Thapsigargin Causes Ca2+ Release and Collapse of the Membrane Potential of Trypanosoma brucei Mitochondria in Situ and of Isolated Rat Liver Mitochondria*

Thapsigargin, previously reported to release Ca2+ from non-mitochondrial stores of different cell types, as well as nigericin, were found, when used at high concentrations, to release Ca2+ and collapse the membrane potential of Trypanosoma brucei bloodstream and procyclic trypomastigotes mitochondria in situ. At similarly high concentrations (>lo p ~ ) , thapsigargin was also found to release Ca" and collapse the membrane potential of isolated rat liver mitochondria. These results indicate that care should be taken when attributing the effects of thapsigargin in intact cells to the specific inhibition of the sarcoplasmic and endoplasmic reticulum Ca2+-ATPase family of calcium pumps. In addition, we have found no evidence for an increase in intracellular Ca" by release of the ion from intracellular stores by nigericin, measuring changes in cytosolic Ca2+ by dual wavelength spectrofluorometry in fura-2-loaded T. brucei bloodstream trypomasti- gotes or measuring Ca2* transport in digitonin-per-meabilized cells. The use of digitonin to

Thapsigargin, previously reported to release Ca2+ from non-mitochondrial stores of different cell types, as well as nigericin, were found, when used at high concentrations, to release Ca2+ and collapse the membrane potential of Trypanosoma brucei bloodstream and procyclic trypomastigotes mitochondria in situ. At similarly high concentrations (>lo p~) , thapsigargin was also found to release Ca" and collapse the membrane potential of isolated rat liver mitochondria. These results indicate that care should be taken when attributing the effects of thapsigargin in intact cells to the specific inhibition of the sarcoplasmic and endoplasmic reticulum Ca2+-ATPase family of calcium pumps. In addition, we have found no evidence for an increase in intracellular Ca" by release of the ion from intracellular stores by nigericin, measuring changes in cytosolic Ca2+ by dual wavelength spectrofluorometry in fura-2-loaded T. brucei bloodstream trypomastigotes or measuring Ca2* transport in digitonin-permeabilized cells.
The use of digitonin to permeabilize the plasma membrane of different trypanosomatids (1)(2)(3)(4)(5)(6)(7)(8) has allowed the identification of two intracellular Ca2+ pools. Ca2+ uptake by the first pool is inhibited by antimycin A, FCCP,' and ruthenium red and stimulated by respiratory substrates, phosphate, and acetate. This pool has a high capacity and low affinity for Ca2+ and is able to buffer external Ca2+ at concentrations in the range of 0.6-0.7 p~ (1)(2)(3)(4)(5)(6)(7)(8). These are characteristics typical of mitochondria (9). Ca2+ uptake by the other intracellular pool is inhibited by high concentrations of sodium vanadate and anti-calmodulin agents and stimulated by ATP. This pool has a low capacity and a high affinity for Caz+ and is able to buffer external Ca2+ at concentrations in the range of 0.05-0.1 pM * This work was supported in part by grants from the Midwest University Consortium for International Activities (to R. D. and S. N. J. M.), from the University of Illinois Research Board (to S. N. J. M.), from the Conselho Nacional de Desenvolvimento Cientifico e Tecnolbgico (CNPdRHAE), and from the Fundaclo de Amparo a Pesquisa do Estado de S l o Paulo, Brazil (to A. E. V.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. methoxyphenylhydrazone. (1)(2)(3)(4)(5)(6)(7)(8). These are characteristics typical of the endoplasmic reticulum (10). Recently, the presence of a third CaZ+ pool sensitive to changes in intracellular pH has been suggested in Trypanosoma brucei bloodstream trypomastigotes (11) on the basis of the changes observed in the fluorescence of fura-2loaded cells when nigericin was included in the incubation medium. In contrast, the sesquiterpene lactone and tumor promoter thapsigargin was shown to apparently release Ca" from a non-mitochondrial pool insensitive to nigericin (12). Nigericin has also been shown to increase cytosolic Ca2+ in Leishmania donouani promastigotes (13). Since we have been unable to detect any increase in intracellular Ca2+ by thapsigargin in Trypanosoma cruzi amastigotes and epimastigotes using fura-2-loaded cells (7), we re-examined the effect of nigericin and thapsigargin on Ca2+ homeostasis in T. brucei using fura-2-loaded and digitonin-permeabilized cells.
