Mitochondrial Protein Sorting

The mitochondrial outer membrane contains two distinct machineries for protein import and protein sorting that function in a sequential manner: the general translocase of the outer membrane (TOM complex) and the sorting and assembly machinery (SAM complex), which is dedicated to β-barrel proteins. The SAMcore complex consists of three subunits, Sam35, Sam37, and Sam50, that can associate with a fourth subunit, the morphology component Mdm10, to form the SAMholo complex. Whereas the SAMcore complex is required for the biogenesis of all β-barrel proteins, Mdm10 and the SAMholo complex play a selective role in β-barrel biogenesis by promoting assembly of Tom40 but not of porin. We report that Tom7, a conserved subunit of the TOM complex, functions in an antagonistic manner to Mdm10 in biogenesis of Tom40 and porin. We show that Tom7 promotes segregation of Mdm10 from the SAMholo complex into a low molecular mass form. Upon deletion of Tom7, the fraction of Mdm10 in the SAMholo complex is significantly increased, explaining the opposing functions of Tom7 and Mdm10 in β-barrel sorting. Thus the role of Tom7 is not limited to the TOM complex. Tom7 functions in mitochondrial protein biogenesis by a new mechanism, segregation of a sorting component, leading to a differentiation of β-barrel assembly.

membrane, and matrix (4 -11). Preproteins carrying cleavable amino-terminal targeting signals (presequences) are transferred from the TOM complex to the presequence translocase of the inner membrane (TIM23 complex) and then sorted to matrix and inner membrane. Many mitochondrial proteins, however, are synthesized without cleavable presequences, including all outer membrane proteins and the majority of intermembrane space and inner membrane proteins. These proteins use distinct import and sorting machineries. Thus, in addition to the presequence pathway, three sorting pathways for non-cleavable mitochondrial precursor proteins have been identified, each of them starting at the TOM complex. Many hydrophobic inner membrane proteins, including the abundant metabolite carriers, are imported via small Tim proteins of the intermembrane space and the twin pore translocase of the inner membrane (TIM22 complex) (8,12). Small proteins of the intermembrane space use a special import and assembly machinery that includes Mia40 and the sulfhydryl oxidase Erv1 (13)(14)(15)(16)(17). Outer membrane proteins are also initially transported via the TOM complex; however, the TOM complex is not able to integrate ␤-barrel proteins into the membrane. Therefore, the outer membrane contains a separate sorting and assembly machinery (SAM complex) that directs membrane integration and assembly of ␤-barrel proteins, including the abundant proteins porin and Tom40 (8, 18 -24).
The precursors of ␤-barrel proteins are first imported by the TOM complex to the intermembrane space side of the outer membrane. The small Tim proteins, which function in a chaperone-like manner to prevent aggregation of proteins in the aqueous intermembrane space, guide not only the transfer of hydrophobic inner membrane proteins but also that of outer membrane ␤-barrel proteins (8, 24 -27). The ␤-barrel proteins are then inserted into the outer membrane by the SAM complex. The SAM core complex contains three subunits, Sam35 (Tob38/Tom38), Sam37 (Mas37/Tom37), and Sam50 (Tob55/ Omp85) (18, 28 -33). Very recently, a fourth unexpected subunit of the SAM complex was identified (34). Mdm10 has been known for its involvement in mitochondrial morphology and distribution (35,36). Mdm10 associates with a fraction of SAM core complexes to form a larger SAM holo complex. Whereas the SAM core complex is required for sorting of all ␤-barrel outer membrane proteins analyzed, the SAM holo complex is dedicated to assembly of the TOM complex, i.e. Mdm10 is selectively required for the final steps of assembly of Tom40 but not porin (34).
The TOM complex consists of seven different subunits, three receptor proteins, the central channel-forming protein Tom40, and three small Tom proteins (7,(37)(38)(39)(40). Nuclear-encoded  precursor proteins are initially recognized by the receptors  Tom20 or Tom70 and are then transferred to the TOM core  complex. The core complex contains the central receptor  Tom22 and the channel Tom40 as well as Tom5, Tom6, and  Tom7. The three small Tom proteins are involved in protein import and stability of the TOM complex (41)(42)(43)(44)(45)(46)(47)(48)(49)(50)(51)(52)(53), although their molecular function is only understood in part. In particular, the function of Tom7, which is well conserved in evolution, is unclear as various findings with pleiotropic and in part contradictory results have been reported. Tom7 was found to function at the trans side of the TOM complex by increasing the efficiency of protein transport to internal mitochondrial compartments (42,46,51), but the strongest import defect of mitochondria lacking Tom7 was observed for the precursor of porin (42,49,53), indicating a role of Tom7 in biogenesis of outer membrane proteins. A conceptual problem arose, however, when the role of Tom7 in biogenesis of Tom40 was analyzed. Although Tom40 and porin are both imported via TOM and SAM, the efficiency of Tom40 assembly was even enhanced by the absence of Tom7 (18,48) in contrast to the inhibitory effect on porin assembly. Moreover, yeast cells lacking Tom7 showed altered mitochondrial morphology (34,54), though it has not been analyzed whether the morphology defect of tom7⌬ cells was different from that of deletion mutants of other TOM or SAM components.
