Tim50 in Trypanosoma brucei Possesses a Dual Specificity Phosphatase Activity and Is Critical for Mitochondrial Protein Import*

Background: Tim50 is a preprotein receptor for the TIM23 complex in mitochondria. Results: T. brucei Tim50 (TbTim50), a protein phosphatase, is involved in the regulation of VDAC expression, interacts with TbTim17, and imports precursor proteins into mitochondria. Conclusion: TbTim50 is evolutionarily distinct from Tim50 in other eukaryotes. Significance: Characterization of TbTim50 is critical for understanding mitochondrial protein import in T. brucei. In eukaryotes, proteins are imported into mitochondria via multiprotein translocases of the mitochondrial outer and inner membranes, TOM and TIM, respectively. Trypanosoma brucei, a hemoflagellated parasitic protozoan and the causative agent of African trypanosomiasis, imports about a thousand proteins into the mitochondrion; however, the mitochondrial protein import machinery in this organism is largely unidentified. Here, we characterized a homolog of Tim50 that is localized in the mitochondrial membrane in T. brucei. Similar to Tim50 proteins from fungi and mammals, Tim50 in T. brucei (TbTim50) possesses a mitochondrial targeting signal at its N terminus and a C-terminal domain phosphatase motif at its C terminus. Knockdown of TbTim50 reduced cell growth and inhibited import of proteins that contain N-terminal targeting signals. Co-immunoprecipitation analysis revealed that TbTim50 interacts with TbTim17. Unlike its fungal counterpart but similar to the human homolog of Tim50, recombinant TbTim50 possesses a dual specificity phosphatase activity with a greater affinity for protein tyrosine phosphate than for protein serine/threonine phosphate. Mutation of the aspartic acid residues to alanine in the C-terminal domain phosphatase motif 242DXDX(V/T)246 abolished activity for both type of substrates. TbTim50 knockdown increased and its overexpression decreased the level of voltage-dependent anion channel (VDAC). However, the VDAC level was unaltered when the phosphatase-inactive mutant of TbTim50 was overexpressed, suggesting that the phosphatase activity of TbTim50 plays a role in regulation of VDAC expression. In contrast, phosphatase activity of the TbTim50 is required neither for mitochondrial protein import nor for its interaction with TbTim17. Overall, our results show that TbTim50 plays additional roles in mitochondrial activities besides preprotein translocation.


DXDX(V/T) 246 abolished activity for both type of substrates. TbTim50 knockdown increased and its overexpression decreased the level of voltage-dependent anion channel (VDAC).
However, the VDAC level was unaltered when the phosphataseinactive mutant of TbTim50 was overexpressed, suggesting that the phosphatase activity of TbTim50 plays a role in regulation of VDAC expression. In contrast, phosphatase activity of the TbTim50 is required neither for mitochondrial protein import nor for its interaction with TbTim17. Overall, our results show that TbTim50 plays additional roles in mitochondrial activities besides preprotein translocation.
Trypanosoma brucei is the causative agent of human African trypanosomiasis (African sleeping sickness), which occurs primarily in sub-Saharan Africa (1). According to the World Health Organization, about 60 million people in Africa are at risk for this disease. The parasite also causes a disease known as nagana, which affects livestock (2), thus affecting economic development in this region.
Similar to other eukaryotes, most of the mitochondrial proteins in T. brucei are encoded in the nucleus (3,4). After synthesis on cytoplasmic ribosomes, these proteins must be transported into the mitochondrion. Therefore, mitochondrial protein import is a vital process for growth and survival of this parasite. Although a great deal of research has been done on the mitochondrial protein translocases (TOM 2 and TIM), in fungi and higher eukaryotes (3,5), analogous complexes are not well delineated in trypanosomatids.
Recent studies have characterized an archaic type T. brucei mitochondrial protein (ATOM) that mediates the transport of nucleus-encoded proteins through the mitochondrial outer membrane (6). ATOM performs similar functions of Tom40 found in higher eukaryotes. However, it is still controversial whether ATOM is related to Tom40 or is a bacterial ortholog of Omp85 (6,7).
Tim17 is the most conserved among all known Tom and Tim proteins (8). We have characterized Tim17 in T. brucei (TbTim17) and found it to be an essential integral inner membrane protein (9). Instead of three homologous proteins, such as Tim17, Tim22, and Tim23, that are found in fungi and higher eukaryotes (3)(4)(5), trypanosomatids possess only one homolog of this family, namely Tim17.
In fungi, humans, and plants, mitochondrial inner membrane and matrix-localized proteins are translocated by two TIM complexes (i.e. TIM23 and TIM22) (3,5). The fungal TIM23 complex consists of the core components Tim17, Tim23, and Tim50 and a dynamic component, Tim21 (3,5,10). Both Tim23 and Tim17 have four transmembrane (TM) domains centrally located in these proteins, which are integrated into the mitochondrial inner membrane, leaving both the hydrophilic N and C termini in the intermembrane space (11). In contrast, Tim50 has a single TM near the N terminus and a large hydrophilic C-terminal domain that is exposed within the intermembrane space (12,13). The TIM23 complex is responsible for translocating preproteins that contain an N-terminal targeting sequence to the mitochondrial matrix (3,5). It also translocates some mitochondrial inner membrane proteins that possess an additional sorting signal (14,15). Tim17 associates with Tim23 and plays a role in lateral sorting of preproteins (10). Tim23 forms the channel found in the TIM23 complex (16). In general, Tim50 (i) acts as a receptor for TIM23 substrates (17,18); (ii) mediates translocation of presequence-containing proteins from TOM to the TIM23 complex (17,18); and (iii) gates the TIM23 translocase (19). Tim17, Tim23, and Tim50 are essential in fungi and animals (3,5,12,17,18). Tim21 is not essential and transiently interacts with the TOM complex to facilitate transport of the preproteins from the TOM complex to the TIM23 complex (10,20). It also interacts with the respiratory chain complexes to enable preproteins to cross the mitochondrial inner membrane (10,21).
