| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Instituto de Biología y Genética Molecular (IBGM), Universidad de Valladolid y Consejo Superior de Investigaciones Científicas, Departamento de Fisiología y Bioquímica, Facultad de Medicina, E-47005 Valladolid, Spain; and
* Instituto Teófilo Hernando, Departamento de Farmacología y Terapéutica, Facultad de Medicina, Universidad Autónoma de Madrid, E-28029 Madrid, Spain.
1Correspondence: IBGM, Departamento de Fisiología y Bioquímica, Facultad de Medicina, E-47005 Valladolid, Spain. E-mail: jgsancho{at}ibgm.uva.es
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: calcium • mitochondria • aequorin • nucleus
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Since Ca2+ may act on many different effectors at different subcellular locations, considerable attention has been paid to the spatiotemporal profiles of [Ca2+]c transients. Diffusion of Ca2+ through the cytosol, cytosolic Ca2+ buffering, and Ca2+ transport by organella are essential for shaping the [Ca2+]c transients. Changes of [Ca2+] inside organella may also be relevant to determine cell responses (2
3
4)
.
Cytosolic Ca2+ buffering and diffusion in bovine chromaffin cells have been studied by Neher and co-workers (5
, 6)
. The cytosol has a total binding capacity of 4 mM and the endogenous Ca2+ buffer is poorly mobile, fast (association rate=10-8 M-1s-1), and with low affinity for Ca2+ (dissociation constant=100 µM). The activity coefficient of the endogenous buffer is 
1/40. Cytosolic buffering slows down Ca2+ mobility to reach an apparent diffusion of 10-7 cm2s-1. The 2-dimensional apparent diffusion coefficient is 40 µm2/s and shows inhomogeneities at the nuclear envelope and at the plasma membrane (7)
. Brief opening of VOCC generates microdomains of high [Ca2+]c near the mouth of the channel that can be detected in Ca2+ imaging measurements (8)
. In such microdomains, Ca2+ can reach concentrations as high as 10 µM and perhaps 100 µM (2
, 9)
. Because of rapid diffusion of Ca2+ toward the surrounding cytosol, [Ca2+]c microdomains are highly restricted in time and space (2
, 10)
.
The Ca2+ fluxes relevant to determine [Ca2+]c transients after stimulation of chromaffin cells are entry and extrusion at the plasma membrane and uptake and release at the cytoplasmic organella. Among the last, the endoplasmic reticulum (ER) and mitochondria seem to be the most directly involved. Secretory granules contain large amounts of calcium, but exchange through their membrane is too slow to contribute to [Ca2+]c transients (but see ref 11
).
Ca2+ entry through VOCC and its subsequent clearance have been studied extensively in bovine and rat chromaffin cells (5
, 6
, 12
, 13)
. Membrane depolarization to 0 mV elicits Ca2+ currents peaking near 800 pA and deactivating with half-time constants of 300–500 ms. In bovine chromaffin cells, a 0.5 s stimulation period typically elicits a mean ICa of 250 pA. In terms of Ca2+ flow, this current is equivalent (for a 15 µm diameter cell) to 700 µmol/l of cells/s (5)
. Measurements of 45Ca uptake by bovine chromaffin cells depolarized with high K+ (59 mM) was linear during the first 5 s at an estimated rate of 0.7·10-15 mol per cell/s, which is equivalent to 400 µmol/l of cells/s (14)
.
Plasma membrane Ca2+ extrusion is due to operation of a plasma membrane Ca2+ ATPase and an Na+/Ca2+ exchange system. The joint action of both transport systems has been estimated to decrease [Ca2+]c at a maximal rate of 0.2 µM/s, equivalent to 20 µmol/l cells/s, in rat chromaffin cells at 27°C (12
, 13)
.
Most of the information on transport parameters by organella in intact cells is indirect, inferred from their effects on [Ca2+]c. Maximal uptake by ER is considered to contribute little, perhaps 1–1.5 µM/s (equivalent to 40–60 µmol/l cells/s), to changes in [Ca2+]c during stimulation of bovine (6)
and rat (13)
chromaffin cells. It has been shown recently that cytosolic Ca2+ may trigger Ca2+-induced Ca2+ release (CICR) from the ER of bovine chromaffin cells (15)
.
