Selective Transfer of Calcium from an Acidic Compartment to the Mitochondrion of Trypanosoma brucei

Organelle compartments are used by cells as reservoirs of exchangeable Ca2+ and as Ca2+ buffers. The following study uses recombinant aequorins (CYT-AEQ and MT-AEQ) to measure the dynamics of Ca2+ flux between organelles in procyclic forms of the pathogenic protozoan, Trypanosoma brucei. Emphasis is placed on the exchange between an acidic Ca2+ reservoir and the mitochondrion. The mammalian mitochondrial targeting sequence was functional in trypanosomes as determined by immunoblots, immunolocalizations, and the observation that MT-AEQ was in a compartment whose Ca2+ uptake was inhibited 82% with carbonyl cyanide p-trifluoromethoxyphenylhydrazone and KCN. The resting level of free calcium ion concentration in the mitochondrion ([Ca2+]mit) was slightly higher than that in the cytoplasm ([Ca2+]cyt) (400 ± 50 nm and 290 ± 40 nm, respectively). Melittin (125 nm) disrupted Ca2+homeostasis by inducing Ca2+ influx across the plasma membrane. [Ca2+]cyt became slightly elevated to 410 ± 100 nm, whereas [Ca2+]mit was selectively increased approximately 12-fold, with a broad peak at 4.8 ± 1.9 μm. At the peak, the mitochondrion contained approximately three times more free Ca2+ than the cytosol. However, mitochondrial retention of the Ca2+ was transient. Similar selective transport into the mitochondrion was observed when Ca2+ efflux from an acidic compartment was induced with monensin (2 μg/ml) in the presence of 5 mm EGTA. [Ca2+]cyt was transiently elevated to 400 ± 50 nm, whereas [Ca2+]mit was elevated to 3.3±1.3 μm. When cells were treated sequentially with monensin (2 μg/ml) and then melittin (200 nm), mitochondrial Ca2+ transport was normal. However, [Ca2+]cyt became elevated to a level that was 1.4-fold higher than with melittin alone. Overall, these data demonstrate that the trypanosome mitochondrion is not a reservoir of exchangeable Ca2+ in the resting cell. However, Ca2+ is selectively channeled to the mitochondrion from the plasma membrane or acidic Ca2+ storage compartment. Additionally, the acidic compartment contributes to maintenance of Ca2+ homeostasis in response to melittin.

Cell survival is dependent upon the ability to receive signals from the environment and initiate appropriate changes in cell activity. This property of cells is especially important in parasitic organisms with digenetic life cycles where multiple environments are encountered and the environment may contain mechanisms to actively seek and destroy the intruder. Trypanosomes of the brucei group are flagellated protozoa that produce lethal infections in humans and livestock throughout sub-Saharan Africa. During the course of the trypanosome infection, the invading cells must adapt to life in the mammalian and insect hosts, as well as to changing environments within these hosts. We are interested in the signal mechanisms within Trypanosoma brucei that allow them to coordinate life cycle events. Emphasis is placed on Ca 2ϩ pathways because Ca 2ϩ can function as a pluripotent regulatory molecule. Conversely, prolonged exposure to elevated [Ca 2ϩ ] cyt can result in cell death (1,2). Trypanosome survival therefore depends upon redundant energy-dependent processes that carefully control cytosolic free Ca 2ϩ concentrations ([Ca 2ϩ ] cyt ) 1 (3)(4)(5)(6)(7)(8). The present study examines within intact trypanosomes the relative importance of several Ca 2ϩ -transporting organelles.