Homologs of Tim17, Tim23, and Tim50 are structurally and functionally conserved from fungi to mammals (22,23). Several reports have indicated that in higher eukaryotes, Tim50 plays additional roles besides mitochondrial protein import, including effects on steroidogenesis (24), development (25,26), apoptosis (27), and the maintenance of mitochondrial outer membrane integrity (27,28). However, the mechanism by which Tim50 influences these processes is unclear.
Here we show that T. brucei possesses a homolog of Tim50 that is essential for the import of presequence-containing nucleus-encoded mitochondrial proteins. TbTim50 interacts with TbTim17 in vivo. TbTim50 also possesses a dual specificity phosphatase activity, with a greater affinity for tyrosine phosphate than threonine phosphate. We identified the important structural motif and critical amino acid residues for the protein phosphatase activity of TbTim50 and found that TbTim50 phosphatase plays a role in maintaining the level of VDAC.

EXPERIMENTAL PROCEDURES
Strains, Media, and Growth-The procyclic form of T. brucei 427 (29-13) cells, which are resistant to hygromycin and neomycin (G418) and express the tetracycline repressor gene (tetR) and T7RNA polymerase (T7RNAP), were grown in SDM-79 medium containing 10% heat-inactivated fetal bovine serum and antibiotics (hygromycin, 50 g/ml; G418, 15 g/ml) (29). Cell growth was assessed by inoculating the procyclic form at a cell density of 2 ϫ 10 6 /ml in fresh medium containing antibiot-ics in the presence or absence of doxycycline. Cells were counted at different growth time points with a Neubauer hemocytometer. The log of the cumulative cell number was plotted against time of incubation in culture.
Sequence and Computational Analysis-The amino acid sequences of Tim50 proteins from different species were used as queries to search the T. brucei genome database (GeneDB) using BLAST (30) analysis. A sequence comparison among Tim50 proteins from different species was performed using the ClustalW multiple sequence alignment program (31) in MacVector, Inc., version 10.0. The prediction of the transmembrane domain of TbTim50 was performed using the TMpred prediction program (32). Assessment for the presence of a mitochondrial targeting signal was performed using MitoProt (33). Mitochondrial localization was also predicted by Mito-Carta (34) and iPSORT (35). Three-dimensional structure alignment was performed using a Cn3-D software program (36).
Generation of the Inducible TbTim50 RNAi and TbTim50overexpressing Cell Lines-To prepare the construct for TbTim50 double-stranded RNA expression, a 505-bp fragment of the coding region (residues 336 -839) of TbTim50 was PCRamplified from T. brucei genomic DNA using the TaqPCR Master Mix Kit (Qiagen). Sense and antisense primers (supplemental Table 1) containing BamHI and HindIII restriction sites were custom-synthesized. The amplified product was cloned into the BamHI and HindIII restriction sites of a tetracyclineinducible dual promoter plasmid vector, p2T7 Ti -177 (37). This construct generated TbTim50 double-stranded RNA from two opposing tetracycline-regulated T7 promoters and expressed the phleomycin-resistant gene constitutively for selection purposes. The TbTim50 overexpression construct was generated using the modified pLEW100 -3HA vector (a gift from Dr. Xiaoming Tu) (38). The entire open reading frame (ORF) of TbTim50 was PCR-amplified from T. brucei genomic DNA using primers shown in supplemental Table 1. The PCR product was cloned between the HindIII and XhoI restriction sites. From this construct, the C-terminal 3HA-tagged Tim50 was expressed upon induction with doxycycline. The construct for TbTim50 phosphatase-inactive mutant was generated also using the pLEW100 -3HA vector, after generation of the TbTim50 D242A and D244A mutant by site-directed mutagenesis (described below). The constructs for TbTim50 RNAi, TbTim50-3HA, and TbTim50 phosphatase-inactive mutant were verified by sequencing. The purified plasmid DNA was linearized by NotI. The linearized plasmid was used for transfection into the procyclic 29-13 cell line expressing T7 polymerase and tetracycline repressor proteins according to standard protocols (29).
RNA Isolation and Northern Analysis-RNA was isolated from the procyclic form of the trypanosomes, harvested at different time points of growth with or without induction of TbTim50 RNAi, using TRIzol (Invitrogen) according to the manufacturer's protocol. Northern blot analysis was performed as described previously (9). Specific probes were made using a random primer labeling protocol (Invitrogen) from the cDNA fragments generated by PCR using the same primer pairs used Tim50 in T. brucei FEBRUARY 1, 2013 • VOLUME 288 • NUMBER 5 for preparation of RNAi constructs. Ethidium bromide-stained ribosomal RNA served as a loading control.
Expression and Purification of Recombinant TbTim50 (rTb-Tim50) and TbTim17 (rTbTim17)-The ORFs of TbTim50 and TbTim17 were subcloned into a pQE30 (Qiagen) bacterial expression vector at the BamHI and HindIII site and transfected into Escherichia coli strain M15 (Qiagen). The His6tagged recombinant proteins were expressed upon induction with 0.5 mM isopropyl ␤-D-1-thiogalactopyranoside for 4 h at 30°C. Cells were harvested from 100 ml of culture, and proteins were solubilized in 4 ml of lysis buffer (50 mM Tris-HCl, pH 8.0, 300 mM NaCl, 0.1% BME, 2 mM EDTA, 4 mM MnCl 2 , 1.0 mg/ml lysozyme, 10 g/ml RNase, 5 g/ml DNase, 1% Triton X-100, 10 mM imidazole, 1 g/ml leupeptin, and 1 mM PMSF). After incubation for 30 min on ice, solubilized proteins were clarified by centrifugation at 13,000 ϫ g for 30 min at 4°C. The supernatant was loaded onto a pre-equilibrated nickel-nitrilotriacetic acid (Ni-NTA)-agarose column. The column was washed with 3 volumes of lysis buffer without lysozyme and the bound proteins were eluted with the same buffer containing 250 mM imidazole. Samples were analyzed by Coomassie Brilliant Blue staining to ensure purity of the eluted proteins. The purified proteins were also tested by immunoblot analysis using anti-TbTim50, anti-TbTim17, and anti-RGS-His (Qiagen) antibodies (data not shown).
