) Same as Fig. 3A, where [P_i] is 5 mM. (B) In both the 10-mM and 1-mM [P_i] groups, there were no significant differences in response to added CaCl₂ or ADP versus the corresponding 5-mM [P_i] group. For P < 0.05, # indicates state 3 versus states 2 and 4 respiration; ^∗ indicates the HP and LP versus CON groups at the same respiration state and CaCl₂. See Fig. 3 legend for additional details.

In a few studies, 1 mM MgCl₂ was added to the buffer (Fig. S1 A in the Supporting Material). Calculated matrix [Mg²⁺] (see Supporting Material, Methods) was 0.51 ± 0.03 mM without added MgCl₂ and 0.85 ± 0.02 mM after adding 1 mM MgCl₂ to the buffer. In the presence of MgCl₂ the same amount of added CaCl₂ resulted in a lesser rise in [Ca²⁺]_m; however, adding ADP caused a proportional increase in [Ca²⁺]_m when more CaCl₂ was added to the buffer to counter the Mg²⁺-inhibited transport of Ca²⁺ through the CU (Fig. S1 A).

Effect of ADP versus ATP on mitochondrial matrix volume versus [Ca²⁺]

Added ADP resulted in a transient volume decrease of 7% (photon count) with a corresponding 51% increase in [Ca²⁺]_m in the 25 μM CaCl₂ group, whereas the maximal volume increase after valinomycin (VAL) was 52% with a corresponding decrease in [Ca²⁺]_m after VAL of only 18% (Fig. S1 B). The ratio of volume to [Ca²⁺]_m during responses to VAL over ADP—(52/18) divided by (7/51) = 0.05—indicated that this small decrease in volume caused by addition of ADP could account for only ∼5% of the increase in [Ca²⁺]_m.

Effect of blocking ADP and ATP transport on matrix free [Ca²⁺]

Mitochondria were treated with OMN or CATR, which inhibit the F₁F₀-ATPase and the AAC respectively, to evaluate the effect of the organic phosphates ADP and ATP on [Ca²⁺]_m. In the presence of OMN, [ADP]_m and [ATP]_m change when ADP is added, because ATP is exchanged for ADP at the AAC. With blocked phosphorylation of ADP by OMN, matrix [P_i] is expected to remain unchanged. In the presence of CATR, however, there can be no changes in [ADP]_m or [ATP]_m during added buffer ADP, because the AAC is blocked.

OMN and CATR had no effect on [Ca²⁺]_m when 0, 10, or 25 μM CaCl₂ was added; thus, there was no effect of these drugs on matrix Ca²⁺ uptake (Fig. 6), but the response to ADP after OMN was a small but significant increase in [Ca²⁺]_m. Moreover, this increase in [Ca²⁺]_m was sustained beyond the period when ADP would have been completely phosphorylated. This increase in [Ca²⁺]_m may indicate limited ADP entry into the matrix in exchange for ATP efflux. The [Ca²⁺]_m response to ADP was abolished after CATR as both ATP efflux and ADP influx were blocked.

Figure 6.

Effect of ADP on [Ca²⁺]_m during inhibited phosphorylation and blocked ADP/ATP transport. (A) Same as Fig. 3A. (B) Adding ADP after CaCl₂ and blocking F₁F₀-ATPase with OMN caused a significant increase in [Ca²⁺]_m after addition of both 10 and 25 μM CaCl₂. (C) Adding ADP after CaCl₂ when the ADP/ATP carrier was blocked with CATR inhibited the ADP-induced increase in [Ca²⁺]_m. For P < 0.05, # indicates state 3 versus states 2 and 4 respiration; ^∗ indicates the OMN and CATR versus CON groups at the same respiration state and CaCl₂. See Fig. 3 for additional details.

Effect of increasing [Ca²⁺]_m on the NADH redox state

PA increased NADH in all groups. Adding either 10 or 25 μM CaCl₂ did not change NADH compared to vehicle (CON, Fig. 7 A). Adding ADP caused a transient and reversible oxidation of NADH (i.e., a decrease in signal) in all groups, except in the OMN and CATR groups (data not shown), because ADP transport and phosphorylation were blocked in these groups. NADH returned to pre-ADP levels as ADP was phosphorylated to ATP. The presence of RR or a different [P_i] did not affect redox state (data not shown).

Figure 7.

