The Mia40 substrate Mix17 exposes its N-terminus to the cytosolic side of the mitochondrial outer membrane

The mitochondrial contact site and cristae-organizing system (MICOS) is essential for the organization of mitochondrial architecture (John et al., 2005; Rabl et al., 2009; Harner et al., 2011; Hoppins et al., 2011; von der Malsburg et al., 2011) and forms contacts between the mitochondrial inner membrane (MIM) and the mitochondrial outer membrane (MOM). The importance of the MICOS complex is reflected in its conservation from yeast to mammals (Alkhaja et al., 2012). Mammalian subunits of MICOS have six homologs in the yeast Saccharomyces cerevisiae: Mic10, Mic12, Mic19, Mic26, Mic27 and Mic60 (Harner et al., 2011; Hoppins et al., 2011; von der Malsburg et al., 2011). Five of these MICOS components are integral MIM proteins facing into the intermembrane space (IMS) and follow the presequence import pathway (Harner et al., 2011; Hoppins et al., 2011; von der Malsburg et al., 2011; Ueda et al., 2019). Only the Mic19 component of MICOS, an IMS protein that is peripherally associated with the MIM, uses the Mia40-dependent import pathway (Ueda et al., 2019).

The mammalian MICOS complex contains additional subunits. Recently, the Mic14 protein, also termed CHCHD10, was shown to interact with Mic60 (also known as mitofilin), Mic19 and Mic25 (Genin et al., 2016). Interestingly, Mic14 is highly conserved. It has an ortholog in yeast, Mix17, the association of which with MICOS has not been analyzed so far. Additionally, its function remains unclear. The deletion of MIX17 negatively affects mitochondrial respiration by an unknown mechanism (Longen et al., 2009). Similar to Mic14, Mix17 contains a CHCH domain with the twin CX₉C motif of typical Mia40 substrates and its import has been shown to be Mia40 dependent (Gabriel et al., 2007; Longen et al., 2009; Gornicka et al., 2014). In contrast to typical Mia40 substrates with twin CX₉C motifs, Mix17 contains additional conserved sequence features, which might play a crucial role for its biogenesis pathway.

Here, we analyzed the protein interactions of Mix17, its submitochondrial location and its import pathway as an important step to elucidate the functional role of Mix17 in mitochondria. In contrast to the reported interaction of Mic14 and MICOS (Genin et al., 2016), yeast Mix17 does not stably associate with MICOS. Interestingly, we demonstrate that Mix17 is accessible to proteases added to intact mitochondria. Unlike all other Mia40 substrates, Mix17 apparently inserts into the MOM, suggesting that it is a novel kind of Mia40 substrate. We show that Mix17 spans the MOM, most likely through the TOM complex. Moreover, the N-terminus of Mix17 is essential for the insertion of Mix17 into the MOM and, in addition, improves the import of Mix17 into mitochondria.

The interaction of the potential mammalian MICOS subunit Mic14 is still a matter of debate (Genin et al., 2016; Burstein et al., 2018), which might be explained by the suggested cell type-specific function of the protein (Burstein et al., 2018). As this protein is highly conserved from yeast to humans (Fig. S1) (Longen et al., 2009), we first addressed whether Mix17 interacts with the MICOS complex in the unicellular organism S. cerevisiae.

To this end, we performed immunoprecipitation using mitochondria of yeast strains expressing 3×HA-tagged versions of Mic10 or Mic60. With both tagged proteins, we were able to co-isolate known MICOS subunits, indicating that the complex remained stable under these conditions. Mix17, however, did not coprecipitate with Mic10 or Mic60 (Fig. 1A). Next, we tested whether it was possible to co-isolate MICOS subunits via Mix17. Therefore, we purified Mix17 using an antibody specific to the endogenous protein. Although we could efficiently immunoprecipitate Mix17, the MICOS subunits Mic10, Mic26, Mic27 and Mic60 were not co-isolated (Fig. 1B). Taken together, we conclude that Mix17, at least under the conditions tested, does not stably associate with MICOS in yeast.

