Uncoupler-stimulated release of Ca2+ from Ehrlich ascites tumor cell mitochondria

ARCHIVES

OF BIOCHEMISTRY

AND

BIOPHYSICS

Vol. 240, No. 2, August 1, pp. 723-733, 1985

Uncoupler-Stimulated Release of Ca2+ from Ehrlich Ascites Tumor Cell Mitochondria’ G. FISKUM*,’ *Department

of Biochemistry, and TEdward

AND

R. S. COCKRELLt

George Washington University School of Medicine, Washington, A. Doisy Department of Biochemistry, St. Louis University School of Medicine, St. Louis, Missouri 6310,$

Received November

D.C. 20037,

5, 1984, and in revised form March 27, 1985

Ruthenium red-insensitive, uncoupler-stimulated release of Ca2+ from Ehrlich ascites tumor cell mitochondria is much slower than from rat liver mitochondria under comparable conditions. In the presence of Pi and at moderate or high Ca2+ loads, ruthenium red-insensitive Ca 21 e&ix elicited with uncoupler is approximately 20 times more rapid for rat liver than Ehrlich cell mitochondria. This is attributed to resistance of tumor mitochondria to damage by Ca2+ due to a high level of endogenous Mg 2+ that also attenuates Ca2+ efflux. Calcium release from rat liver and tumor mitochondria is inhibited by exogenous Mg2+. This applies to (a) ruthenium red-insensi.tive spontaneous Ca 2+ efflux associated with Ca2+ uptake and uncoupling, and (b) ruthenium red-insensitive Ca2+ release stimulated by uncoupling agent. The endogenous Mg ‘+ level of Ehrlich tumor mitochondria is approximately three times that of rat liver mitochondria. Endogenous Ca2+ is also much greater (six fold) in Ehrlich tumor mitochondria compared to rat liver. Despite the quantitative difference in endogenous Mg2+, the properties of internal Mg2+ are much the same for rat liver and Ehrlich cell mitochondria. Ehrlich ascites tumor mitochondria exhibit slow, metabolically dependent Mg2+ release and rapid limited release of Mg2+ during Ca2+ uptake. Both have been observed with rat liver and other types of mitochondria. The proportions of apparently “bound” and “free” Mg2+ (inferred from release by the ionophore, A23187) do not differ significantly between tumor and liver mitochondria. Thus, the endogenous Mg2+ of tumor mitochondria has no unusual features but is simply elevated substantially. Ruthenium red-insensitive Ca2+ efflux, when expressed as a function of the intramitochondrial Ca2+/Mg2+ ratio, is quite similar for tumor and rat liver. It is proposed, therefore, that endogenous Mg2’ is a major regulatory factor responsible for differences in the sensitivity to damage by Ca2+ and Ca2+ release by Ehrlich ascites tumor mitochondria compared to mitochondria from normal tissues. Q 1985 Academic

Press, Inc.

Mitochondria from certain tumor cells, notably those of Ehrlich ascites cells, differ in the unusual facility with which they accumulate and retain Ca2+ without ‘This research was supported by funds from USPHS Grants CA 11766 to R.S.C. and CA 32946 to G.F. ’ To whom correspondence should be addressed.

uncoupling [(l-4), reviewed in Refs. (5, S)]. Passive release of Ca2+ from Ehrlich cell mitochondria also has certain unusual features in comparison to mitochondria from normal tissues such as rat liver (7). Calcium release that is sensitive to ruthenium red inhibition (presumably simple reversal of the Ca2’ uptake pathway) is virtually the same for Ehrlich cell and 723

0003-9861/85 $3.00 Copyright All rights

0 1985 by Academic Press. Inc. of reproduction in any form reserved.

