The C-terminal Region of TIM17 Links the Outer and Inner Mitochondrial Membranes in Arabidopsis and Is Essential for Protein Import*

The translocase of the inner membrane 17 (AtTIM17-2) protein from Arabidopsis has been shown to link the outer and inner mitochondrial membranes. This was demonstrated by several approaches: (i) In vitro organelle import assays indicated the imported AtTIM17-2 protein remained protease accessible in the outer membrane when inserted into the inner membrane. (ii) N-terminal and C-terminal tagging indicated that it was the C-terminal region that was located in the outer membrane. (iii) Antibodies raised to the C-terminal 100 amino acids recognize a 31-kDa protein from purified mitochondria, but cross-reactivity was abolished when mitochondria were protease-treated to remove outer membrane-exposed proteins. Antibodies to AtTIM17-2 inhibited import of proteins via the general import pathway into outer membrane-ruptured mitochondria, but did not inhibit protein import via the carrier import pathway. Together these results indicate that the C-terminal region of AtTIM17-2 is exposed on the outer surface of the outer membrane, and the C-terminal region is essential for protein import into mitochondria.

Mitochondria were derived from a single endosymbiotic event that occurred approximately one billion years ago (1,2). The loss of genes from the mitochondrial progenitor and transfer of genes to the host nucleus require that mitochondria rely on the import of nuclear-encoded cytosolic-synthesized proteins (3). This necessitated the development of specific targeting signals to direct proteins to the mitochondrion and an import apparatus to recognize and translocate the mitochondrial subset of proteins from all cytosolic-synthesized proteins (4 -6). The completion of several genome sequencing projects allows a comparative genomic approach to be taken to identify and characterize the import apparatus of various organisms using the extensively studied yeast system as a template (7)(8)(9)(10).
The import of proteins into mitochondria is achieved by a number of oligomeric protein complexes on the outer and inner mitochondrial membranes. A single translocase of the outer membrane (TOM) 1 recognizes all proteins destined for mitochondria, mediates their passage through the outer membrane, and subsequently passes them to other proteins complexes (11,12). ␤-Barrel proteins destined for the outer membrane are passed to the sorting and assembly machinery of the outer membrane (SAM) (13)(14)(15)(16). Signal anchored proteins in the outer membrane are imported via a distinct pathway (17). The general import pathways describes the import of proteins targeted to the translocase of the inner membrane 17:23 (TIM17: 23) complex that is responsible for the import of proteins that contain N-terminal, generally cleavable, targeting sequences termed presequences. Proteins can be sorted to the inner membrane by this pathway via conservative or stop-transfer mechanisms (7,11). The carrier import pathway describes the insertion of inner membrane proteins that contain internal targeting signals and that contain multiple transmembrane regions. These proteins are passed from TOM to small intermembrane space proteins, which act as chaperones to mediate transfer to the TIM22 complex on the inner membrane (18). TIM22 and TIM23 are voltage-activated channels on the inner membrane that are activated by internal and N-terminal targeting signals respectively (10,19).
Comparison of the protein import apparatus of yeast with that in mammals and plants indicated that the basic machinery is generally well conserved between diverse eukaryotic lineages (8,9,20,21). However differences do exist in specific components. Notably, orthologs for TOM70, TIM12, -18, -54, and TOB38 have not been identified in the Arabidopsis genome (22). Additionally, even when components appear shared at a functional level between organisms, they can be unrelated at a sequence level e.g. TOM20 from plants is not orthologous to yeast TOM20 as it displays low significant sequence similarity and significant structural differences (23). Interestingly, plant TOM22 lacks the cis-receptor domain that in yeast is a binding point for precursor proteins that have been recognized by TOM20. This modification may have been necessary to ensure the specificity of targeting to mitochondria and plastids in plant cells (24,25). After initial recognition of precursor proteins by TOM20 in yeast they are proposed to be passed onto a series of binding sites located in the TOM and TIM17:23 complexes. This model was initially called the acid chain hypothesis but subsequently modified to the binding chain hypothesis to account for the important role of hydrophobic as well as ionic interactions (26 -29). Characterization of TIM17 and TIM23 from Arabidopsis indicated that although they are clearly orthologous to yeast import components they appear to be quite distinct. Structurally Arabidopsis TIM23 lacks a complete preprotein and amino acids transporter (PRAT) domain and does not contain the heptad leucine repeat domain (20,26,30). Furthermore, TIM17 contains a significant extension at the C-terminal end not present in fungi and mammals, but genome sequence information indicates this extension is present in other plant TIM17. The Arabidopsis orthologs can only complement yeast mutants when these differences are changed to resemble the yeast structure (30). This suggests that these components play different or additional roles in plants.