African trypanosomes exhibit some unusual features when considering organelle Ca 2ϩ transport. In the present report, emphasis is placed on in vivo measurement of Ca 2ϩ transport by the mitochondrion and acidic Ca 2ϩ storage compartment. In mammalian cells, the mitochondrion is used as a Ca 2ϩ buffer to protect against large changes in [Ca 2ϩ ] cyt (9). Typically, isolated mammalian mitochondria can lower medium Ca 2ϩ to a set point of approximately 700 nM (10). However, recent studies with targeted aequorins demonstrate that mammalian mitochondria selectively receive Ca 2ϩ from other organelles perhaps by sensing microdomains of elevated Ca 2ϩ in the vicinity of the organelle (11)(12)(13)(14)(15). By contrast with mammalian cells, T. brucei are primitive organisms, and sequence comparisons suggest that their lineage represents the oldest branch of eukaryotic cells that contain a mitochondrion (16). The mitochondrion of T. brucei contains several novel features, including catenated maxicircle and minicircle DNA, and a requirement for RNA editing to express mitochondrion-encoded genes (17,18). During the trypanosome life cycle, the mitochondrion undergoes extensive developmental changes (19). In the mammalian host, the mitochondrion lacks cytochromes and Krebs cycle enzymes (19). Instead, the mitochondrion contributes to NADH oxidation by means of an alternative oxidase (20). A membrane potential is maintained in bloodstream forms at the expense of ATP (21) or through activity of the alternative oxidase (22). By contrast, the procyclic or insect form of the trypanosome contains a large reticulated mitochondrion that extends throughout the length of the organism. In procyclic forms a mitochondrial membrane potential is maintained by a complete respiratory chain (19). Permeabilized cells have been used to demonstrate that mitochondria from both life cycle stages can accumulate large quantities of Ca 2ϩ with a set point around 700 nM (4). However, the relative importance of the trypanosome mitochondrion as a reservoir of stored Ca 2ϩ has not been adequately explored in either life cycle stage. Yet it has been proposed that Ca 2ϩ release from the trypanosome mitochondrion might alter [Ca 2ϩ ] cyt when resting cells are treated with thapsigargin (23) or nigericin (23). These conjectures are only true if the level of Ca 2ϩ in the resting mitochondrion is high enough to affect cytoplasmic Ca 2ϩ levels.
In addition to the mitochondrion, we have described an acidic compartment that is nonmitochondrial in nature (7). We showed that this compartment released Ca 2ϩ in response to the K ϩ /H ϩ antiporter, nigericin. Enough Ca 2ϩ was released to elevate [Ca 2ϩ ] cyt by 3-fold above the resting level (7), suggesting that large quantities of exchangeable Ca 2ϩ were stored in this compartment. Later it was shown that the Na ϩ /H ϩ antiporter, monensin, also released Ca 2ϩ from the acidic compartment (24). Ca 2ϩ transport into the acidic compartment required a vanadate-sensitive Ca 2ϩ -ATPase (8), and Ca 2ϩ release occurred via a Ca 2ϩ /nHϩ exchanger when the organelle pH gradient was collapsed (24). This organelle has also been referred to as the acidocalcisome (8). To date, no evidence exists demonstrating that the T. brucei acidocalcisome is a separate or novel organelle. In mammalian cells, acidic Ca 2ϩ pools have been described (25,26), and it is presumed that these pools reside within the lysosome or endocytic vesicles. Histochemical and x-ray microprobe analysis of T. brucei organelles demonstrate that large amounts of Ca 2ϩ are stored in a compartment that may be lysosomal in origin (27). Whether the acidic compartment contributes to Ca 2ϩ homeostasis in situ is not known. Moreover, the fate of Ca 2ϩ released from this compartment is also not known.
In the present study, interactions between the mitochondrion and acidic Ca 2ϩ storage compartment are investigated. Melittin is used to initiate Ca 2ϩ influx across the plasma membrane (28), whereas monensin is used to disrupt the acidic Ca 2ϩ storage compartment (24). Organelle Ca 2ϩ transport is measured during the return to homeostasis. To measure organelle Ca 2ϩ transport in vivo, the Ca 2ϩ -sensitive photoprotein aequorin has been targeted to the mitochondrial matrix space or expressed without a localization signal to accumulate in the cytoplasm. When reconstituted in vivo with coelenterazine, recombinant aequorins have recently been used to monitor free Ca 2ϩ concentrations in the nucleus (29 -32), cytoplasm (33), mitochondria (11)(12)(13)(14)(15), and ER (34) of a wide range of mammalian cells. We have previously used targeted aequorins to quantify Ca 2ϩ content of the trypanosome nucleus (35). In the present study, targeted aequorins are used to quantify [Ca 2ϩ ] mit in resting and stimulated cells. We report that the mitochondrion does not function as a reservoir of exchangeable Ca 2ϩ in the resting cell. However, during the signaling process, the majority of Ca 2ϩ that enters the cell from across the plasma membrane or is released from the acidic compartment selectively enters the mitochondrion. The mitochondrion retains this Ca 2ϩ transiently. When the mitochondrion releases this sequestered Ca 2ϩ , other homeostatic organelles must return [Ca 2ϩ ] cyt to the resting level. None of these results were predicted from studies with permeabilized cells.

