Conserved regions of budding yeast Tim22 have a role in structural organization of the carrier translocase

Mitochondria are indispensable organelles required for diverse cellular process, including metabolism, bioenergetics, and signaling (Nunnari and Suomalainen, 2012; Sokol et al., 2014; Kang et al., 2018; Pfanner et al., 2019). Optimal mitochondrial health depends on accurate and efficient transport of ∼1500 mitochondrial proteins from the cytosol to their destined mitochondrial sub-compartments (Mootha et al., 2003; Rhee et al., 2013; Hung et al., 2014; Vögtle et al., 2017). The import of mitochondrial proteins is mediated by the coordinated action of specialized translocase machinery present in the outer (OM) and inner mitochondrial membrane (IM) (Baker et al., 2007; Neupert and Herrmann, 2007; Endo and Yamano, 2009; Wiedemann and Pfanner, 2017). The TOM complex present in the OM acts as a general entry gate for most of the nuclear-encoded mitochondrial proteins (Dukanovic and Rapaport, 2011; Bausewein et al., 2017; Araiso et al., 2019). The IM possesses two distinct types of machinery; the presequence translocase (TIM23 complex) and the carrier translocase (TIM22 complex). The TIM23 complex mediates import of presequence-containing proteins into the matrix or the IM, while the TIM22 complex facilitates the IM insertion of more complex multispanning membrane proteins that have internal hydrophobic targeting sequences (Sirrenberg et al., 1996; Rehling et al., 2004; Chacinska et al., 2009; Schmidt et al., 2010; van der Laan et al., 2010; Dudek et al., 2013; Schendzielorz et al., 2018). Substrates of the TIM22 pathway include members of mitochondrial metabolite carriers as well as subunits of the translocases, such as Tim23, Tim22, and Tim17 (Endres et al., 1999; Leuenberger et al., 1999; Paschen et al., 2000).

In yeast, the TIM22 complex is a 300 kDa molecular machinery consisting of a membrane module and a peripheral module (Koehler, 2004; Sokol et al., 2014). The membrane module contains four membrane-integral components, namely Tim22, Tim54, Tim18 and Sdh3, whereas the peripheral module includes small Tim proteins Tim9, Tim10 and Tim12 (Jarosch et al., 1996, 1997; Kerscher et al., 1997; Adam et al., 1999; Koehler et al., 2000; Kovermann et al., 2002; Gebert et al., 2011; Stojanovski et al., 2012). Tim22 exists as a dimer in vivo, forming a twin-pore protein-conducting channel, while Tim18 and Sdh3 are required for assembly of the TIM22 complex (Rehling et al., 2003; Gebert et al., 2011; Okamoto et al., 2014). Tim54 is suggested to function as an adaptor protein for docking of small Tim proteins onto the TIM22 core complex (Hwang et al., 2007; Wagner et al., 2008). Tim9 and Tim10 form a soluble heterohexameric complex of 70 kDa in the intermembrane space (IMS), but a portion also associates with the TIM22 complex through interaction with Tim12 (Adam et al., 1999; Webb et al., 2006; Gebert et al., 2008). The family of small Tim proteins contains two additional chaperone components, Tim8 and Tim13, which mediate the import of few non-carrier proteins, such as Tim23 (Koehler et al., 1999; Leuenberger et al., 1999). Recent findings highlight that even porin (also known as VDAC) interacts with the TIM22 complex to promote carrier protein biogenesis (Ellenrieder et al., 2019). Although the function of the TIM22 complex is conserved from yeast to human (yeast Tim22 shares 41% sequence similarity with its human counterpart), the composition of the human TIM22 complex partially differs from that of yeast. The human TIM22 complex includes two metazoan-specific membrane components, acylglycerol kinase (AGK) and TIM29, which are required for the integrity and translocase activity of the complex (Callegari et al., 2016; Kang et al., 2016, 2017; Vukotic et al., 2017). Intriguingly, the human TIM22 complex is also shown to be associated with the mitochondrial contact site and cristae organizing system (MICOS) complex, required for the efficient import of carrier proteins (Callegari et al., 2019). Furthermore, mutations in different components of the TIM22 pathway are implicated in various pathological conditions. Mutations in TIM8A result in Mohr–Tranebjaerg syndrome (a progressive neurodegenerative disorder), whereas mutations in AGK are associated with Sengers syndrome (Jin et al., 1999; Koehler et al., 1999; Roesch et al., 2002; Kang et al., 2017; Vukotic et al., 2017). Recently, pathogenic variants in TIM22 were identified and linked with the development of mitochondrial myopathy (Pacheu-Grau et al., 2018).

Despite the identification of multiple interacting subunits of the TIM22 complex in the last two decades, their precise recruitment to the TIM22 channel, and the connection between this process and pathophysiology, remains poorly understood. This study uncovers a novel set of conditional mutants from conserved regions of Tim22, which are specifically impaired in the association with other subunits of the TIM22 complex. Our genetic and biochemical analyses provide novel insights into the defined functions of the conserved regions of Tim22 in the assembly of the carrier translocase, lateral pre-protein sorting, and overall mitochondrial integrity.

The predicted topology of Tim22 closely resembles that of Tim23 and Tim17, containing four transmembrane regions (TM1, TM2, TM3 and TM4) with both amino and carboxyl termini exposed towards the IMS (Fig. 1A) (Wrobel et al., 2016). However, the precise organization of Tim22 protein as a core channel in modulating the recruitment of partner subunits for translocase activity is still elusive due to the lack of suitable conditional mutants that are specifically defective in the complex assembly.