Here we characterize the topology of TIM17 from plants in particular reference to the large C-terminal extension that is absent in yeast and mammalian TIM17. We show that this extension is located in the outer membrane and is essential for import of proteins via the general import pathway in plants.

MATERIALS AND METHODS
Isolation of Mitochondria and in Vitro Import of Radiolabeled Precursor Proteins-Precursor proteins for the alternative oxidase (AOX) (31), the F A d subunit of ATP synthase (F A d) (32), adenine nucleotide translocator (ANT), phosphate translocator (Pic) (33), and translocase of the inner membrane 17 (AtTIM17-2) (30), were produced in a coupled TNT transcription-translation system according to the manufacturer's instructions (Promega).
Percoll-purified Arabidopsis mitochondria were isolated according to Millar et al. (34). Outer membrane-ruptured mitochondria (Mit*OM) were prepared for two purposes: (i) to carry out in vitro imports to test for insertion into or across the inner membrane, and (ii) after the import assay to test for the intraorganelle location of imported protein. Rupture of the outer membrane allowed access of added protease to intermembrane space components or inner membrane proteins exposed to the intermembrane space. Mit*OM were prepared by resuspending 100 g of mitochondrial protein in 10 l of SEH buffer (250 mM sucrose, 1 mM EDTA, 10 mM Hepes, pH 7.4) and then adding 155 l of 20 mM Hepes, pH 7.4 and incubating on ice for 20 min. To restore osmolarity, 25 l of 2 M sucrose and 10 l of 3 M KCl were added and mixed, re-pelleted, and washed in import buffer. This procedure removes cytochrome c but not outer membrane components as determined by blotting with outer membrane markers (35). These outer membrane-ruptured mitochondria were either used in in vitro import assays or were subjected to protease treatment. Import assays into mitochondria were carried out with 200 g of mitochondrial protein in each import assay. Import assays were carried out at 26°C for 20 min unless otherwise stated. Import assays were carried out in 15-ml polypropylene tubes in a volume of 200 l in 0.3 M sucrose, 50 mM KCl, 10 mM MOPS, pH 7.2, 5 mM KH 2 PO 4 , 1% (w/v) bovine serum albumin, 1 mM MgCl 2 , 1 mM methionine, 0.2 mM ADP, 0.75 mM ATP, 5 mM succinate, and 5 mM dithiothreitol. Assays were stopped by placing on ice for 3 min, dividing samples into two 100-l aliquots, and adding proteinase K (PK) to one aliquot to a final concentration of 40 g/ml. PK digestion was stopped by the addition of phenylmethylsulfonyl fluoride to 2 mM after 15 min. Mitochondria were diluted into 1 ml of ice-cold import buffer and pelleted by centrifugation for 3 min in a microcentrifuge. Mitochondrial pellets were resuspended in SDS-PAGE gel sample buffer, subjected to SDS-PAGE, gels were dried, and products were visualized by exposing to a BAS TR2040 plate for 24 h and reading in a BAS 2500 Bio imaging analyzer (Fuji, Tokyo). A similar procedure was utilized for import into outer membraneruptured mitochondria except that the mitochondrial outer membrane was ruptured as outlined above before adding to the assay mixture. Likewise these assays were similarly treated with PK unless otherwise stated. Import assays in the presence of difference amounts of malonate, a competitive inhibitor to complex II, were carried out by first adding the desired amount of malonate to mitochondria and then the import buffer and equilibrated for 3 min on ice prior to addition of precursor protein and commencement of import assay (32). For chase experiments import assays were kept on ice with added precursor, washed in ice-cold import buffer without precursor protein, and resuspended in 200 l of import buffer without precursor. These mitochondria were incubated at 26°C for 20 min as for a standard import assay.