Ca2+ transport by mitochondria has received renewed attention in the last few years, both because of possible participation in shaping [Ca2+]c transients and because changes of intramitochondrial [Ca2+] ([Ca2+]M) seem important by themselves for the regulation of cell functions, such as the rate of respiration or programmed cell death (16
, 17)
. Ca2+ is taken up through the Ca2+ uniporter, a low-affinity/high-capacity system driven by the mitochondrial membrane potential (18)
. Ca2+ exit from mitochondria through an Na+/Ca2+ exchanger and through an Na+-independent system, the first being dominant in the adrenal gland (18
, 19)
. The activity coefficient of Ca2+ inside the mitochondrial matrix seems to be very low, in the 1/1000 range (20
21
22)
. It has been shown that mitochondria can contribute to clearing cytosolic Ca2+ loads in rat (13
, 21)
and cow (6)
chromaffin cells. Herrington et al. (13)
reported mitochondrial uptake rates of (in terms of changes of [Ca2+]c) 0.4 to 0.7 µM/s at [Ca2+]c concentrations of 0.5–2 µM. Xu et al. (6)
report much larger rates: 120 µM/s at saturating [Ca2+]c concentrations (200 µM). These differences are consistent with the [Ca2+]c dependence of the activity of the mitochondrial Ca2+ uniporter (23
, 24)
.
We have recently shown the validity of using targeted aequorins with different Ca2+ affinities to monitor directly changes of [Ca2+] inside ER and mitochondria of living chromaffin cells (15
, 23
24
25)
. Taking advantage of this technology, we attempt here a quantitative explanation of fluxes among the extracellular, cytosolic, and organellar compartments during depolarization of chromaffin cells leading to Ca2+ entry through VOCC.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Cells were plated on 12 mm glass poly-D-lysine-coated coverslips (1–5x105 cells; 0.5 ml) and maintained at 37°C in a humidified atmosphere of 5% CO2. [Ca2+]c imaging in cells loaded with either Fura-2 (26)
or fura-4F (24)
was as described previously. Recombinant aequorins containing targeting sequences to several subcellular locations (27)
have been used. Both wild-type aequorin and mutated (Asp119
Ala), low Ca2+ affinity aequorin (AEQmut; ref 28
) were used. For aequorin expression, cells were infected with a defective herpes simplex virus type 1 (HSV-1) containing the corresponding chimeric aequorin gene and cultured for 12–24 h prior to measurement. Viruses expressing mitochondrial aequorin (mitAEQ), low Ca2+ affinity mutated mitochondrial (mitAEQmut), or ER (erAEQmut) aequorins have been described (15
, 23
, 24)
. Nuclear (nucleoplasmin) and cytosolic aequorin cDNAs were obtained from Molecular Probes (Eugene, OR) and cloned in the pHSVpUC plasmid. Packaging and titration of the pHSVnucAEQ (nuclear) and pHSVcytAEQ (cytosolic) viruses were performed as reported before (25)
. The multiplicity of infection was from 0.01 to 0.1 for batch luminescence measurements (15
, 24)
and from 0.3 to 1 for bioluminescence imaging.
Measurements of aequorin bioluminescence and NAD(P)H fluorescence
Cells expressing apoaequorins were incubated for 1–2 h at room temperature with 1 µM coelenterazine. Coelenterazine n was used to reconstitute mitAEQmut in order to decrease further Ca2+ affinity (29)
. Batch cell aequorin photoluminescence measurements were performed as described (15
, 24)
and calibrations in [Ca2+] were done using published constant values (30)
. For bioluminescence imaging (31
, 32)
, cells were placed into a perfusion chamber thermostatized to 37°C under a Zeiss Axiovert S100 TV microscope (objective, Fluar 40x oil, 1.3 n.a.) and perfused at 5–10 ml/min with test solutions prewarmed at 37°C. The standard incubation medium had the following composition (in mM): NaCl, 145; KCl; 5; CaCl2, 1; MgCl2, 1; glucose, 10; sodium-HEPES, 10, pH 7.4. At the end of each experiment, cells were permeabilized with 0.1 mM digitonin in 10 mM CaCl2 to release all residual aequorin counts. Images were taken with a Hamamatsu VIM photon-counting camera handled with an Argus-20 image processor and integrated for 10 s periods. Photons/cell in each image were quantified using the Hamamatsu Aquacosmos software. Total counts per cell ranged between 3 · 103 and 3 · 105 and noise was (mean±SD) 1 ± 1 c.p.s. per typical cell area (3000 pixels). Values are referred to the whole cell area. Data were first quantified as rates of photoluminescence emission/total c.p.s remaining at each time and divided by the integration period (L/LTOTAL in s-1). Calibrations in [Ca2+] were performed using published constant values (30)
. A transmission image was also taken at the beginning of each experiment.