Materials-Monensin was from BIO MOL Research Laboratories.
Melittin, FCCP, ionomycin, lanthanum chloride, nigericin, mowiol, and Ponceau S were from Sigma. Coelenterazine and the mtAEQ vector were from Molecular Probes. Glutathione-Sepharose and pGEX-5X-3 were from Pharmacia Biotech Inc. The PCR primers, 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium, were from Life Technologies, Inc. Goat anti-rat IgG alkaline phosphatase was from Boehringer Mannheim. Ribi's adjuvant was from Ribi ImmunoChem Research, Inc. The rabbit anti-cytochrome c 1 antibodies were kindly provided by S. Hajduk (University of Alabama, Birmingham, Alabama). The procyclic cell line was kindly provided by E. Pays (Free University of Brussels, Belgium).
Organisms-Procyclic forms of T. brucei brucei were derived from AnTat1.1E bloodstream forms and were used throughout the study. Cells were maintained in SDM 79 medium (36) with 15% fetal calf serum at 27°C and 6% CO 2 .
PCR-The mtAEQ plasmid was described by Rizzuto et al. (11) and was purchased from Molecular Probes. The CYT-AEQ cDNA was PCR amplified from the mtAEQ plasmid with upstream primer 5Ј-ATACTC-GAGATGAAGCTTACATCAGACTTCGAC-3Ј and downstream primer 5Ј-ATACCCGGGTTTATTTAGGGGACAGCTCCACCGTA-3Ј. The PCR reaction consisted of 6 ng of mtAEQ plasmid DNA, 1 M primers, 0.4 mM dNTPs, 1.5 mM Mg 2ϩ , 1 ϫ Taq polymerase buffer, and 5 units of Taq polymerase. Amplification used the following program: 30 s at 94°C, 30 s at 42°C, and 45 s at 72°C for 30 cycles.
GST-Aequorin Fusion Protein-Aequorin cDNA without the COX VIII mitochondrial targeting sequence was excised from the mtAEQ vector with HindIII and BamHI. The blunt ended fragment was cloned into the SmaI site of the expression vector pGEX-5X-3. The GSTaequorin fusion protein was affinity purified from bacterial extracts with a glutathione-Sepharose affinity column following the manufacturer's instructions.
Transformation of T. brucei-Procyclic forms of T. brucei were electroporated with 30 -50 g of the appropriate MluI linearized vectors as described previously (35). Transformants were selected with hygromycin B and cloned once by limiting dilution.
Cell Fractionation-Trypanosomes were fractionated as described by others (37). Briefly, 150 ml of procyclic cultures were harvested at a density of 1.5-2 ϫ 10 7 cells/ml and resuspended in 50 ml of 0.15 M NaCl, 20 mM glucose, 20 mM Na 2 HPO 4 , pH 7.9. The trypanosomes were pelleted again and resuspended in 2.5 ml of 0.25 M sucrose, 2 mM EDTA and 20 mM MOPs, pH 7.2 (SME) and placed in the cold chamber of a N 2 bomb at 1,500 p.s.i. for 15 min. Upon release from the chamber, nuclei and unbroken cells were pelleted at 1,000 ϫ g for 15 min. The 1,000 ϫ g supernatant was centrifuged at 17,000 ϫ g for 15 min, and this pellet was washed one time. The washed 17,000 ϫ g pellet was resuspended in 100 l of SME and is referred to as the mitochondrial fraction. The initial 17,000 ϫ g supernatant was further centrifuged at 100,000 ϫ g for 1 h to separate small organelles from the cytosol. The 100,000 ϫ g supernatant was the cytosolic fraction.