We report here that neither thapsigargin nor nigericin, when used at low concentrations, are able to increase the intracellular Ca" concentration of either T . brucei bloodstream or procyclic trypomastigotes by the release of Ca2+ from intracellular stores. In addition, we report that high concentrations of thapsigargin and nigericin are able to release Ca2+ from T. brucei and rat liver mitochondria, probably as a result of the collapse of their membrane potential.

MATERIALS AND METHODS
Culture Methods-T. brucei procyclic forms (ILTar 1 procyclics) were grown at 28 'C in medium  supplemented with hemin (7.5 mg/liter) and 10% heat-inactivated fetal calf serum. Two to three days after inoculation, cells were collected by centrifugation and washed twice in Dulbecco's phosphate-buffered saline. T. brucei bloodstream forms (monomorphic strain 427 from clone MITat 1.4, otherwise known as variant 117) were isolated from infected mice or rats as described previously (15) and kept in separation buffer containing 44 mM NaC1, 55 mM glucose, 57 mM Na2HP04, and 3 mM NaH2P04, pH 8.0, until use, and after loading with fura-2. The final concentration of cells was determined using a Neubauer chamber.
The protein concentration was determined by the biuret assay (16) in the presence of 0.2% deoxycholate. Rat liver mitochondria were isolated as described before (17).
Chemicals-ATP, calcium ionophore A23187, sodium orthovanadate, FCCP, succinate, arsenazo 111, EGTA, Triton X-114, nigericin, and digitonin were purchased from Sigma. Thapsigargin was purchased from Sigma (T9033, more than 99% pure, as indicated by the manufacturer) or from LC Services. Identical results were obtained when thapsigargin from different sources was used. fura-B/AM was from Molecular Probes, Inc., Eugene, OR. All other reagents were analytical grade.
Spectrofluorometric Determinations-fura-2 determinations were performed as described before (6). Concentrations of the ionic species and complexes at equilibrium were calculated by employing an iterative computer program as described before (6).
Determination of Caz+ Movements-Variations in free Ca2' concen-Effect of Thapsigargin on Mitochondria 8565 trations were followed by measuring the changes in the absorbance spectrum of arsenazo 111 (18), using the SLM Aminco DW2000 spectrophotometer at the wavelength pair 675-685 or with a calciumselective electrode (2). No free radical formation from arsenazo 111 occurred under the conditions used (19,20). The calibrations were performed by the sequential addition of known concentrations of EGTA. The initial Ca2+ concentration in the solution was obtained by atomic absorption spectroscopy and the Ca2+ concentration after each EGTA addition was calculated by employing an iterative computer program as described before (6,8).
Estimation of Mitochondrial Membrane Potential-These measurements were made using safranine 0 (5) or a TPP+-selective electrode in combination with a calomel reference electrode (2) as described in the references.

RESULTS
It has been reported (11,12) that nigericin and thapsigargin are able to increase intracellular Ca2+ in T. brucei bloodstream trypomastigotes by releasing it from intracellular stores. We therefore investigated the effect of these compunds on fura-  (11,12). No effects were detected when thapsigargin (8 PM) was added in the presence of high extracellular calcium (not shown). Taken together, these results indicate that the increase in cytosolic Ca" induced by addition of nigericin in the presence of high external Caz+ is probably not due to Ca2+ release from intracellular stores.
T o further demonstrate that neither nigericin nor thapsigargin was able to increase [Ca2+Ii through Ca2+ release from intracellular stores of T. brucei bloodstream trypomastigotes, we used the digitonin-permeabilization technique described previously (6).  T. brucei bloodstream trypomastigotes. A fast decrease in Ca2+ concentration started immediately after addition of digitonin and lowered the ambient free Ca2+ concentration to at least 0.05-0.1 p~ (Fig. 2, trace a). The subsequent addition of nigericin (2.75 p M ) or thapsigargin (8 p M ) in any order ( Fig.  2 and not shown) did not cause any change in the ambient Ca2+ concentration. In contrast, addition of the calcium ionophore A23187 released the Ca2+ taken up, as well as the endogenous Ca2+. Vanadate (Fig. 2, trace b ) totally inhibited Ca2+ uptake and slowly released the accumulated Ca2+. Further addition of thapsigargin did not change this slow rate of Ca2+ release caused by vanadate, whereas addition of calcium ionophore rapidly released the Ca2+ taken up and the endogenous Ca2+.