Mitochondrial NAD(P)H fluorescence was measured using the same set up as for aequorin with excitation at 340 ± 10 nm and emission at 450 ± 40 nm. The integration period was 6 s. For these experiments, 1 mM pyruvate was added to the standard medium in order to keep the cytosolic NAD in the oxidized state.
Coelenterazines and the acetoxymethyl (AM) esters of fura-2 and fura-4F were obtained from Molecular Probes. CGP37157 was from Tocris (Bristol, UK). Other reagents were of the highest quality available from Sigma (St. Louis, MO) Calbiochem (San Diego, CA), or Merck (Rahway, NJ).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
illustrates a typical experiment in which the cells were stimulated by two consecutive high K+ pulses. The peak [Ca2+]c reached 4.8 µM in this experiment (Fig. 1A
). In five similar experiments, the average [Ca2+]c peak (expressed as 
[Ca2+]c; mean±SD; n=213 cells) was 4.05 ± 1.41 µM. The rates of [Ca2+]c change are shown in Fig. 1B
. The increase of [Ca2+]c reached a maximum rate of 1.1 µM/s during stimulation. In five similar experiments, the average value was (mean±SD) 1.15 ± 0.15 µM/s. In two similar experiments performed at 22°C, changes of [Ca2+]c were similar in both peak amplitude (mean±SD, 4.05±1.51 µM; n=75) and kinetics (the maximum rate of [Ca2+]c increase ranged between 0.7 and 1 µM/s in four independent measurements; results not shown).
|
The rate of [Ca2+]c increase measured here is much smaller than expected from typical ICa or 45Ca uptake measurements. For an activity coefficient of 1/40 for the cytosolic Ca2+ buffers (5
, 6)
, Ca2+ entry would be expected to increase [Ca2+]c at a rate of 10–17.5 µM/s, which is one order of magnitude larger than measured in cells loaded with fura-4F. These results suggest that most Ca2+entering the cell during the 10 s stimulation period is taken up into organella.
Once the stimulation pulse ends, Ca2+ is cleared from the cytosol at a rate of 0.7 µM/s, which declines quickly as [Ca2+]c approaches its prestimulation value (Fig. 1B
). At this stage, Ca2+ clearance (unopposed by Ca2+ entry) reflects the joint action of uptake into organella and extrusion through the plasma membrane.
Uptake of Ca2+ by the ER
The uptake of Ca2+ by the ER can be monitored precisely using targeted aequorins with low Ca2+ affinity (15
, 25
, 29
, 33)
. Figure 2
A shows the time course of ER refilling in Ca2+-depleted chromaffin cells upon incubation with Ca2+-containing medium, measured at 22°C. On addition of Ca2+, [Ca2+]ER increased to 500–600 µM in 2–3 min. The maximum rate of the [Ca2+]ER increase was 5.5 µM/s. In 12 similar experiments, the maximum rate was (mean±SD) 4.5 ± 1.3. This value decreased to 0.8 µM/s in cells loaded with BAPTA and increased to 9.0 ± 3.2 (n=8) when cells were depolarized with high K+ solution during the refilling period (data not shown). At 37°C, uptake is fourfold faster (33)
, which gives a maximum figure of 36 µM/s. Assuming an activity coefficient of 1/20 and a volume of 10% relative to the cytosol (3
, 34)
, this value would be equivalent to 72 µmol/l cells/s, which is still one order of magnitude smaller than the value estimated for Ca2+ influx through VOCC. When Ca2+ influx takes place in cells with Ca2+-filled ER, net Ca2+ release rather than Ca2+ uptake is produced (Fig. 2B
), an expression of the activation of the CICR mechanism (15)
. We conclude that the ER could not be responsible of the buffering of [Ca2+]c during stimulation of bovine chromaffin cells. Based on measurements of [Ca2+]c clearance and the effects of inhibitors of SERCA ATPases, Neher and co-workers (6)
and Hille and co-workers (13)
arrived to the same conclusion in bovine and murine chromaffin cells, respectively.
|
Uptake of Ca2+ by mitochondria and nucleus
We have shown before that the bovine chromaffin cell mitochondria accumulate large amounts of Ca2+ during stimulation with either high K+ solutions, acetylcholine, or caffeine (23)
. The Ca2+ load was taken up selectively into a mitochondrial pool amounting to 50% of the total (pool M1); the other half (pool M2) takes up much smaller amounts (23)
. A similar conclusion was reached by Pivovarova et al. (35)
for rat chromaffin cells by studying mitochondrial calcium content by electron microscopy X-ray microanalysis.