Site-directed Mutagenesis-The ORF of TbTim50 was subcloned in a TOPO PCR4 vector (Invitrogen). Site-specific mutagenesis of D242A and D244A was performed using primers designed by QuikChange Primer Design Tools (Agilent Technologies) and a QuikChangeII site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. Desired mutations were confirmed by sequencing. The mutated ORF of TbTim50 was then subcloned into pQE30 vector within BamHI and HindIII sites for expression of the recombinant protein.
Generation of TbTim50 Specific Antibodies and Immunoblot Analysis-Polyclonal antibodies against the TbTim50 protein were generated by Antagene Inc. (Sunnyvale, CA), using a peptide consisting of amino acids 70 -86 of the N terminus as the antigen. The synthetic peptide was purified and conjugated to keyhole limpet hemocyanin using cysteine. The keyhole limpet hemocyanin-conjugated TbTim50 peptide was then used to generate specific antibodies in rabbits. Antiserum against TbTim50 was affinity-purified using the TbTim50 peptide as the ligand. Preimmune serum collected from the same rabbits was used as a control. Total cellular proteins and proteins from isolated mitochondria were analyzed by SDS-PAGE (10 or 12%) and transferred to nitrocellulose membrane at 4°C in transfer buffer (25 mM Tris-HCl, 192 mM glycine, 20% (v/v) methanol, pH 8.3) and 100 V, as described previously (9). Blots were treated with purified TbTim50 polyclonal antibodies (1:100 dilution) and antibodies against T. brucei serine/threonine protein phosphatase 5 (TbPP5) (39), porin (VDAC) (40), Tim17 (TbTim17) (9), mitochondrial Hsp70 (mHsp70) (41), and ATP/ ADP carrier protein (TbAAC) (9), all at a 1:1000 dilution in TBST (10 mM Tris-HCl, 150 mM NaCl, pH 8.0). A monoclonal antibody against T. brucei ␤-tubulin was used at a 1:10,000 dilution in TBST (42). A monoclonal antibody against trypanosome alternative oxidase (TAO) (43) and a mouse monoclonal antihemagglutinin antigen (HA) epitope (clone 12CA5, Roche Applied Science) antibody were used at a 1:50 and 1:1000 dilution in TBST, respectively. Blots were treated with the appropriate secondary antibody and developed using an ECL system (GE Healthcare or Thermo Fisher Scientific Inc). The intensity of the bands was quantified using ImageJ (44) and normalized with the corresponding ␤-tubulin bands.
Subcellular Fractionation-Fractionation of TbTim50-3HA-expressing cells and parental procyclic cells were performed as described (45). Briefly, 2 ϫ 10 8 cells were resuspended in 500 l of SEMP buffer (20 mM MOPS/KOH, pH 7.4, 250 mM sucrose, 2 mM EDTA, 1 mM PMSF) containing 0.03% digitonin and incubated on ice for 5 min. The cell suspension was then centrifuged for 5 min at 6800 ϫ g at 4°C. The resultant pellet was the crude mitochondrial fraction, and the supernatant contained the soluble cytosolic proteins.
Isolation of Mitochondria-Mitochondria were isolated from the parasite after lysis by nitrogen cavitation in isotonic buffer as described previously (9,46). The isolated mitochondria were stored at a protein concentration of 10 mg/ml in SME buffer (250 mM sucrose, 20 mM MOPS-KOH, 2 mM EDTA, pH 7.4) containing 50% glycerol at Ϫ70°C. Before use, mitochondria were washed twice with 9 volumes of SME buffer to remove glycerol.
Sodium Carbonate Extraction of Mitochondria-Isolated mitochondria (100 g) from T. brucei were treated with 100 mM Na 2 CO 3 (100 l) at pH 11.5 for 30 min on ice (9,46). The supernatant and pellet fractions were collected after centrifugation and analyzed by SDS-PAGE and immunoblotting.
Co-immunoprecipitation-Mitochondrial proteins (500 g) isolated from wild type control cells and procyclic cells expressing the TbTim50-3HA were solubilized with lysis buffer (10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM PMSF, 1% digitonin, and 3 mg/ml BSA). The proteins were then subjected to immunoprecipitation with anti-HA-conjugated Sepharose beads or anti-TbTim17-coupled agarose beads at 4°C overnight. The immunoprecipitation product was washed sequentially with lysis buffer and basal buffer (10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM PMSF). Precipitated proteins were analyzed by SDS-PAGE (12%) and immunoblotting. In order to test the binding of rTbTim50 with TbTim17, we mixed equal volumes of detergent (1% digitonin)-solubilized mitochondrial extract with both the purified wild type and mutant rTbTim50. Proteins were precipitated with Ni-NTA-agarose. After washing, proteins were eluted from the agarose beads by boiling with Laemmli buffer. Precipitated and unprecipitated proteins were analyzed by SDS-PAGE and immunoblotted using RGS-His 6 , TbTim17, and TAO antibodies as probes.
MitoTracker Red Staining, in Situ Immunofluorescence, and Confocal Microscopy-Cells were harvested and suspended in fresh culture medium at a density of 4 -5 ϫ 10 6 ml Ϫ1 . Cells were stained with MitoTracker Red, fixed, and permeabilized as described previously (9). Briefly, MitoTracker Red CMXROS (Invitrogen) was dissolved in DMSO at a concentration of 1 mM and added to a final concentration of 0.5 M. The mixture was incubated at 37°C for 10 min. Cells were washed and incubated in fresh culture medium for an additional 30 min. After another Tim50 in T. brucei wash in PBS, the cell suspension was spread on polylysinecoated slides. Cells were then washed twice with PBS and fixed in 3.7% paraformaldehyde at 4°C for 5 min. After blocking with 5% nonfat milk for 30 min, anti-HA antibodies at a dilution of 1:1000 in PBS were applied to the slide for 1 h. Slides were then washed with PBS, and fluorescein isothiocyanate (FITC)-conjugated mouse IgG was applied as a secondary antibody. All images were acquired with a Nikon A1R laser-scanning confocal microscope and oil immersion objective lens. MitoTracker Red was excited at 561 nm, and the emitted light was collected between 570 and 620 nm. FITC was excited at 488 nm, and the emitted light was collected between 500 and 550 nm.