724

FISKUM

AND

COCKRELL

EXPERIMENTAL PROCEDURES rat liver mitochondria (7). However, uncoupler-stimulated release via a ruthePreparation ofmitochondria. Ehrlich Lettre diploid nium red-insensitive pathway is approxiascites tumor cells were kindly provided by Dr. mately 15 times more rapid for rat liver Charles Wenner, Roswell Park Memorial Cancer than tumor mitochondria [e.g., 108 neq Institute (Buffalo, N. Y.). Cells were maintained in Swiss albino mice (ICR strain; min-’ mg-’ compared to 8 neq mind1 mg intraperitoneally protein,-’ Ref. (7)]. These rates can be the Hilltop Laboratories, Scottsdale, Pa.). Routine cell same but only if tumor mitochondria are transfers were performed at intervals of 6-10 days. loaded with three to four times the Ehrlich cells were isolated by a modification of the method of Pedersen and Morris (15) as detailed amount of Ca2+ as rat liver mitochondria elsewhere (8) consisting of three washes with 144 (7). No explanation for the resistance of mM NaCl plus 10 mM Na-Mops’ (pH 7.4). tumor mitochondria to uncoupling by Ca2+ Mitochondria from Ehrlich ascites tumor cells nor the attenuation of Ca2+ efflux has were prepared by a modification of the Nagarse been provided. It is noteworthy that the method used for L1210 ascitic leukemia cell mitoresistance to uncoupling and efflux rates chondria (2) as described previously (8). This profor either ruthenium red-sensitive or ru- cedure yields well-coupled mitochondria, i.e., uncouthenium red-insensitive “pathways” in vi- pling agent (FCCP) stimulates respiration lo-fold tro are virtually the same as in situ, i.e., with pyruvate plus malate or approximately &fold with glutamate plus malate, @-hydroxybutyrate, or compared to mitochondria within digisuccinate (plus rotenone) with medium containing tonin-treated Ehrlich ascites cells (8). This and 1.0 125 mM choline-Cl, 5 mM Tris-glycylglycine, indicates neither is significantly altered mg BSA/ml (pH 7.0). Release of respiratory control in the course of isolating tumor mitochonby ADP is almost negligible (RCI values of 1.1 to dria. 1.2) for these preparations. The slow state-3 rate One explanation for the unusually slow with ADP is most likely due to inhibition of phosruthenium red-insensitive Ca2+ efflux in phorylation or ATP/ANP exchange by a high endogenous Ca*+ [see Experimental Results and Refs. vitro and in situ is quantitative or quali(1% 1711. tative differences in endogenous regulaAssay techniques. Calcium fluxes were determined tory factors retained by Ehrlich ascites spectrophotometrically by the murexide indicator cell mitochondria in vitro. Exogenous Mg2+ opposes the activating influence of Ca2’ technique (18). Absorbancy changes at 540-510 nm were measured with an Aminco DW-2 dual-waveupon processes associated with uncoupling length spectrophotometer. Calibrations with Ca2+ (9) and/or Ca2+ release (10-14). Since were corrected for binding (chelation by succinate Ehrlich cell mitochondrial Ca2+ accumuor phosphate, metabolically independent binding by lation is relatively insensitive to Mg2+ mitochondria, etc.) Controls for light scattering were compared to rat liver (8), this differential routinely performed by omitting murexide. The basic suspending medium contained 250 mM Mgzf sensitivity has been explored in the present studies. The endogenous Mg2+ of sucrose, 10 mM Tris-Mops (pH 7.2), 30 pM murexide, Ehrlich ascites tumor mitochondria is 2.5 and oligomycin (6 pg/ml). Where indicated, 2.5 mM was included. For respiration-detimes that of rat liver (129 neq versus 49 Tris-phosphate was empendent Ca”+ uptake, 5 mM Tris-succinate neq/mg protein). This can explain why ployed and air was constantly blown over the susruthenium red-insensitive Ca2+ efflux difpension. It should be mentioned that 95% oxygen fers from that of rat liver mitochondria. was found to inhibit maximum Ca2+ uptake by rat If rates are expressed as functions of liver but not tumor mitochondria (data not preinternal Cazf/Mg2+ quotients, the differsented). In all measurements, the temperature was ences between tumor and rat liver largely maintained at 25°C and mitochondria were pretreated with rotenone (0.2 pg/mg protein). The cudisappear. It is suggested the endogenous Mg2+ also protects tumor mitochondria from spontaneous release of accumulated 3Abbreviations used: BSA, bovine serum albuCa2+ and uncoupling. The characteristics min; EGTA, ethylene glycol bis(&aminoethyl ether)of the endogenous Mg2+ of tumor mitoN,N’-tetraacetic acid; FCCP, earbonylcyanide-pchondria have been examined and found trifluoromethoxyphenylhydrazone; Hepes, 4-(2-hyto be the same as endogenous Mg2+ of rat droxyethyl)-1-piperazineethanesulfonic acid; Mops, liver according to several criteria. 4-morpholinepropanesulfonic acid.

