Identification of Mammalian TOM22 as a Subunit of the Preprotein Translocase of the Mitochondrial Outer Membrane*

A mitochondrial outer membrane protein of ∼22 kDa (1C9-2) was purified from Vero cells assessing immunoreactivity with a monoclonal antibody, and the cDNA was cloned based on the partial amino acid sequence of the trypsin-digested fragments. 1C9-2 had 19–20% sequence identity to fungal Tom22, a component of the preprotein translocase of the outer membrane (the TOM complex) with receptor and organizer functions. Despite such a low sequence identity, both shared a remarkable structural similarity in the hydrophobicity profile, membrane topology in the Ncyt-Cin orientation through a transmembrane domain in the middle of the molecule, and the abundant acidic amino acid residues in the N-terminal domain. The antibodies against 1C9-2 inhibited the import of a matrix-targeted preprotein into isolated mitochondria. Blue native polyacrylamide gel electrophoresis of digitonin-solubilized outer membranes revealed that 1C9-2 is firmly associated with TOM40 in the ∼400-kDa complex, with a size and composition similar to those of the fungal TOM core complex. Furthermore, 1C9-2 complemented the defects of growth and mitochondrial protein import in Δtom22 yeast cells. Taken together, these results demonstrate that 1C9-2 is a functional homologue of fungal Tom22 and functions as a component of the TOM complex.

Most mitochondrial proteins are encoded by the nuclear genome and are synthesized in the cytosol as preproteins. They are guided to the mitochondrial surface by cytoplasmic chaperones (1) and are then transported to the intramitochondrial compartments by the preprotein import machinery of the outer and inner membranes: the TOM 1 and TIM complexes, respectively (2)(3)(4)(5)(6)(7)(8). Extensive genetic and biochemical studies in Sac-charomyces cerevisiae and Neurospora crassa have identified components of these complexes. The S. cerevisiae TOM complex is composed of at least nine proteins, Tom71, Tom70, Tom40, Tom37, Tom22, Tom20, Tom7, Tom6, and Tom5. Tom70, Tom37, Tom22, and Tom20 function as import receptors. Tom71 has strong similarity to Tom70 and is weakly associated with the TOM complex (9), although its function is unclear. Tom40 is deeply embedded in the outer membrane in a predicted ␤-barrel structure and functions as the central component of the translocation channel (10 -14). Tom6 and Tom7 modulate the dynamics of the TOM channel (15,16). Tom5 is tightly associated with Tom40 and represents the connecting link between import receptors and the translocation channel (17). Thus, Tom40, Tom22, and three smaller Tom proteins form the general preprotein import pore (the TOM core complex) of ϳ400 kDa (14,18). A recent study revealed that Tom22 not only functions as the import receptor, but also regulates the TOM complex organization (19). The mitochondrial inner membrane has at least two separate import machineries. The Tim23-Tim17 system mediates, in conjunction with Tim44 and mHsp70 in the matrix, mitochondrial transport of matrix-targeted preproteins (8,20). The Tim54-Tim22-Tim18 system acts with the small Tim proteins in the intermembrane space to mediate insertion of the metabolite carrier proteins, such as the ADP/ATP carrier, as well as several TIM components, such as Tim23, Tim22, and Tim17, into the inner membrane (21)(22)(23)(24)(25)(26)(27).
Despite extensive knowledge of fungal systems, relatively little is known about the import machinery of mammalian mitochondria. Several mammalian counterparts have been identified (for review, see Ref. 28): TOM20 (29 -32); TIM17, TIM23, and TIM44 (33); and DDP1, a homologue of Tim8 (34). In addition, TOM34 and metaxin have been identified as unique components of the mammalian mitochondrial import system. Human TOM34 was cloned using a degenerate tetratricopeptide repeat sequence present in Tom70 and Tom20 (35). Metaxin has a 25% sequence identity to yeast Tom37 in the N-terminal region, although whether it is a mammalian counterpart of Tom37 is not yet known (36). TOM34 and metaxin might function as the import receptor. Furthermore, we have recently identified and characterized the rat homologue of Tom40 2 and rat TOM70. 3 This study reports the identification and characterization of a mammalian homologue of Tom22. This protein was found serendipitously during the study of the diphtheria toxin recep-tor complex using monoclonal antibodies prepared against the membrane fraction of Vero cells as antigens (37,38). We found that one of the monoclonal antibodies reacted with the ϳ22-kDa mitochondrial outer membrane protein 1C9-2. The cDNA was cloned based on the partial amino acid sequence of the purified protein. Its primary sequence was 19 -20% identical to that of fungal Tom22, the multifunctional regulator of the protein translocase complex of the mitochondrial outer membrane (the TOM complex). Despite this low sequence identity, 1C9-2 and fungal Tom22 shared characteristic structural similarities in the hydropathy profile, the membrane topology, and the distribution of acidic amino acid clusters in the extramembrane domains. Furthermore, 1C9-2 is a component of the mammalian TOM complex containing TOM40 and several components of smaller molecular size, and the anti-1C9-2 antibody inhibited mitochondrial protein import. Most notably, it complemented the functional defect in ⌬tom22 yeast cells. We conclude that 1C9-2 is the functional homologue of Tom22.