Figure 3
A shows new measurements performed at the single cell level. Cells expressing the mitochondria-targeted aequorin were repeatedly stimulated with high K+ solution or caffeine and then permeabilized with digitonin to burn up the remaining aequorin. Representative images corresponding to each pulse are shown at the top; photoluminescence emission (expressed as L/LTOTAL, s-1; see Materials and Methods) and aequorin consumption by one of the cells are shown below. About 40% of the aequorin was consumed during the first stimulus, but subsequent stimuli had a much smaller effect. Permeabilization with digitonin burned up the remaining half of the aequorin pool. This contrasts with the changes in [Ca2+]c accomplished by repeated stimulation with high K+, which were reproducible (Fig. 1A
; see also ref 23
). We performed another series of experiments using nucleus-targeted aequorin (Fig. 3B
). Here the behavior was the same as that found for [Ca2+]c: each stimulus produced similar aequorin consumption (5% of the total) and luminescence emission. Similar results were reported for the cytosolic aequorin (not shown).
|
Figure 4
A shows the average consumption of the mitochondria-targeted aequorin (±SE) of 55 cells present in the same microscope field together with the responses of four individual cells selected to illustrate the range of variation observed. Even though there were differences among the extent of the responses of individual cells, in all cases the response to the second stimulus was much smaller than to the first one. Figure 4B
shows calibrated [Ca2+]M responses for the same experiment. The first Ca2+ peak reached 10 µM; the ensuing ones consistently reached only 2–4 µM. Using aequorin reconstituted with coelenterazine n, which has a lower affinity for Ca2+ (measuring range 1–100 µM, ref 23
), behavior regarding aequorin consumption was similar (Fig. 4C
). The calibrated signal (Fig. 4D
) now revealed a [Ca2+]M peak of 80 µM for the first stimulus; the second reached only 8 µM and subsequent ones where in the 2–4 µM range, which is barely detectable for this aequorin–coelenterazine combination.
|
We conclude from the above experiments that Ca2+ is taken up during the first stimulus into a fraction of mitochondria (pool M1, amounting to 50% of the total mitochondrial pool) at a high enough concentration to burn out all its aequorin. Because of this, the ensuing high K+ stimuli do not produce light emission from this pool. The other half of the mitochondria (pool M2) takes up much smaller amounts of Ca2+, resulting in a modest light emission that is seen uncontaminated during the last stimuli. The aequorin contained into pool M2, however, is burned out upon permeabilization with digitonin (Fig. 3A
).
The extent of the uptake of Ca2+ into pool M1 can be quantified using a mutated, low Ca2+ affinity aequorin reconstituted with coelenterazine n (mitAEQmut) that can report Ca2+ concentrations as high as 10-3 M for several minutes (23)
. Due to its lower Ca2+ affinity, this aequorin is burned only partly during each stimulation with high K+. Figure 4E
shows the average response (±SE) of 10 single cells to a challenge with high K+, as reported by mitAEQmut reconstituted with coelenterazine n. [Ca2+]M reached a peak of 350 µM within 10–15 s, then decreased slowly (the decrease of [Ca2+]M is especially slow because it was performed at 20°C, see below), remaining above 300 µM for at least 1 min. According to our estimates (see Materials and Methods), > 90% of the wild-type aequorin should be consumed by < 1 s when reconstituted with normal coelenterazine and within 1 min when reconstituted with coelenterazine n.
Figure 5
A shows the time course of Ca2+ uptake, measured with mitAEQmut at 37°C, by the mitochondria of pool M1 during repeated stimulation with high K+ solution. [Ca2+]M increased to 300–400 µM by the end of the stimulation period. The rate of uptake reached maximum values of 40–50 µM/s (Fig. 5B
). In 18 similar measurements, the maximum rate of [Ca2+]M increase was (mean±SD) 56 ± 12 µM/s. Assuming an activity coefficient of 1/1000 and a relative mitochondrial volume of 4%, this Ca2+ load would be equivalent to 1100 µmol/l cells/s (taking into account that only 50% of the mitochondria take up Ca2+). This value is similar to the one estimated for Ca2+ influx through VOCC, suggesting that most of the Ca2+ load taken up by the cell during stimulation can accumulate into mitochondria.
|
To study the kinetics of Ca2+ transport through the uniporter, we measured mitochondrial uptake in bovine chromaffin cells permeabilized by brief treatment with digitonin and perfused with different Ca2+ buffers (see Materials and Methods). Results are shown in Fig. 5C
. The uptake was very slow at [Ca2+]c below 2 µM, but increased steeply at higher concentrations. The line is the best fit to the equation:
 |
, 19
, 36
) and the one estimated by Xu et al. (6)
from measurements of the decrease of [Ca2+]c in intact bovine chromaffin cells (40 µM). The value of Vmax would be equivalent to 6320 µmol/l cells/s, not far from the 4800 value estimated by Xu et al. (6)
at 200 µM [Ca2+]c.