FACS Analysis-Cells were harvested and suspended in fresh culture medium at a density of 4 -5 ϫ 10 6 ml Ϫ1 . Cells were stained with MitoTracker Red as described above. Cells were centrifuged, washed, and resuspended in cold PBS and stored at 4°C until FACS analysis. Fluorescence intensity was measured with a FACSCalibur (BD Biosciences) analytical flow cytometer using absorption at 578 nm and emission at 599 nm. CellQuest software (BD Biosciences) was used to analyze the results.
In Vitro Transcription and Translation-The ORFs for cytochrome oxidase subunit IV (COIV), AAC, and VDAC were subcloned in a pGEM-4Z (Promega Corp.) vector as described previously (46,47). The ORF of mitochondrial RNA-binding protein 2 (MRP2) (48)was amplified from genomic DNA of T. brucei using primers found in supplemental Table 1. The PCR product was subcloned into the pGEM-4Z (Promega Corp.) vector between HindIII and BamHI. The TAO-dihydrofolate reductase (DHFR) fusion construct was also generated in the pGEM-4Z (Promega Corp.) vector. The TAO ORF was amplified from the TAO cDNA clone (pTAO25) (43) using forward and reverse primers containing the HindIII and BamHI restriction sites at the 5Ј-ends, respectively. The mouse DHFR ORF was PCR-amplified using the pQE16 vector (Qiagen) as the template and the forward and reverse primers containing BamHI and EcoRI restriction sites at the 5Ј-ends, respectively. The PCR products were then ligated and cloned into the pGEM-4Z (Promega Corp.) vector between the HindIII and EcoRI sites to form the TAO-DHFR fusion construct. Radiolabeled precursors COIV, AAC, VDAC, MRP2, and TAO-DHFR were synthesized in vitro using a quick coupled transcription/ translation rabbit reticulocyte system (TNT, Promega Corp.) according to the manufacturer's protocol. In Vitro Import Assay-Radiolabeled precursor proteins were used for in vitro import into isolated mitochondria from T. brucei as described previously (9,46,47). Proteins were separated by SDS-PAGE and transferred to nitrocellulose membrane. After transfer, the blot was dried at 37°C for 30 min and exposed to x-ray film (Kodak BioMax MR film, Sigma-Aldrich) for detection of radioactive proteins. For MRP2, COIV, and TAO-DHFR, the processed matured proteins of smaller molecular mass were generated during import. Postimport sodium carbonate extraction was performed for AAC and VDAC to remove unimported and peripherally associated proteins; residual proteins were considered imported protein.
Phosphatase Assay-Phosphatase activity was measured in phosphatase buffer (50-l reaction mixtures containing Tris acetate, pH 7.4, 1 mM EDTA, 1 mM EGTA, 0.1% BME, and 0.1% ethanol). The substrate for phosphatase activity, p-nitrophenyl phosphate (pNPP), was used at a concentration of 400 mM. Aliquots of 0 -60 g of enzyme (purified recombinant proteins) were preincubated for 1 min at 30°C. The reaction was initiated by adding a mixture of pNPP at 30°C. The reactions were quenched by adding 450 l of 0.25 N sodium hydroxide. Release of p-nitrophenol was determined by measuring the phosphatase activity at A 410 . The extinction coefficient for p-nitrophenol used was 17.8 ϫ 10 3 mole/liter. Protein-tyrosine phosphatase activities of the recombinant Tim50 proteins were measured against 0 -500 M phosphotyrosyl peptides, END-pYINASL, and protein serine/threonine phosphatase activity was measured against phosphothreonyl peptides, RRApTVA, using a protein phosphatase assay system (Promega), according to the manufacturer's protocol.

RESULTS
T. brucei Contains a Homolog of Tim50-Through BLAST analysis of the T. brucei genome (58), using known Tim50 proteins as the query sequences, we identified a gene, Tb927.3.2110, with a high probability score. Tb927.3.2110 encodes a protein of predicted molecular mass of 47.2 kDa. We termed this homolog TbTim50. Using several bioinformatics applications, we compared the primary and secondary structure of TbTim50 to Tim50 proteins that have been studied previously in other organisms. ClustalW alignment of Tim50 protein sequences revealed that TbTim50 shares an overall 13.3-16.7% identity and 23.9 -31.1% similarity with Tim50 proteins from other species (supplemental Fig. 1 and supplemental Table 2). The primary sequence of TbTim50 most strongly resembled that of human Tim50 with an identity of ϳ17% and a similarity of ϳ31%. The C-terminal half was more conserved than the N-terminal half of Tim50 proteins among different species. The alignment also revealed a unique region, 259 DLV-LDVRMENTSTTVY 274 , in the center of the protein (supplemental Fig. 1). Orthologs of TbTim50 are conserved in other kinetoplastid parasites, such as Trypanosoma cruzi and different Leishmania species as found by analysis using GeneDB.
A major characteristic of the Tim50 protein is a C-terminal domain phosphatase motif, which is similar to the transcription factor IIF (TFIIF)-stimulated CTD phosphatase. The TFIIFstimulated CTD phosphatase belongs to a special group of protein phosphatases that dephosphorylates the C-terminal domain of the RNA polymerase II subunit (49). The CTD phosphatase-like domain has also been found in proteins at different subcellular locations (49,50). During BLAST analysis, we found eight T. brucei genes besides Tb927.3.2110 that encode proteins containing the TFIIF-stimulated CTD phosphatase motif. The gene identification numbers for these are Tb927. 10 individually using Tim50 proteins from all of the organisms as a query. Proteins encoded by Tb09.160.4460 and Tb09.160.4480 are the same; therefore, these are two copies of a gene. Further analysis revealed that the protein encoded by Tb09.160.4460/ Tb09.160.4480 genes is not involved in mitochondrial protein import in T. brucei (data not shown). Based on these results, we concluded that Tb927.3.2110 is the closest homolog of Tim50 in T. brucei.