Tagging of AtTIM17-2 and TIM23 with c-Myc and His 6 and in Vitro Import-The Rapid Translation System RTS-500 (Roche Applied Sci-ence) was used for in vitro production of epitope-tagged AtTIM17-2. The coding region was cloned into PIVEX2.3-MCS (Roche Applied Science) ensuring that the C-terminal His 6 tag was in the correct reading frame. Addition of the N-terminal c-Myc tag (MEQKLISEEDL) was carried out using PCR with the N-terminal primer containing the coding sequence for the c-Myc tag. The resulting product was cloned into PIVEX2.3-MCS and confirmed by DNA sequencing. The proteins were synthesized in the RTS-500 system according to the manufacturer's instructions for 24 h at 30°C; the reaction vessel was mixed with magnetic agitation at 120 rpm. The synthesized protein was purified via two rounds of nickel affinity chromatography according to the manufacturer's instructions (Qiagen). The final elution buffer was 8 M urea, 0.1 M NaH 2 PO 4 , 10 mM Tris-HCl, pH 4.3. The protein was stored at Ϫ80°C until use. Import assays using this precursor protein were carried out by thawing the purified protein followed by sonication for 10 s for three times on a microtip setting of 4 (Branson sonifier B-12). This protein was immediately added to the in vitro import assay; 4 l was added to a 400-l assay to dilute the urea to below 100 mM so as not to inhibit the import assay. The import assays were treated as outlined above except that after SDS-PAGE, the products were detected by Western blot analysis using antibodies to c-Myc and His 6 (Roche Applied Science).
Production of Antibodies and Western Blot Analysis-Immunodetection of various proteins was carried out by transferring proteins separated by SDS-PAGE to supported nitrocellulose using a semidry blotter (Amersham Biosciences). Proteins were probed with a variety of antibodies and visualized and quantitated using chemiluminescence (Roche Applied Science) using a LAS 1000 (Fuji, Tokyo). Antibodies to Arabidopsis TOM20 (36), Uncoupling protein (UCP) (37), cytochrome c (BD Pharmingen), and AtTIM17-2 were used to probe membranes. Inoculating rabbits with the C-terminal 100 amino acids of TIM17 as antigen produced the AtTIM17-2 antibodies (38). The C-terminal extension of AtTIM17-2 (amino acids 143-243) was cloned into pDest15 (Invitrogen) protein expression vector using the Gateway Cloning Technology (Invitrogen). A PreScisson protease cleavage site (LEVLFQGP) was introduced inbetween the N-terminal GST tag and TIM17. Recombinant GST-LEVLFQGP-TIM17 expression was induced in the Escherichia coli strain BL21(DE3)pLys5 via isopropyl-1-thio-␤-D-galactopyranoside and purified using glutathione-Sepharose. TIM17 was released from the GST tag by digestion with PreScisson protease (Amersham Biosciences), as the recombinant protein remained bound to glutathione-Sepharose. The 15-kDa protein was dialyzed in phosphate-buffered saline and sequenced by mass spectrometry to confirm identity prior to inoculation of rabbits.

In Vitro Imported TIM17 Remains Sensitive to Externally
Added Protease-TIM17 is encoded by three genes in Arabidopsis, designated AtTIM17-1 (At1g20350), AtTIM17-2 (At2g37410), and AtTIM17-3 (At5g11690) (9,30). Although all predicted proteins display 50% sequence similarity with yeast TIM17 (158 amino acids in length) the predicted proteins differ in length; AtTIM17-1 is 218 amino acids, AtTIM17-2 is 243 amino acids, and AtTIM17-3 is 133 amino acids. Sequence alignments show AtTIM17-1 and AtTIM17-2 both contain long C-terminal extensions not present in AtTIM17-3 or the yeast TIM17. The AtTIM17-2 transcript was expressed at 10-fold higher levels than the other two genes in all Arabidopsis tissues examined and its product, AtTIM17-2, was the only TIM17 paralog identified in isolated Arabidopsis mitochondria by tandem mass spectrometry (39). Therefore we investigated the intraorganelle location and function of AtTIM17-2 with particular reference to its C-terminal extension that appears to be plant specific. In vitro mitochondrial uptake of radiolabeled AtTIM17-2 resulted in a protease-protected product with an apparent molecular mass of 31 kDa. An additional band of 28 kDa was generated on protease treatment of mitochondria compared with the precursor alone (Fig. 1A, lanes 1-3). The generation of this 28-kDa product required an inner membrane potential as increasing malonate concentration to 50 mM, to dissipate the membrane potential (⌬), resulted in no protease generated product (Fig. 1, lane 9). Notably, the protease-protected product of 31 kDa was still evident in the absence of a ⌬ (Fig. 1, lanes 8 and 9). We also tested the import of a protein imported via the carrier import pathway, ANT. The N-terminal extension of the plant ANT was cleaved generating a mature form of 31 kDa from a 39-kDa precursor (33,35). At 50 mM malonate, no processing of ANT to the mature form was detected suggesting ⌬ dependence. AOX, a protein imported via the general import pathway, was imported and processed as previously documented except at 50 mM malonate, where no import occurred because of the lack of ⌬ (Fig. 1A) (31). Thus we concluded that all precursors tested were imported in a membrane potential-dependent manner. The protease-protected forms of AtTIM17-2 and ANT at 50 mM malonate likely represent stage III import intermediates (see below). Translo-cation of proteins via the carrier pathway to a protease insensitive location in the absence of a membrane potential has been shown in mitochondria (40,41) and a reconstituted vesicle system (42). From these observations it appeared that a portion of AtTIM17-2 remained protease accessible when imported into the inner mitochondrial membrane.
