Two driving forces energize precursor translocation across the inner mitochondrial membrane. Although the membrane potential (Δψ) is considered to drive translocation of positively charged presequences through the TIM23 complex (presequence translocase), the activity of the Hsp70-powered import motor is crucial for the translocation of the mature protein portion into the matrix. In this study, we show that mitochondrial matrix proteins display surprisingly different dependencies on the Δψ. However, a precursor’s hypersensitivity to a reduction of the Δψ is not linked to the respective presequence, but rather to the mature portion of the polypeptide chain. The presequence translocase constituent Pam17 is specifically recruited by the receptor Tim50 to promote the transport of hypersensitive precursors into the matrix. Our analyses show that two distinct Δψ-driven translocation steps energize precursor passage across the inner mitochondrial membrane. The Δψ- and Pam17-dependent import step identified in this study is positioned between the two known energy-dependent steps: Δψ-driven presequence translocation and adenosine triphosphate–driven import motor activity.
About two thirds of mitochondrial precursor proteins use N-terminal presequences as targeting signals (Vögtle et al., 2009). The presequence translocase (TIM23 complex) mediates transport of these precursors across the inner membrane (Neupert and Herrmann, 2007; Schulz et al., 2015). Initially, precursors are transported from the cytosol into the intermembrane space by the TOM complex in the outer membrane and are passed to the TIM23 complex (Chacinska et al., 2009; Endo and Yamano, 2010). Presequence-containing precursors can be subdivided into two classes: (1) precursors fully translocated across the inner membrane into the matrix and (2) precursors released from the translocase into the lipid phase of the inner membrane (inner membrane sorting). Precursor transport across the inner membrane is initially driven by the mitochondrial membrane potential (Δψ) that acts on the positively charged presequences (Schleyer et al., 1982; Roise and Schatz, 1988; Martin et al., 1991; Chacinska et al., 2009; Endo and Yamano, 2010; Schulz et al., 2015; Turakhiya et al., 2016). The Δψ draws the presequence of the polypeptide chain through the protein-conducting channel by electrophoretic force. This energy suffices to direct laterally sorted precursors into the inner membrane (van der Laan et al., 2007). However, translocation into the matrix requires ATP hydrolysis by the presequence translocase-associated motor (PAM), in addition to the Δψ (Neupert and Brunner, 2002; Schulz et al., 2015).
The presequence translocase consists of a channel-forming module formed by Tim23 and Tim17. Tim50 acts as the receptor for presequences (Meinecke et al., 2006; Qian et al., 2011; Schulz et al., 2011). In addition to these essential proteins, Tim21 and Mgr2 are also constituents of the TIM23 complex. Tim21 is specific to the “motor-free” state of the translocase and enables its association with proton-pumping respiratory chain complexes (van der Laan et al., 2006). Mgr2 is positioned at the lateral gate of the translocase to regulate inner membrane sorting (Gebert et al., 2012; Ieva et al., 2014) and participates in the dynamics of the mitochondrial import motor (Schulz and Rehling, 2014). For transport of matrix proteins, the import motor is recruited to the TIM23 complex. Its central force-generating constituent is mtHsp70 (Ungermann et al., 1994; Mapa et al., 2010). Whereas Tim44 positions mtHsp70 at the channel exit for precursor engagement (Liu et al., 2003), the Pam16/18 complex regulates its ATPase activity (D’Silva et al., 2003; Truscott et al., 2003; Kozany et al., 2004). In addition, Pam17 was suggested as a subunit or assembly factor of the import motor (van der Laan et al., 2005; Hutu et al., 2008). However, its molecular function has remained enigmatic.
In this study, we demonstrate that mitochondrial precursors differ significantly with regard to their Δψ requirement for import. In contrast to the current view of how Δψ energizes the translocation process, we find that a precursor’s hypersensitivity to the reduction of Δψ (Δψ hypersensitivity) is independent of its presequence but rather is linked to the mature portion of the polypeptide chain. Pam17 recruitment by the receptor Tim50 is specifically required for the import of these Δψ-hypersensitive precursors but is largely dispensable for import of precursors with low Δψ sensitivity. Accordingly, the Δψ energizes a second step in matrix translocation in a presequence-independent manner.