EXPERIMENTAL PROCEDURES
Monoclonal Antibody-BALB/c mice were immunized by subcutaneous injection of the alkali-extracted membrane fraction from Vero cells (monkey kidney-derived cells) (37). Spleen cells from the immunized mice were fused with X63-Ag8-653 mouse myeloma cells as described previously (38), and the hybridoma producing an antibody reacting with an ϳ22-kDa mitochondrial protein was selected. We refer to this antigen as 1C9-2.
Immunofluorescence Microscopy-Vero cells were plated at a density of 1 ϫ 10 5 cells/well the day before staining. The cells were incubated with 300 nM MitoTracker Green FM (Molecular Probes, Inc.) for 40 min, fixed in 3.7% formaldehyde in phosphate-buffered saline, and permeabilized with 0.1% Triton X-100 in phosphate-buffered saline. The cells were incubated with anti-1C9-2 monoclonal antibody for 30 min at room temperature and then with Cy3-conjugated donkey anti-mouse IgG antibody for 30 min at room temperature.
Purification and Amino Acid Sequencing of Monkey 1C9-2-Vero cells (1 ϫ 10 9 ) were disrupted by freeze-thawing in 20 mM Tris-HCl buffer (pH 7.4) containing a complete protease inhibitor mixture (Roche Molecular Biochemicals) and centrifuged at 1,700 ϫ g for 10 min. The precipitate was washed with the same buffer; incubated with 20 mM Tris-HCl buffer (pH 7.4) containing 150 mM NaCl, 60 mM octyl glucopyranoside, and the complete protease inhibitor mixture at 0°C for 2 h; and centrifuged at 27,000 ϫ g for 30 min. The supernatant was diluted with the same volume of acetonitrile containing 0.1% trifluoroacetic acid and applied to a Waters Sep-Pak tC 2 cartridge. The cartridge was washed with 50% acetonitrile containing 0.1% trifluoroacetic acid and eluted with 60% acetonitrile containing 0.1% trifluoroacetic acid. The eluate was diluted with 0.5 volume of deionized water containing 0.1% trifluoroacetic acid and subjected to reverse-phase high pressure liquid chromatography (HPLC) using a 5TMS-MS column (4.6 ϫ 50 mm; Nacalai Tesque) at a flow rate of 1 ml/min with a 40 -80% acetonitrile linear gradient containing 0.1% trifluoroacetic acid. The 1C9-2-enriched fraction, as detected by Western blotting, was lyophilized and separated by SDS-polyacrylamide gel electrophoresis (PAGE) (15% gel). After staining with Coomassie Brilliant Blue, the 1C9-2 band was excised and subjected to in-gel digestion with trypsin (39). The digested sample was resolved by reverse-phase HPLC using a 5C18-AR II column (2.1 ϫ 150 mm; Nacalai Tesque) at a flow rate of 0.2 ml/min with a linear gradient from deionized water containing 0.085% trifluoroacetic acid to 42% acetonitrile and 18% 2-propanol containing 0.0718% trifluoroacetic acid. The separated peptides were subjected to amino acid sequence analysis using an Applied Biosystems 473A Protein Sequencer.
cDNA Cloning of Human 1C9-2-5Ј-RACE and 3Ј-RACE were performed using a cDNA library from CMK cells (human myeloid leukemia cell line) as the template and primers that were designed according to the human expressed sequence tag clones (GenBank TM /EBI Data Bank accession number AA316462 for the longest clone): 5Ј-GCCCCTGGCA-TTCCTCCTGAGAGCCC-3Ј for the first amplification of the 5Ј-end, 5Ј-GGTCCCAATCCACAAGGCTGCCCTGG-3Ј for the second amplification of the 5Ј-end, 5Ј-CCCGAAAGGCGACGCGGAGAAGCC-3Ј for the first amplification of the 3Ј-end, and 5Ј-GAGACCCTGTCGGAGAGAC-TATGG-3Ј for the second amplification of the 3Ј-end.