To further characterize the location of imported AtTIM17-2 we increased the amount of added protease and also ruptured the outer membrane after the import assay but prior to protease treatment (Fig. 1B). Increasing of protease to four times the amount normally used to remove unimported precursor proteins did not further convert the 31-kDa AtTIM17-2 precursor FIG. 1. Import of radiolabeled AtTIM17-2 into mitochondria. A, import of AtTIM17-2, ANT, and AOX into mitochondria with increasing amounts of malonate. Import across or into the inner membrane was assessed by the fact that the highest amount of malonate used (50 mM) could prevent the generation of a processed product for AOX (precursor 36 kDa, mature 32 kDa), and ANT (precursor 39 kDa, mature 31 kDa), and thus we concluded that these proteins were not inserted into or across the inner membrane. In the case of AtTIM17-2, no protease-generated product of 28 kDa was generated at 50 mM malonate (lanes 8 and 9), and thus the generation of this product depended on the presence of a membrane potential. The amount of malonate present in each import assay is indicated in millimolar concentrations. B, import of AtTIM17-2 and AOX into mitochondria and treatment with increasing amounts of protease after the import assay. Lane 1, translation mixture alone. Lane 2, precursor protein incubated with mitochondria under conditions that support import. Lane 3, as lane 2 with PK added to a final concentration of 40 g/ml. Lane 4, PK added to 80 g/ml. Lane 5, PK added to 160 g/ml. Lane 6, mitochondria were lysed by osmotic shock after the import assay to produce outer membrane-ruptured mitochondria (Mit*OM). Lane 7, as lane 6 with Mit*OM treated with PK added to 40 g/ml. Lane 8, Mit*OM treated with PK added to 80 g/ml. Lane 9, Mit*OM treated with PK added to 160 g/ml. The additional products generated by translation initiation at internal methionines are shown on the left side of the figure with apparent molecular masses of 27.5, 24.5, 23, and 18 kDa. The band with an apparent molecular mass of 21 kDa, marked with an asterisk and generated upon import into mitochondria, likely represents an endogenous proteolysis product of these aberrant translation products (see text for details). C, chase of bound precursor protein into mitochondria. Precursor proteins were incubated with mitochondria on ice, washed, and resuspended, and the import assay was allowed to proceed as normal. Lane 1, precursor protein alone. Lane 2, precursor protein incubated with mitochondria on ice for 20 min. Lane 3, as lane 2 with reaction treated with PK added to a final concentration of 40 g/ml. No imported products are evident for AOX and only a small amount of the 31-kDa AtTIM17-2 product is detected. No protease-generated product of 28 kDa was detected. Lane 4, mitochondria from lane 2 were pelleted and resuspended in import buffer (no precursor protein was added), and the import assay was allowed to proceed at 26°C for 20 min. Lane 5, as lane 4 with PK added to a final concentration of 40 g/ml. Over 90% of the bound AtTIM17-2 31-kDa protein was converted into the protease-generated 28-kDa product, the 32-kDa AOX mature form was protease-protected. D, antibodies to TOM20, cytochrome c (Cyt c), and UCP were used to monitor the rupture of mitochondria.
form to the 28-kDa form. Confirmation that the imported At-TIM17-2 was located in the inner membrane was demonstrated by the fact that upon rupture of the outer membrane a 16-kDa inner membrane protease protected form was evident as previously reported (30). AOX was used as a control because it is located on the inner surface of the inner membrane, and mature imported protein was protected from increasing protease concentration in both intact and outer membrane-ruptured mitochondria (Fig. 1B). In addition to the 28 kDa protein generated by externally added protease, a product with an apparent molecular mass of 21 kDa was generated during incubation of AtTIM17-2 with mitochondria (Fig. 1B, lane 2, marked with an asterisk). This product was partially sensitive to protease treatment and is probably attributed to translation initiation at internal methionines. Previously we have demonstrated that the four radiolabeled bands with apparent molecular masses of 31, 27.5, 24.5, and 23 kDa, which are present in the translation mixture of AtTIM17-2, represent translations commencing at methionine residues 1, 25, 57, and 80, respectively. Additionally the 18-kDa product represents translation commencing at methionine residue 143 (30). The translation products of 27.5, 24.5, and 23 kDa still contain transmembrane regions 3 and 4, which contain the mitochondrial targeting activity of At-TIM17-2 (30). Thus the 21-kDa product generated upon incubation with mitochondria likely represents these imported products that have been subjected to endogenous proteolysis. Like the full-length translation product of 31 kDa, the 21-kDa product is only partially sensitive to externally added protease in intact mitochondria. The 18-kDa translation product beginning at methionine 143 has no mitochondrial targeting activity (30).