Next, we analyzed the location of Mix17 in isolated mitochondria. Surprisingly, we found that Mix17, similar to the MOM marker Tom70, is easily degradable by proteinase K (PK) in intact mitochondria (Fig. 2A). Addition of PK resulted in the generation of two smaller fragments, which could still be recognized by an antibody against the C-terminal Myc tag on Mix17. This indicates that Mix17 spans the MOM, exposing its N-terminus to the cytosol (Fig. 2A). As a control for intact mitochondria, the MIM protein Tim50 was stable at these conditions. When we treated osmotically swollen mitochondria (SW) with PK, Mix17 and Tim50 were completely undetectable (Fig. 2A). The strong signal reduction of Mix17 upon re-isolation of swollen mitochondria as well as its presence in the soluble fraction (S) upon alkaline extraction suggests that Mix17 spans the MOM, possibly in a proteinaceous environment, rather than being integrated into the lipid bilayer of the membrane (Fig. 2A). As expected, the integral membrane proteins Tim50 and Tom70 stayed in the membrane protein fraction (M) upon alkaline extraction and were not lost upon swelling and re-isolation (Fig. 2A). Importantly, we obtained virtually the same results for the endogenous Mix17 protein in wild-type mitochondria, indicating that the Myc tag on Mix17 did not alter its topology (Fig. 2B). Next, we confirmed our results with another protease. Mix17 contains six arginine residues in its N-terminal segment (Fig. S1), which should be recognized by trypsin if the N-terminus is exposed to the cytosol. Indeed, we detected a faster migrating fragment upon trypsin treatment of intact mitochondria. At the same time, Tom70 was degraded but Tim50 stayed intact (Fig. 2C). We then asked whether we could reproduce this stunning result following the in vitro import of radioactively labeled Mix17 precursor protein. Incubation of the Mix17 precursor with wild-type mitochondria allowed its efficient import into isolated mitochondria. Moreover, addition of trypsin to intact mitochondria after the import reaction resulted in the generation of the same fragment as detected for the endogenous protein (Fig. 2D).

In summary, our results showed that Mix17 spans the MOM in an N_out–C_in topology. Moreover, the loose association of Mix17 to the MOM indicates that it might be embedded in the MOM in a proteinaceous environment.

Next, we set out to analyze which factors are important for the import of Mix17 and its insertion into the MOM. In line with previous results (Gabriel et al., 2007; Gornicka et al., 2014), we show that Mia40 is essential for the import of Mix17 despite its uncommon mitochondrial sublocation for a Mia40 substrate. Import of Mix17 was strongly impaired in mitochondria depleted of Mia40 (Fig. 3A). The import of Mia40 substrates depends on a conserved CX₉C motif, which is also present in Mix17 and its homologs (Fig. S1). Disulfide bonds are formed between these cysteines by the MIA40 system, which is essential for the import of this class of proteins (Mesecke et al., 2005; Hell, 2008; Koehler and Tienson, 2009; Riemer et al., 2009; Stojanovski et al., 2012; Edwards et al., 2020). In accordance with this, inhibition of the formation of a disulfide bond by dithiothreitol (DTT) strongly reduced the import efficiency of Mix17 into wild-type mitochondria (Fig. 3B). The introduction of point mutations into Mix17 had an even more obvious effect. The import of the Mix17 precursor protein in which the cysteines of the twin CX₉C motif were replaced by serines (Mix17 C4S) was almost completely inhibited (Fig. 3C). Finally, we tested whether Mia40 is also essential for the import of Mix17 in vivo. We isolated mitochondria from wild-type cells and Mia40-depleted cells (Mia40↓) and tested for the steady-state levels of Mix17 in comparison to those of control proteins and those in wild-type mitochondria. Interestingly, Mix17 was virtually absent upon downregulation of Mia40, similar to the Mia40 substrate Tim13. In contrast, the levels of mitochondrial proteins that do not depend on Mia40 did not change substantially (Fig. 3D). Taken together, and in line with previous results (Gabriel et al., 2007; Gornicka et al., 2014), we demonstrate that the disulfide relay system and the twin CX₉C motif of Mix17 are essential for its import into mitochondria.

Mix17 contains, in addition to the twin CX₉C motif, several characteristic sequence features. Its N-terminus has a predicted probability to be a mitochondrial targeting signal (Claros and Vincens, 1996). Therefore, we asked whether the N-terminus plays a crucial role for the import of Mix17 and its insertion into the MOM. Presequence proteins harboring a mitochondrial targeting signal depend on the mitochondrial membrane potential for their import. Thus, we first analyzed whether the import of Mix17 is membrane potential dependent.