CALCIUM

RELEASE

FROM

vette contents were mixed continuously by means of a magnetic stirring device that gave mixing times of less than 2.0 s. Magnesium fluxes were measured spectrophotometrically with Erichrome Blue SE (Plasmacorinth B) by dual-wavelength measurements at 575-540 nm (19). Magnesium was calibrated and light-scattering controls were performed as for Ca’+. The reaction medium was identical except 30 pM Plasmacorinth B was substituted for murexide. Oxygen utilkzation was measured polarographieally with a Clarke (membrane) electrode (20). Atomic absorption measurements of Ca2+ and Mg2+ were performled with a Perkin-Elmer atomic absorption spectrometer (Model 303) on 10% nitric acid extracts of mitochondria. Mitochondrial protein contents were determined by the method1 of Lowry et al. (21), with BSA as the standard. Materials. M:urexide, Plasmacorinth B, oligomyein, antimyein, rotenone, oxidizable substrates (pyruvate, malate, glutamate, @-hydroxybutyrate, and succinate), and BSA (Fraction V) were purchased from Sigma Chemical Company, (St. Louis, MO.). Nagarse was obtained from the Enzyme Development Corporation, (New York, N. Y.). Ruthenium red was purchased from Sigma and recrystallized by the method of Luft (22). The concentration of stock solutions was determined spectrophotometrically using a millimolar extinction coefficient of 68.5 at 532 nm (22). The ionophore A23187 was contributed by Eli Lilly, and FCCP was provided by Dr. P. G. Heytler, E. I. DuPont de Nemours and Company, (Wilmington, Del.). RESULTS

The resistance of Ehrlich ascites tumor cell mitochondria to uncoupling by Ca2+ is evident in measurements of respiratory control, maximum Ca2+ uptake, and spontaneous release of accumulated Ca2+. As shown in Fig. 1, a loss of respiratory control (uncoupling) in a sucrose medium is evident
TUMOR

MITOCHONDRIA

b.RatLlver 1111 nw d+/m9

725

B.TvnOr 334 ne4 d+/m9

FIG. 1. Release of respiratory control by Ca*+ in rat liver and tumor mitochondria. The experimental conditions are provided under Experimental Procedures. In addition to the basic medium, 2.5 mM Trisphosphate was included. The quantities of Ca2+ added are indicated as well as respiration rates in parentheses (ngm-atom 0 min’mg-i) after each Ca” addition.

for rat liver mitochondria (six preparations) are 0.68 peq Ca2+/mg (no additions), 1.22 peq Ca2+/mg (plus 2 mM Mg2+), and 1.74 peq Ca2+/mg (plus Mg2+ and 1 IrIM ATP). Ehrlich cell mitochondria are stimulated much less [about 25%, Ref. (S)] by Mg2+, i.e., they accumulate about the same Ca2+ as rat liver supplemented with Mg2+, whereas Ca2+ uptake is enhanced by adenine nucleotides by about the same factor [1.5 fold, Ref. (S)]. The influence of these factors upon maximum Ca2+ uptake is qualitatively the same as their effect upon retention of coupling, i.e., respiratory control. As shown in Fig. 2, the basal respiration of tumor mitochondria increases sharply at approximately 1.1 peq Ca2+/mg (for this preparation). Magnesium has no appreciable effect whereas Mg2+ plus ATP extends the range over which respiratory control is maintained (to >2.2 peq Ca2+/ mg). The respiratory control of rat liver mitochondria is similar to maximum Ca2+ uptake in that both exogenous Mg2+ and ATP suppress uncoupling by Ca2+ and Mg 2+ has a more pronou need effect than with tumor mitochondria (data not shown). Uncoupling and cessation of net Ca2’ uptake are also associated with spontaneous Ca2+ release. In the case of rat liver

726

FISKUM

0

AND

I

I

I

I

200

400

600

800

COCKRELL

I

I

I

I

I

I

Co** Added

b-teqlmg pmt.)

FIG. 2. Retention of coupling following Ca2+ accumulation by tumor mitochondria. additions (200 pM) were made as in Fig. 1B to rotenone-treated tumor mitochondria plus oligomycin (2 pg/mg), MgClz (2 mM) and Tris-ATP (1 mu) were added initially

(Fig. 3), Ca2+ loss is prevented by exogenous Mg2+ which therefore promotes further Ca2+ uptake. The release of Ca2+ under these conditions is largely insensitive to ruthenium red whereas exogenous Mg2+ blocks release completely (Fig. 3). Spontaneous Ca2+ release from tumor mitochondria only occurs at much greater Ca’+ loads (typically about 1.3 peq/mg). Spontaneous Ca2+ release is complex since it can be limited by the Ca2+ permeability, coupling, proton permeability, etc. These complications are minimized by including uncoupling agent as in previous studies (7). Increasing H+ conductance with protonophores stimulates Ca2+ release. The rate of ruthenium red-sensitive Ca2+ release presumably reflects operation of the Ca2+ uniport in reverse. Ruthenium redinsensitive Ca2+ efflux (which dominates spontaneous release, Fig. 3) is most likely determined by the Ca2+ permeability, although the nature of this “pathway” is controversial. Previous comparisons of Ca2+ efflux (7) were extended to take into account not only the Ca2+ dependence but also possible influences of endogenous divalent cations in view of results with Mg”. Phosphate was examined since it significantly affects

I

1000 1200 1400 1600 1800 2000 2200

Multiple Ca2+ (0.6 pg/mg), as indicated.

