Control of Cytosolic Free Calcium by Intracellular Organelles in Bovine ...

Vol. 262, No. 9, Issue of March 25, pp. 4053-4058, 1987 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society of Biological Chemist% Inc.

Control of Cytosolic Free Calcium by Intracellular Organelles in Bovine AdrenalGlomerulosa Cells EFFECTS OF SODIUM AND INOSITOL1,4,5-TRISPHOSPHATE* (Received for publication, August 4, 1986)

Michel F. RossierS, Karl-Heinz Krausep, P. Daniel Lewpll, Alessandro M. Capponi, and Michel B. Vallotton From the Diviswns of Endocrinology and $Infectious Diseases, University Hospital, CH-1211 Geneva 4, Switzerland

The regulation of cytosolic free Ca2+ concentration renal glomerulosa cells is mediated by a transient increase in ([Ca”+],)by intracellular organelles was studied in per[Ca2+lC, due to Ca2+mobilization from intracellular pools (1meabilized bovine adrenal glomerulosa cells. 3). Potassium ion (K+), another potent secretagogue for alTwo compartments,withdistinctcharacteristics, dosterone, exerts its effect by activating voltage-dependent were able to pump Ca2+. 1) A first pool, sensitive to Ca2+ channels(1). rutheniumredand presumably mitochondrial, reAn early event in the stimulation of steroidogenesis by AI1 quired respiratory chain substrates to maintain [Caz+Icaround 700 nM. Ca2+efflux from this compart- is the activation of polyphosphoinositide hydrolysis (3-5) by ment was activated by Na+ (EDao = 5 mM). Inositol a specific phospholipase C, leading to the formation of two 1,4,5-trisphosphate (IPS)had no effect on this pool. 2) second messengers: inositol 1,4,5-trisphosphate (or IP3) (6,7) A second nonmitochondrial pool required ATP to lower and diacylglycerol (8,9). These two products have been dem[Ca2+],to about 200 nM and released Ca2+transiently onstrated to activate two pathways; diacylglycerol acts as a upon addition of IPS. When the two systems were al- natural activator of protein kinase C (8), whichcouldbe lowed to work simultaneously, the nonmitochondrial responsible for the sustained phase of aldosterone production pool regulated [Ca”’], and IPS released Ca” in a concentration-dependent manner (ECao= 0.6 ”). Under (3, lo), while IP, mobilizes Ca2+from intracellular pools (6) these conditions the mitochondria seemed Ca2+ de- and appears responsible for the early onset of aldosterone pleted. Upon repeated stimulations withIPS,a marked production. The exact localization of the IP,-sensitive intracellular Ca2+ attenuation of the response was observed. This phenomenon was due toCa2+sequestration by a nonmito- pool is not yet clearly defined, and the role of the mitochondria chondrialIP3-insensitive pool. Neitherdantrolene in regulating the resting level of [Ca2+Icis still controversial (200 WM) nor 8-(N,N-diethylamino)octyl-3,4,S-trime- (11-14). Isolated mitochondria from rat insulinoma have been thoxybenzoate (10MM)were ableto abolish IP3-induced used to measure Ca2+transport and to investigate the influCa2+release, though both compounds efficiently inhib- ence of the cationic background on Ca2+handling (15). Reited aldosterone production in intact cells stimulated cently, it was demonstrated by 45Ca2+ fluxes analysis that AI1 with angiotensin I1 (10JIM) or K+(12 mM). p-trifluoromethoxyphenylhydrazone, a and carbonyl cyanide These results suggest that in permeabilized adrenal glomerulosa cells: 1)the nonmitochondrial pool is re- proton ionophore acting on mitochondria, mobilize Ca2+from sponsible for buffering [Ca2+],and for releasing Ca2+ different pools in adrenal glomerulosa cells (16) and that[”PI in response toIPS;2) at resting [Ca”+], levels, the IP, binds to a specific receptor present on microsomes in mitochondria appear Ca2+ depleted; 3) when [Ca”’], adrenal cortical cells, hepatocytes, and neutrophils (17, 18). rises above their set point, the mitochondria accumuIt isnow generally accepted that theendoplasmic reticulum late Ca2+as a function of [Na+],; 4) the mitochondria is the site of action for IP3, and a direct action of IP3 on Ca2+ are not involved in the desensitization mechanism of mobilization was demonstrated in various permeabilized cells the response to IP3. such as pancreatic acinar cells (19-21), RINm5F insulinoma cells (22, 23), hepatocytes (24, 25), pituitary GH, cells (26, The steroidogenic response to angiotensin I1 (AII)’ in ad- 27), 3T3 fibroblasts (28), neutrophils (29), parathyroid cells ________~ (30), and platelets (31). Saponin-permeabilized adrenal glo* This work wassupported by Grants 3.914.083 and 3.990.084 from merulosa cells were also used to demonstrate the effect of IP3 the Swiss National Science Foundation, by the Swiss Foundation for Cardiology, and by the Foundation C. and E. de