Submitochondrial Fractionation-Rat liver mitochondria (40) were swollen in 10 mM HEPES-KOH (pH 7.3) containing 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and the protease inhibitor mixture (10 g/ml each chymostatin, pepstatin, antipain, and leupeptin) on ice for 30 min and then sonicated. The solution was layered over a 12-ml linear gradient of 0 -1.6 M sucrose in 10 mM HEPES-KOH (pH 7.3) containing the protease inhibitor mixture. After centrifugation at 100,000 ϫ g for 18 h, 0.5-ml fractions were collected from the top of the tube.
Determination of Topology of 1C9-2 in Mitochondria-1C9-2 carrying the C-terminal FLAG epitope tag was used to determine the membrane orientation of 1C9-2. A linker encoding the 8-amino acid FLAG tag sequence (DYKDDDDK) was cloned into the XbaI-ApaI site of pRC/ CMV (Invitrogen) to obtain pRC/CMV/FLAG. cDNA encoding human 1C9-2 was inserted into the HindIII-XbaI site of pRC/CMV/FLAG to generate pRC/CMV/h1C9-2/FLAG. L cells (mouse fibroblast-derived cells) were plated at a density of 5 ϫ 10 5 cells/10-cm dish the day before transfection. The cells were transiently transfected with 20 g of pRC/ CMV/h1C9-2/FLAG per dish using calcium phosphate. Two days after transfection, the cells were collected, homogenized in 10 mM HEPES-KOH (pH 7.4) containing 0.22 M mannitol and 0.07 M sucrose by repeated aspiration through a 27-gauge needle, and centrifuged at 1,000 ϫ g for 10 min. The supernatant was divided into five aliquots and centrifuged at 10,000 ϫ g for 10 min to obtain the mitochondrionenriched precipitates. One aliquot was used for a protein assay using Coomassie protein assay reagent (Pierce), and the other aliquots were suspended in 10 mM HEPES-KOH (pH 7.4) containing 0.22 M mannitol and 0.07 M sucrose; in 10 mM HEPES-KOH (pH 7.4); or in 10 mM HEPES-KOH (pH 7.4) containing 0.22 M mannitol, 0.07 M sucrose, and 1% Triton X-100. The mixtures were incubated with 20 g/ml proteinase K at an enzyme/protein ratio of 1:50 on ice for 30 min. After boiling for 5 min, the reaction mixtures were analyzed by SDS-PAGE and immunoblotting using mouse anti-FLAG monoclonal antibody (Sigma) or anti-1C9-2 monoclonal antibody.
Blue Native Gel Analysis-Blue native PAGE was performed essentially as described previously (41,42). The mitochondrial outer membranes (50 g) were solubilized with 50 l of 10 mM HEPES-KOH (pH 7.4) containing 2% (w/v) digitonin, 200 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, and 10% (v/v) glycerol at 0°C for 30 min, and insoluble material was removed by centrifugation for 15 min at 100,000 ϫ g. The supernatant was mixed with 5 l of sample buffer (100 mM BisTris (pH 7.0) containing 5% Coomassie Brilliant Blue G-250 and 500 mM 6-aminocaproic acid) and electrophoresed through a 5-16% polyacrylamide gradient gel. Each lane of the first dimension gel was excised and subjected to second dimension Tricine/SDS-PAGE.