Ca2+ efflux (12-14, 23) and the release of other ions including Mg’+ (24,25). Calcium influx and efflux values as functions of added, i.e., accumulated Ca2+, are summarized in Figs. 4 A-D with Pi present (note the two- to fourfold difference in abscissas for tumor and liver data plots). The Ca2+ influx rates are more rapid for tumor mitochondria (Fig. 4A); however, this is presumably due to the higher maximal rate of respiration (see insert of Fig. 4A in which respiration rates are expressed as the percentage of the maximum rate with uncoupling agent). At saturating extramitochondrial Ca” concentrations, the rate of H+ ejection limits Ca2+ influx rather than the Ca2+ uniport activity (26). The rates of Ca2+ release with uncoupling agent alone are approximately 1.5 to 3 times more rapid for rat liver than tumor mitochondria (Fig. 4B; 50 and 30 neq min-’ mg-’ at 100 neq Ca2+/ mg or 220 and 60 neq min-’ mg-’ at 400 neq Ca2+/mg). The corresponding rates for ruthenium red-sensitive and -insensitive components of the total Ca2+ efflux were determined by difference. They indicate that the more rapid release from liver mitochondria can be almost entirely accounted for by much greater rates of

CALCIUM

RELEASE

FROM

Co’+ 1

a+

+MIJ T AAs,0-S,0’$()3 1

b+ 150secn FIG. 3. The influence of Mg2+ on Ca2+ uptake and spontaneous release. The conditions were the same as described in Fig. 1A except that 30 pM murexide was included. Each Ca*’ addition was 87 neq/mg protein to rat liver mitochondria (5.0 mg prot/ml) plus MgCl* (10 IIIM) and ruthenium red (2 pM) as indicated. A change in absorbance of 0.03 equals a change in total [Ca”] of 325 pM.

ruthenium red-insensitive Ca2+ efflux. Ruthenium red-sensitive Ca2+ efflux is the same (30 neq min-’ mg-’ for both at 100 neq Ca2+/mg) or comparable (80 and 53 neq min-l mg-l at 400 neq Ca’+/mg) for liver and tumor mitochondria. Ruthenium red-insensitive Ca2+ differs substantially (18 and 1 neq min-’ mg-’ at 100 neq Ca2+/ mg and 140 and 7 neq min-’ mg-’ at 400 neq Ca2+/mg). Therefore, the total Ca2+ efflux from liver mitochondria is moderately faster than from tumor mitochondria; however, this is due to a nearly 20-

TUMOR

MITOCHONDRIA

727

fold more rapid rate of ruthenium redinsensitive Ca2+ release. The much slower ruthenium red-insensitive Ca2+ efflux rates for tumor mitochondria evidently are not due to lower matrix free Ca2+ concentrations since the efflux rates with A23187 (Fig. 4D) are either comparable (180 and 340 neq min-’ mg-’ at 100 neq Ca2+/mg) or somewhat greater than (750 and 575 neq min-’ mg-’ at 400 neq Ca2+/mg) those for rat liver mitochondria. Calcium fluxes were measured without added Pi for different Ca2+ loads as in the experiments of Figs. 4A-D. In the absence of exogenous Pi, liver mitochondria accumulated only 50-60 neq Ca2+/mg protein. Flux values from these experiments for a Ca2+ load of 50 neq/mg are compiled in Table I. Calcium influx was slower in the absence of Pi for both tumor and liver mitochondria, yet still more rapid for tumor mitochondria. Efflux rates were much greater for all conditions. This is presumably due to a higher matrix free [Ca”‘] in the absence of Pi. This interpretation is based upon the much more rapid A23187-stimulated Ca*+ release, especially for rat liver (Table I). Similarly, ruthenium red-sensitive Ca2+ efflux is about five times more rapid for liver than tumor. Thus, although the rate of ruthenium red-insensitive Ca2+ efflux is ninefold greater in rat liver than tumor mitochondria, this very likely involves differences in matrix free [Ca2+] as well as Ca2+ permeability. Results without added Pi are more difficult to interpret for this reason. The difference between the ruthenium red-insensitive Ca2+ efflux of tumor and liver mitochondria appears to be greater in the presence of Pi. However, this conclusion must be qualified due to uncertainties in the interpretation of data obtained without exogenous Pi. If tumor mitochondria contain a higher endogenous Pi, this could explain the apparently higher matrix free [Ca”‘] of liver mitochondria; however, this was not measured. Nevertheless, in the presence of Pi, the major difference in Ca2+ release is the sluggish ruthenium red-insensitive Ca2’ efflux from tumor mitochondria. This cannot be attributed to a lower internal free

728

FISKUM

AND

COCKRELL

Rot Liver

Cap’ Added (n-eq/mg)

Ca2’ Added h-eq/mg)