Effects of adding CaCl₂ and ADP on mitochondrial bioenergetics. All data correspond to [Ca²⁺]_m data displayed in Fig. 3A (CON). (A) PA increased NADH as the TCA cycle was activated. Adding CaCl₂ did not significantly alter NADH values. (B) ADP transiently oxidized mitochondria (i.e., reduced NADH (A)) as energy contained in ΔΨ_m was consumed, as shown by the transiently lowered ΔΨ_m. (C) PA induced a mild alkalinization; adding CaCl₂ did not further alter the matrix pH. ADP transiently reduced matrix pH as matrix proton influx temporarily exceeded proton pumping. (D) O₂ consumption (respiration) increased on addition of PA, and more so on addition of 25 μM CaCl₂. Adding ADP markedly enhanced O₂ consumption, but adding CaCl₂ did not produce any additional effect on respiration. See Table 1 for summary data and statistics on respiration.

Effect of increasing [Ca²⁺]_m on ΔΨ_m

Energizing mitochondria with PA increased ΔΨ_m in all groups. Adding CaCl₂ had no significant effect on ΔΨ_m (CON, Fig. 7 B). Adding ADP caused a transient and reversible partial depolarization of ΔΨ_m in all groups, as ADP was transported into the matrix and phosphorylated to ATP, which was also not affected by added CaCl₂. ΔΨ_m was not affected by RR or at different buffer [P_i] or [Mg²⁺] (data not shown). The ADP-induced depolarization did not occur when ADP and ATP transport was prevented by CATR or when phosphorylation of ADP to ATP was blocked by OMN (data not shown). CCCP maximally depolarized the IMM.

Effect of increasing [Ca²⁺]_m on matrix pH

Matrix pH increased on energizing mitochondria with PA. Adding CaCl₂ had no significant effect on matrix pH compared to vehicle (CON, Fig. 7 C). Adding ADP acidified the matrix as protons enter the matrix through the F₁F₀ATP-ase to generate ATP. This alteration in pH by ADP was not affected by adding RR or by altering buffer [P_i], but was inhibited by CATR or OMN (data not shown). Adding CCCP caused matrix acidification, as this protonophore facilitates transport of protons across the IMM (data not shown).

Effect of increasing [Ca²⁺]_m on respiration

Respiration rates in states 2 and 4 were significantly increased only at 518 nM [Ca²⁺]_m (Fig. 7 D and Table 1); state 3 respiration was higher but unaltered at any [Ca²⁺]_m. At 183 nM [Ca²⁺]_m, respiration rates in states 2 and 4 were not significantly altered. Therefore, the respiratory control index (RCI) remained unchanged at 183 nM [Ca²⁺]_m, and slightly decreased at 518 nM [Ca²⁺]_m, due to the change in state 4 respiration. Adding MgCl₂ had no added effect on state 3 or state 4 respiration (Fig. S1 C). Respiratory effects of [Ca²⁺]_m were also not altered at any [P_i] (data not shown).

Table 1.

O₂ consumption in different respiratory states

	State 2	State 3	State 4	RCI
0 CaCl2	0.91 ± 0.04	14.07 ± 0.58	1.01 ± 0.05	13.6 ± 0.41
10 CaCl2	0.87 ± 0.04	14.15 ± 0.58	1.08 ± 0.02	13.0 ± 0.36
25 CaCl2	1.04 ± 0.03∗	14.44 ± 0.48	1.19 ± 0.02∗	12.1 ± 0.28∗

Oxygen consumption was measured in μmol O₂/h/mg protein. Adding CaCl₂ (485 nM [Ca²⁺]_e) significantly increased O₂ uptake in states 2 and 4 and lowered RCI (state 3/state 4 ratio).

Discussion

Changes in matrix free [Ca²⁺]_m during respiration

This study demonstrates that in energized heart mitochondria, a transient ADP influx coupled to instantaneous ATP efflux markedly increases [Ca²⁺]_m. Moreover, this study demonstrates that acute physiological changes in [Ca²⁺]_m have little effect on mitochondrial bioenergetics other than to slightly increase resting-state respiration. The ADP-induced rise in [Ca²⁺]_m may be in large part due either to the lower Ca²⁺ buffering capacity of ADP versus ATP or to an ADP-induced protein release of stored matrix Ca²⁺. This transient increase in [Ca²⁺]_m is not measurably altered by Mg²⁺ or P_i,, and is little affected by an ADP-induced matrix contraction, or by ADP-induced uptake of matrix Ca²⁺ via the CU.

The uptake of [Ca²⁺]_m at a given [Ca²⁺]_e (Fig. 4) supports earlier studies (34,35) showing that Ca²⁺ influx through the CU causes Ca²⁺ to accumulate in the matrix, as predicted by the Nernst equation. However, the twofold gradient between [Ca²⁺]_m and [Ca²⁺]_e observed in this study after adding CaCl₂ (114 or 485 nM [Ca²⁺]_m) is smaller than that reported in earlier studies (34,35). Differences in species, isolation techniques, energy state, buffering capacities, equilibration rate, and measurement and calibration techniques for ionized [Ca²⁺] could underlie some of the differences.