In combination, these data suggested the undigested 31-kDa form of AtTIM17-2 could represent one of two forms, either it is an imported product located in the inner membrane with a fraction also in the outer membrane, or the protease-protected 31-kDa form is a stage III import intermediate of the carrier import pathway not yet inserted into the inner membrane, as suggested by the fact it is imported into a protease-insensitive location in the absence of a ⌬ (Fig. 1) (40 -42). To confirm the latter we bound AtTIM17-2 to mitochondria and then chased it into the inner membrane (Fig. 1C). Incubation of AtTIM17-2 with mitochondria at 4°C resulted in only a very small amount of 31-kDa AtTIM17-2 been protected from protease digestion, which suggested external binding to mitochondria (Fig. 1C,  lane 3). Similarly, no import of AOX was observed under these conditions because no protease-resistant products were detected (Fig. 1C, lane 3). These mitochondria were washed in import buffer without precursor protein and incubated at 26°C for 20 min. It was observed that ϳ90% of the bound 31-kDa AtTIM17-2 protein was converted to the 28-kDa form. Thus we concluded that the 28-kDa form of AtTIM17-2 represents the truly imported form, and the protease-protected 31-kDa band is a stage III-like import intermediate that is imported across the outer membrane to a protease-insensitive location but not inserted into the inner membrane.
The integrity of mitochondria, the rupture of the outer membrane and efficacy of the protease treatments performed in these experiments was monitored using antibodies raised to TOM20 (outer membrane marker), cytochrome c (intermembrane space marker), and the uncoupling protein (UCP; inner membrane marker). UCP was resistant to protease and outer membrane rupture, cytochrome c was only digested by protease upon outer membrane rupture, whereas TOM20 was digested by protease in intact and outer membrane-ruptured mitochondria (Fig. 1D).
The C-terminal Region of AtTIM17-2 Is Protease-sensitive in the Outer Membrane-To determine which region of AtTIM17-2 was protease-sensitive in mitochondria, the N terminus was tagged with c-Myc and the C-terminal with His 6 . These proteins were expressed in an in vitro translation system, purified using nickel affinity chromatography, denatured in urea, and used in import assays with purified mitochondria. The imported and protease-protected proteins were detected by Western blotting (Fig. 2A). Protease treatment resulted in the removal of the C-terminal His 6 tag from AtTIM17-2 because no signal was detected when anti-His 6 antibodies were used to detect imported products ( Fig. 2A). In contrast, antibodies to c-Myc detected a protease product of 31 kDa compared with the undigested form of 34 kDa. The small differences in apparent molecular mass between the tagged proteins (Fig. 2) and radiolabeled proteins (Fig. 1) are caused by the presence of the tags. Deletion of the 100 amino acids at the C-terminal region of AtTIM17-2 resulted in both the C-terminal and N-terminal tags being protease-protected after import. We conclude that the C-terminal His 6 tag can be efficiently imported across the outer membrane and that the 100 amino acids of the C-terminal region of AtTIM17-2 were responsible for the protease sensitivity in the outer membrane.
To confirm the location of the C-terminal region of At-TIM17-2 in the outer membrane we also raised antibodies to the C-terminal 100 amino acids. Purification and protease treatment of mitochondria were carried out. Antibodies to the outer membrane protein import receptor, TOM20, detected a single band with an apparent molecular mass of 23 kDa, which is sensitive to the lowest amount of protease added (Fig. 2B). Antibodies raised to the 100 C-terminal amino acids of At-TIM17-2 detected a protein with an apparent molecular mass of 31 kDa. This product was also sensitive to protease digestion but was somewhat more protected than TOM20 (Fig. 2B). Typically, 75% of the signal was removed by digestion with 40 g/ml of proteinase K and increasing PK to 80 g or above resulted in 90% digestion. However, complete digestion was not achieved with higher amounts of added protease. This indicates that a small proportion (ϳ10%) of AtTIM17-2 is not exposed in the outer membrane. The intactness of the outer membrane was confirmed because cytochrome c was protected even with higher protease concentrations (Fig. 2B). Rupture of the outer membrane resulted in complete digestion of TOM20, AtTIM17-2, and cytochrome c. Note that the antibody raised against AtTIM17-2 will only recognize the C-terminal region and not the inner membrane-located region. Attempts to raise antibodies to the predicted inner membrane regions of At-TIM17-2 have proven unsuccessful to date.