Reuter. The costa of on intracellular pools loaded with 45Ca2+(3). In some cells, a publication of this article were defrayed in part by the payment of marked attenuation of the response was observed after several page charges. This article must therefore be hereby marked “adver- additions of IP, (27, 29, 31) while in some others no such tisement” in accordance with 18U.S.C. Section 1734 solelyto indicate desensitization occurred (23, 24, 30). this fact. Dantrolene, askeletal muscle relaxant (32), has been shown $ To whom correspondence should be addressed Division of Ento suppress partially the spontaneous release of Ca2+ from docrinology, University Hospital, CH-1211 Geneva 4, Switzerland. isolated sarcoplasmic reticulum (33) and to inhibit glucose1Recipient of a Max Cloetta career development award. The abbreviations used are: AII, angiotensin 11; [Ca2+],,cytosolic mediated Ca2+efflux and insulin secretion in rat pancreatic free calcium concentration; IPB, myoinositol 1,4,5-trisphosphate; islets (34). In adrenal glomerulosa cells it was reported that pro- dantrolene inhibits AII-induced, but not K+-stimulated, Ca2+ TMBS, 8-(N,N-diethylamino)octyl-3,4,5-trimethoxybenzoa~ tein kinase C, calcium-sensitive phospholipid-dependent protein ki- efflux and AII-stimulated aldosterone production (35). In the nase; Hepes: 4-(2-hydroxyethyl)-1-piperazineethanesulfonicacid; CCCP, carbonyl cyanide m-chlorophenylhydrazone; BSA, bovine absence of extracellular Ca2+(or in the presence of dihydroserum albumin; EGTA, [ethylene bis(oxyethylenenitrilo)]tetraacetic pyridine Ca2+ antagonists (3611, no response to AI1 was obacid. served in dantrolene-treated cells in terms of either Ca2+ ~

~~

~~~

~~

~

~~

~~~

~~~~~

~~

~

~

4053

4054

Control of [Ca2+1, in Permeabilized Glomerulosa Cells

More than 85% of the electrode response to additions of Ca2+ ina cell-free medium was achieved within 10 s. The cells were incubated at 30 "C in400plofbuffer I containing 0.1 p M rotenone, 1 p~ oligomycin, and an ATP-regenerating system (2 mM Mg-ATP,10 mM phosphocreatine, and 8 units/ml creatine kinase) and supplemented or not with mitochondrialsubstrates (5 mM succinate, 5 mM pyruvate). It was ascertained in a cell-free medium that none of the TMB-8 (8-(N,N-diethylamino)octyl-3,4,5-trimethoxyben-tested agents interfered withthe electrode signal. In addition, these zoate) was first reported to be a nonselective inhibitor of minielectrodes have been shown to be lo6 times more selective for contraction in skeletal and smooth muscle (37).Since then, Ca2+than for M P (41). The variations of [Ca2+]wererecorded continuously. With this method, it is assumed that the [Ca2+]measTMB-8 has been widely described as a n "intracellular Ca2+ ured at steady state is equivalent to the cytosolic free calcium conantagonist" and used in several preparations as a probe for centration, [Ca2+],. ea2+-mediatedprocesses (see for example Ref. 38).Recently, Functional Tests-For static incubations, the cellswere resusit was demonstrated in adrenalglomerulosa cells that TMB- pended in medium 199containing 0.2% BSA at a concentration of 5 8 is without effect on either AII-stimulated 45Ca2+efflux or X 106 cells/ml.Theywere first preincubatedfor 15 min at room the response of permeabilized cells to IPSbut affects aldoster- temperature with or without the tested inhibitors, dantrolene (200 p ~ and ) TMB-8 (200 pM), and thereafter incubated under constant one production induced by both AI1 and K+ and by phorbol agitation for 45 min at 37 "C with 95% C02, 5% O2 in the presence esters (39). of various stimulatorsof steroidogenesis. At the end of the incubation In the present study we have investigated therole of intra- period, the cells were sedimented by centrifugation, andthe aldostecellularorganelles and their interactions in buffering the rone content of the media was measured by direct radioimmunoassay [Ca"], in permeabilized adrenal glomerulosa cells, and we as previously described(42). Analysis of Data-Results are expressed as mean f standard error have characterized the action of IP3 on Ca2+ mobilization of the mean,unlessotherwise stated. Statistical significancewas from intracellular pools. This system was also employed to assessed by the Student's t test. show the lack of a specific inhibitory effect of dantrolene and TMB-8 on the IP3-induced Ca2+release. RESULTS efflux or aldosterone production; it was concluded that AI1 acts by mobilizing 50% of the necessary Ca2+from a dantrolene-sensitive intracellular pool (2, 35). In addition, in saponin-permeabilized glomerulosa cells an inhibitory effect of Ca2+release and ea2+ loading dantrolene on both Ips-induced into intracellularpools has been found (3).