Using TMpred prediction software, hydropathy analysis of TbTim50 indicated the presence of a single TM domain (amino acid residues 285-310), consisting of 25 amino acids in the C-terminal region of the protein. The position of the TM domain in TbTim50 is atypical, given that all Tim50 proteins from other eukaryotes possess a single TM domain that is located in the N-terminal region (Fig. 1A). Together, these results indicated that the membrane topology and some regions of the primary structure of TbTim50 are distinct from Tim50 proteins in fungi and animals ( Fig. 1B and supplemental Fig. 1).
TbTim50 Is Located in the Mitochondrial Membrane-Analysis of TbTim50 by MitoProt, a software program that predicts mitochondrial targeting signals and cleavage sites, identified an N-terminal sequence of 22 amino acids that resembles a mitochondrial targeting signal (51). In TbTim50, this sequence consists of six basic amino acid residues; the putative cleavage site possesses an "R (Ϫ2)" motif (52). MitoCarta and iPSORT software programs also predicted its mitochondrial localization.
To confirm the subcellular localization of TbTim50, we analyzed total, cytosolic, and mitochondrial fractions by immuno-blot analysis using anti-TbTim50 antibody. Fig. 2A shows that TbTim50 was enriched in the mitochondrial fraction, although a smaller proportion of this protein was detected in the cytosolic fraction. This portion might represent contamination of the mitochondrial fraction, or it is possible that a small fraction of TbTim50 was released into the cytosol during cell fractionation. The size of the protein band detected by this antibody is ϳ52 kDa, which is in contrast to the calculated size of the protein of 47.2 kDa. The reason for this discrepancy is not clear at this time. However, it could be due to a post-translational modification of TbTim50. As expected, other mitochondrial proteins, such as VDAC (an outer membrane protein), TAO and TbTim17 (inner membrane proteins), and mHsp70 (a matrix protein), were found exclusively in the mitochondrial fraction ( Fig. 2A). In these studies, the cytosolic protein, TbPP5, was used as a marker for the cytosolic fraction. Sodium carbonate extraction of isolated mitochondria demonstrated that TbTim50 was present primarily in the alkali-resistant membrane pellet, similar to two other integral membrane proteins, VDAC and TAO, suggesting that it is an integral membrane protein. As expected, mHsp70 is primarily found in the alkalisoluble fraction (Fig. 2B).
TbTim50-3HA Is Targeted to the Mitochondria in T. brucei-The C-terminal 3HA-tagged TbTim50 was expressed in T. brucei from an inducible expression vector. After induction, a protein of ϳ54 kDa was expressed. The anti-HA monoclonal antibody recognized this protein only in cells that were induced with doxycycline (Fig. 3A). Anti-TbTim50 antibody also recog-

Tim50 in T. brucei
nized this ectopically expressed protein along with the endogenous TbTim50, as expected. Overexpression of TbTim50 did not show any significant effect on cell growth (data not shown). Subcellular fractionation followed by immunoblot analysis of the total, cytosolic and mitochondrial fractions using anti-HA and anti-TbTim50 antibodies showed that TbTim50-3HA was localized in the mitochondrial fraction, similar to the endogenous protein (Fig. 3B). In situ immunofluorescence microscopy results also confirmed that the C-terminally 3HA-tagged TbTim50 is localized in T. brucei mitochondrion because it showed a complete overlap of anti-HA and MitoTracker stains only in cells that expressed TbTim50-3HA but not in the wild type control (Fig. 3C).
Knockdown of TbTim50 by RNAi Suppressed Cell Growth and the Mitochondrial Membrane Potential-To understand the function of TbTim50, we induced the expression of TbTim50-specific double-stranded RNA in the procyclic form of the parasite with doxycycline. The region of TbTim50 chosen for double-stranded RNA expression was specific for TbTim50 and not present in other genes that also encode proteins containing the CTD phosphatase-like domain. Northern blot analysis revealed that there was ϳ80% reduction in the transcript of TbTim50 at day 2, indicating successful knockdown of TbTim50 (Fig. 4A). To confirm this knockdown at the protein level, whole-cell lysates were analyzed after induction with doxycycline. Immunoblot analysis with anti-TbTim50 antibody revealed that the protein was reduced ϳ50 and ϳ70% by day 2 and 4, respectively (Fig. 4B). These results also confirmed that the 52-kDa protein recognized by anti-TbTim50 antibody is the product of Tb927.3.2110.
To determine the effect of TbTim50 knockdown on parasite growth, cells were counted at different time points after induction with doxycycline. Knockdown of TbTim50 reduced cell growth within 6 days but did not totally cease growth (Fig. 4C). These results indicate that TbTim50 is required for optimal cell growth and suggest that cells may still be able to grow at a slower rate even with a partial loss of TbTim50.
To evaluate the effect of knockdown of TbTim50 on the mitochondrial membrane potential, we chose to quantify and analyze the uptake of MitoTracker Red in cells using flow cytometry. This analysis revealed that the mitochondrial membrane potential was unaffected within 2 days but was reduced about 25% at 4 days after induction of TbTim50 RNAi in comparison with uninduced cells (Fig. 4, D and E). As a control, the mitochondrial membrane potential of the parental wild type cells was also assessed by MitoTracker staining followed by FACS analysis. As expected, the mitochondrial membrane potential of these cells was not altered by doxycycline treatment but was reduced more than 50% after treatment with carbonyl cyanide m-chlorophenyl hydrazone (CCCP) (Fig. 4F).