In this report we have addressed the apparently conflicting results on the function of Tom7. We observed that cells lacking Tom7 displayed significantly stronger defects in mitochondrial morphology than cells lacking the other small Tom proteins, Tom5 and Tom6, pointing to a possible relation of Tom7 to morphology components. A direct comparison of the assembly pathways of Tom40 and porin in tom7⌬ and mdm10⌬ mitochondria revealed antagonistic effects of Tom7 and Mdm10. We show that Tom7 exerts an inhibitory effect on Mdm10 by promoting its segregation from the SAM holo complex into a low molecular mass form, implying a new mechanism for mitochondrial protein sorting.
Microscopy-Yeast mutants and corresponding wild-type strains were grown at 28°C to mid-log phase, resulting in an optical density of 0.5-1. Cells from this preculture were inoculated into fresh YPD or YPG medium and grown overnight at 28 or 37°C to an optical density of 0.5-1. Aliquots of each strain were taken and analyzed for mitochondrial morphology. Either yeast strains had been transformed with an expression vector encoding a mitochondria-targeted version of the enhanced green fluorescent protein b 2 ⌬-EGFP (34) and selected on minimal medium (60) or mitochondria were stained with 150 nM 3,3Јdihexyloxacarbocyanine iodide (DiOC 6 ; Molecular Probes) for 15 min in growth medium (61,62). Cells were centrifuged at 400 ϫ g for 5 min, washed, resuspended in 10 mM HEPES, pH 7.4, and directly mounted in 0.5% (w/v) low melting agarose (FMC Bioproducts) on a microscope slide.
Images of cells with representative phenotypes were taken using an upright confocal microscope (Carl Zeiss AG) equipped with a differential interference contrast optic, a ϫ63 objective (Plan Apochromat, numeric aperture 1.4, Carl Zeiss AG), and LSM 510 Meta software. Fluorescent mitochondria were illuminated by an argon laser (Lasos GmbH) using an HFT 488 filter and an LP 505 filter. Fluorescent sections of the mitochondria were recorded in the direction of the z-axis by a step size of 0.13 m. The recorded z-sections were assembled to threedimensional images using Imaris 4.2 (Bitplane AG) software, and isosurface projections of the mitochondria were generated with comparable thresholds and system variables.
Miscellaneous-For Western blot analysis, blue native gradient gels or standard SDS gels were transferred onto polyvinylidene fluoride membranes by semi-dry blotting, followed by immunodecoration with specific antisera and the enhanced chemiluminescence system (ECL; GE Healthcare). Signals were quantified using NIH ImageQuant.

RESULTS
Differential Influence of Tom7, Tom6, and Tom5 on Mitochondrial Morphology-In yeast wild-type cells, mitochondria form a tubular network (63)(64)(65)(66). We directly compared the mitochondrial morphology of yeast mutants of each of the small Tom proteins by confocal laser scanning microscopy. When yeast cells were grown at 28°C, the majority of tom7⌬ cells displayed aberrant (aggregated/giant) mito-chondrial shapes, whereas most tom5⌬ and tom6⌬ cells showed a normal tubular mitochondrial morphology like wild-type cells ( Fig. 1 and Table 1). Upon growth of the cells at elevated temperature (37°C), aggregated/giant mitochondrial shapes were observed in the majority of cells from deletion mutants of all three small Tom proteins (Table 1). We compared the tom mutants to deletion mutants of the SAM core and SAM holo complexes. The morphology defect of sam37⌬ cells (34) was only observed at high temperature and not at the usual low growth temperature, whereas mdm10⌬ cells displayed giant mitochondria at any growth temperature as reported (35, 67) ( Fig. 1; Table 1). Thus, the expression of morphological phenotypes (alteration of mitochondrial shape) in tom7⌬ cells showed more similarity to that in mdm10⌬ cells than to those in mutants of the other small Tom proteins or Sam37. These findings raised the possibility that Tom7 exerts a function with relation to mitochondrial morphology components.