Immunolocalization-Coverslips were coated with 0.1% polylysine for 30 min, rinsed once in distilled H 2 O, and air dried. Procyclic cells were centrifuged at 1500 ϫ g for 5 min and suspended in phosphatebuffered saline (10.1 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , 138 mM NaCl, and 2.7 mM KCl, pH 7.4) at a density of 1.3 ϫ 10 8 cells/ml. The cell suspension (100 l) was allowed to settled on the coverslip for 3 h at 4°C. All subsequent steps were performed in phosphate-buffered saline. Cells on the coverslip were fixed for 10 min with 3% paraformaldehyde and permeabilized with 0.1% Triton X-100 for 10 min. Fixed and permeabilized trypanosomes were washed two more times and were incubated with purified rat anti-aequorin antibody (described below) diluted 1:10 in phosphate-buffered saline plus 0.1% bovine serum albumin for 1 h at 37°C. The cells were washed three times and incubated with FITC-or rhodamine-conjugated secondary antibodies plus 0.1% bovine serum albumin for 1 h at 37°C. The coverslips were set in mounting medium as described previously (38), except 0.2 g/ml mowiol replaced the polyvinyl alcohol. Cells were photographed with Tmax p3200 film using 8 -10-s exposures and a Microphot-FX microscope.
Antibodies against the GST-Aequorin Recombinant Protein-Rats were immunized with GST-aequorin fusion protein reconstituted to a final concentration of 1 mg/ml in Ribi's adjuvant. For immunolocalization, rat antiserum was affinity purified by the method of Olmstead (39) using recombinant GST-aequorin as the immunoabsorbant. Then, the antibodies were preabsorbed with normal procyclic cells that had been fixed for 10 min with 3% paraformaldehyde and permeabilized for 10 min with 0.1% Triton X-100. The procyclics and the attached antibodies were centrifuged at 10,000 ϫ g for 5 min and discarded. Antibodies in the supernatant were further preabsorbed with a 50-fold excess of recombinant GST. The purified anti-aequorin antibodies were stored in 0.02% sodium azide and 0.1% bovine serum albumin at Ϫ20°C. Luminescence-Transformed cells at a density of 1.5-2 ϫ 10 7 cells/ml were incubated with 2.5 M coelenterazine for 3 h in SDM79 and 15% fetal bovine serum. The cells were then pelleted and resuspended at the same density in reaction buffer (116 mM NaCl, 5.4 mM KCl, 0.8 mM MgSO 4 , 5.5 mM D-glucose and 50 mM Hepes, pH 7.0) to remove the excess coelenterazine. Photon pulses from the Ca 2ϩ -dependent oxidation of coelenterazine were detected with a home built luminometer. The luminometer consisted of a stirred room temperature macrochamber located 7 cm from the unfiltered opening to a Hamamatsu R1572P low dark current photomultiplier tube. Mirrors around the sample compartment directed light to the photomultiplier tube opening. Output from the photomultiplier tube was collated with a photon counting board (Photon Technology International) and processed with software v.2060 from the Delta Scan dual wavelength flourimeter (Photon Technology International). The luminescence was monitored at the rate of five data points/s. The cells were lysed with 0.1% Triton X-100 and 10 mM CaCl 2 at the end of every experiment to discharge the remaining aequorin. The luminescence detected after that time was considered background. To calculate the Ca 2ϩ concentration from the aequorin luminescence, raw data were smoothed, and the background was subtracted from every data point. The data were saved as ASCII files and were processed off-line by a Fortran program in which the kinetic equation [Ca 2ϩ ] ϭ [10 (log␣-9) ] 1 ⁄2 (14) was applied, where ␣ equals the rate of photon emission at each time divided by the total remaining photons.

RESULTS
Recombinant Aequorin Constructs-The mtAEQ insert cDNA includes the localization signal of human mitochondrial cytochrome c oxidase subunit VIII ligated in frame with the aequorin cDNA, as described by Rizzuto et al. (11). The mtAEQ vector was digested with EcoRI, blunt-ended with Klenow fragment, and cloned into the EcoRV site of the trypanosome expression vector, pTSA-HYG2 (40). The resulting trypanosome expression vector was called pMT-AEQ. The insert for plasmid pCYT-AEQ was prepared by PCR using pMT-AEQ as a template (35). The amplified apoaequorin cDNA encodes the fulllength amino acid sequence of mature apoaequorin, except the PCR primer replaced the N-terminal Val in processed apoaequorin with the Met start codon to initiate protein translation. The CYT-AEQ cDNA was ligated into the XhoI/EcoRV sites of the pTSA-HYG2 vector. The new vector was named pCYT-AEQ.