Matrix-destined precursor proteins display differential dependencies on Tim50
Tim50 is the major presequence receptor of the TIM23 complex in the inner mitochondrial membrane and regulates gating of the Tim23 pore. A surprising and still unresolved observation is that a loss of Tim50 leads to robust import defects for matrix proteins, but has a much lesser effect on precursors sorted into the inner membrane (Geissler et al., 2002). To assess the function of Tim50 in protein transport, we isolated mitochondria from Saccharomyces cerevisiae cells in which TIM50 was under control of the GAL1 promoter. Growing yeast in glucose-containing medium represses the GAL1 promoter and concomitantly blocks transcription of TIM50. To avoid secondary effects, levels of Tim50 were maintained at 20% of the wild-type (WT) amount. Under these conditions, the protein levels of other TIM23 complex components were similar between Tim50-depleted and WT mitochondria (Fig. 1 A). Because Tim50 regulates Tim23 channel activity, we assessed Δψ in mitochondria with reduced amounts of Tim50 using a potential-sensitive dye. Tim50-depleted and WT mitochondria were equally able to quench the fluorescent dye, indicating that the mitochondrial Δψ was not affected under our conditions of controlled Tim50 depletion (Fig. 1 B). We assessed the import capacity of matrix proteins in Tim50-depleted mitochondria with in vitro transport assays. Import of F1β and the model matrix protein b2(167)Δ–dihydrofolate reductase (DHFR), which consists of an N-terminal portion of cytochrome b2 fused to DHFR, was strongly affected, whereas inner membrane–sorted precursors (e.g., cytochrome c1) were much less Tim50 dependent, as previously reported (Fig. 1, C and D; and Fig. S1 A; Geissler et al., 2002). Considering that Tim50 serves as a presequence receptor in the intermembrane space and hence acts upstream of the import motor, the stronger reliance of motor-dependent substrates on Tim50, compared with motor-independent precursors, is surprising and still lacks an explanation (Geissler et al., 2002). To better understand the Tim50 dependence of mitochondrial proteins, we imported other matrix-targeted precursor proteins, like Tim44, a component of the import motor, and F1α, a soluble subunit of the F1Fo-ATP synthase. Surprisingly, the import of both precursor proteins was only mildly affected in Tim50-depleted mitochondria (Fig. 1, E and F). In fact, the import efficiency was similar to the mild import defect observed for inner membrane–sorted precursor proteins (Geissler et al., 2002).
To ascertain that the observed import defects were directly linked to Tim50 function, we screened for temperature-conditional tim50 mutants (Fig. S1 B). The mutant tim50-19 was selected for analysis because purified mitochondria exhibited WT-like Δψ (Fig. S1 C). Upon heat shock of isolated mitochondria, import of F1α was clearly more efficient than F1β in tim50-19 mitochondria (Fig. S1, D and E). Importantly, steady-state levels of mitochondrial proteins in tim50-19 mitochondria were similar to WT (Fig. S1 F). We conclude that impairment of Tim50 function, either by depletion of the protein or by mutagenesis of the TIM50 gene, affects the import of different matrix proteins to different extents.