Tom7, Mdm10, and Sam37 Exert Different Effects on the Assembly Pathways of Tom40 and Porin-As opposing effects of Tom7 on the sorting and assembly pathways of Tom40 and porin have been reported (48,49,53), we asked whether they were related to the effects of Mdm10 on sorting of those ␤-barrel proteins (34). We thus directly compared ␤-barrel assembly in tom7⌬ mitochondria and mdm10⌬ mitochondria, and for comparison also in sam37⌬ mitochondria as a mutant affecting the SAM core complex (Fig. 2,  A and B). Mitochondrial precursor proteins were synthesized and radiolabeled in reticulocyte lysates. Mitochondria were isolated from yeast cells grown at low temperature to minimize indirect pleiotropic effects that are often observed upon growth of mutant cells at elevated temperature (13,18,34). The precursor proteins were imported into the isolated mitochondria, and complexes were analyzed by blue native electrophoresis (18, 34, 48, 49).  tom5⌬, tom6⌬, tom7⌬, sam37⌬, and mdm10⌬ strains were grown at 28°C to mid-log phase. Three-dimensional images of fluorescent mitochondria were recorded by confocal laser scanning microscopy. Projections of the surface of mitochondria were generated with Imaris 4.2. DIC, differential interference contrast. Bar, 1 m.
Assembly of Tom40 occurs in distinct steps via an assembly intermediate I to an assembly intermediate II and the mature TOM complex ( Fig. 2A) (18,34,48). The formation of assembly intermediate I occurred with approximate wild-type efficiency in tom7⌬ mitochondria and mdm10⌬ mitochondria, whereas it was strongly impaired in sam37⌬ mitochondria ( Fig. 2A, lanes 4, 10,  and 16; Fig. 2C, columns 1-3). The subsequent assembly steps of Tom40, however, were differently affected in tom7⌬ mitochondria and mdm10⌬ mitochondria. Whereas the lack of Mdm10 led to an accumulation of assembly intermediate II and strong inhibition of formation of the mature TOM complex, the lack of Tom7 showed the opposite effect ( Fig. 2A, lanes 6 and 12;  Fig. 2C, columns 4, 5, 7, and 8) (34,48). Thus the formation of mature TOM complex was faster in tom7⌬ mitochondria than in wild-type mitochondria. In sam37⌬ mitochondria, as expected due to the impaired formation of assembly intermediate I, the formation of assembly intermediate II and mature TOM complex was similarly impaired ( Fig. 2A, lanes 17 and 18;  Fig. 2C, columns 6 and 9) (18). Thus, sam37⌬ mitochondria are already impaired in early steps of Tom40 assembly as is characteristic for the components of the SAM core complex (18, 28 -33), whereas Tom7 and Mdm10 affect the late steps of Tom40 assembly but in an opposing manner.
We then compared the assembly of radiolabeled porin in the three mutant mitochondria. Mature as-FIGURE 2. Differential effects of mitochondria lacking Tom7, Mdm10, or Sam37 on the assembly pathways of Tom40 and porin. A, import of 35 S-labeled precursor of Tom40 into isolated yeast mitochondria. Isolated mitochondria (50 g of protein) from tom7⌬, mdm10⌬, and sam37⌬ yeast strains and the corresponding wild-type (WT) strains were incubated with 35 S-Tom40 at 25°C for the indicated times. Mitochondria were washed and resuspended in lysis buffer containing 1% (w/v) digitonin. After blue native electrophoresis, gels were dried and radiolabeled proteins were visualized by digital autoradiography. B, import of 35 S-labeled porin into isolated mitochondria. Experiments were performed as in panel A, except that 35 S-labeled porin was used and the lysis buffer contained 0.5% (w/v) digitonin. C, quantitative comparison of import efficiencies. Radiolabeled proteins of Tom40 and porin in complexes shown in panel A or panel B were quantified using ImageQuant 5.2 (GE Healthcare). Import times used for quantification were Tom40 assembly I, 5 min; Tom40 assembly II and TOM complex, 50 min; porin assembly, 15 min.

TABLE 1 Mitochondrial morphology of Saccharomyces cerevisiae mutant cells
Yeast cells were grown overnight at 28 or 37°C to mid-log phase. Mitochondria were fluorescently labeled with the dye DiOC 6 or by using mitochondria-targeted enhanced green fluorescent protein (b 2 ⌬-EGFP). Shown are the total number of cells analyzed, the percentage of cells with normal (tubular) mitochondrial morphology, and the percentage of cells with aberrant mitochondrial morphology.