We observed similar import efficiencies in the presence and absence of the membrane potential (Fig. 4A). The integration of the N-terminus of Mix17 into the MOM was also not affected in the absence of the membrane potential. Imported Mix17 was still accessible to trypsin. Next, we asked whether the N-terminal segment per se is important for the import of Mix17. Therefore, we deleted the first 24 amino acids of Mix17 and tested its import efficiency compared to that of full-length Mix17. The truncated Mix17 Δ1–24 variant showed a reduced import efficiency to 20% of that of the full-length protein (Fig. 4B). Of note, the radioactive signal of Mix17 Δ1–24 lysate was considerably stronger than that of the full-length version. When we treated mitochondria with trypsin after the import reaction, we were not able to detect a reduction in the size of Mix17 Δ1–24 (Fig. 4B). This was expected, as the truncation led to the loss of the trypsin cleavage site that is required for detection of the trypsin accessibility. To be able to analyze the insertion of Mix17 Δ1–24 into the MOM, we repeated the experiment using PK instead of trypsin. Again, we detected an import defect for the truncated version. Similar to previous experiments (Fig. 2A), treatment of mitochondria with PK generated two fragments of imported full-length Mix17. Such fragments were not observed following the import of Mix17 Δ1–24. Importantly, both PK-generated Mix17 fragments migrated considerably faster upon SDS-PAGE than Mix17 Δ1–24. If Mix17 Δ1–24 inserted into the MOM, this variant should have been accessible to the added PK (Fig. 4C). This strongly indicates that the N-terminus of Mix17 is essential for the insertion of Mix17 into the MOM. Consistently, we did not detect clipping of Mix17 Δ1–24 by PK or trypsin when we tested mitochondria isolated from a Δmix17 yeast strain expressing Mix17 Δ1–24–FLAG in the protease accessibility assays (Fig. 4D). Thus, the N-terminus appears to be also crucial in vivo for the insertion of Mix17 into the MOM. In summary, we obtained evidence that the N-terminus of Mix17 is important for its integration into the MOM and supports the import of Mix17 into mitochondria.

The ability of Mix17 Δ1–24 being imported into mitochondria but not being exposed to the cytosol raises an intriguing question. Is full-length Mix17 first completely imported into mitochondria and then inserted into the MOM? To address this question, we generated a Mix17 variant that was fused to the N-terminal bipartite presequence of the cytochrome b₂ preprotein (encoded by CYB2). This bipartite presequence drives import of the cytb₂ precursor protein by the TIM23 presequence translocase into mitochondria and the release of the precursor protein as an intermediate form into the MIM (Hartl et al., 1987; Glick et al., 1992). The intermediate form is cleaved to release the mature form into the IMS (Schneider et al., 1991). Upon import of the radiolabeled Cytb₂(1–84)–Mix17 fusion protein into isolated mitochondria, we could detect an intermediate form and the release of Mix17 as its mature form into the IMS. Importantly, the intermediate form was completely imported and protected from tryptic degradation, indicating that it is present in the IMS without exposing its N-terminus to the cytosol. The mature form, however, was entirely accessible to trypsin (Fig. 5), indicating its insertion into the MOM after the complete import of the fusion protein and the subsequent release of Mix17 into the IMS. Thus, we conclude that Mix17 can insert into the MOM after being completely imported into the IMS.

Two MOM complexes were shown to be able to insert proteins into the MOM: the TOB/SAM complex and the MIM complex (Kozjak et al., 2003; Paschen et al., 2003; Becker et al., 2008; Papić et al., 2011). Thus, we next analyzed whether one of these complexes might be responsible for the insertion of Mix17 into the MOM.

First, we tested a potential role of the TOB/SAM complex in the insertion process by depletion of its essential core component Tob55 (also known as Sam50). Virtually complete depletion of Tob55 (Tob55↓ strain) did not affect the steady-state level of Mix17, although it already resulted in reduced levels of the known substrates Tom40 and Por1 (Fig. 6A). When we tested the protease accessibility of Mix17 in the absence of Tob55, we also did not detect any difference compared to that in wild-type cells (Fig. 6B). These results indicate that the TOB/SAM complex is not crucial for the membrane insertion of Mix17.