Antibodies to the C Terminus of AtTIM17-2 Specifically Inhibit the General Import Pathway-The antibodies raised against the C-terminal 100 amino acids of AtTIM17-2 did not inhibit protein import into intact mitochondria, perhaps because of difficulties in the antibodies accessing the outer membrane component of At-TIM17-2 (data not shown). Therefore we tested whether the antibody could inhibit import into outer membrane-ruptured mitochondria. Import was conducted with two proteins by the general import pathway (AOX, F A d) and one carrier protein via the carrier import pathway (Pic) into outer membrane-ruptured mitochondria (Fig. 2C, lane 3) (33,35). Preincubation of outer membrane-ruptured mitochondria with preimmune serum had no affect on the import of these proteins (Fig. 2C, lanes 4 and 5). However antibodies to AtTIM17-2 completely abolished the import of AOX and F A d (Fig. 2C, lanes 6 and 7). In contrast no inhibition of import of Pic was observed (Fig. 2C, lanes 6 and 7). We used this precursor because it also contains an N-terminal cleavable extension that is removed upon import, but nonetheless is imported via the carrier pathway (33). Thus the C-terminal 100 amino acids of TIM17-2 are essential for protein import via the general import pathway but are not necessary for import via the carrier pathway.
Import via the General Import Pathway Requires an Outer Membrane Protease-sensitive Component-The binding chain hypothesis proposes that there are a series of binding sites on components of the import apparatus that achieve the unidirectional transport of precursor proteins across both mitochondrial membranes (29). It has been demonstrated in yeast that TIM17:23 can recognize and import proteins via the general import pathway when the outer membrane and the TOM complex are disrupted or absent (43,44). Rupture of the outer membrane of plant mitochondria followed by import assays indicated that import via the general import pathway (AOX) and carrier import pathway (ANT) was still possible (Fig. 3A). This import was reduced in abundance compared with intact mitochondria (Fig. 3, A and C). For the carrier import pathway we have previously shown that import can be stimulated to levels observed with intact mitochondria by re-addition of intermembrane space fractions (35). However it is possible that import is still taking place via the TOM complex in the attached but disrupted outer membrane. To ensure the TOM complex was bypassed, we pretreated mitochondria with protease that resulted in complete abolition of import via the general and carrier import pathways (Fig. 3B, lanes 4 and 5, 11  and 12). When the outer membrane was ruptured, import of ANT was restored (Fig. 3B, lanes 6 and 7), thus bypassing the FIG. 2. Immunodetection of AtTIM17-2 in mitochondria and immunoinhibition of protein import. A, AtTIM17-2 was tagged at the N terminus with c-Myc and at the C terminus with His 6 . Also tagged in a similar manner was AtTIM17-2 that had the coding region for the 100 C-terminal amino acid-deleted AtTIM17-2⌬143-243. The proteins were expressed in an in vitro expression lysate, purified via nickel affinity chromatography, and imported into mitochondria. The imported products were detected with antibodies to c-Myc and His 6 . Lane 1, precursor protein detected with c-Myc antibody. Lane 2, precursor protein incubated with mitochondria under conditions that support import and precursor detected with c-Myc antibodies. Lane 3, as lane 2 except that mitochondria were treated with PK to a final concentration of 40 g/ml. The TIM17 precursor is reduced in apparent molecular mass from 34 to 31 kDa. Lanes 4 -6 as for lanes 1-3 except that antibodies to His 6 were used to detect precursor proteins. In the case of AtTIM17-2, no product was detected on protease treatment (lane 6). B, antibodies raised against the 100 amino acids of AtTIM17-2 were used to probe mitochondria and outer membrane-ruptured mitochondria (Mit*OM) treated with increasing amounts of protease as indicated. C, effect of AtTIM17-2 antibodies on import of AOX, F A d, and Pic precursor proteins into outer membrane-ruptured mitochondria. Lane 1, precursor protein alone. Lane 2, precursor protein incubated with mitochondria under conditions that support import. Lane 3, as lane 2 except outer membrane-ruptured mitochondria were used in the in vitro import assay and treated with PK. Lanes 4 and 5, as lane 3 except that outer membrane-ruptured mitochondria were incubated with 100 and 200 g of preimmune serum (PI) prior to addition of the uptake assay. Lanes 6 and 7, as lanes 4 and 5 except 100 and 200 g of antibodies to AtTIM17-2 was added to outer membrane-ruptured mitochondria prior to addition of the uptake assay. AtTIM17-2 Ab, AtTIM17-2 antibodies. TOM complex could be achieved with a protein imported via the carrier import pathway. In contrast no import of AOX could be detected with protease-treated outer membrane-ruptured mitochondria. Thus a protease-sensitive outer membrane-exposed component was essential for import via the general import pathway.