Complementation of ⌬tom22 Yeast Cells-For constitutive expression of 1C9-2 in yeast cells, cDNA encoding human 1C9-2 was inserted downstream of the alcohol dehydrogenase promoter of the yeast expression vector pMD288 (2, TRP1) to create pMD288/h1C9-2. The yeast strain MNMS-1C, which has a chromosomal TOM22 gene disrupted by HIS3 that is rescued by pYE-Ura3:TOM22, in which TOM22 is placed under a GAL1 promoter (43), was transformed with either pMD288 or pMD288/h1C9-2 and selected using synthetic medium plates (lacking His, Ura, and Trp) containing 2% galactose. The transformants were streaked onto the synthetic medium plates (lacking His, Ura, and Trp) containing either 2% galactose or glucose and incubated at 30°C for 3 days. To completely deplete the plasmid-borne yeast Tom22, the yeast cells harboring pMD288 or pMD288/h1C9-2 grown on the first 2% glucose-containing synthetic medium plate were streaked onto the second 2% glucose-containing synthetic medium plate and incubated at 30°C for 3 more days. Similarly, the cells grown on the first 2% galactose-containing synthetic plate were replated onto the second 2% galactose-containing synthetic medium plate and incubated at 30°C for 3 more days. The effect of the expression of human 1C9-2 on mitochondrial import of pre-Hsp60 was examined as follows. MNMS-1C cells harboring pMD288 or pMD288/h1C9-2 were grown in 2% galactosecontaining synthetic medium at 30°C for 12 h. The cells were harvested by centrifugation, suspended in 2% galactose-or glucose-containing synthetic medium, and incubated at 30°C for 72 h with medium exchanges at 24 and 48 h of incubation. Cells were subjected to SDS-PAGE followed by Western blotting with monoclonal antibodies against human 1C9-2 or Hsp60 or with polyclonal antibodies against yeast Tom22.

RESULTS
Isolation of cDNA Clone Encoding 1C9-2-We obtained a monoclonal antibody reacting with an ϳ22-kDa protein (1C9-2) of mammalian cell mitochondria during the course of screening monoclonal antibodies raised against the membrane fraction of Vero cells (monkey kidney-derived cells) for those recognizing the diphtheria toxin receptor complex. Immunofluorescence microscopy revealed that 1C9-2 colocalized with MitoTracker (Fig. 1A). Western blot analysis with this antibody revealed that 1C9-2 was expressed ubiquitously in various human tissues as well as in several cultured cell lines (Fig. 1B). 1C9-2 was purified from cultured Vero cells, and amino acid sequences of several trypsin-digested fragments were determined. A data base search revealed that the sequences (boxed in Fig. 2A) are all found in human expressed sequence tag clones (GenBank TM /EBI Data Bank accession number AA316462 for the longest clone). Because the assembled sequences of these clones did not contain an in-frame initiator methionine, 5Ј-RACE and 3Ј-RACE were performed using a cDNA library of CMK cells (human myeloid leukemia-derived cells). Human 1C9-2 cDNA thus obtained coded for a 15521-Da protein containing 142 amino acid residues (Fig. 2A). The Nterminal region, which was missing in the data base, was characteristic in that it contains highly negative charges. The entire sequence of 1C9-2 was not found in the cDNA data banks, but the gene encoding the entire sequence was deposited in the GenBank TM /EBI Data Bank as clone 508I15, which lies within human chromosome 22q12-13 (accession number AL021707). Comparison of both nucleotide sequences revealed that the 1C9-2 gene consists of four exons (Fig. 2B). The exon junctions are shown in Fig. 2A (arrowheads). The data base search for homologous proteins revealed a Caenorhabditis elegans protein with 27% sequence identity (accession number AAB71053), which was annotated to have sequence similarity to Tom22 from S. cerevisiae. The predicted sequence of human 1C9-2 showed overall identities of 19 and 20% to Tom22 from S. cerevisiae and N. crassa, respectively. Although the sequence identity was low, these proteins shared a significant structural similarity in the distribution of clusters of acidic amino acid residues along the extramembrane domains (Fig. 3A) and in the hydropathy profile (Fig. 3B). The hydropathy plot shows the presence of two hydrophilic segments (residues 1-82 and 102-142), which are separated by a hydrophobic segment of 19 amino acid residues. Of note, the N-terminal segment is extremely rich in acidic amino acid residues (21 negative charges in 82 total residues). The C-terminal tails of N. crassa and S. cerevisiae Tom22, however, have an overall negative charge with a net charge of Ϫ5. In marked contrast, 1C9-2 has a C-terminal tail with a neutral net charge.