The small, steady decline in [Ca²⁺]_e after addition of ADP (Fig. 2) could be due to increased sequestration of matrix free Ca²⁺ by matrix proteins. The continuous decrease in [Ca²⁺]_e without any further change in [Ca²⁺]_m after a rapid rise in [Ca²⁺]_m upon addition of 25 μM CaCl₂ to the buffer (485 nM [Ca²⁺]_e) is an interesting phenomenon that might also be explained by accumulating matrix Ca²⁺ storage in the form of Ca-PO₄ complexes (26), such as Ca₃(PO₄)₂, as Ca²⁺ enters the matrix through the CU. Formation of these complexes could result in a steady-state [Ca²⁺]_m during continued matrix Ca²⁺ uptake, as additional Ca²⁺ entering the matrix will precipitate with PO₄³⁻. Although there is no direct evidence for matrix Ca-PO₄ complex formation, there are indications that it occurs (36–38). However, we could not observe a clear difference in [Ca²⁺]_m at different buffer [P_i], perhaps because of the low, but physiologic, [Ca²⁺]_m.

Adding ADP at any [Ca²⁺]_m caused a further 1.6- to 2-fold increase in [Ca²⁺]_m during state 3 compared to state 2 respiration (Fig. 4). This effect of ADP on [Ca²⁺]_m was more than additive at higher [Ca²⁺]_m and is unlikely to be due to activation of the CU by ADP because the ADP-induced increases in [Ca²⁺]_m were not appreciably reduced by RR (Fig. 3). Although ADP transiently reduced matrix volume, this was not in itself sufficient to account for the much larger increases observed in [Ca²⁺]_m. Further, it is possible that a change in light scattering during state 3 respiration may also arise from a change in mitochondrial shape. Overall, these data suggest that matrix contraction by ADP is not an important factor.

There was no clear correlation among the different buffer [P_i]s and increases in [Ca²⁺]_m on ADP addition. The relationship between [Ca²⁺]_m and [P_i] has been shown previously (26), although at a relatively high [Ca²⁺]_m, and that study did not examine for any difference in [Ca²⁺]_m during state 3 respiration. It was not feasible to use P_i carrier blockers in our experiments to examine the importance of matrix [P_i] on [Ca²⁺]_m during state 3 respiration, because these blockers affect ADP/ATP transport as well (39).

In contrast to [P_i], the differential binding of ADP and ATP to Ca²⁺ apparently plays an important role in the ADP-induced increase in [Ca²⁺]_m, as shown by the slight increase of [Ca²⁺]_m in the OMN versus the CATR group. The lack of an increase in [Ca²⁺]_m on ADP addition after blocking the AAC shows that the ADP-induced increase in [Ca²⁺]_m is not due to contamination of ADP with divalent cations that can bind to indo-1. The small but significant increase in [Ca²⁺]_m in the OMN group shows that a change in [ADP]_m, without altering matrix [P_i], can change [Ca²⁺]_m. As discussed below, ADP binds to Ca²⁺ with a 10-fold lower affinity than does ATP; therefore, a decrease in matrix ATP due to ADP/ATP exchange alone may explain some of the increase in [Ca²⁺]_m. Another possible explanation for the ADP effect is release of ionized Ca²⁺ from proteins similar to calsequestrin, a release mechanism that is observed in the sarcoplasmic reticulum (40).

To better understand how ATP, ADP, and P_i affect [Ca²⁺]_m, it is first necessary to understand how these molecules are taken up into mitochondria and how they differentially bind to Ca²⁺. ADP³⁻ enters mitochondria via the AAC in exchange for ATP⁴⁻. Since this transport is electrogenic, and constitutes a net influx of one positive charge, it is driven by, and uses the energy of, the ΔΨ_m (41–43). In the matrix, ADP is phosphorylated to ATP by the F₁F₀ATP-ase, a process driven by proton flux into the matrix. The P_i carrier is responsible for electroneutral cotransport of P_i⁻/H⁺ (or P_i⁻/OH⁻ antiport), which is driven by the proton gradient, ΔpH (44–46).