FIG. 4. Calcium influx and efflux in the presence of Pi as functions of the added Ca”. The levels of total Ca2+ added (and accumulated since uptake was complete) are indicated on the figures. In (A), respiration-driven Ca2+ uptake rates (solid lines) are compared to the rates of oxygen consumption (dashed lines) measured separately for single additions of Ca*+. The latter are expressed as the percentage of the maximum respiration rate with uncoupling agent, FCCP. The basic media are described under Experimental Procedures. After completion of Ca2’ uptake, Ca” red (4 pM) was added efllux was elicited with FCCP (4 FM; B). When included, ruthenium immediately before FCCP (C) and A23187 (2 pg/ml) was added simultaneously with FCCP (D). The final mitochondrial protein concentrations were typically 2.5-3.5 mg/ml for tumor mitochondria and 4.2-5.2 mg/ml for liver mitochondria. Each symbol represents a different mitochondrial preparation.

[Ca”‘] since the release by A23187 was either comparable or faster for tumor compared to rat liver mitochondria (Fig. 4D). Calcium release under a variety of conditions is inhibited by Mg2+ and, in the present studies, spontaneous Ca2+ efflux associated with uncoupling was inhibited by exogenous Mg 2+ (Fig. 3). Whether or not this also applied to ruthenium redsensitive and -insensitive Ca2+ release with uncoupling agent was determined next. Net Ca2’ uptake was adjusted so that ruthenium red-insensitive Ca2+ efflux was the same for liver and tumor mitochondria, i.e., the Ca2+ load was four times as great for tumor (see Fig. 4C). Under these conditions, uncoupler-stimulated ruthenium red-sensitive Ca2+ release is comparable and stimulated somewhat by Mg2+

in liver but inhibited in tumor mitochondria (Table II). This difference cannot be explained at this time. Ruthenium redby insensitive Ca2+ release is inhibited Mg2+ to the same extent (approximately 70%) in liver and tumor. Because inhibition by Mg2+ was immediate (preincubation was not required) and the Mg2+ permeability is minimal for rat liver and tumor mitochondria (described later), exogenous Mg2+ most likely acts either by binding to the external inner membrane or by preventing loss of endogenous Mg2+. Since release of endogenous Mg2+ and adenine nucleotides is associated with Ca2+ cycling and uncoupling (14), the latter seemed more likely. Therefore, the contents and properties of endogenous Mg2+ (and Ca2+) were compared for liver and tumor mitochondria.

CALCIUM

RELEASE

FROM

TUMOR

729

MITOCHONDRIA

+Ruthetium Red,FT c 146

Ca2+ Added (n-eq/mg)

Caz+ Added (n-eq/mg)

FIG. I-Continued

The results of Fig. 5 illustrate certain features of the endogenous Mg” of tumor mitochondria compared to rat liver. A slow Mg2+ loss occurs which is “metabolically dependent” (24) since it is inhibited by antimycin. This Mg2+ release is stimulated by phosphate (not shown) as demonstrated previously with heart mitochondria (24). Representative rates for this Mg2+ efflux from tumor mitochondria were 3.1 (-Pi) and 4.3 neq Mg2+ min-’ mg protein-’ (+Pi). The corresponding values for rat liver were 0.5 and 1.6 neq min-’ mg protein-‘, respectively. Therefore, Mg2+ loss by this process, although slow, was actually more rapid for tumor mitochondria. This is probably due to the higher endogenous Mg2+ in tumor mitochondria, as will be discussed. Magnesium loss by this process was too slow to significantly affect the internal Mg2+ level during Ca2+ transport measurements. However, net efflux of Mg2+ accompanying Ca2+ uptake by tumor mitochondria (Fig. 5) or rat liver (results not shown) was significant. The extent of this Mg2+ release

was limited and only a small fraction of the Ca2’ accumulated (approximately 5%). Finally, tumor mitochondria contain a large pool of presumably “free” Mg2+, i.e., released by the ionophore A23187 (Fig. 5)

TABLE

I

COMPARISONOFC~~+INFLUXANDEFFLUX BETWEEN TUMOR AND LIVER MITOCHONDRIA Ca2+ Influx or Efflux Rate (neq Caa+ min-’ mg protein-‘)

Influx R.R. sensitive efflux R.R. insensitive efflux A23187-stimulated efflux

Tumor

Rat liver

540 113 7 230

340 540 60 875

Note. The conditions were identical to those described in Fig. 4 except that Tris-phosphate was omitted from the reaction medium. The Ca’+ load for efflux measurements was 50 neq Ca’+/mg protein (R.R., ruthenium red).