Relaxation of [Ca2+]c peaks
When Ca2+ entry through VOCC ceases, [Ca2+]c begins to decrease as a result of multiple fluxes. Plasma membrane Ca2+ ATPase and Na+/Ca2+ exchanger extrude Ca2+ from the cytosol to the extracellular medium. We have not performed specific measurements of these fluxes, but an upper limit of 0.8 µM/s can be estimated from the maximum rate of [Ca2+]c decrease (Fig. 1C
). The mean value from six independent measurements was (mean±SD) 0.73 ± 0.07, which is equivalent to 29 µmol/l cells/s. Herrington et al. (13)
have estimated a value of 0.3 µM/s in rat chromaffin cells at 27°C. Assuming a Q10 value of 3–4, this value is comparable to our estimate.
If this maximum rate was sustained for a few seconds, the return of [Ca2+]c to the prestimulation level would be extremely fast. However, the rate of [Ca2+]c decline slows down quickly (Fig. 1C
). Desaturation of the plasma membrane Ca2+ extrusion mechanism may help to slow down relaxation. However, the major determinant of this [Ca2+]c stabilization is that once it declines below 2 µM, plasma membrane Ca2+ extrusion is almost matched by Ca2+ release from the organella (see below).
Ca2+ release from the ER is comparatively quite slow. Upon treatment of chromaffin cells with cyclopiazonic acid, we find a maximal rate of [Ca2+]ER decrease of 1.3 µM/s, which is equivalent to 2.6 µmol/l cells/s (15)
. Release from mitochondria is much faster. Upon termination of the stimulus with high K+, [Ca2+]M decreases at a maximal rate of 13–18 µM/s (Fig. 5B
). In five similar experiments, the maximum rate of [Ca2+]M decrease was (mean±SD) 14.2 ± 3.4 µM/s. Since exit takes place from only one-half of the mitochondrial pool, it would be equivalent to 280 µmol/l cells/s. This rate decreases quickly as [Ca2+]M declines (Fig. 5B
).
Kinetics of Ca2+ release from mitochondria loaded with Ca2+ is shown in Fig. 6
. Traces A and B compare the time course of the [Ca2+]M changes in control cells and in cells treated with CGP37157, an inhibitor of the mitochondrial Na+/Ca2+ exchanger. Traces C and D show the estimated rate constants and E a plot of exit against [Ca2+]M. As shown before (23
, 24)
, CGP37157 slowed down Ca2+ efflux. In control cells, mitochondrial Ca2+ exit showed saturation at the higher [Ca2+]M and a sigmoidal shape at the lower [Ca2+]M. In the presence of CGP37157, saturation and sigmoidicity both seemed to disappear (Fig. 6E
). When we constructed a similar plot with the average results of 11 control experiments, the resulting data points could be well fitted by the following equation:
 |
|
The efflux of Ca2+ from loaded mitochondria was also highly sensitive to temperature (23
; compare Fig. 4E
and Fig. 5A
). At 22°C, both sigmoidicity and saturation were lost, although CGP37157 still produced a large inhibition of mitochondrial exit (Fig. 6F
).
Changes in mitochondrial NADH fluorescence upon stimulation with high K+
Several mitochondrial NADH dehydrogenases are activated by Ca2+ in the micromolar range (19)
. To check whether the increase in [Ca2+]M we observed is sufficient to produce this effect, we measured endogenous NAD(P)H fluorescence (37)
in chromaffin cells stimulated with high K+ solution. In these experiments, pyruvate (2 mM) was added to the incubation solution in order to keep cytosolic NADH oxidized throughout the reaction catalyzed by lactate dehydrogenase. At the end of the experiment, the cells were perfused with 2 mM NaCN to obtain full reduction of the mitochondrial NADH by blocking electron transfer to oxygen. Figure 7
summarizes the results. Images taken at representative times during the experiment and the average trace (±SE) of 10 single cells are shown. Stimulation with high K+ increased the NADH fluorescence by 40%, and this increase declined slowly after removal of the stimulus. The treatment with CN- increased the NAD(P)H fluorescence by 100%; this increase reverted quickly upon removal of the poison.