Next, we analyzed whether the MIM complex is important for the topology of Mix17. Again, we did not detect reduced protein levels of Mix17 in the MIM1 deletion strain Δmim1. The steady-state levels of Tom20, the membrane integration of which is supported by Mim1, however, were strongly reduced (Fig. 6C). Moreover, Mix17 was accessible to proteases in mitochondria isolated from the Δmim1 mutant similar to mitochondria from wild type (Fig. 6D). Based on these results, we conclude that neither the TOB/SAM complex nor the MIM complex play a crucial role in the insertion of Mix17 into the MOM.

Our previous results suggested that Mix17 integrates into the MOM in a proteinaceous environment rather than into the lipid bilayer (Fig. 2A). Thus, we next asked whether the conserved hydrophobic segment of Mix17 is required for its membrane insertion. For this, we deleted the hydrophobic segment of Mix17 and tested whether this Mix17 Δ53–80 variant could span the MOM. Following its in vitro import, this variant was still accessible to PK and trypsin, indicating that the conserved hydrophobic segment is not required for the exposure of the N-terminus of Mix17 to the cytosol (Fig. 7A). In addition, this result supports our previous data, suggesting that Mix17 integrates into a protein complex than into the lipid bilayer of the MOM.

As Mix17 did not integrate into the TOB/SAM or the MIM complex, it is possible that Mix17 uses the Tom40 pore of the TOM complex to span the MOM. To address a possible role of Tom40 in the MOM insertion of Mix17, we asked whether Tom40 interacts with endogenous Mix17. To this end, we chromosomally tagged Tom40 with a C-terminal 3×HA tag and performed immunoprecipitation with anti-HA agarose beads. Strikingly, immunoprecipitation of Tom40–3×HA revealed the co-isolation of endogenous Mix17 from chemically crosslinked mitochondria. The co-isolation was specific as Mix17 was not bound to the beads when wild-type mitochondria were used. As a control, the import receptor Tom70, a component of the TOM complex, specifically co-purified with Tom40–3×HA, in contrast to Aac2, an MIM protein (Fig. 7B). To further substantiate the specificity of the interaction of Mix17 and Tom40, we performed immunoprecipitation of Mix17 (Fig. 7C). Interestingly, this again revealed the successful co-isolation of Tom40. Aac2 as well as Tom70, however, could not be co-purified, suggesting that Mix17 is most likely in close proximity to the protein import pore but not to the receptor (Fig. 7C). Moreover, we found that the interaction could only be detected using chemically crosslinked mitochondria, indicating a transient rather than a stable interaction. As we had previously observed that the N-terminal truncation variant Mix17 Δ1–24–FLAG was not protease accessible in mitochondria in contrast to Mix17–FLAG (Fig. 4D), we analyzed whether this N-terminal truncation variant interacts with Tom40. Consistent with its protease accessibility, we found that Mix17 Δ1–24–FLAG did not co-isolate with Tom40–3×HA in contrast to the full-length Mix17 protein (Fig. 7D). Of note, the steady-state levels of Mix17 Δ1–24–FLAG were considerably lower compared to those of the full-length variant, potentially due to the reduced import or stability of the N-terminal truncation mutant (Fig. 4B,C).

In summary, Mix17 is able to interact with the TOM complex in the MOM at endogenous protein levels. Mix17 Δ1–24–FLAG was not protease accessible in intact mitochondria and did not interact with Tom40, suggesting that the TOM complex mediates the insertion of the N-terminus of Mix17 into the MOM.

In the present study, we analyzed the submitochondrial location and the biogenesis of the highly conserved protein Mix17. We show that Mix17 spans the MOM with an N_out–C_in topology. Our data suggest a working model in which the N-terminus of Mix17 is important for the biogenesis of Mix17. First, it supports the import of Mix17 into the mitochondrial IMS (Fig. 8, step 1), where the disulfide bonds of Mix17 are formed by the MIA40 system (Fig. 8, step 2). Second, it is essential for its exposure to the cytosolic side of the MOM (Fig. 8, steps 3a and 3b). Mix17 interacts with Tom40, which is likely to be important for its exposure to the cytosol. This interaction might take place in the protein-conducting pore or at the protein–lipid interface of Tom40. Mix17 appears to associate with Tom40 dynamically rather than permanently (Fig. 8, steps 3a and 3b).