Two Distinct Systems Are Able to Control [Ca2+],-After permeabilization with digitonin, glomerulosa cellswere resusCollagenase type I, digitonin, Hepes, Mg-ATP, phosphocreatine, pended in an "intracellular medium" (buffer I) mimicking the creatine kinase, succinate, pyruvate, oligomycin, rotenone, antimycin, ionic composition of the cytosol. When permeabilized cells CCCP, TMB-8, and BSAwere obtained from Sigma. Medium 199 was purchased from Seromed and EGTA from Fluka. Quin 2 (free were added to the medium containingmitochondrialsubacid) was purchased from Behring Diagnostics, ruthenium red from strates, in the absence was of ATP, a decrease of[ea"'] Aldrich, and sodium dantrolene from Norwich Eaton Pharmaceutical observed, indicating a pumping of Ca" by cell organelles (Fig. co. lA).After this initial phase, thecells maintained a set point Preparation of Isolated Bovine Adrenal Glomerulosa Celk-Isolated around 700 nM [ea'+], even when variations were induced by bovine adrenal glomerulosa cells wereprepared by collagenase digestion and mechanical dispersion as previously described(1).After two additions of Ca2+or of a chelator of Ca2+(quin 2 free acid). Addition of IP3 under these conditions was without effect washes with medium 199, the isolated cells were permeabilized by two different techniques. (Fig. L4, inset). T h e pool responsiblefor this control appeared Cell Permeabilization-In most experiments, the cells were per- to be mitochondrial, because ruthenium red, a specific inhibmeabilized with digitonin as follows. The cells were resuspendedat a itor of mitochondrial Ca2+ uptake, abolished this set point concentration of about 5 X lo6 cells/ml (2 mgof protein/ml) in a and Ca2+was immediately released into the medium. If the buffer containing 20 mM Hepes, 137 mM NaCl, 3 mM KCl, 4 mM MgC12,10.6 mM D-glucose, 5 mM Na2ATP,1 mM EGTA, 0.05% BSA medium was then supplemented with an ATP regenerating at pH 7.2. The cells were then incubated for 10 min at 37 "C in the system, the cells regained the ability to pumpCa", bringing presence of 15 p~ digitonin added from a freshly prepared solution. it down to an even lower concentration of about 200 nM. A The permeabilization was assessed with trypan blue; more than 70% new steady statewas reached, which was sensitive toIPS; this of the cells stained with the dye, while this ratio amounted to less agent released Ca2+transiently froma nonmitochondrial pool. than 10% in control nontreated cells. The cellswere then washed A second addition of IP3 1 min later had little effect and twice at 4 'C in a second buffer containing 20 mM Hepes, 5 mM, showed that the increase in [Ca"] was not due t o Ca2+ NaCl, 115 mM KCl, 5 mM NaHC03, 1 mM KH&0' 4, 1 mMMgC12, 0.05%BSA, at pH 7.2 (buffer I). In the experiments performed inthe contamination. Addition of vanadate, a known inhibitor of The two absence of Na+ andM P , a similar bufferwas used lackingthese two ATPases, induced an irreversible release of ea2+. cations. Thecells were kept on iceat a concentration of 8 X lo6 cells/ pools were shown tobe vesicular, becausethey released rapidly 50 pl. all their Ca2+content into the medium in the presence of 5 Alternatively, inthe experiments dealing with the effect of dantro- p~ ionomycin (data not shown). lene and TMB-8, the cells were rendered leaky by using a high voltage When the medium contained an ATP-regenerating system electric field discharge (40). In this case, the cells were resuspended in a buffer containing 20 mM Hepes, 137 mM NaCl, 3 mM KCl, 5.6 from the beginning of the experiment (Fig. l B ) , the cells pumped Ca2+ immediatelydown to 200 nM. IP3induced a net mM D-glucose, and 0.05% BSA, pH 7.2, and incubated for 15 min at 37 'C in the presence of 1 mM EGTA. The cells were then washed release