Tim50 is important for Pam17 recruitment
To assess whether depletion of Tim50 affects TIM23 complex organization, we immunoisolated the TIM23 complex with Tim23-specific antibodies. No significant differences of the tested motor or core complex constituents associated with Tim23 were apparent. However, although steady-state levels of Pam17 were WT-like in Tim50-depleted mitochondria (Fig. 1 A), the amount of Pam17 copurified with Tim23 was drastically reduced (Fig. 2 A). A lack of Pam17 leads to a matrix protein import defect (van der Laan et al., 2005; Schiller, 2009). However, only a small set of model matrix proteins had been tested as import substrates. We therefore assessed whether import defects of pam17Δ mitochondria resembled those of Tim50-depleted mitochondria. To exclude unspecific effects, we confirmed that the Δψ was not affected in pam17Δ mutant mitochondria (Fig. 2 C) and in intact yeast cells (Figs. 2 B and S1 G). As reported, import of b2(167)Δ-DHFR and F1β was strongly affected in pam17Δ mitochondria (Fig. 2, D and E; van der Laan et al., 2005). In addition, import of Pam18 and Atp14 depended on Pam17 (Fig. S2, A and C). In contrast, import of F1α, Tim44, Atp5, and Mdj1 was only mildly affected in the absence of Pam17 (Fig. 2, F and G; and Fig. S2, B and D). Interestingly, this differential matrix import phenotype resembled the defects observed in mitochondria with reduced Tim50 levels (Fig. 2 H). To exclude the possibility that the observed differences in pam17Δ mitochondria were caused by different dependencies of the precursors on import motor activity, we imported F1α and F1β into mitochondria isolated from temperature-conditional mtHsp70 (Ssc1 in yeast) mutant cells (ssc1-3). For this, ssc1-3 cells were grown at a permissive temperature, and the phenotype was induced by shifting purified mitochondria to a nonpermissive temperature before the import reaction. After heat inactivation of mtHsp70, the import of both precursors, F1α and F1β, was compromised, demonstrating that both precursors depend to the same extent on motor function (Fig. S2, E and F). To assess whether Pam17-dependent precursors also accumulated in vivo, we generated cell lysates from WT and pam17Δ mutant cells. As expected, we found that the precursor of Atp14 was detectable in pam17Δ cells. Moreover, the amount of Pam18 was drastically reduced in the pam17Δ mutant (Fig. S2 G).
Pam17 affects matrix protein import independent of motor function
The selective role of Pam17 in matrix protein import suggested that Pam17 participates in import motor function (van der Laan et al., 2005). We therefore directly assessed the inward-driving force generated by the motor in WT and pam17Δ mitochondria. To this end, radiolabeled b2(167)Δ-DHFR was imported into mitochondria in the presence of methotrexate (MTX). MTX leads to a tight folding of the C-terminal DHFR moiety. The unfolded N-terminal part of the precursor is imported through the TOM and TIM23 complexes, whereas the DHFR moiety is pulled tightly against the TOM complex by the import motor. After initial import, the Δψ was dissipated, and proteinase K (PK) was added after different time points. In the absence of a Δψ, the import motor prevents the precursor from backsliding. Because the tightly folded DHFR moiety only becomes protease accessible if the precursor slides back, this assay enables an estimation of the pulling force of the import motor (Voisine et al., 1999). Remarkably, pam17Δ mitochondria did not display a pulling defect when the model protein b2(167)Δ-DHFR was used, whereas mutant mitochondria affected in Pam16 function showed a clear pulling defect for this precursor (Fig. 3 A). Our previous work showed that pam17Δ mitochondria displayed a defect in this assay when inner membrane–sorted b2(220)-DHFR was used (van der Laan et al., 2005). Because of the presence of a heme-binding domain, import of this precursor into mitochondria is motor dependent. Interestingly, mitochondria lacking Mgr2 also display protease sensitivity of the accumulated b2(220)-DHFR, caused by an accelerated release from the TIM23 complex into the inner membrane (Ieva et al., 2014; Schulz and Rehling, 2014). Surprisingly, Mgr2 levels were strongly reduced in mitochondria lacking Pam17 (Fig. S3 A). Hence, we compared the ability of mgr2Δ mitochondria to prevent backsliding of the two precursors b2(167)Δ-DHFR and b2(220)-DHFR. A significant amount of the b2(220)-DHFR intermediate became protease accessible in mgr2Δ mitochondria, whereas the protease sensitivity of b2(167)Δ-DHFR was similar in WT and mgr2Δ mitochondria (Fig. 3 B; Ieva et al., 2014). We concluded that the apparent pulling defect in pam17Δ was likely an indirect defect caused by loss of Mgr2 in these mitochondria.