Phosphorylation of ADP and dephosphorylation of ATP are dependent on the mitochondrial bioenergetic state (e.g., ΔpH) and ATP utilization by the cell, respectively. ADP levels rise in the matrix when extramatrix ADP is added and the ATP that is formed is ejected rapidly from the matrix by the AAC; matrix ADP levels remain elevated until all ADP is phosphorylated to ATP (4,47). It is important to note that the dissociation constants (K_ds) for ATP and ADP (Table S1) indicate a 10-fold greater binding affinity of Ca²⁺ and Mg²⁺ with ATP⁴⁻ over ADP³⁻, and a twofold greater binding affinity of Mg²⁺ versus Ca²⁺ to these phosphates. The binding affinities of these phosphate entities are uniformly high for H⁺ and very low for K⁺. Our premise is that these differences in binding affinities for ATP and ADP largely account for the ADP-induced increase in [Ca²⁺]_m.

The results of these experiments imply a small role for P_i and a larger role for the ADP/ATP ratio in buffering of Ca²⁺ by mitochondria. Nevertheless, the ADP-induced increase in [Ca²⁺]_m may be ascribed in part to other mechanisms that we explored. Our experiments performed with RR indicate that the ADP-induced Ca²⁺ flux through the CU cannot explain the phenomenon. The NCE was not active in the Na⁺-free buffer as verified by the NCE blocker CGP 37157. However, as there are currently no known blockers for Na⁺-independent Ca²⁺ exchange (NICE) such as Ca²⁺/H⁺ (CHE), it is not possible to exclude a decrease in Ca²⁺ efflux through NICE during state 3 respiration as an alternative explanation for the net increase of Ca²⁺ flux into the matrix (J_net = J_CU − J_NICE) that causes the increase in [Ca²⁺]_m. Because Ca²⁺ efflux through the NICE is dependent on both ΔpH ([H⁺]_e − [H⁺]_m) and Δ[Ca²⁺] ([Ca²⁺]_m − [Ca²⁺]_e) (20,48), we would not expect a major change in Ca²⁺ efflux through the NICE during state 3 respiration, since Δ[Ca²⁺] increases (Fig. 4) and ΔpH decreases (Fig. 7 C).

Another possible factor for the ADP-induced increase in [Ca²⁺]_m is the change in pH upon ADP addition. During state 3 respiration, pH_m decreases, albeit slightly (Fig. 7 C), due to influx of protons through the F₁F₀-ATPase (49). This transient acidification might affect the K_d between indo-1 and Ca²⁺ (50). However, the measured pH change was minimal (0.03 pH unit) and the reported changes in K_d were measured with a ΔpH of 1 pH unit. We expected the effect of ΔpH on [Ca²⁺]_m to be minimal in this study because of 1), the small change in pH_m, 2), the difference in [Ca²⁺]_m after ADP addition in the presence of low buffer P_i (no significant change in pH compared to control; data not shown), and 3), the increase in [Ca²⁺]_m after ADP addition in the presence of OMN (no significant change in pH on ADP addition; data not shown). Finally, the ADP-induced increase in [Ca²⁺]_m could be due to an alteration in Ca²⁺ buffering capacity by TCA cycle intermediates, which can bind to Ca²⁺ as well (51).

Changes in [Ca²⁺]_m and mitochondrial bioenergetics

These experiments did not show significant changes in mitochondrial bioenergetics when CaCl₂ was added to the buffer except for small increases in respiration in states 2 and 4, and only after 25 μM CaCl₂. Possible explanations for the increased resting-state respiration are opening of Ca²⁺-dependent K⁺ channels (31,52) or proton cycling through the putative NICE (48). In a previous study (53), in which respiration was measured in permeabilized cardiac cells, it was reported that adding CaCl₂ increased state 2 respiration, but decreased state 3 respiration. However, other studies report that an increase in [Ca²⁺]_m is associated with an increase in state 3 as well as state 2 respiration (5,54,55). In many of these studies, though, higher concentrations of CaCl₂ might have induced a large increase in respiration due to other factors, such as mPTP opening. In this study, irreversible mPTP opening did not occur, as there was no decrease in ΔΨ_m at the highest [Ca²⁺]_m observed with ADP. The small increase in O₂ consumption on CaCl₂ addition was not accompanied by other changes in bioenergetics, indicating that any slight uncoupling effect of Ca²⁺ could easily be corrected by increasing TCA cycle activity.

Many studies have shown a correlation between mitochondrial NADH and [Ca²⁺]_m (1–3,6). A possible explanation for the lack of an increase in NADH with increased [Ca²⁺]_m is that the highest [Ca²⁺]_m only just reached the K_0.5 for activation of TCA cycle dehydrogenases of ∼1 μM (55,56) during state 3 respiration, a period in which there is much fluctuation in bioenergetics. Other studies also question the role of Ca²⁺ in activation of NADH-producing dehydrogenases (57,58). Although these studies did not disprove the hypothesis of Ca²⁺ activation of TCA dehydrogenases, they did ascribe the changes in bioenergetics at least in part to other mechanisms (e.g., Mg²⁺ and ADP/ATP ratio), indirectly altered by Ca²⁺.