Intracellular Localization of 1C9-2-We examined the intracellular localization of 1C9-2 in rat liver cells by Western blotting using anti-human 1C9-2 polyclonal antibody. As shown in Fig. 4A, this antibody recognized 1C9-2 in rat liver cells, although with a lower reactivity. Using cDNA cloning, we confirmed the presence of the rat counterpart with a sequence FIG. 4. Subcellular and submitochondrial localization and membrane topology of 1C9-2. A, subcellular localization 1C9-2 in rat liver cells. 10 g each of the mitochondrial, microsomal, and cytosolic fractions were resolved by SDS-PAGE and subjected to Western blot analysis with the indicated antibodies. Because only antibodies against rat liver marker proteins were available to us, we used rat liver for the fractionation experiments. Several nonspecific bands were observed with polyclonal antibodies against human 1C9-2 because of poor reactivity with the rat counterpart. k, kilodaltons; MSF(L), mitochondrial import stimulation factor large subunit. B, submitochondrial localization of 1C9-2. The rat liver submitochondrial fractions obtained after sucrose density gradient centrifugation were resolved by SDS-PAGE and then subjected to Western blot analysis with the indicated antibodies. T, unfractionated membranes (2.5 g) from rat liver mitochondria. C, topology of 1C9-2 in the mitochondrial outer membrane as probed by proteinase K susceptibility. C-terminal FLAG-tagged 1C9-2 was expressed in L cells as described under "Experimental Procedures." The mitochondrion-rich fraction was isolated; divided into aliquots; and treated with (lanes 2-4) or without (lanes 1) 20 g/ml proteinase K at 0°C for 30 min in isotonic buffer (lanes 2), in hypotonic buffer (lanes 3), or in isotonic buffer containing 1% Triton X-100 (lanes 4). Each reaction mixture was divided into two aliquots and resolved by SDS-PAGE. The gels were subjected to Western blot analysis using anti-FLAG monoclonal antibody (left panel) or anti-1C9-2 monoclonal antibody (right panel). Note that the anti-1C9-2 monoclonal antibody did not react with endogenous 1C9-2 in L cells. identity of 93.6%. 4 1C9-2 cofractionated with rat TOM40, the central component of the protein import channel of the mitochondrial outer membrane, but not with microsomal cytochrome P450 (44) or with a large subunit of the mitochondrial import stimulation factor in the cytosol (40,45) (Fig. 4A). Submitochondrial fractionation using sucrose density gradient centrifugation indicated that 1C9-2 cosedimented with rat TOM40, but not with rat TIM23, a component of the protein import machinery of the inner membrane (33) (Fig. 4B).
We then analyzed the topology of 1C9-2 within mitochondria using proteinase K as the probe (Fig. 4C). For this purpose, mouse L cells were used because anti-1C9-2 monoclonal antibody does not recognize rodent 1C9-2 (data not shown), but detects only human 1C9-2 expressed in L cells. Human 1C9-2 carrying a FLAG tag at the C terminus was expressed in L cells, and the mitochondrion-rich fraction was prepared. As shown in Fig. 4C (lanes 2), proteinase K treatment produced a 1C9-2 fragment with slightly increased mobility by removal of the small N-terminal region while the C-terminal FLAG tag remained attached to the fragment. The fragment that had lost the entire extramitochondrial segment could not be detected for unknown reasons. Under hypotonic conditions, which disrupt the outer membrane but maintain the integrity of the inner membrane (33), proteinase K digestion produced an ϳ17-kDa fragment that had lost the FLAG tag (Fig. 4C, lanes 3), and this fragment was completely digested by proteinase K in the presence of Triton X-100 (lanes 4). These results indicated that 1C9-2 is inserted into the outer membrane with the N-terminal portion protruding to the mitochondrial surface and the Cterminal portion protruding to the intermembrane space (Ncyt-Cin orientation). Thus, 1C9-2 has the same membrane topology as fungal Tom22. The ϳ17-kDa fragment seems to be protected against proteinase K within the TOM complex (see below) until the complex is dissociated by Triton X-100.