Pam17 dynamically associates with the TIM23 core complex, and its recruitment is triggered by presequence recognition (Popov-Celeketić et al., 2008; Lytovchenko et al., 2013). However, once a precursor spans the TOM and TIM23 complexes and engages the import motor, Pam17 is released from the translocation intermediate (Fig. S3 B). We asked whether differences between the import efficiencies of Tim44 and F1α, and of F1β and b2(167)Δ-DHFR, were caused by distinct properties of their presequences. To test this hypothesis, the presequences of F1α and F1β were swapped. If the presequence is the determining factor, the import defects should be reversed (Fig. 3 C). However, the presequence swap did not alter the import defects observed in mitochondria lacking either Tim50 or Pam17 (Fig. 3, D and E). Thus, the presequences do not determine the differential Tim50 and Pam17 dependence of the precursors.
Two matrix protein classes display distinct Δψ dependencies
Because a presequence swap between F1α and F1β did not affect their import properties in pam17Δ mitochondria, we investigated other factors that might be responsible for the disparities in import efficiency. Previous analyses demonstrated that the unfolding of precursors did not rescue import into pam17Δ mitochondria (Schiller, 2009), indicating that the folding state of the preprotein does not influence its Pam17 dependency. Therefore, we analyzed whether the tested precursors displayed characteristic differences with regard to their Δψ dependence. We imported precursors into isolated mitochondria in the absence or presence of increasing amounts of the protonophore carbonyl cyanide m-chlorophenyl hydrazone (CCCP). For all tested proteins, the import efficiency decreased with lower Δψ, as expected. However, the import of F1α and Tim44 was significantly more efficient at low Δψ than the import of F1β and b2(167)Δ-DHFR (Fig. 4, A–F). Compared with F1α and Tim44, F1β and b2(167)Δ-DHFR displayed Δψ hypersensitivity. Intriguingly, the Δψ-hypersensitive precursors were also Pam17 and Tim50 dependent. This effect could also be recapitulated in vivo. When yeast cells were treated with increasing amounts of CCCP to gradually decrease the Δψ, the Pam17-dependent precursor Atp14 accumulated at a lower CCCP concentration than the Pam17-independent precursor of Mdj1 (Fig. S3 C).
The current concepts of presequence-mediated protein import into mitochondria state that the Δψ acts on positively charged residues of the presequence, and thereby drives the initial import of preproteins in an electrophoretic manner until the import motor engages with the preprotein. Accordingly, the Δψ dependence of a precursor should be mainly determined by presequence properties. However, we show that a swap of presequences between F1α and F1β does not affect the observed import phenotype in mitochondria affected in Pam17 or Tim50 function. We therefore tested whether a presequence swap can reverse Δψ hypersensitivity of the precursor proteins. To this end, we performed CCCP titration experiments for the mature portion of F1β carrying the F1α presequence (pF1α-F1β) and the mature portion of F1α fused to the presequence of F1β (pF1β-F1α). Astonishingly, the presequence did not alter the Δψ dependence of the imported precursor, but rather the mature portion of the polypeptide determined Δψ hypersensitivity (Fig. 4, G and H). To address whether a depletion of the Δψ affects the association of Pam17 with the translocase, we immunoisolated the TIM23 complex in the presence of CCCP. At a concentration of 40 µM, when import of Δψ hypersensitivity precursors is significantly affected, Pam17 remained bound to TIM23 (Fig. S3 D). We also analyzed the association of Pam17 with Tim23 by chemical cross-linking in the presence or absence of Δψ (Hutu et al., 2008). The reported Pam17–Tim23 cross-link was not affected by CCCP addition (Fig. S3 E). Accordingly, the observed Δψ hypersensitivity of precursors is not linked to a dissociation of Pam17 from the import machinery. We conclude that in addition to the universal Δψ-driven translocation of the presequence, a second Δψ-dependent translocation step promotes the transport of the mature portion in a class of mitochondrial matrix proteins.