|
Modeling redistribution of Ca2+ among different cell compartments after stimulation of chromaffin cells
Figure 8
A compares fluxes (in µmol/l cells/s) estimated for entry through VOCC, mitochondrial uptake through the Ca2+ uniporter, and pumping by the plasma membrane and the SERCA ATPases at different [Ca2+]c. It is clear that Ca2+ entry overwhelms the pumps and [Ca2+]c must rise quickly. Above 1 µM, the mitochondrial uniporter begins to dominate Ca2+ transport and at 10 µM it becomes equal to Ca2+ entry through VOCC. Therefore, a steady state with no further change in [Ca2+]c would be predicted at this concentration. We measured, however, smaller [Ca2+]c concentrations during stimulation (Fig. 1)
. The maximal rate of mitochondrial uptake measured (56 µM/s=3300 µmol/l cells/s into one-half of the mitochondrial pool) requires [Ca2+]c values of 20–40 µM (Fig, 8A; see also ref 23
). On the other hand, we identify a second mitochondrial pool grouping the remaining 50% of mitochondria that takes up Ca2+ at rates corresponding to [Ca2+]c concentrations of only 3–4 µM (23)
. The simplest explanation for these results is that [Ca2+]c is not homogeneous but that cytosolic domains with different Ca2+ concentrations are established. Entry through VOCC amounts to only 10% of the Vmax estimated for the mitochondrial uniporter (Fig. 8A
). This means that as little as 10% of the mitochondria would be able to take up all the Ca2+ load entering through VOCC provided a near-saturating [Ca2+]c is reached at the cytosolic face of this mitochondrial pool. This would slow the progress of the Ca2+ wave toward the core of the cell.
|
Figure 8B
shows the predictions of a simplified model (in the upper part) that includes the transport parameters proposed here for the organella when Ca2+ entry is sustained for a few seconds. The [Ca2+]c concentrations graded from the subplasmalemmal region to the core of the cell within a range of more than one order of magnitude (from 0.3 to 20 µM). This was because most of the Ca2+ load was taken up into the mitochondria physically closer to the plasma membrane. Figure 8C
shows individual flows among the different compartments. After stimulation, it can be seen that uptake through the mitochondrial uniporter quickly becomes dominant and matches plasma membrane entry.
By the end of the stimulation period, [Ca2+]c decreases quickly to the submicromolar range (Fig. 8B
). This is because Ca2+ clearance from the cytosol, mainly by mitochondria, is no longer counteracted by influx through the plasma membrane. Once [Ca2+]c drops below 1 µM, extrusion by the plasma membrane and ER Ca2+ ATPases become the dominant flows and release from mitochondria through the exchanger becomes larger than uptake through the uniporter (Fig. 8C
). The decrease in [Ca2+]c then becomes much slower because extrusion from the cytosol is almost matched by mitochondrial release. During this period, [Ca2+]c concentrations are graded from the cell core to the subplasmalemmal region (Fig. 8B
) because plasma membrane extrusion takes place from this last area and Ca2+ released from mitochondria must diffuse toward there before being cleared.
ER plays a relatively modest role in this model. It takes up a minor fraction of Ca2+ during stimulation, thus contributing to sharpen the [Ca2+]c domains, especially at the lower [Ca2+]c regions (cell core). After stimulation, ER continues accumulating Ca2+ as long as [Ca2+]c is kept high by mitochondrial release. At longer periods (not shown), ER slowly releases its Ca2+ load. To keep it simpler, we have not included CICR in modeling results of Fig. 8
. even though we have been able to document experimentally its operation in bovine chromaffin cells (15)
. When included, CICR sharpened the [Ca2+]c domains observed during stimulation, as release happened preferentially from ER areas close to the plasma membrane while ER from the core was taking up Ca2+ from the surrounding cytosol. This Ca2+ diffused throughout the ER matrix to be released again near the plasma membrane.
Table 1
summarizes the main phases that can be defined from this model regarding typical [Ca2+]c values and fluxes at cell, mitochondria and ER membranes (see below).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 10)
. The smaller [Ca2+]c ranges reached and the spatiotemporal restrictions may preclude a major contribution of organellar transport to shaping of the [Ca2+]c peaks under this approximation. Our data, on the contrary, stress the role of organellar transport in the clearance of cytosolic Ca2+. It is unclear which type of stimulus should be considered closer to physiological. Short entry would simulate better low frequency–action potentials whereas sustained entry may be closer to bursts of action potentials, as is the case at the splanchnic nerve–chromaffin cell synapse (38)
, or to maintained action of acetylcholine or other chemical stimuli.