1C9-2 Is a Component of the ϳ400-kDa TOM Complex Containing Tom40 -In fungal mitochondria, Tom22 forms the TOM core complex of ϳ400 kDa with Tom40 and three smaller Tom proteins (14,18,46). Because the results described above suggest that 1C9-2 is the mammalian counterpart of fungal Tom22, we examined whether it is contained in the complex associating with several mammalian TOM components identified so far. The outer membrane of rat liver mitochondria was solubilized with 2% digitonin; subjected to blue native PAGE, which allows for separation of the protein complex under native conditions (41,42,46); and then resolved in the second dimension by Tricine/SDS-PAGE (Fig. 5). The polyclonal antibody against 1C9-2 recognized rat 1C9-2. 1C9-2 was electrophoresed slightly faster in the Tricine/SDS-PAGE system than in the Tris/SDS-PAGE system for unknown reasons (see Fig. 4A for comparison). As is clearly shown, rat TOM40 migrated as a complex of ϳ400 kDa, and 1C9-2 was exclusively contained in this complex. The interaction between rat TOM40 and 1C9-2 in the digitonin-solubilized membrane was also demonstrated by immunoprecipitation with anti-rat TOM40 antibody (data not shown). On the other hand, the rat TOM20 and TOM70 import receptors mostly dissociated from the complex (Fig. 5).
Anti-1C9-2 Polyclonal Antibody Inhibits Preprotein Import into Mitochondria in Vitro-After confirming 1C9-2 to be a component of the mammalian TOM complex associating with rat TOM40, we then examined inhibition of protein import using antibodies against 1C9-2 with the mitochondrion-enriched fraction prepared from HeLa cells. Polyclonal antibodies raised against 1C9-2 inhibited the import of pre-Su9-DHFR (fusion protein between the presequence of F 0 -ATPase subunit 9 and dihydrofolate reductase) to a weak but significant extent (Fig. 6).
1C9-2 Complements Defects of Growth and Mitochondrial Protein Import in ⌬tom22 Yeast Cells-To further confirm the function of 1C9-2 in vivo, we tested whether 1C9-2 could complement functional defects in ⌬tom22 yeast cells. ⌬tom22 yeast cells carrying the GAL1 promoter-regulated expression plasmid for yeast Tom22 (pYE-Ura3:Tom22) (43) were transformed either with the alcohol dehydrogenase promoter-driven expression vector for 1C9-2 (pMD288/h1C9-2) or with the empty vector. As shown in Fig. 7A, the ⌬tom22 yeast cells harboring pMD288/h1C9-2 grew on a glucose-containing synthetic medium plate, whereas those harboring the empty vector (pMD288) did not. We then examined whether 1C9-2 complemented the mitochondrial protein import defect in ⌬tom22 cells. As shown in Fig. 7B, when ⌬tom22 cells harboring pMD288 were grown in glucose-synthetic medium, the precursor form of Hsp60 accumulated in the cells. In marked contrast, the precursor form of Hsp60 was expressed at a significantly reduced level in ⌬tom22 cells harboring pMD288/h1C9-2. Western blot analysis revealed that 1C9-2 was expressed only in the pMD288/h1C9-2-transformed cells (Fig. 7B). Endogenous Tom22 was undetectable under these conditions. We conclude that 1C9-2 complements the defect of mitochondrial function in ⌬tom22 yeast cells. DISCUSSION We have identified an ϳ22-kDa human mitochondrial outer membrane protein (1C9-2) as a component of the TOM complex. The amino acid sequence of 1C9-2 has an overall identity of 19% to S. cerevisiae Tom22. Despite the low identity of the primary sequence, they have several structural similarities: transmembrane orientation, hydropathy profile, and distribution of highly acidic amino acid regions in the N-terminal segment. Consistent with these features, 1C9-2 complemented the defects of both growth and protein import in ⌬tom22 yeast cells. We thus conclude that 1C9-2 is the functional homologue of Tom22. To our knowledge, this is the first report on the identification of mammalian TOM22 as a subunit of the TOM complex, and we refer to this protein as human TOM22. We also cloned rat TOM22 cDNA, which codes for a 142-amino acid protein with a sequence identity of 93.6% to human TOM22. Blue native PAGE of the digitonin-solubilized outer membrane of rat liver mitochondria revealed that rat TOM22 is firmly associated with rat TOM40 and forms the ϳ400-kDa TOM complex. In addition to these components, several components ϳ5-10 kDa in molecular size, OM10, OM7.5, and OM5, have 4 H. Suzuki, K. Saeki, and K. Mihara, unpublished observations. FIG. 5. Mammalian TOM complex as analyzed by two-dimensional blue native PAGE. The outer membranes of rat liver mitochondria were solubilized by 2% digitonin and subjected to first dimension blue native PAGE and then to second dimension Tricine/SDS-PAGE. After electroblotting the filter was cut between the 25-and 33-kDa molecular size markers. The upper filter was incubated sequentially with antibodies against rat TOM40 and TOM70, whereas the other half was incubated sequentially with antibodies against 1C9-2 and rat TOM20. The immunoreacted proteins were visualized by enhanced chemiluminescence. A nonspecific band cross-reacted with antirat TOM40 IgG and is shown by an asterisk.