Intracellular Localization of Aequorin Proteins-Location of the CYT-AEQ and MT-AEQ proteins was initially verified with immunoblots of cell fractions. Cells were homogenized with a N 2 bomb, and the mitochondrial fraction (17,000 ϫ g pellet) and cytosolic fraction (100,000 ϫ g supernatant) were obtained. Total protein in each fraction was separated by SDS-polyacrylamide gel electrophoresis (Fig. 1A) and analyzed by Western blot procedures with antibodies against a GST-aequorin fusion protein (Fig. 1, B and C). In cells transformed with pCYT-AEQ, a cross-reacting protein of appropriate size (22 kDa) was detected in the whole cell fraction (Fig. 1B, lane a) and was enriched in the cytosolic fraction (lane b) but was not present in the mitochondrial fraction (lane c). The distribution of CYT-AEQ in cytosol is consistent with the fact that CYT-AEQ protein does not contain any organelle import signal. In cells transformed with pMT-AEQ, recombinant aequorin migrated with an apparent mass of 24 kDa (Fig. 1C, lanes a and c), consistent with retention of the 33-amino acid signal sequence. Proteolytic processing of this sequence would produce a protein with a predicted mass of 22 kDa. The total protein amount of the mitochondrial fraction (lane c) was only half of the other fractions on the immunoblot. However, the antibody reaction was more intense, demonstrating that MT-AEQ protein was concentrated in the mitochondrial fraction. By contrast, the cytosolic fraction did not contain any cross-reacting proteins (lane b). Cytochrome c 1 is a mitochondrial protein, and antibodies against cytochrome c 1 of T. brucei also only cross-reacted with proteins in the 17,000 ϫ g fraction (data not shown).
Because the mitochondrial fraction was crude and contained several different organelles, immuonolocalization was used to verify the distribution of the recombinant MT-AEQ protein (Fig. 2). Transformants were fixed with 3% paraformaldehyde and permeabilized with 0.1% Triton X-100. The affinity purified and preabsorbed GST-aequorin antibodies were detected with FITC-conjugated secondary antibodies. No fluorescence was detected in control procyclic cells ( Fig. 2A), which is consistent with the absence of cross-reacting proteins in the immunoblot (Fig. 1B, lane a). By contrast, the antibodies localized MT-AEQ protein to an extensive tubular network throughout the cell (Fig. 2B), which is similar in appearance to published diagrams of the single mitochondrion of procyclic form T. brucei (41). To verify that the tubular network was indeed the mitochondrion, co-localization was performed with rabbit anti-cytochrome c 1 antibodies and rhodamine-conjugated secondary antibodies. MT-AEQ and cytochrome c 1 exhibited identical distribution throughout the cells (Fig. 2, C and D). Therefore, immunolocalization and cell fractionation suggest that the recombinant MT-AEQ protein is selectively translocated into the mitochondrion. By contrast, the CYT-AEQ protein was not detected by immunolocalization, probably because the protein was dispersed over a larger area. In Vivo Reconstitution of Aequorin-Functional aequorins were reconstituted by incubation of procyclic transformants with coelenterazine for 3 h. When 2 ϫ 10 7 cells in the reaction buffer were lysed with 0.1% Triton X-100 supplemented with 10 mM Ca 2ϩ , the recombinant aequorin was completely discharged by the high Ca 2ϩ concentration in the buffer (10 mM). The peak rates of light emission recorded in the pCYT-AEQ and pMT-AEQ were 300,000 and 400,000 photons/s, respectively (Fig. 3, A and B). These large total rates of emission relative to the base line (approximately 30 photons/s) allowed the signal to be used for the quantitation of free calcium concentrations. Similar recombinant aequorins have been used extensively by others to measure free calcium concentrations (11-15, 29 -34).