We find several new properties of mitochondrial Ca2+ transport systems acting on intact cells using targeted low Ca2+ affinity aequorins, which are able to measure near millimolar [Ca2+]M. For example, maximum capacity of the mitochondrial uniporter approaches 6000–7000 µmol/l cells/s, a value well above most previous proposals. Previous estimates of Ca2+ accumulation by mitochondria of rat chromaffin cells (13)
or frog sympathetic neurons (35
, 39
, 40)
are in the range of 50–200 µmol/l cells/s at [Ca2+]c concentrations of 0.5–2 µM. However, the uniporter shows second-order kinetics with regard to Ca2+ (18)
; since these concentrations are well below saturation of the transport system (K50=10–40 µM), it is not possible to extrapolate reliable Vmax values from these data. Increasing [Ca2+]c over a much wider range (
200 µM) by photon-induced Ca2+ release from DM-dinitrophen, Xu et al. (6)
estimated in bovine chromaffin cells Vmax values of 4800 µmol/l cells/s, a value close to the one reported here. Dialysis of cytosolic components in cells maintained under whole cell patch, which decreases eightfold the activity of the uniporter (21)
, could also contribute to explain differences with previous results. Therefore, we believe that the high rates of uptake reported here are representative of the real values in intact cells. This implies that mitochondria can take up most of the Ca2+ load entering the cells through VOCC if cytosolic Ca2+ concentrations high enough to activate the uniporter are built up. The idea that mitochondria can contribute to clearing Ca2+ loads from the cytosol has already been proposed for chromaffin cells (6
, 21
, 23)
and pancreatic acinar cells (41
, 42)
, but we provide a solid quantitative argument.
Our model predicts that a fraction of the mitochondrial pool closer to the plasma membrane can sink most of the Ca2+ load entering through VOCC during several seconds (Fig. 8)
. This results in the generation of subcellular domains with different [Ca2+]c. Thus, high Ca2+ domains are restricted to the subplasmalemmal region, whereas low Ca2+ domains locate at the core of the cell. Mitochondria act as biosensors to discriminate high and low Ca2+ domains. Second-order kinetics of the uptake through the uniporter sharpens differences in [Ca2+]c above and below a threshold located near its K50, 10-5 M. Taking advantage of the fact that the native, high Ca2+ affinity aequorin is fully burned up within a few seconds into a high Ca2+ (>10-5 M) environment, we have defined two mitochondrial pools, M1 and M2, each amounting to 50% of the total, which differ by their rate of Ca2+ uptake after cell stimulation (23)
. Pool M1 takes up Ca2+ at rates of > 50 µM/s (Fig. 5)
corresponding to [Ca2+]c concentrations of > 20 µM (Fig. 8A
). Pool M2 takes up Ca2+ at only 0.3 µM/s, corresponding to 2–3 µM [Ca2+]c (23)
. Electron microscopy X-ray microanalysis of frog sympathetic neurons after stimulation with high K+ also revealed the existence of two mitochondrial pools with different Ca2+ contents (35)
. In pancreatic acinar cells, entry of Ca2+ through plasma membrane also caused preferential Ca2+ uptake into subplasmalemmal mitochondria (42)
.
Continuity of the intramitochondrial space seems much greater than previously thought (43
, 44)
. Thus, it is conceivable that the Ca2+ load that fills the M1 pool may have been taken up by a smaller fraction of mitochondria close to the plasma membrane Ca2+ channels, then diffused throughout the mitochondrial matrix to invade a larger fraction of the mitochondrial pool. According to the figures used here for Vmax of the uniporter and Ca2+ entry through VOCC, the full Ca2+ load could be taken up by only 10% of the mitochondrial pool if a high enough [Ca2+]c is reached. Assuming a spherical shape for the cells and uniform distribution of mitochondria, 10% of the mitochondrial space would occupy only 0.25 µm beneath the plasma membrane in a 15 µm diameter cell.
Neher and co-workers first elaborated the idea that transient opening of VOCC during action potentials should generate high Ca2+ microdomains near the channel mouth. Such microdomains are highly restricted in time (ms scale) and space (nm scale) due to dissipation by rapid diffusion of Ca2+ to the surrounding cytosol. The existence of such microdomains as well as the constrictions imposed by diffusion and binding to cytosolic Ca2+ buffers has been convincingly documented in bovine chromaffin cells (2
, 10)
. The high Ca2+ subcellular domains proposed here are maintained for seconds because of the pump/leak steady state established between Ca2+ entry and mitochondrial uptake. We have no indication of the size of such domains, although they would probably be bigger than microdomains. When modeling, the ratio between the Vmax of the uniporter and the diffusion coefficient is important in determining the shape of the [Ca2+]c gradation, which becomes sharper the larger this ratio is. For these reasons, the present model should be regarded as a crude approximation. For example, denser packing of the organella or slower diffusion rates at the subplasmalemmal region would render gradation of Ca2+ domains steeper. Previous observations on irregular progression of the Ca2+ wave at this region (7)
suggest that such physical constraints may apply.