That an increase in [Ca²⁺]_m can activate TCA dehydrogenases in the matrix to enhance respiration was proposed long ago (55,56). Our study provides an alternative hypothesis to the correlation between work load (changes in NADH/NAD⁺) and [Ca²⁺]_m. We clearly observed an increase in [Ca²⁺]_m due to activation of oxidative phosphorylation with addition of ADP, and we believe this is due in part to a decrease in matrix Ca²⁺ buffering capacity of ADP versus ATP. It is conceivable that an increase in [Ca²⁺]_m is a result of enhanced phosphorylation-induced respiration, rather than enhanced respiration being a result of increased [Ca²⁺]_m. Our study was not designed to disprove or prove the hypothesis of Ca²⁺ and its role in the regulation of mitochondrial bioenergetics; however, this alternative hypothesis should be further explored.

[Ca²⁺]_m and mitochondrial pH

Matrix pH was not significantly affected by increased [Ca²⁺]_m. It was anticipated that [Ca²⁺]_m would compete with protons for binding with ATP, ADP, and P_i, and result in matrix acidification, just as acidification results in an increase in Ca²⁺ through buffering pathways (59). The amount of protons that dissociate from these buffering sites, when these sites bind to Ca²⁺, apparently can be sufficiently buffered in the matrix to maintain pH. We observed that [H⁺]_m varies between 10 and 100 nM (pH between 7 and 8), which is about one order of magnitude lower than the changes in [Ca²⁺]_m in this study. This indicates a higher matrix buffering capacity for protons than for Ca²⁺. Apparently the increase in [Ca²⁺]_m is insufficient to produce a significant decrease in the H⁺ buffering capacity of the matrix. Otherwise, it is likely that an increase in proton expulsion coupled to electron transfer corrected for this increase in [H⁺]_m, as respiration was slightly increased by CaCl₂ addition. We have shown in preliminary experiments (unblocked NCE) that pH decreases with the addition of higher [CaCl₂] (50 or 100 μM) (J. Haumann, A. Camara, and D. Stowe, unpublished observations). This decrease in pH was not blocked by cyclosporin A, an mPTP inhibitor, indicating that mPTP opening did not cause the acidification.

Summary and Conclusions

To our knowledge, the up to twofold increase in [Ca²⁺]_m during oxidative phosphorylation under conditions of blocked CU and inactive NCE has not been reported previously. This study demonstrates the importance of changes in ADP phosphorylation on matrix free [Ca²⁺] and suggests a change in matrix buffering of Ca²⁺ or release of matrix stores of Ca²⁺ as possible mechanisms for the ADP-induced increase in [Ca²⁺]_m. This observed increase may be dependent on buffer conditions and the different binding constants of Ca²⁺ to ATP, ADP, P_i, and matrix proteins. Changes in redox state, matrix pH, buffer [P_i], and [Mg²⁺]_m, did not alter [Ca²⁺]_m. Activation of the CU, and contraction of the matrix volume by ADP did not appear to account substantially for this effect of ADP. The conventional postulation is that increased [Ca²⁺]_m enhances respiration via NADH-linked substrates; we show, alternatively, that ADP-induced stimulation of respiration enhances [Ca²⁺]_m. Computer modeling (23) of the dynamics of Ca²⁺ flux and mitochondrial buffering mechanisms will help to confirm our findings, or to iteratively delineate the relative importance of each of the several proposed mechanisms that may modulate [Ca²⁺]_m between states 3 and 4.

The important and dependent relationship between [Ca²⁺]_m and [ADP], [P_i], and [ATP] should lead to a reexamination of the relationship between [Ca²⁺]_m and control of respiration. Although increased [Ca²⁺]_m slightly enhanced respiration in states 2 and 4 (NCE blocked), the increase in [Ca²⁺]_m via reduced binding in the matrix during state 3 respiration did not enhance respiration. We could not demonstrate any Ca²⁺-induced changes in ΔΨ_m or NADH, but we found a small decrease in RCI, indicating slight uncoupling. The evidence that ΔΨ_m and NADH were not different in the OMN and CATR groups suggests that the changes observed resulted from an altered buffering effect due to differences in the ADP and ATP concentrations or to release of Ca²⁺ from stores, and less from increased Ca²⁺ influx through stimulation of the CU, or decreased Ca²⁺ efflux through NCE.