Differential Protease Sensitivity of Import and the Abundance of TOM20 and TIM17-2-Digestion of mitochondria with very limiting amounts of protease indicated that whereas TOM20 was sensitive to very low amounts of protease (0.5-1 g/ml), AtTIM17-2 was barely digested with these amounts (Fig. 4A). This differential sensitivity was reminiscent of the lower sensitivity of the import of the F A d subunit than of other proteins into plant mitochondria by the general import pathway (32). Keeping in mind the inhibition of import observed above with antibodies to the C-terminal 100 amino acids of AtTIM17-2, and this difference in protease sensitivity in intact mitochondria between TOM20 and AtTIM17-2, we carried out imports into mitochondria pretreated with different amounts of protease. This was carried out to determine whether loss of import observed upon protease digestion of intact mitochondria could be correlated with loss of TOM20 or AtTIM17-2, or attributable to both. Purified mitochondria were treated with protease and tested for the loss of TOM20 or AtTIM17-2 (Fig. 4,  B and C). Import assays into protease-treated mitochondria indicated that the lowest amount of protease (0.5 g/ml) re-sulted in only 40% the amount of import compared with untreated mitochondria. Doubling the amount of protease (1 g/ ml) resulted in only 8% the amount of import compared with untreated mitochondria (Fig. 4B). This closely correlated with the observed decreases of TOM20 abundance to 29 and 12% of controls following these protease treatments (Fig. 4C). In contrast the import of F A d was more resistant to protease treatment. Loss of import of F A d correlated with the loss of At-TIM17-2 (Fig. 4, B and C). Treatment of mitochondria with 4 g/ml PK resulted in import of F A d being reduced to only 46% (Fig. 4B) and the amount of AtTIM17-2 reduced to 42% as determined by Western blotting (Fig. 4C). DISCUSSION The complete sequencing of the Arabidopsis genome allows a comparative genomic approach to be taken to identifying the components of the mitochondrial import apparatus using the model yeast system as a template (7,10,45). Using sequence similarity as a primary tool it has been possible to quickly identify many of these components in other organisms, albeit organism-specific components will not be identified using this approach (8,9,22). In Arabidopsis and other organisms it has emerged (in contrast to yeast) that many of the import components are encoded by small gene families (8,9). Evidence is emerging that these genes are expressed in a tissue, developmental, or inducible manner and thus the functional charac- FIG. 3. Import of precursor protein into mitochondria and outer membrane-ruptured mitochondria. A, import of ANT and AOX into mitochondria and outer membrane-ruptured mitochondria. Lane 1, precursor protein. Lane 2, precursor protein incubated with mitochondria, an additional 31-kDa mature form is evident. Lane 3, as lane 2 with PK added to 40 g/ml. Lanes 4 and 5, as lanes 2 and 3 except that mitochondria that had the outer membrane-ruptured were used in import assays. Import was reduced (see C). Lanes 6 -10, as lanes 1-5 except that the AOX precursor protein was used. B, import of ANT and AOX into mitochondria, mitochondria that were treated with PK to 40 g/ml, and outer membrane-ruptured mitochondria where the mitochondria were treated with PK to 40 g/ml prior to rupture of the outer membrane. Lane 1, precursor protein. Lane 2, precursor protein incubated with mitochondria. Lane 3 as lane 2 with PK added to 40 g/ml after the import assay. Lanes 4 and 5, as lanes 2 and 3 except that mitochondria were treated with PK to 40 g/ml prior to import assay, no imported products are detected. Lane 6, as lane 4, except that the outer membrane of protease-treated mitochondria was ruptured prior to the import assay, a mature form of 31 kDa is visible. Lane 7, as lane 6, with Mit*OM treated with PK to 40 g/ml after the import assay. Lanes 8 -14, as lanes 1-7, except that AOX was used in the import assay. No import was detected into PK-pretreated mitochondria or outer membrane-ruptured mitochondria that had been treated with PK to 40 g/ml. C, quantitation of import into mitochondria, outer membrane-ruptured mitochondria (Mit*OM), and PK-treated mitochondria that had outer membrane-ruptured (PK-Mit*OM).
terization of these components needs to be undertaken in each organism (39,46,47). This is especially true when orthologous components have additional protein extensions or domains not present in yeast, such as in TIM17 in Arabidopsis.