Knockdown of TbTim50 Increased and Overexpression of TbTim50 Decreased the Level of VDAC-To understand the effect of knockdown or overexpression of TbTim50 on other nucleus-encoded mitochondrial proteins, the levels of such proteins were analyzed by immunoblot analysis using mitochondria from uninduced controls, TbTim50 knockdown, and TbTim50-overexpressed cells (Fig. 5, A and B). The TbTim50 level was reduced by 47% in mitochondria isolated from   FEBRUARY 1, 2013 • VOLUME 288 • NUMBER 5

Tim50 in T. brucei
TbTim50 knockdown cells but was increased about 60% in mitochondrial samples from cells where TbTim50 was overexpressed. Surprisingly, the level of VDAC was significantly upregulated by 69% (Fig. 5, A and C) due to TbTim50 knockdown and was decreased about 80% (Fig. 5, B and D) due to TbTim50 overexpression. Knockdown and overexpression of TbTim50 had a minimal effect on the expression level of the other mitochondrial proteins, such as TAO, TbTim17, AAC, and mHsp70 (Fig. 5, A-D). Although ␤-tubulin is not a mitochondrial protein, it is always present in our mitochondrial preparation as an associated protein and was therefore used as a loading control. These results indicated that there is a connection between TbTim50 and VDAC. There are no reports of such a link between these two proteins in other systems.
TbTim50 Is Required for the Import of N-terminal Signal Sequence Containing Nucleus-encoded Mitochondrial Proteins-To evaluate whether TbTim50 is involved in mitochondrial protein import, in vitro import assays were performed using isolated T. brucei mitochondria and radiolabeled precursor proteins. To determine the substrate specificity of TbTim50, we selected four nucleus-encoded mitochondrial proteins in T. brucei, MRP2, COIV, AAC, and VDAC, as well as a fusion protein, TAO-DHFR. MRP2 is an N-terminal signal-containing mitochondrial matrix protein. COIV is a mitochondrial inner membrane protein containing an N-terminal targeting signal.
TAO-DHFR is a fusion protein composed of mitochondrial inner membrane protein TAO and a cytosolic protein, DHFR, from mice. We used this protein as one of our substrates because it can be radiolabeled at a much higher intensity due to the presence of more methionine residues. Therefore, we could see results after a much shorter exposure period compared with the times required for endogenous proteins, such as MRP2 and COIV. AAC is also a mitochondrial inner membrane protein that does not have an N-terminal targeting signal, whereas VDAC is a mitochondrial outer membrane protein. In fungi and higher eukaryotes, it has been shown that Tim50 is responsible for the import of N-terminal presequence-containing proteins that are translocated via the TIM23 complex. The carrier family proteins, such as AAC, are imported by the TIM22 complex; thus, they do not depend on TbTim50. Therefore, in vitro import analysis of the above mentioned proteins would reveal if TbTim50 has a particular specificity with regard to any of these substrate proteins.
Reduction of TbTim50 by RNAi inhibited import of MRP2 about 80% within 10 min (Fig. 6, A and F). Import of COIV was also inhibited due to induction of TbTim50 RNAi for 48 h; however, the percentage of inhibition (60% compared with controls) was less dramatic than that for MRP2 (Fig. 6, B and G). Import analysis of the TAO-DHFR in control mitochondria revealed that this fusion protein was imported in a time-dependent manner and that the signal sequence of TAO was pro- cessed as expected (Fig. 6C). Further TAO-DHFR import was strongly inhibited in TbTim50 knockdown mitochondria (Fig.  6, C and H). In contrast, the imports of AAC and VDAC were minimally affected by knockdown of TbTim50 (Fig. 6, D, E, I,  and J). These results indicate that TbTim50 is necessary for import of the N-terminal signal-containing nucleus-encoded mitochondrial proteins.
TbTim50 Interacts with TbTim17 in Vivo-In fungi, Tim50 is one of the core components of the TIM23 complex. The intermembrane space domain of ScTim50 directly interacts with the N-terminal region of ScTim23 protein. Because TbTim17 is the only homolog of Tim17/23 family protein and is also involved in the import of N-terminal signal-containing proteins in T. brucei (9), we wanted to assess, using co-immunoprecipitation assays, whether TbTim50 interacts with TbTim17 in vivo. We observed that the anti-HA monoclonal antibody efficiently pulled down TbTim50-3HA as well as a significant portion of TbTim17 from mitochondrial samples that were prepared from TbTim50-3HA-expressing cells but not from control mitochondrial samples (Fig. 7A). As a negative control, we reprobed these blots with anti-TAO monoclonal antibody, and no protein was detected in the precipitated samples. A reverse pulldown using anti-TbTim17 antibody also showed that ectopically expressed TbTim50 was precipitated along with TbTim17 (Fig. 7B). TAO was not detected in the immunoprecipitate either from control or TbTim50-3HA-containing mitochondrial proteins (Fig. 7B). These results indicate that TbTim50 and TbTim17 specifically interact in vivo.
TbTim50 Possesses a Dual Specificity Phosphatase Activity-In order to determine if TbTim50 possessed phosphatase activity, we purified rTbTim50 from E. coli using Ni-NTA-agarose chromatography. We also purified recombinant rTbTim17 under similar conditions as a control. Each purified protein showed a single protein band of the expected size on a Coomassie-stained gel (Fig. 8A). Phosphatase assays using pNPP revealed that rTbTim50 was capable of releasing inorganic phosphate from this substrate and that the activity showed a linear relationship with time (Fig. 8B). The activity reached a maximum at 15 min. The phosphatase activity also increased linearly with an increasing amount of rTbTim50 (Fig. 8C). In contrast to rTbTim50, rTbTim17 and heat-killed rTbTim50 did not show any phosphatase activity (Fig. 8, B and C). The rTbTim50 exhibited a pH optimum (i.e. pH 5) similar to the reported value for recombinant human Tim50 (27,28). Interestingly, previous studies show that fungal Tim50 did not have any phosphatase activity (53). To determine substrate specificity, we performed phosphatase assays using two peptides, phosphotyrosyl and phosphothreonyl peptides (Fig. 8E). rTbTim50 showed activity with both substrates; however, it had a higher affinity for phosphotyrosyl peptide than phosphothreonyl peptide.