The mitochondrial Δψ is a crucial driving force for inner membrane translocation that acts on precursor proteins in the vicinity of the inner membrane. We report that mitochondrial matrix proteins do not display a uniform dependency on Δψ. Although import of some matrix-targeted proteins is already severely impaired when the Δψ is slightly reduced, other precursors are hardly affected until Δψ is strongly compromised. Previous work defined the concept of two energy-dependent translocation stages (Chacinska et al., 2009; Neupert, 2015; Schulz et al., 2015). First, the Δψ acts on the positive charges of the presequence to drive precursor translocation in an electrophoretic manner. Second, mtHsp70 engages with precursors to energize the ATP-dependent unfolding of precursor proteins and their matrix translocation. We now show that the observed Δψ hypersensitivity of the precursors cannot be alleviated by replacement of the presequence with that of a less Δψ-dependent precursor protein. Our findings reveal an unexpected shortcoming in the current concepts of mitochondrial protein import. They confine a critical Δψ dependency also to the mature portion of matrix-targeted proteins. Accordingly, precursor translocation is facilitated by two independent Δψ-driven steps.
We show that Pam17 specifically promotes the transport of Δψ-hypersensitive precursors into the matrix and that the presequence receptor Tim50 is critical for efficient Pam17 recruitment to the TIM23 channel. Consequently, mitochondria affected in Tim50 function mimic the pam17Δ phenotype with regard to import defects for Δψ-hypersensitive precursors. Because the Pam17 dependency of precursor proteins is determined by their mature domain and because Tim50 likely contains more than one precursor protein binding site (Qian et al., 2011; Schulz et al., 2011; Lytovchenko et al., 2013; Rahman et al., 2014), it is tempting to speculate that Tim50 may not only interact with presequences, but also with segments of the mature region of precursors. Because there are no specific mutations defined that affect Tim50’s interaction with a presequence, the question of whether this domain also acts on the mature part can currently not be assessed experimentally. Our findings suggest that Tim50 may translate information about the nature of the precursor protein into an energy supply requirement. Pam17 has been considered to affect the functionality of the import motor (van der Laan et al., 2005; Hutu et al., 2008; Schiller, 2009). This model is based on matrix protein import defects observed in pam17Δ mutant mitochondria and the protease sensitivity of arrested b2(220)-DHFR (van der Laan et al., 2005). However, in this study, we show that the matrix protein import defect in pam17Δ is specific for Δψ-hypersensitive precursors and that the observed protease sensitivity of b2(220)-DHFR is likely caused by a loss of Mgr2, leading to enhanced lateral release of this precursor into the inner membrane. We therefore conclude that motor function is not rate limiting to the import of matrix proteins in pam17Δ mitochondria, but that the matrix protein import defect is linked to a Δψ-sensitive translocation step particularly affecting those precursors that are Δψ hypersensitive.
How does the Δψ affect the transport of Δψ-hypersensitive precursor proteins? Work by Huang et al. (2002) demonstrated that the Δψ participates in the unfolding of the mature portion of precursor proteins. However, this Δψ-mediated unfolding activity has been shown to depend on the precursor’s presequence. In line with this, chemical unfolding of the precursor does not alleviate the pam17Δ import defect (Schiller, 2009). Accordingly, the Pam17 dependence and Δψ hypersensitivity of precursors does not reflect an increased unfolding requirement during import. Regarding the order of transport events at the presequence translocase, the second Δψ-dependent transport step succeeds a presequence-dependent translocation event, but precedes the motor-requiring transport stages. Pam17 was shown to directly interact with the central core component, Tim23 (Hutu et al., 2008; Popov-Celeketić et al., 2008). Of note, Δψ affects the conformation of the Tim23 protein independently of presequence binding (Malhotra et al., 2013). In this light, we propose that the Tim23 channel gating activity contributes significantly to the passage of Δψ-hypersensitive precursors. This interpretation could link the transport reaction to an unresolved behavior of the Tim23 channel. Although the Tim23 channel is activated by presequences at low Δψ, in the absence of presequences the channel displays a biphasic activity pattern. At low Δψ, the gating activity of the Tim23 channel is low; however, it becomes activated at high Δψ (Truscott et al., 2001; van der Laan et al., 2007). Our analyses are in agreement with the idea that Δψ-hypersensitive precursors especially depend on this second, presequence-independent channel activation step, which may be facilitated by Pam17 recruitment to the import channel.