We previously reported that upon import, AtTIM17 appeared to be protease-sensitive in intact mitochondria, suggesting that some region may be located in the outer membrane (30). Here we demonstrated that the C terminus is located in the outer membrane using epitope tagging and antibodies raised against this region. Notably upon import we could chase ϳ90% of precursor protein into this protease accessible form (Fig. 1C), which was similar to the amount of AtTIM17-2 that was protease accessible in mitochondria (i.e. digestion of intact mitochondria resulted in the loss of ϳ90% of the signal detected with antibodies to the C-terminal 100 amino acids, Fig. 2B). Based on this chase experiment, the undigested or proteaseresistant forms of AtTIM17-2 normally observed with import of radiolabeled precursor proteins likely represent stage III import intermediates (10). This designation was also consistent with the appearance of this form in the absence of a membrane potential (Fig. 1A, lanes 8 and 9).
A physical link between the outer and inner membrane via AtTIM17 suggests a novel role in the import of proteins. Import of proteins into plant mitochondria that had the outer membrane-ruptured (but not removed) can take place via the general and carrier import pathways (Fig. 3A) (33,35). However, prior protease digestion of mitochondria with 40 g/ml PK followed by outer membrane rupture abolished the import of proteins via the general import pathway, but not via the carrier import pathway (Fig. 3, B and C). Thus an outer membrane protease-sensitive component was essential for import via the general import pathway in Arabidopsis. Antibodies to the Cterminal 100 amino acids of AtTIM17-2 inhibited import into outer membrane-ruptured mitochondria of proteins that are imported via the general import pathway (Fig. 4B). Together these lines of evidence suggest that a portion of the C-terminal 100 amino acids of AtTIM17-2 plays a novel and direct role in protein import, either binding precursor proteins or interacting with components that bind precursor and undertake an essential role in transferring them to TIM17:23.
The reason why plants have this important function associated with AtTIM17-2 is not clear. We have previously demonstrated that AtTIM17-2 can complement a yeast mutant but only when the C-terminal 100 amino acids were removed (30). Thus in yeast, this extension appears to interfere with the inner membrane function of TIM17. Interestingly, the binding partner for TIM17, known as TIM23, has an N-terminal region located in the outer membrane in yeast (44), but TIM23 does not contain this extension in the Arabidopsis sequence (30). The C-terminal extension in TIM17 in plants may conceivably play an early role in the import process; after the precursors are bound to TOM20 they could be passed to AtTIM17-2. The correlation of import of the F A d precursor with AtTIM17-2 supports this possibility. However, notably, AtTIM17-2 could not compensate for removal of TOM20 by low amounts of protease during the import of AOX (Fig. 4B) indicating that At-TIM17-2 is unlikely to be able to act as a primary receptor for proteins imported via the general import pathway. FIG. 4. Differential protease sensitivity in import of proteins and import components. A, protease sensitivity of AtTIM17-2 compared with TOM20. Isolated mitochondria were treated with increasing amounts of protease as indicated and probed with antibodies to TOM20, AtTIM17-2, and cytochrome c. B, import of AOX and F A d into mitochondria treated with limiting amounts of protease. Mitochondria were treated with the amounts of PK indicated, washed extensively, and then used with in vitro uptake assays. After the protein uptake assay mitochondria were split into two halves, and one treated with PK to a final concentration of 40 g/ml to assess the amount of protein uptake into mitochondria. The figures next to the import panels indicate the amount of import, 100% import represent the amount of import that was observed when mitochondria were not pretreated with any protease, i.e. normal import assay. The amount of import into mitochondria treated with different amounts of protease was expressed relative to mitochondria that were not pretreated with PK. C, Western blot analysis of mitochondria treated with limiting amounts of protease and used with in vitro uptake assays as outlined in C. The amount of TOM20, TIM17, and Cyt c detected by immunoblotting is indicated. The amount of each of these proteins detected in mitochondria not treated with any protease was set to 100%, and the amount in protease-treated mitochondria expressed relative to this amount. Note that as TOM20 and TIM17 are exposed on the outer membrane treatment with 40 g/ml of PK after import completely removes these epitopes and thus quantification for these proteins is only possible in samples not treated with PK post import.