Ca 2ϩ Measurements in the Trypanosome Mitochondrion-The contributions of the trypanosome mitochondrion to the propagation of Ca 2ϩ signals and to Ca 2ϩ homeostasis were investigated. The average resting level of [Ca 2ϩ ] cyt was 290 Ϯ 40 nM (n ϭ 68), whereas the average resting level of [Ca 2ϩ ] mit was 400 Ϯ 50 nM (n ϭ 36). To evaluate Ca 2ϩ transport properties of the mitochondrion in intact cells, Ca 2ϩ homeostasis was disrupted in two distinct ways: either by influx across the plasma membrane or by release from intracellular pools. Melittin is an amphiphilic peptide that causes Ca 2ϩ to selectively enter the cell across the plasma membrane of T. brucei bloodstream forms (28). When procyclic cells were treated with 200 nM melittin, the pCYT-AEQ transformants show an immediate spike (possibly an injection artifact), followed by a transient elevation from the resting level of Ca 2ϩ to 1.2 Ϯ 0.4 M (n ϭ 16) (Fig. 3C). Homeostatic pathways decreased [Ca 2ϩ ] cyt over a 60-s period to 570 Ϯ 160 nM (n ϭ 16). The same dose of melittin caused a similar injection spike in [Ca 2ϩ ] mit (Fig. 3D). The [Ca 2ϩ ] mit was then rapidly increased up to 8.4 Ϯ 2.7 M (n ϭ 13), which was approximately 7-fold higher than the peak value of [Ca 2ϩ ] cyt . A new steady level of [Ca 2ϩ ] mit at 1.50 Ϯ 0.5 M (n ϭ 11) was reached. The elevation in [Ca 2ϩ ] mit was not the result of a direct interaction between melittin and the mitochondrion. The calcium channel blocker lanthanum completely inhibited Ca 2ϩ influx into the cytosol (dashed curve in Fig. 3C). Under these conditions, melittin was without effect on [Ca 2ϩ ] mit (dashed curve in Fig. 3D), showing that the rise in [Ca 2ϩ ] mit was dependent upon Ca 2ϩ influx across the plasma membrane. Finally, Fig. 3 (E and F) shows that the luminescence of MT-AEQ recombinant protein originated from the mitochondrial matrix. When the mitochondrial electrochemical potential was disrupted with the respiratory inhibitor KCN and the mitochondria uncoupler FCCP, the increase in [Ca 2ϩ ] mit evoked by melittin was reduced by 82% (n ϭ 11). By contrast, KCN and FCCP did not affect the melittin-induced change in [Ca 2ϩ ] cyt , which remained at 1.0 Ϯ 0.6 M (n ϭ 15).
A low dose of melittin was used to illustrate the Ca 2ϩ transporting abilities of the mitochondrion in T. brucei when Ca 2ϩ in the environment was lower than the proposed set point of 700 nM (4). Treatment with 125 nM melittin produced a transient injection spike followed 7 s later by a barely detectable change in [Ca 2ϩ ] cyt to a new value that was just 410 Ϯ 100 nM (n ϭ 13) (Fig. 4A). By contrast, [Ca 2ϩ ] mit became transiently elevated with a broad peak of 4.8 Ϯ 1.9 M, which was 11-fold higher than the level of [Ca 2ϩ ] cyt (Fig. 4B). These data demonstrate that the mitochondrion of T. brucei is extremely sensitive to small changes in [Ca 2ϩ ] cyt when Ca 2ϩ entry is induced across the plasma membrane. The total amount of free Ca 2ϩ that enters the mitochondrion is approximately three times greater than the total amount of Ca 2ϩ that accumulates in the cytosol.