Mix17 contains the twin CX₉C motif of a class of typical Mia40 substrates (Gabriel et al., 2007; Longen et al., 2009). We show that this motif is indeed crucial for import of Mix17. Moreover, we confirm previous reports of Mia40-dependent import of Mix17 into mitochondria (Gabriel et al., 2007; Gornicka et al., 2014). However, we demonstrate here that Mix17 is the first Mia40 substrate that can insert into the MOM from the IMS to reach the cytosol. Of note, Mix17 has been previously described as a protease-resistant IMS protein (Gabriel et al., 2007; Gornicka et al., 2014). It is tempting to speculate that these different results are caused by individual experimental conditions or different states of mitochondria. A similar observation has been made before for Tim23, a core component of the TIM23 translocase. In contrast to previous studies (Kübrich et al., 1994), the N-terminus of Tim23 was reported to be protease accessible in intact mitochondria (Donzeau et al., 2000). It was suggested that this depends on the concentration of salt present in the experimental setup (Donzeau et al., 2000). In addition, the protease accessibility and, thus, the topology of Tim23 were shown to respond to the translocation activity of the TIM23 complex (Popov-Celeketic et al., 2008). Such a dependence of the topology on experimental and physiological conditions might also apply to Mix17.

Our results in vitro and in vivo demonstrate that the N-terminus of Mix17 is crucial for the insertion into the MOM. This appears to be independent of the presence of multiple positive charges in the N-terminus (data not shown). These positive charges are characteristic for mitochondrial targeting sequences. Likewise, Mix17 has been predicted to be a protein with such an N-terminal presequence (Claros and Vincens, 1996). However, in line with Gornicka et al. (2014), we observed that Mix17 did not require the mitochondrial membrane potential for its import into mitochondria, in contrast to typical presequence proteins employing the TIM23 pathway. Moreover, the Mix17 C4S variant that could not use the Mia40 import pathway was not imported into mitochondria. Apparently, the N-terminus cannot function as a typical presequence that is sufficient to mediate import into mitochondria. Nonetheless, we show that it strongly improves the import of Mix17, most likely due to its N-terminal positive charges. A similar observation was made for the import of the MICOS subunit Mic19. In contrast to typical Mia40 substrates, the presence of an unfolded DUF domain has been suggested to decrease the import efficiency of Mic19 (Ueda et al., 2019). In this case, myristoylation of the N-terminus counteracts this effect by mediating binding to the MOM and Tom20 (Ueda et al., 2019). Such inhibition of the import might also occur for Mix17, which has, for a Mia40 substrate, a remarkably large N-terminal segment. This segment is largely unfolded according to AlphaFold prediction (Jumper et al., 2021; Varadi et al., 2024). Interestingly, there are different results for Mic14, the mammalian ortholog of Mix17, on whether the N-terminus is required for import into mitochondria (Burstein et al., 2018; Lehmer et al., 2018). To address the function of mammalian Mic14, it will be crucial to elucidate whether Mic14 also exposes its N-terminus across the MOM.

Tom40 forms the central pore of the TOM complex, the main entry gate for import of proteins into mitochondria (Neupert and Herrmann, 2007; Wiedemann and Pfanner, 2017; Araiso et al., 2022). Interestingly, it has been reported that strongly overexpressed Mix17 can be co-purified with Tom40 during its import as a translocation intermediate (Gornicka et al., 2014). Our data suggest that endogenous Mix17 is embedded in the MOM in a proteinaceous environment after its import. Although we cannot rule out the existence of another yet unknown interaction partner of Mix17 in the MOM, we present the first evidence that Tom40 might provide this proteinaceous environment to mediate the exposure of Mix17 to the cytosol. First, we found that endogenous Mix17 interacts with Tom40. Second, the interaction with Tom40 was strongly reduced upon expression of the N-terminal truncation mutant Mix17 Δ1–24, which cannot be exposed to the cytosol. This interaction was only detected upon chemical crosslinking of mitochondria, indicating a transient or dynamic rather than a stable interaction. In agreement with a dynamic interaction, a significant fraction of Mix17 was released from mitoplasts generated by swelling of mitochondria. Interestingly, Mix17 could be crosslinked to Tom40 but not to Tom70. These results are consistent with the hypothesis that Mix17 spans the TOM pore formed by Tom40. Still, it is possible that Mix17 interacts with Tom40 at the protein–lipid interface of the TOM complex. Moreover, it cannot be excluded at this point that Tom40 transiently mediates the insertion and assembly of Mix17 into its proteinaceous environment in the MOM. The mode of interaction of Mix17 with Tom40 and its effect on the import capacity of the TOM complex are important questions to be addressed in further studies.