Nigericin is known to uncouple oxidative phosphorylation in rat liver mitochondria when used at high concentrations (1-10 PM) (21-23). Previous works on the effects of nigericin on intracellular Ca2+ homeostasis in L. donouani (13) and T. brucei (11,12) reported the use of 4 p~ (13) and 1-2.75 p~ (11,12) nigericin. When we examined the effect of similar concentrations of nigericin on the mitochondrial Ca2+ uptake (in the presence of vanadate, Ref. 6) by digitonin-permeabilized T. brucei bloodstream trypomastigotes, we observed that addition of the drug after a steady state was attained led to Ca2+ release (not shown). Even thapsigargin (8 p~) was able to release Ca2+ from these mitochondria in situ (not shown).
We have demonstrated previously (6) that T. brucei bloodstream trypomastigotes mitochondria in situ only take up Ca2+ in the presence of ATP but not in the presence of respiratory substrates (6). In order to better analyze the effects of nigericin and thapsigargin on mitochondria, we studied the effect of these drugs on T. brucei procyclic trypomastigotes mitochondria in situ. These mitochondria are able to take up Caz+ in the presence of respiratory substrates and in the absence of ATP (6)) and therefore there is no interference with Ca2+ uptake by other cellular pools. Fig. 3 shows Ca2+ uptake by digitonin-permeabilized T. brucei procyclic trypomastigotes in the presence of succinate. When the cells were added to the reaction medium, a decrease in Ca2+ concentration started after a period of about 30 s and continued until the ambient free Ca2+ concentration was lowered to about 0.7-0.8 pM, in agreement with a previous report (6). The subsequent addition of antimycin A was followed by the release of all the Ca2+ taken up. If nigericin was added instead of antimycin, a higher Ca2+ release than that observed with antimycin A was detected, thus indicating, as expected (21-23), a clear effect on these mitochondria in situ. Fig. 4, trace A , shows that thapsigargin (12.5 phi) also released Ca2+ from the mitochondria in situ of digitonin-permeabilized procyclic trypomastigotes. In this experiment vanadate was included to inhibit Caz+ uptake by the non-mitochondrial pool (6). Addition of FCCP after thapsigargin did not release any additional Ca2+, but addition of the calcium ionophore A23187 caused the release of endogenous Ca2+ (Fig. 4, trace a). The Ca" release caused by thapsigargin (Fig. 4, trace a ) was similar to that caused by antimycin A ( A A , Fig. 4, truce b). When FCCP was added to the incubation medium (Fig. 4, trace B ) Ca2+ uptake was completely inhibited. Addition of thapsigargin under these conditions did not release any endogenous Ca2+, whereas the calcium ionophore A23187 released a considerable amount of endogenous Ca2+, as it has been reported previously (6).
Since Ca2+ release from these mitochondria in situ could be due to the collapse of their membrane potential, we investigated the effect of nigericin and thapsigargin on the mitochondrial membrane potential of digitonin-permeabilized T. bruceiusing the safranine 0 method (5,7,8). increase in absorbance at the wavelength pair 511-533 compatible with the stacking of the dye to the energized mitochondrial inner membrane. Addition of thapsigargin (12.5 p~) caused a rapid and extensive decrease in the membrane potential which was completed by the further addition of antimycin A. Fig. 6 shows that addition of different concentrations of nigericin to T. brucei procyclic (Fig. 6A) or bloodstream (Fig.  6 B ) trypomastigotes mitochondria in situ caused a dosedependent collapse of their membrane potential similarly to that caused by FCCP, valinomycin, or, in the case of bloodstream trypomastigotes mitochondria ( 6 ) , to oligomycin (Fig.  6B, dashed line).
Thapsigargin has been shown to cause the release of Ca2+ from the endoplasmic reticulum in several different cell types (including platelets, lymphocytes, neutrophils, macrophages, hepatocytes, adrenal chromaffin, and parotid acinar cells) (24) without affecting their mitochondrial activity (25). It has been indicated that this is because thapsigargin has a remarkable specificity for the sarcoplasmic and endoplasmic reticulum Ca2+-ATPase family of calcium pumps (26). Since our results indicated that high concentrations of thapsigargin were able to release Ca2+ from mitochondria of T. brucei in situ, we further investigated the characteristics and presence of this effect in isolated rat liver mitochondria to verify if this effect of thapsigargin was specific for T. brucei mitochondria.