Transfer of Ca 2ϩ from an Acidic Compartment to the Mitochondrion-Experiments were initiated to determine if the mitochondrion responded with equal sensitivity when Ca 2ϩ was released from an internal storage site as when Ca 2ϩ entry was induced across the plasma membrane. In T. brucei, an acidic compartment contains a large pool of exchangeable Ca 2ϩ . Both nigericin (7) and monensin (24) were shown to release Ca 2ϩ from this acidic compartment. The mechanism involves H ϩ efflux from the compartment coupled with Ca 2ϩ / nH ϩ exchange to restore the pH gradient (24). In the present study, cells were treated with 2 g/ml monensin in the presence of 5 mM EGTA to chelate the extracellular Ca 2ϩ . Lower resting levels of Ca 2ϩ in both mitochondria and cytosol were observed with EGTA. The [Ca 2ϩ ] cyt and [Ca 2ϩ ] mit were 260 Ϯ 50 nM (n ϭ 16) and 340 Ϯ 30 nM (n ϭ 18), respectively. Upon addition of 2 g/ml monensin, [Ca 2ϩ ] cyt was elevated to 400 Ϯ 50 nM (n ϭ 9) and then returned to the basal level (Fig. 5A). By contrast, [Ca 2ϩ ] mit was altered dramatically by monensin, rising from the resting level to 3.3 Ϯ 1.3 M (n ϭ 16) (Fig. 5B). The peak value of [Ca 2ϩ ] mit was approximately 7-fold higher than [Ca 2ϩ ] cyt . These data demonstrate that the release of Ca 2ϩ from the acidic compartment changes [Ca 2ϩ ] mit much more than [Ca 2ϩ ] cyt . The channeling of Ca 2ϩ from the acidic compartment to the mitochondrion may represent a chemical signaling process in which organelles can communicate with each other by inducing selective changes in the Ca 2ϩ content.
To show that the transient properties of Ca 2ϩ influx into the mitochondrion were not the result of damage to the mitochondrion and to observe changes in cellular homeostasis after the acidic compartment was compromised with monensin, melittin was added to cells after the monensin injection. If the mitochondrial membrane is damaged by monensin, it would not efficiently transport Ca 2ϩ when melittin is applied. For these experiments 1 mM Ca 2ϩ was added to the medium (Fig. 6). The injection of 2 g/ml monensin caused a transient peak in [Ca 2ϩ ] cyt to 560 Ϯ 100 nM (n ϭ 13) (Fig. 6A), and the second addition of 200 nM melittin elevated [Ca 2ϩ ] cyt to the peak value of 1.7 Ϯ 0.9 M (n ϭ 13). This value of [Ca 2ϩ ] cyt is 1.4-fold higher than is observed with melittin alone (Fig. 3B). At the  (Fig. 6B). After the [Ca 2ϩ ] mit returned to a steady level, the second injection of 200 nM melittin increased [Ca 2ϩ ] mit to an average peak value of 7.3 Ϯ 1.3 M (n ϭ 9), which indicated that the mitochondrion was not damaged by the monensin treatment.
Overall, these data demonstrate that large quantities of Ca 2ϩ selectively accumulate in the mitochondrion when homeostasis is disrupted by Ca 2ϩ influx across the plasma membrane or Ca 2ϩ is released from an acidic compartment. The total amount of Ca 2ϩ in the mitochondrion at these times exceeds the quantity in the cytoplasm. These data are consistent with a central role for the mitochondrion in maintaining Ca 2ϩ homeostasis in the primitive protozoan, T. brucei. In addition, Ca 2ϩ transfer to the mitochondrion from the acidic compartment indicates a mechanism by which one organelle might use chemical messengers to regulate the activity of another compartment. DISCUSSION Ca 2ϩ homeostasis is critical for cell survival. Homeostatic pathways protect the cell from toxic effects of excess Ca 2ϩ and serve as the source of Ca 2ϩ pulses during the signal process. In the present study, targeted aequorins were used to directly measure Ca 2ϩ concentrations inside the cellular organelles of live trypanosomes. This approach allows the dynamic interactions between homeostatic compartments to be studied. A mammalian targeting sequence was shown to efficiently target the Ca 2ϩ -sensitive photoprotein aequorin to the mitochondrion of T. brucei. It has recently been shown that a yeast presequence and sequences from trypanosome proteins can also direct reporter proteins into the trypanosome mitochondrion (37,(43)(44)(45). Transport requires ATP, a membrane potential and a protein component on the mitochondrial surface (37,43). Our study shows that the human localization sequence also works in trypanosomes and that the signal sequence is not proteolytically processed upon entry into the mitochondrial matrix space. Evidence of correct localization comes from: (a) immunoblots, (b) immunolocalization, (c) inhibition of aequorin luminescence with FCCP and KCN, and (d) the quantitatively different signal from CYT-AEQ and MT-AEQ, suggesting that they are not in the same compartment of the cell.