After stimulation, two phases can be defined in the relaxation of the [Ca2+]c peak (Table 1)
. In the early poststimulation period (PS1), [Ca2+]c drops quickly to concentrations below 1 µM mainly because of rapid clearance from the cytosol by mitochondrial uptake. After this initial drop, [Ca2+]c decreases much more slowly, as extrusion through the plasma membrane is counteracted by mitochondrial Ca2+ release (late poststimulation period; PS2 in Table1). This phase is usually evident in fura-2 measurements as a long period when [Ca2+]c remains moderately increased above resting levels. It has been reported that collapsing the mitochondrial membrane potential by addition of protonophores induces a [Ca2+]c increase by mitochondrial Ca2+ release, but only for a short period after the end of stimulation (13
, 21)
. We have documented recently that protonophores may induce rapid release of mitochondrial Ca2+ by reversal of the Ca2+ uniporter, but this requires relatively high [Ca2+]c (24)
, a condition accomplished only during period PS1. During the PS2 period, [Ca2+]c is graded from the cell core to the subplasmalemmal region. In the last region [Ca2+]c drops very quickly to the resting level, thus securing the end of exocytosis, by action of the plasma membrane Ca2+ extrusion mechanisms. At the cell core, [Ca2+]c remains discretely high for many seconds because of the relatively slow release of the mitochondrial load. Perhaps this maintained increase helps mobilize new secretory vesicles from the reserve pool toward the plasma membrane, thus becoming ready to be used for the next exocytotic episode (2)
.
Exit of the Ca2+ load from mitochondria occurs primarily through the Na+/Ca2+ exchanger (Fig. 6
; ref 19
). We also document some new properties for the in situ operation of this transport system. First, activity is much faster than most of previous estimates in isolated mitochondria (18
, 19)
. The same reasons cited above for the uniporter may apply here. Moreover, the mitochondrial Na+/Ca2+ exchanger is very sensitive to temperature (Fig. 6)
, which was 20°C in most of the previous estimates. Another new finding was the sigmoidal, second-order kinetics of [Ca2+]M. The activity of this transport system had been reported to show Michaelian first-order kinetics with regard to Ca2+ in experiments with isolated mitochondria (18)
. Again, temperature may explain some of the discrepancies, as sigmoidicity disappears at 20°C (Fig. 6)
. Sigmoidicity also disappeared in the presence of the inhibitor CGP37157 (Fig. 6)
, thus excluding an artifact generated by the measuring procedure.
Second-order kinetics fits well with the high rates of mitochondrial Ca2+ clearance observed at high [Ca2+]M and the rapid decline when [Ca2+]M decreases (Fig. 6)
. Accordingly, [Ca2+]M would remain in the micromolar range several minutes after the stimulus. Stimulation of pyruvate- and 
-ketoglutarate-dehydrogenase by Ca2+ has been reported to be complete at > 3 µM (16
, 19)
. Thus, residual Ca2+ would keep these dehydrogenases active during the poststimulus period, as suggested by NAD(P)H fluorescence measurements (Fig. 7)
. The stimulation of respiration may help to provide energy for clearing the Ca2+ load. Isocitrate dehydrogenase shows stimulation by Ca2+ at higher concentration ranges (2x10-5 to 1x10-4 M; ref 19
) so that activity of this enzyme may change during period PS1.
In summary, our results suggest a highly structured spatiotemporal organization of Ca2+ signals originated by sustained Ca2+ entry through VOCC in adrenal chromaffin cells. High [Ca2+]c domains suitable for triggering exocytosis are generated at the subplasmalemmal region. Domains with a smaller [Ca2+]c increase but more sustained in time (adequate for mobilizing the reserve pool of secretory vesicles; ref 2
) are generated at the core regions of the cytosol and at the nucleus. Mitochondria seem to be essential for shaping these local Ca2+ domains. On the other hand, Ca2+ uptake by mitochondria activates NADH-dehydrogenases, thus tuning up respiration to match the increased local energy needs (16)
. Full relaxation of the [Ca2+]c peak requires clearance of the mitochondrial Ca2+ load through the Na+/Ca2+ exchanger, which proceeds slowly at low micromolar [Ca2+]M levels. This keeps respiration stimulated and core [Ca2+]c discretely high for a long period after stimulation, acting as a kind of memory that may help to secure restoration of the initial conditions.
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 28, 2001. Accepted for publication November 6, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