Materials and methods
Yeast growth and handling
Yeast strains were grown in YP medium (1% yeast extract and 2% peptone) containing 2% glucose (YPD) or 3% glycerol (YPG) at 30°C. pam17Δ (van der Laan et al., 2005), pam16-1 (Frazier et al., 2004), and corresponding WT strains were grown at 25°C, except for the experiment described in Fig. S2 G, for which pam17Δ and WT cells were grown at 30°C. mgr2Δ and the corresponding WT strain were shifted to 39°C for 24 h before harvesting (Gebert et al., 2012). A strain in which the TIM50 gene was under the control of a GAL1 promoter was described previously (Geissler et al., 2002). For down-regulation of Tim50 expression, yeast cells were precultured in YP medium containing 2% galactose, 1% raffinose, and 3% lactate, pH 5.0, and subsequently grown for 38 h at 30°C in YP medium containing 3% lactate, pH 5.0, and 0.2% glucose. In the absence of galactose and presence of glucose, the GAL1 promoter is repressed. Consequently, Tim50 levels are reduced as a result of protein turnover and cell division.
For the generation of the temperature-sensitive allele tim50-19, a plasmid-encoded WT TIM50 allele was replaced in the corresponding gene deletion strain derived from S. cerevisiae YPH499 by gap repair and shuffling of a TIM50 version obtained by error-prone PCR. Temperature-conditional alleles were selected by comparing the growth of strains at permissive and nonpermissive temperatures. Sequence analysis of the tim50-19 allele revealed four mutations, K93N in the matrix domain, F345L and W376R in the core domain, and I422V in the C-terminal presequence-binding domain.
For detection of import defects in cells, WT and pam17Δ strains were grown in YPD medium at 30°C overnight. The next morning, cells were diluted to OD600 = 0.2 and grown for 10 h in YPG at 30°C. Afterward, cells were harvested and cell lysates were analyzed by SDS-PAGE and Western blotting.
To assess mitochondrial membrane potential in cells, yeast strains were grown in YPG medium supplemented with 20 mg/L adenine at 25°C overnight. The next morning, cells were harvested at OD600 = 2–3, pelleted, and washed with 10 mM Tris/HCl, pH 6.8. Next, cells were incubated with 20 µM CCCP or the same volume of ethanol for 5 min at 25°C before addition of 2 µM JC-1. After 10 min of incubation at 25°C, cells were washed three times with 10 mM Tris/HCl, pH 6.8, and analyzed by microscopy.
For this, a 1:25 dilution of cells in a 384-well microtitre plate was automatically imaged at 30°C on an Imaging Machine 03-dual widefield high content screening microscope (Acquifer) equipped with a white light-emitting diode array for brightfield imaging, a light-emitting diode fluorescence excitation light source, an sCMOS (2,048 × 2,048 pixel) camera, a temperature-controlled incubation chamber, and a stationary plate holder in combination with movable optics. Images were acquired in brightfield and with 470-nm and 590-nm filter cubes (exitation 469/35 nm, emission 525/39 nm, and dichroic 497 nm; and exitation 590/20 nm, emission 628/32 nm, and dichroic 607 nm; respectively) with a 40× CFI Super Plan Fluor ELWD NA 0.60 (Nikon). Integration times were fixed at 200 ms for both fluorescence channels. The focal plane was detected in the brightfield channel using a yeast autofocus algorithm.
Fluorescence was quantified using a Knime (Berthold et al., 2009) pipeline to automate segmentation of the individual cells in the brightfield channel, followed by quantification of the mean cell fluorescence in both channels.
Import of precursor proteins
Precursor proteins were radiolabeled by translation in the presence of [35S]methionine using rabbit reticulocyte lysate (Promega). Mitochondria were resuspended in import buffer (250 mM sucrose, 10 mM MOPS/KOH, pH 7.2, 80 mM KCl, 2 mM KH2PO4, 5 mM MgCl2, 5 mM methionine, and 3% fatty acid–free BSA) supplemented with 2 mM ATP and 2 mM NADH. Membrane potential was dissipated using a final concentration of 8 µM antimycin A, 1 µM valinomycin, and 20 µM oligomycin. 20-µg/ml PK treatment was performed for 10 min on ice. 2 mM PMSF was added for 10 min on ice to inactivate PK. Mitochondria were sedimented, washed with SEM (250 mM sucrose, 1 mM EDTA, and 20 mM MOPS, pH 7.2) and further analyzed by SDS-PAGE and autoradiography. Quantifications were performed using ImageQuant TL (GE Healthcare) using a rolling ball background subtraction.