Aequorin luminescence can be converted into calculated values for free Ca 2ϩ concentrations using look-up tables or kinetic equations (11-15, 29 -34). We chose an equation described by others (14) because it generated the lowest calculated values for [Ca 2ϩ ]. However, this equation is likely to overestimate the free Ca 2ϩ concentration around 200 nM, because the slope of the curve log␣ versus pCa 2ϩ decreases in this concentration range. Consequently, the values reported with aequorin for basal levels of [Ca 2ϩ ] cyt are higher than those we report using the Fura-2 system (typically in the range of 50 -100 nM) (7).
The resting level of Ca 2ϩ within the mitochondrion varied from 340 to 400 nM depending upon whether the medium was supplemented with EGTA or Ca 2ϩ , respectively. The large single mitochondrion has been estimated by others to occupy approximately 25% of the total cell volume in procyclic cells (42). Consequently, if all of the mitochondrial free Ca 2ϩ were to equilibrate with the cytosol, it would only elevate [Ca 2ϩ ] cyt by around 25 nM. The low level of [Ca 2ϩ ] mit in the resting cell precludes this organelle as a source of Ca 2ϩ for regulatory purposes. However, we demonstrate that the trypanosome mitochondrion in vivo can effectively accumulate Ca 2ϩ , even when [Ca 2ϩ ] cyt is below the set point of 700 -800 nM (4). Similar results have been obtained with mammalian cells by monitoring in situ changes in dihydro-Rhod 2 fluorescence (46) or using targeted aequorins (11)(12)(13)(14). Isolated mitochondria also exhibit the ability to respond rapidly to small Ca 2ϩ pulses (47). The shunting of Ca 2ϩ to the trypanosome mitochondrion occurred whether the Ca 2ϩ pulse originated from the plasma membrane or from an acidic storage compartment. The mechanism of shunting probably involves the sensing of elevated microdomains in the vicinity of the plasma membrane or acidic compartment. The channeling of Ca 2ϩ from the ER to the mitochondrion has been observed in HeLa and other mammalian cells (12,13). The extent of Ca 2ϩ shunting depended upon the proximity of the mitochondrion to the ER membrane (14). ER Ca 2ϩ transport has been reported in permeabilized T. brucei, and a SERCA cDNA encoding a thapsigargin-sensitive ATPase has been cloned (4, 6). However, thapsigargin only released small quantities of Ca 2ϩ from the ER in situ (5). In the present study, thapsigargin did not affect [Ca 2ϩ ] mit as occurs in mammalian cells (data not shown).
The reason why the trypanosome mitochondrion transports Ca 2ϩ is not altogether clear. In mammalian cells, Ca 2ϩ -sensitive dehydrogenases are found in the mitochondrion and Ca 2ϩ import may be a mechanism of stimulating energy metabolism in preparation for a change in cell activity (48). This situation may apply to T. brucei procyclic forms where a complete complement of mitochondrial enzymes occur. However, in bloodstream forms, mitochondrial Ca 2ϩ transport has still been reported (4), although no functional dehydrogenases are found. Therefore, Ca 2ϩ transport into the mitochondrion of these very primitive organisms may regulate other process or be exclusively for the purpose of Ca 2ϩ homeostasis.
Overall, the present report illustrates the dynamic interplay between homeostatic organelles within this important pathogen. During the signal process, the vast majority of Ca 2ϩ accumulates transiently in the mitochondrion until the [Ca 2ϩ ] mit saturates around 8 M. Release of Ca 2ϩ from the mitochondrion does not elevate [Ca 2ϩ ] cyt due in part to the activity of the acidic compartment. Nonetheless, even in the presence of melittin and monensin, [Ca 2ϩ ] cyt still returns to the basal level (Fig.  5A), suggesting that other energy-dependent homeostatic organelles compensate when one system is disrupted. Nonsequestered Ca 2ϩ accumulates in the cytosol, and a portion of this Ca 2ϩ moves into the nucleus (35).