of Ca", which was completely reversedwithin 10 min. a medium Inhibition of ATPases with vanadate led the system toa new twicein the same buffer (without EGTA) and once in containing 250 mM sucrose, 5 mM Hepes, at pH 7.2, and then set point of 450 nM, which was insensitive to IP3 and was resuspended in this nonionic buffer at a concentration of 8 X 10' cells/100 p1 and kept on ice until permeabilization. The cell suspen- abolished by ruthenium red. In order to investigate involvement the of the mitochondria sion was placed in a small Plexiglas chamber(400 pl) and repeatedly exposed to an intense electric field (2 kV/cm) created between two in maintaining a low [Ca2+],(at 200 nM), we have tested two electrodes connectedto a homemade condenser. Each discharge lastedmitochondrial inhibitors of Ca2+ uptake; 1) ruthenium red, 100 p s . The cells were then immediately used for experiments. The known as a specific inhibitor of Ca2+influx; and 2) Na+, which results obtained with both permeabilization techniques, for the set activates Ca2+efflux via a Na+/Ca2+exchange in mitochondria point values and for the magnitude of the response to IPS,were of many excitable tissues (12, 13).These two agents induced identical (n = 15 experiments). from the mitochondria (Fig. 2) only when Measurement of [Ca2+]-Ca2+-selective minielectrodes were pre- a net loss of eaz+ pared and calibrated as previously described by Prentki et al. (15). the nonmitochondrialpool was partially or totally inactivated MATERIALS AND METHODS

Control of [Ca2+lein Permeabitized Glomerulosa Cells

4055

Caz+ accumulated in the mitochondria at micromolar Ca2+ level to be released. Indeed, this small effect elicited by Na+ was completely abolished when the cells pumped Ca2+ in the fl I I 13 presence of a physiological Na+ concentration (5 mM) (not shown). The Effect of Na+ and Mg2+ on the Mitochondrial Set Point-The effect of two important cytosolic cations, Na+ and M$+, onthehandling ofCa2' by mitochondria was investigated. All these experiments were performed in the 1 presence of 1 mM vanadate in order to prevent Ca2+pumping 0.7 by the nonmitochondrialpool. In the absence of M$+ and at low Na+ concentration (1 mM, due to contamination by sodium vanadate), the mitochondria were able to pump Ca2+ down t o a set pointof 560 f 60 nM (n= 6), which was reached 0.2 within 10 min and was maintained against pulses of Ca2+(Fig. 3). Successive increments of [Na'] (2.5 mM) led t o a progres0.1 sive rise of the mitochondrial set point atomaximal value of 1260 f 70 nM. At this point, the mitochondria were still able a set point. This FIG. 1. The two systems capable of regulating [Cas+], in to counteracta pulse of Ca2+ and to maintain permeabilized adrenal glomerulosa cells. Variations of [Ca2+] set point was considerably increased by the addition of 1 mM were measured with a Ca2+-selective minielectrode,at 30 "C, in 0.4 Mg2+ (Fig. 3). The selectivity of the electrode for Ca2+ was ml of buffer I as described under "Materials and Methods." Caz+ confirmed by the observation that 1mM M e did not interfere pumping started with addition of digitonin-permeabilized cells(8 X with the electrode signal in a cell-free medium where Ca2+ lo6 cells). Where indicated, 2 p M CaCIZ,5 p M quin 2 free acid, 1 pM EGTA at M. ruthenium red, 5 pM IPS, or 2 mM sodium vanadate were added to wasbufferedwith When recorded in the presence of 1 mM M e , the whole the incubation medium. An ATP-regenerating system containing 2 mM Mg-ATP, 10mM phosphocreatine,and 8 units/ml creatine kinase [Ca"] tracing was shifted towardhigher values. The effect of was added after 35 min in experiment A and was already present in Na+ on mitochondrial Ca2+ handling is best illustrated Fig. in the medium at the beginning of experiment