Protein complex isolation
Coimmunoprecipitation experiments using Tim23-specific serum were performed essentially as described previously (Herrmann et al., 2001). In brief, mitochondria were resuspended to 1 mg/ml in solubilization buffer (20 mM Tris/HCl, pH 7.4, 150 mM NaCl, 10% glycerol [wt/vol], 1 mM PMSF, and 1% digitonin) and incubated for 30 min on ice. After a clarifying spin, supernatant was loaded on protein A–Sepharose beads cross-linked to Tim23 antibodies, incubated for 1.5 h at 4°C on a rotating wheel, washed 10× (solubilization buffer with 0.3% digitonin), and eluted with a 50-µl double-bed volume of 0.1 M glycine, pH 2.8 (neutralized with 1-M Tris base). To address Pam17 association with the TIM23 complex at low membrane potential, mitochondria were incubated with CCCP in solubilization buffer lacking digitonin for 5 min on ice before solubilization with 1% (wt/vol) digitonin and coimmunoprecipitation.
Membrane potential measurements
Mitochondrial membrane potential was assessed using 3,3′-dipropylthiadicarbocyanine iodide (DiSC3(5)). Mitochondria were suspended in buffer containing 600 mM sorbitol, 1% (wt/vol) BSA, 10 mM MgCl2, and 20 mM KPi, pH 7.4, to a concentration of 166 µg/ml. Changes in fluorescence were recorded using a F-7000 fluorescence spectrophotometer (Hitachi) at 25°C with excitation at 622 nm, emission at 670 nm, and slits of 5 nm. Components were added to the cuvette containing 500 µl of buffer in the following order: DiSC3(5), 83 µg of mitochondria, and valinomycin (5 µl from 100-µM stock in EtOH, 1 µM final) to dissipate the membrane potential. The difference in fluorescence before and after the addition of valinomycin was used to compare relative membrane potential between strains.
Presequence swap of Neurospora crassa F1β and S. cerevisiae F1α was performed by overlap PCR. For pF1α-F1β, the first 38 aa of yeast F1α were fused to the mature part of N. crassa F1β (40–end). For pF1β-F1α, the first 45 aa of F1β were fused to the mature part of F1α (36–end).
Assessing import-driving activity
Import-driving activity was assessed as previously described (Voisine et al., 1999). For import experiments, temperature-sensitive strains were incubated in import buffer for 15 min at 37°C, and subsequently 2 mM ATP, 2 mM NADH, 5 mM creatine phosphate, and 0.01 mg/ml creatine kinase were added. Radiolabeled b2(167)Δ-DHFR was imported at 25°C for 15 min in the presence of 5 µM MTX, and membrane potential was dissipated using 1 µM valinomycin. A sample was taken (Δt = 0 min), the precursor was chased at 25°C, and samples were treated with PK after 1, 3, 6, and 11 min for 15 min on ice. Samples were analyzed by SDS-PAGE and autoradiography. The amount of PK-resistant intermediate was quantified and standardized to the total amount of generated intermediate (Δt = 0) for each strain.
Membrane potential reduction by CCCP titration
For reduction of membrane potential, the protonophore CCCP was used as previously reported (van der Laan et al., 2006). Mitochondria were resuspended in import buffer with 1% BSA and 20 µM oligomycin to prevent regeneration of membrane potential by the reverse function of the F1Fo-ATPase. CCCP was added from a 4 mM stock in EtOH, and mitochondria were incubated for 5 min at 25°C before import.
For in vivo CCCP titration, cells were grown at 30°C overnight in YPD medium. The next day, cells were diluted to OD600 = 0.5 in 2× YPAD, and grown for another 5 h. Then, cells were diluted back to OD600 = 1 in 2× YPAD, and CCCP was added from a 50 mM stock solution. After 30 min of incubation, cells were harvested and cell lysates were analyzed by SDS-PAGE and Western blotting. For quantification, the amount of accumulated precursor at 80 µM CCCP was set to 100%.