Number of cells Tubular Aberrant Number of cells Tubular Aberrant
Thus, our comparative analysis revealed a rather complex situation of how Tom7, Mdm10, and Sam37 influence ␤-barrel assembly in the mitochondrial outer membrane. In short, the SAM core complex with Sam37 as representative component is already required for formation of assembly intermediate I as reported (18, 28 -34, 68), whereas Tom7 and Mdm10 influence the late steps of assembly of Tom40 in an opposing manner. Tom7 and Mdm10 also show antagonistic effects on the assembly of porin, but in this case the stimulatory and inhibitory effects are reversed.
Steady-state Levels of TOM, SAM, and MDM Components Are Not Altered in Mitochondria Lacking Tom7-A possible explanation for the effects of Tom7 on mitochondrial morphology and ␤-barrel sorting would be an influence of Tom7 on the steady-state levels of critical components. We thus systematically compared the steady-state levels of various Tom proteins, Sam proteins, and morphology proteins in the mitochondria used for the protein import experiments. All TOM, SAM, and MDM components analyzed were present in wild-type amounts in tom7⌬ mitochondria like the control proteins of the inner membrane translocase, Tim23 and Tim50 (Fig. 3). We conclude that a lack of Tom7 does not change the steady-state levels of these proteins.
Tom7 Promotes Segregation of Mdm10 into a Low Molecular Mass Complex-As the total levels of Mdm10 were not changed in the tom7 mutant mitochondria we asked whether Tom7 influenced the distribution of Mdm10 in protein complexes. We performed blue native electrophoresis with a gradient gel that permitted the visualization of protein complexes in both the high and low molecular mass range. In wild-type mitochondria, Mdm10 was found in a complex migrating at ϳ350 kDa and in a low molecular mass range at ϳ140 kDa (Fig. 4A, lane 1,  and 4B). The specificity of the antibody reaction was demonstrated by the absence of both complexes in mdm10⌬ mitochondria (Fig. 4A, lane 6). The larger complex was selectively absent in sam37⌬ mitochondria (Fig. 4A, lane 5), confirming that it represented the SAM holo complex (34). In mitochondria lacking Tom5 or Tom6, the ratio between SAM holo complex and the low molecular mass form of Mdm10 was not significantly altered compared with wild-type mitochondria (Fig. 4A,  lanes 1-3). A striking effect, however, was observed for mitochondria lacking Tom7. The amount of Mdm10 found in the SAM holo complex was strongly increased, whereas the mobility of the low molecular mass form of Mdm10 was shifted (Fig. 4A,  lane 4). Thus, the lack of Tom7 promotes a redistribution of Mdm10 from the low molecular mass region to the SAM holo complex. The increased formation of SAM holo complex in tom7⌬ mitochondria was confirmed by the analysis of Sam50 as the ratio of SAM core complex to SAM holo complex on blue native electrophoresis was strongly shifted toward the larger complex in tom7⌬ mitochondria (Fig. 4C).
A two-dimensional analysis (blue native electrophoresis followed by SDS-gel electrophoresis) revealed that a fraction of Tom7 comigrated with the low molecular mass form of Mdm10 (Fig. 4B). The major fraction of Tom7 migrated with the TOM complex as expected, because the TOM complex is an order of magnitude more abundant than Mdm10 (69). The low molecular mass form of Mdm10 is distinct from the low molecular . Steady-state protein levels of tom7⌬ mitochondria compared with wild-type mitochondria. Mitochondria were isolated from wild-type (WT) and tom7⌬ cells. Mitochondrial proteins (g of protein as indicated) were separated by SDS-polyacrylamide gel electrophoresis. Immunodecoration with antibodies against the indicated proteins was performed using the enhanced chemiluminescence system.  Fig. 2A, lane 12) (34).
Second, antibody shift blue native electrophoresis indicated that Tom5 is present in the intermediate II complex (18). We analyzed the assembly intermediate II in mutant mitochondria of the three small Tom proteins and indeed found that the mobility of the intermediate was shifted in tom5⌬ mitochondria, confirming that Tom5 is present in the intermediate (Fig.  4D, lane 2). The mobility of the intermediate II was not altered in mitochondria lacking Tom6 or Tom7 (Fig. 4D, lanes 3 and 4). As a lack of Tom7 alters the gel mobility of the low molecular mass form of Mdm10, it is evident that Tom7 is not present in the assembly intermediate II of the precursor of Tom40.
We conclude that Tom7 promotes the segregation of Mdm10 into a low molecular mass form outside of the SAM complex. Thus, the amount of Mdm10 present in the SAM holo complex is reduced by Tom7, explaining the antagonistic effects of Tom7 and Mdm10 on sorting of mitochondrial ␤-barrel proteins.