been detected in the rat TOM complex. 2 The TOM complex purified from potato mitochondria is ϳ230 kDa in size and has unique subunit compositions: Tom40, Tom20, Tom7, and four other components of smaller molecular size, but the counterpart of Tom22 was absent (47). Therefore, the mammalian TOM complex resembles the fungal TOM complex in size and composition, although the smaller molecular size components remain to be characterized.
The surface-exposed cis-site and the intermembrane spaceexposed trans-site act in series to drive the translocation of preproteins across the outer membrane (48). Tom20 and Tom22 provide the cis-site, and preproteins bind to this site through weak electrostatic interactions (48 -50). The intermembrane space-exposed segment of Tom40 and the C-terminal tail of Tom22 might contribute to the trans-site (51)(52)(53). Highly acidic segments are frequently observed in many components of the mitochondrial import system, and it has been speculated that a preprotein with a basic N-terminal mitochondrion-targeting signal is transported across the outer membrane by sequential binding to a relay of acidic receptor sites (the acid chain hypothesis) (54,55). In fact, the N-terminal cytoplasmic domain of Tom22 is conserved among species carrying highly abundant acidic amino acid residues: the contents of acidic residues in this region are 22.6, 25.2, and 25.6% for N. crassa, S. cerevisiae, and humans, respectively. The conserved highly acidic region might participate as the cis-binding site in the recognition and transfer of the basic N-terminal mitochondrion-targeting signal of the preprotein (54,55), although this issue is still controversial (51,56). The intermembrane spaceexposed C-terminal tail of fungal Tom22 is rich in acidic amino acid residues: the contents of acidic residues are 18.4% (net charge, Ϫ5) and 21.2% (net charge, Ϫ5) for N. crassa and S. cerevisiae, respectively. The significance of the acidic amino acid residues in this region in protein translocation, however, remains controversial (43,51,53,57). In contrast, the C-terminal tail of mammalian TOM22 does not carry net negative charges: the acidic amino acid content is 7.31% (net charge, 0). It should be noted, in this relation, that the C-terminal tail of C. elegans TOM22 carries a net positive charge. Thus, the importance of the C-terminal tail of Tom22 in preprotein import into mitochondria should be further examined. More extensive studies using the purified domain or liposome-reconstituted vesicles with defined TOM components are required.
In summary, we have identified TOM22 in organisms other than S. cerevisiae and N. crassa. These results, in conjunction with the fact that several mammalian homologues (TOM20, TOM40, and several TIM proteins) have been identified, indicate that the mitochondrial import machineries of the outer and inner membranes are evolutionarily conserved among eukaryotic organisms. Subtle variations do exist, however, among different species. Future work is required to clarify the functional significance of these differences among species.

FIG. 7. Complementation of the functional deficiency in ⌬tom22 yeast cells by human 1C9-2.
A, complementation of the growth defect in ⌬tom22 yeast cells. MNMS-1C cells transformed with the empty vector pMD288 or with pMD288/h1C9-2 (the vector harboring human 1C9-2 cDNA) were streaked onto a 2% galactose-or glucosecontaining synthetic medium plate and incubated at 30°C for 3 days. The cells grown on the first 2% galactose-or glucose-containing synthetic medium plate were streaked onto the second 2% galactose-or glucose-containing synthetic medium plate, respectively, and grown at 30°C for 3 more days. B, effect of the expression of human 1C9-2 on mitochondrial import of pre-Hsp60 in MNMS-1C cells. MNMS-1C cells carrying pMD288 or pMD288/h1C9-2 were grown in 2% galactosecontaining synthetic medium at 30°C for 12 h. The cells were isolated by centrifugation, suspended in 2% galactose-or glucose-containing synthetic medium at the same cell density, and incubated at 30°C for an additional 72 h with two medium exchanges. The same number of cells were subjected to SDS-PAGE followed by Western blotting with the monoclonal antibodies against Hsp60 and h1C9-2 and with the polyclonal antibodies against yeast Tom22 (yTom22). p, pre-Hsp60; m, mature Hsp60.