Interestingly, the functions of yeast Mix17 and its mammalian ortholog Mic14 appear to be conserved as both are associated with respiration (Longen et al., 2009; Burstein et al., 2018; Lehmer et al., 2018). Thus, the protein–protein interactions of Mix17 and Mic14 might also be conserved. In line with its enrichment at cristae junctions, Mic14 was shown to interact with MICOS (Bannwarth et al., 2014; Genin et al., 2016). However, this interaction appears to be cell type specific (Burstein et al., 2018). We show that Mix17 does not stably associate with MICOS but is in close proximity to Tom40. Interestingly, in yeast, MICOS was shown to interact with the TOM complex and to contribute to protein import (von der Malsburg et al., 2011; Bohnert et al., 2012). Therefore, it is tempting to speculate that Mix17 is also in the neighborhood of MICOS and might associate with the complex under specific conditions, perhaps to contribute to protein import. It will be an important task to analyze this and the factors contributing to its topology in future studies.

Chromosomal manipulations (knockouts, C-terminal tagging) were done as described before starting from the YPH499 wild-type strain (Longtine et al., 1998; Knop et al., 1999). The Δmix17 deletion strain was generated by replacing the entire coding region through marker cassette. In addition, the yeast wild-type strain D273-10b was used. The genotypes of the strains used in this study are listed in Table S1.

For the generation of plasmids, the MIX17 coding region was amplified by PCR. Point mutations were introduced by site-directed mutagenesis. Detailed information on the primers and restriction sites used are given in Table S2. Of note, for the in vitro synthesis of radiolabeled full-length Mix17, we exchanged the amino acid methionine at position 25 to alanine to omit the presence of an internal translation product that disturbs the analysis of the in vitro import assay.

Yeast cells were grown on YP medium (1% w/v yeast extract, 2% peptone) supplemented with 3% glycerol (YPG), synthetic medium (0.67% yeast nitrogen base) supplemented with 2% glucose (SD), 3% galactose (SGal) or 3% glycerol (SG) or lactate medium (Daum et al., 1982; Izawa and Unger, 2017). The yeast strain Mia40↓ carrying the MIA40 gene under control of the GAL10 promoter was first grown on galactose-containing medium and then transferred to glucose-containing medium for 16.5 h to deplete Mia40 from the cells. The Tob55↓ strain was grown similarly, but shifted for 21 h to deplete Tob55. Yeast wild-type cells and the respective mutants were grown in parallel and kept in logarithmic phase until further application.

Mitochondria were isolated according to established protocols (Daum et al., 1982; Izawa and Unger, 2017). Cells were harvested by centrifugation and washed with distilled water. The cell pellet was resuspended in DTT-containing buffer (10 mM final concentration) followed by incubation for 10 min at 30°C with gentle agitation. Spheroplasts were generated by incubation with zymolyase (200 U per gram wet weight of cells; Amsbio) for 30 min at 30°C with gentle agitation. Spheroplasts were opened by pipetting or douncing in lysis buffer [20 mM MOPS-KOH, pH 7.2, or 10 mM Tris, pH 7.4, and 1 mM EDTA, 0.6 M sorbitol, 0.2% (w/v) bovine serum albumin (BSA), 1 mM phenylmethylsulfonyl fluoride (PMSF)]. After a clarifying spin at 2000 g and 4°C for 5 min, the crude mitochondrial fraction was harvested by centrifugation at 14,000 g and 4°C for 10 min. Mitochondria were resuspended in isotonic SM buffer (0.6 M sorbitol, 20 mM MOPS, pH 7.4) or in SH buffer (0.6 M sorbitol, 20 mM HEPES, pH 7.4), frozen in liquid nitrogen and stored at −80°C.