Aspartate Residues Asp 242 and Asp 244 Are Critical for Phosphatase Activity of TbTim50-Structure-function analysis of CTD phosphatase-like domain-containing proteins, such as SCP1, reported a conserved motif, DXDX(V/T), that is critical for phosphatase activity. The first aspartate in this signature motif undergoes metal-assisted phosphorylation during catalysis (49). By analysis of the protein sequence of TbTim50, we found two such motifs. The first motif, 242 DLDET 246 , is present in the region upstream of the single TM domain of this protein, and the second motif, 346 DLDRV 350 , is located downstream of the TM domain of the protein. Interestingly, these sequence motifs are not fully conserved in Tim50 proteins from fungi and  FEBRUARY 1, 2013 • VOLUME 288 • NUMBER 5 humans, because only a partially conserved motif near the C-terminal region is found in each of these proteins. Moreover, there is no report of the functional significance of this motif in Tim50 from either fungi or humans. In order to understand if the signature motif 242 DLDET 246 serves as the active site of the TbTim50 phosphatase, we replaced two aspartate residues in this motif with alanine by site-directed mutagenesis. The mutated protein was expressed and purified from bacterial lysate (supplemental Fig. 3), and phosphatase activity was assessed using both phosphotyrosyl and phosphothreonyl peptides as substrates. Results showed a total loss of activity due to these mutations ( Table 1), suggesting that either one or both of these aspartate residues (Asp 242 and Asp 244 ) are critical for TbTim50 phosphatase activity. We also found that both wild type and mutant rTbTim50 efficiently co-precipitate TbTim17 from mitochondrial extract (supplemental Fig. 4). Therefore, the loss of activity is not due to incorrect folding of the mutant TbTim50. Together, these results confirmed that 242 DLDET 246 is indeed the active site of this phosphatase and the phosphatase activity is not involved for interaction of TbTim50 with TbTim17.

Tim50 in T. brucei
The Phosphatase Activity of TbTim50 Is Involved in Regulation of VDAC Protein Level but Not in Mitochondrial Protein Import-In order to understand the physiological function of the TbTim50 phosphatase activity in T. brucei, we overex-

Tim50 in T. brucei
pressed the D242A/D244A mutant of TbTim50 in T. brucei. The mutant protein was expressed and targeted to mitochondria similar to the endogenous TbTim50 (Fig. 9A). Overexpression of the D242A/D244A mutant of TbTim50 did not show any significant effect on cell growth (data not shown). Next we compared the level of VDAC in mitochondria after overexpression of either wild type or phosphatase-inactive mutant of TbTim50. Overexpression of the phosphatase-inactive mutant TbTim50 did not show a similar reduction on the steady state level of the VDAC in mitochondria as found when we expressed the wild type TbTim50 (Fig. 9, B and C). There was no effect observed on the levels of TbTim17, AAC, and Hsp70 in mitochondria due to overexpression of the phosphatase-inactive mutant of TbTim50. These results indicate that the phosphatase activity of TbTim50 is possibly involved in post-translational modification of VDAC, which in turn may modulate the level of this protein.
In order to understand if the phosphatase activity plays role on translocation of precursor proteins into mitochondria, we compared the protein import efficiency of mitochondria possessing wild type and mutant TbTim50 using in vitro import assays. We did not find any effect on the import of proteins COIV and AAC into both types of mitochondria (Fig. 9, D-G), suggesting that the phosphatase activity of TbTim50 does not play any role in mitochondrial protein import in T. brucei.

DISCUSSION
TbTim50, the interacting partner of TbTim17, was identified and functionally characterized. Unlike its fungal counterpart, the rTbTim50 showed dual specificity phosphatase activity.
The active site motif of the CTD phosphatase-like domain is conserved in TbTim50. The TbTim50 was required for import of presequence-containing nucleus-encoded proteins into mitochondria. The phosphatase activity of TbTim50 is involved in regulation of the VDAC protein level, but it is not required for mitochondrial protein import.
The components of T. brucei mitochondrial protein translocases are currently being revealed via functional proteomics and reverse genetic approaches performed by different investigators (6,54). Recently, we showed that TbTim17 is present in a larger protein complex in the mitochondrial membrane; we also identified a number of trypanosome-specific proteins that are associated with TbTim17 (54). Here, we identified another interacting partner of TbTim17, namely TbTim50.
In fungi and higher eukaryotes, the N-terminal domain of Tim23 directly interacts with the Tim50 C-terminal domain that is exposed in the intermembrane space (55). Although a true homolog of Tim23 is absent in T. brucei, kinetoplastid parasites like T. brucei possess a single homolog of the Tim17/ 22/23 family of proteins referred as TbTim17 (9). TbTim17 is an important protein translocator in the mitochondrial inner membrane (9). Here, we found that TbTim50 is also involved in mitochondrial protein import and interacts with TbTim17, suggesting that these two proteins are possibly present in the same protein complex. However, determination of whether these two proteins directly or indirectly interact with each other requires further investigation. We observed that only a fraction of TbTim17 was pulled down by an anti-HA antibody from the mitochondrial extracts containing TbTim50-3HA. Thus, it is possible that TbTim50 is only dynamically associated with TbTim17 during protein import. In addition, we did not find TbTim50 in a tandem affinity-purified TbTim17 protein complex (54). This may be because the interaction between these two proteins is weak, and the interaction could not resist the harsh conditions of double affinity chromatography. Furthermore, the mechanism of interaction of these two proteins is further complicated by the predicted membrane topology of TbTim50, which is found to be different from that for Tim50 in other organisms. Therefore, it would be interesting to know which regions of these proteins are involved in protein/protein interaction.