730

FISKUM TABLE

AND

II

THE INFLUENCE OF Mg2+ ON Ca*+ EFFLUX FROM TUMOR AND RAT LIVER MITOCHONDRIA Ca2+ efflux rate (neq Ca2+ min-’ mg-‘) Tumor

Rat liver

R.R. sensitive R.R. insensitive

-Mg2

+Mgs+

60 119

75 28

-Mg*+ 89 125

+Mg’+ 60 45

Note. Mitochondria were loaded with Ca” (359 neq/mg protein for liver and 1113 neq/mg protein for tumor mitochondria) as depicted in Fig. 3. Calcium release was initiated with uncoupling agent (2 PM FCCP) in the absence and presence of MgCla (10 mM added initially) and ruthenium red (2 PM added immediately before uncoupler). The ruthenium redsensitive efflux was determined by subtracting the rate in the presence of ruthenium red from the efflux rate with FCCP only.

and a pool of “bound” Mg2+ not released by A23187. These fractions of Mg2+ and Ca2+ are compared to those of rat liver in Table III. The total endogenous Mg2+ of tumor mitochondria is 2.5 times that of rat liver. The fractions that are “bound and “free” according to release by A23187 are virtually the same for liver and tumor (about 60% free Mg2+). The very high endogenous Ca2+ of tumor mitochondria (6 times that of rat liver, Table III) is not surprising in view of the very slow release of Ca2+ [Ref. (7) and Fig. 41 and the unusual ability of tumor mitochondria to retain accumulated Ca2’ (3, 4). The release of Mg2+ during Ca2+ uptake (Fig. 5) is compared for rat liver and tumor mitochondria in Fig. 6. The Mg2+ released per Ca2+ accumulated is apparently greater for tumor mitochondria than rat liver (Fig. 6). However, the slopes (approximately 1 neq Mgzc released per 20 neq Ca2+ accumulated) are very similar. If the endogenous Ca2+ levels are included, the curve for tumor mitochondria is shifted toward greater Ca2+ loads and the two curves are nearly superimposable. For example, at a total internal Ca2+ (accumulated plus endogenous) of

COCKRELL

200 neq Ca2+/mg protein, the Mg2+ release is 7 and 6 neq/mg for tumor and liver, respectively; at 400 neq Ca2+/mg protein the values are 15 and 13, respectively. Therefore, assuming endogenous and accumulated Ca2+ are equivalent, there is no significant difference between the Ca2’induced Mg2+ loss of tumor and liver mitochondria. The apparently opposing effects of intramitochondrial Ca2+ versus extramitored-insenchondrial Mg2+ on ruthenium sitive Ca2+ release promoted an analysis of Ca2+ et&ix in terms of internal Ca2+/ Mg2+ ratios for liver and Ehrlich ascites tumor mitoehondria. If internal Ca2+ and Mg2+ are antagonistic, then ruthenium red-insensitive Ca2+ elIlux from both types of mitochondria should be relatively similar, providing endogenous Mg2+ and Ca2+ levels are considered. As shown in Fig. 7, the rates of ruthenium red-insensitive Ca2+ efflux expressed as functions of the internal Ca2+/Mg2+ ratio are much more similar than when treated as a function of internal Ca2+ alone. For example, to achieve a given efflux rate (50 neq Ca2+ mini’ mg-‘, Fig. 4C), the Ca2+ load required was 4.7 times greater for tumor mitochondria than liver. In contrast, the Ca2+/Mg2+ ratios for this efRux rate (18 and 12 for tumor and liver, Fig. 7) differed by a factor of only 1.5. Over the entire range of observed ruthenium red-insensitive Ca2+ efflux values, the average Ca2+ accumulated for a given efflux ratio was 4.85 times greater for tumor mitochondria than rat liver. On the other hand, the corresponding Ca2+/Mg2+ ratios for the same efflux values (adjusted for the prevailing endogenous Cap+ and Mg2+ levels) were only 1.6 times greater for tumor than liver mitochondria. The similarity in Ca2+ release by tumor and liver mitochondria according to their respective Ca2’/Mg2+ ratios was surprisingly good considering probable influences of other regulatory factors. DISCUSSION

Previous studies uation of ruthenium

demonstrated red-insensitive

attenCa2+

CALCIUM

RELEASE

FROM

TUMOR

731

MITOCHONDRIA

FIG. 5. Magnesium efllux from tumor mitochondria. Magnesium was determined spectrophotometrically with Plasmacorinth B as described under Experimental Procedures. Measurements were begun within 20 s of addition of mitochondria to the standard reaction medium used for measuring Ca2+, which included 5 mM Tris-sueeinate and 2.5 mM Tris-phosphate. Other additions, indicated on the figure, were antimycin (0.2 pg/ml), CaCl, (164 neq Caa+/mg protein) and A23187 (2 pg/ml). Changes in absorbance of 0.005 and 0.01 correspond to a change in total [Ca’+] of 20 and 40 P’Y[, respectively.