For cross-linking, mitochondria were resuspended in import buffer without BSA to a final concentration of 1 mg/ml. CCCP was added from a 5 mM stock solution, and mitochondria were incubated for 5 min on ice. Next, disuccinimidyl glutarate in DMSO was added to a final concentration of 500 µM, and mitochondria were incubated for 30 min on ice. Afterward, excess cross-linker was quenched using 100 mM glycine, pH 8.0, for 10 min on ice. After reisolation of mitochondria, cross-links were analyzed by SDS-PAGE and Western blotting.
Generation and isolation of the TIM23–TOM supercomplex
A TIM23–TOM supercomplex was isolated essentially as described previously (Chacinska et al., 2003). In brief, recombinant b2(167)Δ-DHFR was imported into mitochondria in import buffer with BSA in the presence of 5 µM MTX for 15 min at 25°C. Next, mitochondria were reisolated, washed with SEM, resuspended to 1 mg/ml in solubilization buffer (20 mM Tris/HCl, pH 7.4, 150 mM NaCl, 10% glycerol [wt/vol], 1 mM PMSF, and 1% digitonin), and incubated for 30 min on ice. After a clarifying spin, complexes were isolated by incubating lysates with Tim23- or Tom22-specific antibodies coupled to protein A–Sepharose beads. Beads were washed 10× with solubilization buffer containing 0.3% (wt/vol) digitonin and eluted with 100 mM glycine, pH 2.8.
Isolation of mitochondria
Mitochondria were isolated essentially as previously described (Meisinger et al., 2006). If not stated otherwise, yeast cells were grown in YPG medium and harvested at OD600 = 2–3. After treatment with buffer A (10 mM DTT and 100 mM Tris/H2SO4, pH 9.4) for 30 min at 30°C, cells were washed and treated with zymolyase buffer (20 mM KPO4, pH 7.4, 1.2 M sorbitol, and 0.57 mg/L zymolyase) for 1–2 h at 30°C. After additional washing in zymolyase buffer without enzyme, cells were resuspended in ice-cold homogenization buffer (600 mM sorbitol, 10 mM Tris/HCl, pH 7.4, 1 g/L BSA, 1 mM PMSF, and 1 mM EDTA) and opened using a cell homogenizer. The mitochondrial fraction was obtained by differential centrifugation, resuspended in SEM buffer, and frozen in liquid nitrogen.
Online supplemental material
Fig. S1 shows corresponding import experiments to Fig. 1 in tim50-19 mitochondria. Moreover, in vivo membrane potential assessment in pam17Δ cells and WT cells are shown. Fig. S2 shows precursor imports in pam17Δ mitochondria and in vivo precursor accumulation, extending Fig. 2 (D–G). Also, F1α and F1β imports in ssc1-3 mitochondria show that both precursors are to the same extent motor dependent for import. Fig. S3 relates to Fig. 3 and shows steady-state protein analysis of pam17Δ mitochondria and the Pam17 association with the TIM23 complex when a precursor spans both TOM and TIM23 complexes. Fig. S3 also shows in vivo CCCP titration and Pam17 association with the TIM23 complex under low-membrane potential conditions supporting data in Fig. 4.
We thank M. Meinecke for discussion, S. Callegari for critical reading, and Á. Farkas for support in data analyses.
This work was supported by the Deutsche Forschungsgemeinschaft (DFG) grants SFB860 to P. Rehling, SFB746 to M. van der Laan, and SFB1190/Z03 to A. Clancy. This work was also supported by the Molecular Biology PhD program at the International Max Planck Research School and Göttingen Graduate School for Neurosciences and Molecular Biosciences DFG grant GSC 226/1 to A.B. Schendzielorz and C. Schulz, the Excellence Initiative of the German federal and state governments EXC 294 BIOSS grant to M. van der Laan, the Centre National de la Recherche Scientifique–Institut National de la Santé et de la Recherche Médicale ATIP-Avenir program grant to R. Ieva, Boehringer Ingelheim Fonds grant to C. Schulz, and the Max Planck Society (to P. Rehling).
The authors declare no competing financial interests.