Depletion of TbTim50 specifically inhibited the import of N-terminal targeting signal-containing proteins. These results are similar to that found for Tim50 proteins in fungi and in humans. Therefore, regarding mitochondrial protein import, TbTim50 function was found to be conserved. A comparison of the predicted three-dimensional structure of TbTim50 with the known structure of the soluble domain of the fungal Tim50 (56) revealed that TbTim50 possesses a conserved presequencebinding region (supplemental Fig. 5A), which supports our observation related to the specificity of import substrates. It has been reported that ScTim50 interacts with ScTim23 via two protruded ␤-strands (56). These two ␤-strands are also structurally conserved in human Tim50 (supplemental Fig. 5B). However, this conserved structure is not found in TbTim50 (supplemental Fig. 5A). Thus, TbTim50 lacks this interacting domain, which is not surprising because T. brucei also lacks a homolog of Tim23. Interestingly, this upstream non-overlap- FIGURE 7. Co-immunoprecipitation of TbTim50 with TbTim17. Mitochondrial proteins (500 g) were solubilized with digitonin (1%) at a protein concentration of 1 mg/ml. The lysate from T. brucei 427 (29-13) (WT Control) and TbTim50-3HA mitochondria were subjected to immunoprecipitation (IP) with anti-HA-Sepharose beads (A) or anti-TbTim17-agarose beads (B). Five percent of the total soluble proteins (Input) and 50% of precipitated proteins (Pellet) were analyzed by SDS-PAGE (12%) and immunoblotted with anti-TbTim17 antibody or anti-HA antibody as indicated. TAO antibody was used as a negative control. FEBRUARY 1, 2013 • VOLUME 288 • NUMBER 5 ping region of TbTim50 possesses a unique sequence (amino acid residues 259 -274), as mentioned earlier. We thus postulated that this structurally distinct region of TbTim50 could be critical for the interaction with TbTim17 and/or other trypanosome-specific protein(s).

Tim50 in T. brucei
A CTD phosphatase-like domain has been found to be conserved in Tim50 proteins characterized thus far (12,13). However, the function of this domain and the active site of this phosphatase have not been determined. Fungal Tim50 does not have any phosphatase activity; however, human Tim50 pos-sesses a phosphatase activity (27) similar to what we found for TbTim50. The active site of the CTD phosphatase, like SCP1, is aspartate-based and contains a conserved motif DXDX(T/V) (49). We found two such perfect motifs, 242 DLDET 246 and 346 DLDRV 350 , in TbTim50. Mutation of the first motif completely abolished the phosphatase activity of TbTim50, suggesting that this motif is the catalytic active site of TbTim50 CTD phosphatase.
We found that the forced expression and knockdown of TbTim50 inversely regulated the level of VDAC in mitochondria. A similar observation has not been reported in other systems. Upon further investigation, we noticed that although VDAC protein levels were significantly changed as a function of TbTim50 levels, the expression of the VDAC transcript remained unaltered, whereas TbTim50 levels increased or decreased (data not shown). Moreover, we found that the reduction of TbTim50 inhibited neither the import of VDAC nor its membrane integration ( Fig. 6E and supplemental Fig. 2). These results suggest that a post-translational modification dependent on TbTim50 expression could be responsible for the stability of VDAC. We presented evidence that TbTim50 possesses an active phosphatase domain. In addition, we showed FIGURE 8. Purification and phosphatase assays of His 6 -tagged recombinant TbTim50. A, recombinant TbTim50 (rTbTim50) and TbTim17 (rTbTim17) were expressed in Escherichia coli and purified by nickel chelation chromatography. Purified proteins (2 g/lane) were analyzed on SDS-PAGE and stained with Coomassie Brilliant Blue. B, the phosphatase activity of rTbTim50 (20 and 40 g) and rTbTim17 (40 g) was measured using pNPP (0.4 M) as the substrate for different time periods (0 -15 min). C, rTbTim50 phosphatase assays were performed using pNPP (0.4 M) as the substrate and varying amounts (0 -60 g) of purified protein for two time points (10 and 15 min). Varying amounts (0 -60 g) of heat-denatured rTbTim50 (heat-killed) were used in parallel as a control. D, rTbTim50 phosphatase activity was measured at various pH values (pH 3-9) using pNPP (0.4 M) as the substrate in the presence of 150 mM NaCl. E, protein phosphatase activity of purified rTbTim50 was measured using phosphotyrosyl and phosphothreonyl peptide as substrates (0 -500 M). The data shown are the averages of triplicate experiments. Error bars, S.D.

TABLE 1
The effect of D242A and D244A mutations on TbTim50 phosphatase activity The recombinant TbTim50 wild type and TbTim50 mutant were expressed in E. coli and purified by nickel chelation chromatography. Protein phosphatase activity of the purified wild type and mutated proteins (40 g) was measured using phosphotyrosyl and phosphothreonyl peptide as substrates (300 M). Assays were done in triplicate, and S.D. values were calculated.

Tim50 in T. brucei
that overexpression of the phosphatase-inactive mutant of TbTim50 in T. brucei was unable to decrease the VDAC protein level in mitochondria, similar to the effect of overexpression of the wild type TbTim50. Therefore, it is possible that TbTim50 either directly or indirectly regulates the phosphorylation status of VDAC, a process known to be a regulatory factor for VDAC stability in other eukaryotic cells (57). It is also possible that dephosphorylation of other proteins by TbTim50 increases the turnover rate of VDAC, thus regulating its expression level in T. brucei. From all of these observations, we speculate that changes in VDAC levels could be a cellular adaptation to deal with stress due to alterations of TbTim50 protein levels in T. brucei. It has been reported in humans and in Drosophila that Tim50 is involved in other functions besides mitochondrial protein import, such as apoptosis and maintenance of mitochondrial outer membrane integrity (24 -28). It has also been reported in Drosophila that different Tim50 isomers are expressed in a tissue-specific manner, and in zebrafish, TbTim50 plays a crucial role during embryonic development (25). However, it is unclear if a Tim50 phosphatase activity plays a role in any of these functions. Here, we showed additional evidence that the phosphatase activity of TbTim50 is involved in the regulation of VDAC protein levels in mitochondria. Because the VDAC levels are dramatically altered in different developmental stages of T. brucei (40), it can be speculated that TbTim50 may also play a role in the developmental regulation of mitochondrial activities in T. brucei critical for parasite survival.