Figs. 7 shows that high concentrations of thapsigargin (>lo p~) caused a dose-dependent Ca2+ release from isolated rat liver mitochondria without a significant inhibition of Ca2+ uptake. The effect of thapsigargin on the membrane potential of isolated rat liver mitochondria is shown in Fig. 8. Collapse of the mitochondrial membrane potential was evident with similar concentrations of thapsigargin than those that caused brucei bloodstream trypomastigotes (Fig. 1). Even when used at high concentrations, neither nigericin nor thapsigargin (Fig. 2) was able to release Ca2+ from non-mitochondrial stores of digitonin-permeabilized T. brucei bloodstream trypomastigotes. When used at a high concentration (2.75 p M ) , nigericin was able to increase [Ca2+Ii (Fig. 1, truces a and b ) , and this increase was higher in the presence of a high extracellular Ca2+ concentration (Fig. 1, trace b), in agreement with a previous report (11). However, at this high concentration, nigericin caused Ca2+ release from the mitochondria in situ of permeabilized cells (Fig. 3) and collapsed their mitochondrial membrane potential (Fig. 6), in agreement with its known uncoupling effect on oxidative phosphorylation of rat liver mitochondria (21-23). On the other hand thapsigargin, when used a t high concentrations (>8 p~) , was unable to increase [Ca2+]i of T. brucei bloodstream trypomastigotes ( Fig. 1) but was able to stimulate Ca2+ release from T. brucei bloodstream and procyclic (Fig. 4) trypomastigotes mitochondria in situ and to collapse their mitochondrial membrane potential (Fig.  5 ) .
Our results differ from the results previously reported on the effect of nigericin (11) and thapsigargin (12) on T. brucei bloodstream trypomastigotes. Differences due to the different strain or differentiation stage of the T. brucei bloodstream trypomastigotes used in those reports cannot be ruled out. In this regard, it should be noted that we have used similar (1 p~, 12) or higher concentrations (8 p~, Figs. 1 and 2) of thapsigargin from the same sources in our experiments with different results (see Fig. 1 and Figs 1 and 2 of Ref. 12). On the other hand, the possible uncoupling effect of nigericin at the very high concentrations used in previous studies in trypanosomatids (4 pM (Ref. 13); 1-2.75 pM (Refs. 11 and 12)) could have led, through ATP depletion, to an increase in [Ca2+Ii.
Taken together our results do not support the hypothesis of the presence of a nigericin-sensitive non-mitochondrial Ca2+ pool in T. brucei long slender bloodstream trypomastigotes (11). In addition, we were unable to find a non-mitochondrial thapsigargin-sensitive Ca2+ pool in T. brucei bloodstream and procyclic trypomastigotes as well as in other trypanosomatids that we have examined thus far, including T. cruzi epimastigotes and amastigotes ( 7 ) and L. donovuni promastigotes.' A number of previous studies (24, 27) have indicated that thapsigargin causes influx of calcium from the extracellular medium across the plasma membrane into the cytoplasm. Most investigators have proposed that this thapsigargin-induced calcium influx results from a thapsigargininduced depletion of intracellular stored calcium, which by some unkown mechanism then causes increased calcium influx (27). Our results favor this hypothesis, since in trypanosomes there is a lack of thapsigargin-induced depletion of intracellular stored calcium and concomitantly no increase in [Ca2+]i through Ca2+ influx.
Another important consequence of the present work concerns the use of thapsigargin as a tool for the study of Caz+ fluxes in intact cells. Thapsigargin was able to cause Ca2+ release from mitochondria in situ of T. brucei as well as from isolated rat liver mitochondria. This indicates that care should A. E. Vercesi and R. Docampo, unpublished results.

Effect of Thapsigargin on Mitochondria
be taken when attributing the effects of thapsigargin in intact cells to the specific inhibition of the sarcoplasmic and endoplasmic reticulum Ca2+-ATPase family of calcium pumps. In this regard the use of thapsigargin in intact cells at concentrations of 10 (27) or 10-50 p M (28) has been reported recently.