Mitochondria were diluted in SM buffer to a concentration of 1 mg/ml. 100 µg of mitochondria was mixed with 100 µl of 200 mM sodium carbonate and incubated for 30 min on ice. Soluble and membrane proteins were separated by centrifugation for 30 min at 91,000 g (TLA55, Beckman Coulter) and 4°C. Soluble proteins in the supernatant were precipitated by the addition of trichloroacetic acid (TCA; 12%). The pelleted membrane proteins and the precipitated soluble proteins were resuspended in SDS sample buffer and analyzed by SDS-PAGE and immunoblotting (Fig. S2). Westerns blots were analyzed using a LI-COR Odyssey CLx scanner and Image Studio Lite software (LI-COR Biosciences; version 5.2.5), Image Scanner III (GE Healthcare) or the ChemiDoc imaging system with Image Lab 6.1 software (Bio-Rad).

50 μg mitochondria were diluted in SM buffer (0.6 M sorbitol, 20 mM MOPS, pH 7.4), swelling buffer (20 mM MOPS, pH 7.4) or lysis buffer [1% (v/v) Triton X-100, 20 mM MOPS, pH 7.4]. PK (final concentration, 12.5 µg/ml; Sigma-Aldrich) or trypsin (final concentration, 12.5 or 25 µg/ml; Sigma-Aldrich) was added as indicated and samples were incubated for 15 min on ice. Proteolysis was stopped by addition of PMSF (final concentration, 2 mM) or soybean trypsin inhibitor (final concentration, 500 µg/ml) followed by incubation for 10 min on ice. Triton X-100-containing samples were precipitated with TCA. Mitochondria and mitoplasts were centrifuged at 17,000 g and 4°C for 20 min, and the pellets were either first subjected to TCA precipitation and then analyzed or directly analyzed by SDS-PAGE and immunoblotting (Fig. S2). Westerns blots were analyzed using a LI-COR Odyssey CLx scanner and Image Studio Lite software (LI-COR Biosciences; version 5.2.5), Image Scanner III (GE Healthcare) or the ChemiDoc imaging system with Image Lab 6.1 software (Bio-Rad).

Radioactively labelled proteins were synthesized in the presence of ³⁵S-methionine in a standard or in a transcription- and translation-coupled reticulocyte lysate system (Promega). Mitochondria (final concentration, 0.5 mg/ml) were incubated in SI buffer (50 mM HEPES/KOH, pH 7.2, 0.5 M sorbitol, 80 mM KCl, 10 mM MgAc, 2 mM KH₂PO₄, 2.5 mM EDTA, 1 mM MnCl₂ and 0.01% fatty acid-free BSA) containing 10 mM creatine phosphate, 0.1 mg/ml creatine kinase, 4 mM NADH and 2 mM ATP for 3 min at 12°C. If membrane potential was dissipated (-ΔΨ), valinomycin (0.5 μM; Sigma-Aldrich), antimycin A (8 μM; Sigma-Aldrich) and oligomycin (20 μM; Sigma-Aldrich) were added in addition. Lysates were added and the reaction was further incubated at 12°C. At the indicated time points, samples were taken and diluted in ice-cold SH buffer (20 mM HEPES, 0.6 M sorbitol, pH 7.4) on ice. Mitochondria were pelleted and resuspended in SH buffer. Non-imported material was degraded by incubation with proteases (PK, 12.5 µg/ml; trypsin, 25 µg/ml) for 20 min on ice. Protease treatment was stopped by the addition of 2 mM PMSF or a 20-fold excess of soybean trypsin inhibitor, followed by incubation for 5 min on ice. Mitochondria were re-isolated by centrifugation for 10 min at 14,000 g and 4°C and analyzed by SDS-PAGE and autoradiography (Fig. S2).

When chemical crosslinking was performed prior to immunoprecipitation, 2 mg mitochondria was diluted in SI buffer [50 mM HEPES-KOH, 0.5 M sorbitol, 80 mM KCl, 10 mM Mg(Ac)₂, 2 mM KH₂PO₄, 2.5 mM EDTA, 2.5 mM MnCl₂, pH 7.2] to a concentration of 1 mg/ml. Dithiobissuccinimidylpropionate (DSP; Thermo Fisher Scientific) was added to a final concentration of 800 μM, followed by incubation for 30 min on ice. The reaction was stopped by the addition of 1 M glycine at pH 8.8 (final concentration of 160 mM), followed by incubation for 10 min on ice.