B. Inset, the lack of effect 4,in which the achieved steady-state [Ca'+], is representedas of IP, on the mitochondrial set point. The abscissa and ordinate units a function of increasing Na' concentration. In theabsence of are the same as in the main figure. M$+, the effect of Na+ was maximal a t 25 mM, with ECso of about 5 mM (n = 5). In another seriesof experiments, various important agents andmediators were testedfortheirabilityto affect the RR = ruthenium red mitochondrial steady state in the absenceof Na+ and M$+; neither cyclic AMP (1mM), cyclic GMP (1mM), nor phorbol Na+ 15 mM myristate acetate (200 nM) had any effect on [Ca"],. The Effect of IP3on Ca2+Release from the Nonmitochondrial Pool-The concentration dependence of the IPSeffect on Ca2+ release from the IPS-sensitive (nonmitochondrial) pool was investigated in detailas depicted in Fig. 5. The height of the Ca2+ peakwas compared withconsecutive additions of known quantities of Ca2+ used as internal standards (Fig. 5, inset). Because the system displayed a marked desensitization to a second challenge with IP3 (see below), the response to each IP3concentration was measured in a separate cell preparation. Computer analysis of the concentration-responsecurve indicated a maximal release of Ca2+ of 1.6 nmol/8 X lo6 cells,

A

ka*4,

b'Y, ce11s 0

10

20

30

40

pN

+ mltnchnndrial substrates

cnn Na'l

Time (min)

FIG. 2. The effect of ruthenium red and Na+ on mitochondrial Caa+ uptake at various Ca" steady states. Digitoninas described in the legend of Fig. permeabilized cells were incubated 1, in the presence of an ATP-regenerating system and of various concentrations of vanadate. No Na+ was present in the medium. Ruthenium red (RR, 2 p M ) and NaCl(15 mM) were addedas indicated.

t f Gait

Mg++

1pM

Id

These results are representative of three separate experiments. by submaximal or maximal concentrations of vanadate. In the absenceof vanadate, the nonmitochondrialpool was fully FIG. 3. The effect of Na+ and Mga+on the mitochondrial set active and itwas not possible t o release Ca2+ from the mitochondria, in spite of the lower [Ca2+],. Thus, the mitochon- point. Digitonin-permeabilized cells were incubated as described in the legend of Fig. 1, in the presence of 1 mM vanadate, in order to drial pool appeared Ca2+ depleted. Asmall effect was elicited inhibit the nonmitochondrialpool.NaClwasadded as indicated. by Na' in the absence of vanadate (Fig. 2). This suggested Pumping of Ca2+in the presence of 1 mM MgClZis shown in the that a minimal amount of Na+ was necessary t o allow the dotted trace.

Control of [Ca2+Icin Permeabilized Glomerulosa Cells

4056

0.2 EL50

=

5 mM Na' Na+

I

1

1

I

I

(n=5)

0

20

10

30

[Na'] mM

FIG.4. The concentration-response curve of the effect of Na+ on the mitochondrial set point. The results are the mean f S.E. of 5 separate experiments performed under conditions similar to those described in the legend of Fig. 3.

r

IP3 1 fl

:&,

Os4-

T

0.3-

0.2

1 pfl 10 pM 0.5 pM

rutheniumred

antimycin CCCP

-

I

I

0

10

I

I

20

30

C

I

40

Time bin)

I

I

0 0

I

I

1

2

I

3

I

I

4

5

FIG.6. The desensitization of IPs-induced Ca" release. Digitonin-permeabilized cells were incubated under the same conditions as those described in the legend of Fig. 5. Where indicated, 10 p~ CaClz or 1 p~ IPS were added to the medium. In panel C, a mixture containing 1 pM ruthenium red, 10 pM antimycin, 0.5 phi CCCP was added 2 min before the first addition of IP3. The tracings are representative of 3-4 similar experiments.

'

(IP3l Yfl

FIG.5. The concentrationdependence of the effect of IPSon

Ca2+mobilization from the nonmitochondrial pool. Digitonin-

-E

-

o

n

a

*

0 inhlbitor

+

+

dantralene 200 TMB-E 200

vfl

~ ~ 0 . 0 5 **D