DISCUSSION
We have found a new function for Tom7. In addition to its role at the trans side of the general translocase TOM (42,46,51), Tom7 plays a specific role in regulating the activity of the SAM complex. Tom7 favors formation of the SAM core complex at the expense of the SAM holo complex by promoting a segregation of Mdm10. Because Mdm10 in the SAM holo complex is only required for the assembly of the TOM complex but not for other ␤-barrel proteins like porin, the seemingly controversial effect of tom7⌬ mitochondria on the biogenesis of Tom40 and porin is now explained by the antagonistic effect of Tom7 on Mdm10. In the absence of Tom7, more Mdm10 is available for the SAM holo complex, and thus the assembly of the TOM complex is faster than in wild-type mitochondria. How can the improvement of porin assembly in mdm10⌬ mitochondria compared with wild-type mitochondria be explained? Meisinger et al. (34) suggest that Mdm10 occupies a site of the SAM core complex that can also be used for assembly of porin. In the absence of Mdm10, this site would be free for the porin pathway. The findings reported here clearly support this model as porin assembly is inhibited in tom7⌬ mitochondria, i.e. under conditions where more Mdm10 is present in the SAM complex.
The lack of Tom7 leads to significant changes of mitochondrial shape with giant/aggregated mitochondria resembling the phenotype of mdm10⌬ cells (35,67). Several Mdm/Mmm proteins have been reported to cooperate with Mdm10 in maintenance of mitochondrial shape, and yeast mutants of these proteins display similar morphological alterations (36,54,67,70). The steady-state levels of Mdm10 and other morphology components were not changed in tom7⌬ mitochondria, excluding the simple explanation that Mdm10 and cooperating morphology components would be lacking in tom7⌬ mitochondria. Because the molecular mechanism of how Mdm/Mmm proteins regulate mitochondrial shape is unknown (71), it is not possible to define the exact reaction chain of Mdm/Mmm and Tom7 in controlling mitochondrial morphology. However, the findings reported provide a hint for a regulatory role for Tom7. Tom7-mediated segregation changes the distribution of Mdm10 between different functional pools. Because only FIGURE 4. Increased formation of SAM holo complex in tom7⌬ mitochondria. A, Tom7 promotes segregation of Mdm10. Mitochondria (100 g of protein) isolated from the indicated yeast strains were lysed in 1% (w/v) digitonin buffer and separated by blue native electrophoresis. Immunodecoration was performed with affinity-purified antibodies against Mdm10. B, twodimensional analysis of Mdm10 and Tom7. Isolated wild-type mitochondria were lysed by digitonin and separated by a first-dimension blue native electrophoresis, followed by a second-dimension SDS electrophoresis. Mdm10 and Tom7 were detected by immunodecoration. Mdm10Ј, fragment of Mdm10 formed upon lysis of mitochondria. C, shift of Sam50 from the SAM core complex in WT mitochondria to the SAM holo complex in tom7⌬ mitochondria analyzed by blue native electrophoresis. The experiment was performed as described for panel A except that affinity-purified antibodies against Sam50 were used. D, the assembly intermediate II of Tom40 contains Tom5, but not Tom6 or Tom7. 35 S-labeled Tom40 precursor was imported into isolated mitochondria of the indicated yeast strains. After accumulation of the assembly intermediate I, the mitochondria were reisolated and the Tom40 precursor was chased for an additional 80 min at 25°C (25). The mitochondria were lysed in digitonin and analyzed by blue native electrophoresis and digital autoradiography.
Mdm10 and no other Mdm protein was present in the SAM complex (34), Mdm10 enriched in the SAM holo complex of tom7⌬ mitochondria is apparently not available for cooperation with other morphology components, providing an explanation for the morphology defects of tom7⌬ mitochondria.
Taken together, the findings presented here and previous studies (18,48,49,53) reveal that each of the three small Tom proteins plays a distinct role in the assembly of outer membrane proteins. Tom5 directly interacts with the precursor of Tom40 to form the assembly intermediate II (18). Tom7 promotes segregation of Mdm10, and thus its function is antagonistic to the role of Mdm10, which is required for the late steps of TOM assembly. Eventually, Tom6 stabilizes the fully assembled TOM complex of 450 kDa (44,53).
In summary, the function of Tom7 is not limited to the TOM complex. Tom7 is involved in the differentiation of ␤-barrel assembly pathways in the mitochondrial outer membrane by modulating the levels of Mdm10 in the SAM complex. We propose that segregation of individual sorting components represents a mechanism to regulate mitochondrial protein import pathways.