efflux from Ehrlich cell mitochondria (7). If the endogenous mitochondrial levels of divalent cations are taken into account, Ca2+ efflux. by this pathway is quite similar for tumor and liver mitochondria. TABLE ENDOGENOW;

III

MITOCHONDRIAL

Tumor mitocbondria Rat liver mitochondria

Presumably antagonism between Ca2+ stimulation and Mg2+ inhibition is manifested as a relatively close correspondence (within a factor of 1.5) between the intramitochondrial Ca2+/Mg2+ ratio and the rate of ruthenium red-insensitive Ca2+

DIVALENT

CATIONS

Mg2+

Ca2+

(nedmg

beq/mg

protein)

protein)

129 (80)

90 (20)

49 (25)

14 (5)

Note. The values in parentheses are for MgZt or Ca*+ released by A23187. The Mg’ and Cazf contents of tumor and rat liver mitochondria were determined by atomic absorption (see Experimental Procedures). The values in parentheses are the quantities of Mg2+ and Ca2+ released with A23187 determined with Plasmacorinth B (as in Fig. 5) and murexide, respectively.

200

400

600

800

1000

Net Ca2' UPtake (neOV3)

FIG. 6. Relationship between Ca*+ uptake and net Mg*+ release. The conditions for determining Mg2+ release during Caa+ uptake were the same as described in Fig. 5.

732

FISKUM

AND

FIG. 7. Relationship between ruthenium red-insensitive Ca2+ efflux and the intramitochondrial Ca2+/ Mg2+ quotient. Ruthenium red-insensitive Ca2+ efflux, measured as in Fig. 4 (with phosphate added), is plotted as a function of the intramitochondrial Ca’+/ Mg*+ ratio derived from atomic absorption measurements (see experimental procedures) and PlasmaCorinth B determinations of Mg*+ loss as in Fig. 6.

release. The elevated Mg2+ of Ehrlich ascites tumor mitochondria cannot be explained at this time. The endogenous Ca2+ and Mg2+ levels may be altered during mitochondrial isolation; however, uncoupler stimulated Ca2’ efflux rates for mitochondria in vitro [Ref. (7) and present studies] and for mitochondria in digitoninpermeabilized Ehrlich cells are nearly identical (8). This argues against any substantial redistribution of divalent cations. Exogenous Mg2+ inhibits spontaneous Ca2+ release under respiring conditions (Fig. 3) and that elicited with uncoupling agent (Table II); however, they may not proceed by the same mechanism. Spontaneous release of accumulated Ca2+ (Fig. 3) associated with uncoupling (Fig. 1) was inhibited immediately and completely by exogenous Mg2+ (Fig. 3). In contrast, Mg2+ only partially inhibited ruthenium redinsensitive Ca2+ efflux stimulated by uncoupling agent (70%, Table III). As noted previously (7), the rate of ruthenium redinsensitive Ca2+ release from liver mitochondria can proceed more rapidly than predicted from the acceleration of respiration. In contrast, the Ca2+release rates for tumor mitochondria are, in general, accommodated for by the Ca2+-dependent increments in respiration rates and therefore compatible with membrane “dam-

COCKRELL

age,” e.g. nonspecific alterations in Ca”+ and H+ permeability. The attenuation of ruthenium red-insensitive Ca2+ efflux from tumor mitochondria in which endogenous Mg2+ appears to play a dominant role may be yet another manifestation of resistance to “damage” by Ca2+.Endogenous Mg2+ may similarly affect, for example, the “resistance” of tumor mitochondria to release of respiratory control by Ca2+[Figs. 1 and 2, Ref. (S)]. This is not as markedly different from that of rat liver mitochondria providing the latter are supplemented with exogenous Mg 2+.Furthermore, added Mg2+ has little effect on tumor mitochondria (Fig. 2). Likewise, “massive Ca2+ loading” of tumor mitochondria is approximately twice that of rat liver mitochondria. Here again, exogenous Mg2+ has only a modest effect upon tumor mitochondria (8) whereas it enhances the maximum Ca2+ uptake of rat liver (see Results) to about the same level as that of untreated Ehrlich ascites tumor mitochondria (1.2-1.3 peq/mg protein). It is proposed that the endogenous Mg2+ content is the common denominator underlying (a) the differential sensitivity of tumor and liver mitochondria to effects of exogenous Mg2+, and (b) the differences in ruthenium red-insensitive Ca2+ efflux stimulated by uncoupling agent. Studies of the influence of Ca2+ efflux inhibitors, e.g., local anesthetics like nupercaine which are effective with liver (27, 28) and rat heart mitochondria (ll), may provide additional clues for identification of differences between Ca2+ transport by tumor mitochondria and mitochondria from other cell types. It is not yet clear whether nupercaine blocks a specific Ca2+ efflux pathway (2’7, 28) or acts indirectly by preventing Ca2+ activation of mitochondrial phospholipase A2 and/or resultant secondary changes in membrane permeability (29). In either event, it has been found that nupercaine supplants the salutary effects of Mg2+ on liver mitochondria with no comparable effect upon either isolated tumor mitochondria or those of digitonin-treated Ehrlich ascites cells (30). Hence, the ele-