Mitochondrial ATP was enzymatically depleted for 10 min at 25°C using 15 U hexokinase (Sigma-Aldrich) and 20 mM glucose or 10 U apyrase (Sigma-Aldrich) per 1 mg mitochondrial protein to protect Mix17 from unspecific degradation. Mitochondria were re-isolated and the pellets were lysed in IP buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM EDTA) containing 1% (w/v) digitonin (Merck) and 2 mM PMSF. Cleared lysates were incubated with anti-HA agarose beads (30 μl beads/1 mg protein; Sigma-Aldrich) or anti-Mix17-conjugated protein A sepharose beads (50 µl beads) for 1.5 h. The beads were washed three times with 500 µl IP buffer containing 0.1% (w/v) digitonin and 1 mM PMSF. Bound proteins were eluted with SDS sample buffer. Given that digitonin disturbs the running behavior at the low molecular mass range, total lysate and unbound material were subjected to TCA precipitation to remove the digitonin when proteins with a molecular mass of ∼10 kDa needed to be detected. Otherwise, samples were directly diluted in SDS sample buffer. Samples were analyzed by SDS-PAGE and immunoblotting (Fig. S2). Westerns blots were analyzed using a LI-COR Odyssey CLx scanner and Image Studio Lite software (LI-COR Biosciences; version 5.2.5), Image Scanner III (GE Healthcare) or the ChemiDoc imaging system with Image Lab 6.1 software (Bio-Rad).

For immunoprecipitation of Mix17, anti-Mix17 antibody or pre-immune serum (Pineda Antikörper Service) was bound to protein A sepharose beads (GE Healthcare) overnight at 4°C. Beads were washed three times with 0.5 ml of 0.2 M boric acid (pH 9.0, NaOH). IgGs were coupled to protein A sepharose beads by incubation with 30 mM dimethylpimelimidate (Sigma-Aldrich) in 0.2 M boric acid (pH 9.0, NaOH) for 30 min at room temperature. Beads were washed three times with 0.5 ml of 0.2 M ethanolamine (pH 8.0, HCl) and crosslinking was quenched by incubation in 1 ml of 0.2 M ethanolamine (pH 8.0, HCl) for 2 h at room temperature. Afterwards, the beads were washed twice with 1 ml IP buffer, once with 1 ml of 0.1 M glycine at pH 2.5, twice with 0.5 ml of 0.1 M Tris/HCl at pH 8.5, and twice with 0.5 ml IP buffer.

The antibody against Mix17 was generated by injecting a peptide comprising amino acids 99–118 into rabbits. For affinity purification of the Mix17 antibody, the antigen was coupled to SulfoLink Coupling Resin (1 mg protein/ml bead volume; Thermo Fisher Scientific) in 50 mM Tris, 5 mM EDTA, pH 8.5. 6 ml of antisera was diluted in 24 ml of 10 mM Tris, pH 7.5, containing 1 mM PMSF, 1× complete protease inhibitor cocktail and 1 mM EDTA. The antisera were applied to the beads and washed successively with 10 mM Tris, pH 7.5, and 10 mM Tris, 0.5 M NaCl, pH 7.5. Elution of the antibody was performed in 100 mM sodium citrate, pH 4.0, followed by 100 mM glycine, pH 2.5. The individual fractions were immediately neutralized by addition of 1 M Tris, pH 8.8. The specificity of the antibody was tested using samples generated from the Δmix17 deletion strain or strains expressing tagged versions of Mix17 and can be seen in Fig. 4D. All additional antibodies used in this study are listed in Table S3.

M.E.H. thanks Dr Michael Kiebler, LMU Munich, for generous and extensive support, and Daniela Rieger for expert technical assistance. K.H. and M.R. thank Professor Andreas Ladurner for his generous support, and Petra Robisch and Alexandra Weinzierl for excellent technical assistance. Moreover, we thank Dejana Mokranjac for helpful discussion and reagents.

The peer review history is available online at https://journals.biologists.com/jcs/lookup/doi/10.1242/jcs.263661.reviewer-comments.pdf