CALCIUM

RELEASE

FROM

vated endogenous Mg2+ of Ehrlich cell mitochondria may inhibit a specific efflux pathway, suppress mitochondrial phospholipase A2 activity, or prevent permeability changes affected by the latter. REFERENCES 1. THORNE, R. W., AND BYGRAVE, F. L. (1974) Nature (Ladon) 248, 349-351. 2. REYNAFARIE, B., AND LEHNINGER, A. L. (1973) Proc. NatL Acad Sci USA 70, 1744-1748. 3. MCINTYRE, H. J., AND BYGRAVE, F. L. (1974) Arch. Biochem Biophys. 165, 744-748. 4. CARPENTIERI, N., AND SORDAHL, L. A. (1975) Cancer Res. 35, 3336-3341. 5. BYGRAVE, F. L. (1976) in Control Mechanisms in Cancer (Criss, W. E., Ono, T., and Sabine, J. R., e&s.), pp. 411-423, Raven Press, New York. 6. BYGRAVE, 1~. L. (1978) BioL Rev. 53, 43-79. 7. FISKUM, G., AND COCKRELL, R. S. (1978) FEBS I&t. 92, 125-128. 8. COCKRELL, R. S. (1981) Arch. Biochem Biophys. 212, 443-451. 9. LEHNINGER., A. L., CARAFOLI, E., AND ROSSI, C. S. (1967) Adv. Enzymol. 29, 239-320. 10. HUNTER, D. R., AND HAWORTH, R. A. (1979) Arch B&hem, Biophys. 195, 468-477. 11. BEATRICE, M. C., PALMER, J. W., AND PFEIFFER, D. R. (1980) J. Biol. Chem. 255, 8663-8671. 12. COELHO, J. L., AND VERCESI, A. E. (1980) Arch B&hem,. Biophys. 204, 141-147. 13. WOLKOWIC~, P. E., AND MCMILLIN-WOOD, J. (1981) Arch. Biochem Biophys. 209, 408-422. 14. ZOCCARATO, F., RUGOLO, M., SILIPRANDI, D., AND

TUMOR

15. 16.

17. 18.

19. 20.

21.

22. 23. 24. 25.

26. 27. 28. 29.

30.

MITOCHONDRIA

733

SILIPRANDI, N. (1981) Eur. J. Biochem 114, 195-199. PEDERSEN, P. L., AND MORRIS, H. P. (1974) J. BioL Chem. 249, 3327-3334. EBOLI, M. L., MALMSTROM, K., GALEOTTI, T., LOPEZ-ALARCON, L., AND CARAFOLI, E. (1979) Cancer Res. 39, 2737-2742. VILLALOBO, A. AND LEHNINGER, A. L. (1980) J. Biol. Chem. 255,2457-2464. SCARPA, A. (1972) in Methods in Enzymology (San Pietro, A., ed.), Vol. 24, pp. 343-351, Academic Press, New York. SCARPA, A. (1974) Biochemistry 13,2789-2794. ESTABROOK, R. (1967) in Methods in Enzymology (Estabrook, R. W., and Pullman, M. E., Eds.), Vol. 10, pp. 41-47, Academic Press, New York. LOWRY, 0. H., ROSEBROUGH, N. J., FARR, A. L., AND RANDALL, R. J. (1951) J. BioL Chem 193, 265-275. LUFT, J. H. (1971) Anat. Rec. 1’71, 347-368. Roos, I., CROMPTON, M., AND CARAFOLI, E. (1980) Eur. J. Biochvm 110, 319-325. CROMPTON, M., CAPANO, M., AND CARAFOLI, E. (1976) B&hem. J. 154, 735-742. SILIPRANDI, D., TONINELLO, A., ZOCCARATO, F., AND SILIPRANDI, N. (1977) B&hem. Biophys. 78, 23-27. Res. Commun HEATON, G. M., AND NICHOLLS, D. G. (1976) B&hem .I 156, 635-646. DAWSON, A. P. SELWYN, M. J., AND FULTON, D. V. (1979) Nature (London) 277, 484-486. DAWSON, A. P., AND FULTON, D. V. (1980) B&hem J. 188, 749-755. BROEKEMEIER, K. M., SCHMID, P. C., SCHMID, H. O., AND PFEIFFER, D. R. (1985) J. Biol. Chem. 260, 105-113. COCKRELL, R. S. (1982) FEBS. L&t. 144, 279-282.