As a first step, we performed a genetic screen using an error-prone PCR library of TIM22 to obtain conditional mutants with compromised growth. Five novel conditional mutants in different regions of the Tim22 protein, namely tim22_K45A/D151A in the IMS region, tim22_G57E in TM1, tim22_K127A in TM2, tim22_G163L in TM3 and tim22_G179L in TM4, were isolated (Fig. 1A,B). Most of the conditional mutations isolated are of residues that are highly conserved across species (Fig. S1). Of the conditional mutants isolated, the IMS and TM4 mutants exhibited a severe growth defect at 37°C in both dextrose and lactate media (Fig. 1C). Additionally, the TM4 mutant showed growth sensitivity even at 34°C (Fig. 1C). The TM1 and TM2 mutants displayed growth defects at 37°C only in lactate medium. Moreover, the TM1 mutant also exhibited a mild cold-sensitive growth phenotype at 24°C (Fig. 1C). Strikingly, the TM3 mutant exhibited a severe cold-sensitive phenotype at 24°C in addition to a mild growth defect at 30°C (Fig. 1C). Conversely, the corresponding single point mutants from the IMS region, namely tim22_K45A and tim22_D151A, did not display any growth defect (Fig. S2). Taken together, these results suggest that the IMS and TM regions of Tim22 play a critical role in its translocase activity.

To assess the effect of mutations on the stability of Tim22, the steady-state protein levels of wild type (WT) and mutants were estimated in isolated mitochondria. Protein levels of Tim22 and Tom70 in most of the mutants were found to be similar to those in WT (Fig. 1D). However, IMS and TM4 mutants showed a drastic reduction in the steady-state levels of Tim22 (Fig. 1D). To eliminate the possibility that the growth defects observed in these mutants were not due to reduced protein levels, the mutant proteins were overexpressed under the control of different centromeric plasmids. Strikingly, even though the resulting steady-state levels of mutant Tim22 proteins were comparable to those in WT, overexpression did not completely rescue the growth defects of the mutants observed at non-permissive temperatures (37°C), when compared to WT (Fig. 1E,F). We further utilized these overexpression strains (IMS_CYC1 and TM4_GPD) for all the downstream analyses to delineate the significance of the IMS and TM4 regions of Tim22. Furthermore, the expression levels of other subunits of the TIM22 complex, as well as of the TIM23 complex, remained unaltered in the isolated mitochondria of all the mutants (Fig. 1E). Taken together, these results indicate that the compromised growth phenotype of tim22 mutants is a consequence of defective translocase function.

To gain insights into the growth defects of tim22 mutants, we first tested the effects of the mutations on maintenance of the architecture of the TIM22 complex by analyzing the stability of the TIM22 complex using blue native PAGE (BN-PAGE). Mitochondria isolated from WT cells grown at 30°C in dextrose medium showed an intact TIM22 complex at 300 kDa when analyzed with antibodies specific for Tim22, Tim18 (Tim18–FLAG) and Tim54. Interestingly, the IMS, TM3, and TM4 mutants exhibited a drastic reduction in levels of the 300 kDa complex (Fig. 2A). Conversely, levels of the 300 kDa TIM22 complex were unaffected in TM1 and TM2 mutants (Fig. 2A). To further asses the role of the TM1 and TM2 regions, mitochondria were isolated from cells grown at 30°C in lactate medium, subjected to BN-PAGE, and the TIM22 complex was immunodetected using anti-Tim22, anti-FLAG, and anti-Tim54 antibodies. Intriguingly, the levels of TIM22 complex were considerably lower in the case of the TM1 mutant, whereas the TM2 mutant showed a stable 300 kDa complex similar to that of WT (Fig. 2B). As a control, the amounts of the respiratory chain complexes (III₂ IV₂ and III₂IV) and TIM23 complex were analyzed and found to be comparable between WT and all the mutants (Fig. 2A,B). These findings suggest that mutations in the IMS and TM regions (except TM2) of Tim22 destabilize the carrier translocase assembly.

Tim22 forms an intramolecular disulfide bond that is crucial for the stability of the TIM22 translocase (Wrobel et al., 2013; Okamoto et al., 2014). Because the IMS mutation identified in our genetic screen resides close to one of the cysteine residues involved in the disulfide-bond formation (Fig. 2C), we examined whether disulfide bond formation is affected in the tim22 mutants. To test this, the migration rates of WT and mutant Tim22 were assessed under reducing and non-reducing conditions in mitochondria isolated from cells grown at 30°C. Upon analysis, WT and mutants of the TM1, TM2, and TM3 regions exhibited faster migration of Tim22 (below 20 kDa) in non-reducing conditions than in reducing conditions, suggesting that Tim22 was able to form a stable disulfide bond (Fig. 2D). However, IMS and TM4 mutants showed a migration rate similar to the reduced form of Tim22 (above 20 kDa), indicating that these mutants are defective in a stable disulfide bond formation (Fig. 2D). Intriguingly, the disulfide bond defect associated with the TM4 mutant suggests that the TM4 and IMS regions of Tim22 are presumably close to each other in the central core channel (Fig. 2E).

As a next step, the recruitment of different subunits to the core TIM22 channel in the conditional mutants was examined by co-immunoprecipitation (Co-IP). Mitochondria from WT and mutants expressing C-terminally FLAG-tagged Tim18 were isolated from cells grown at 30°C in dextrose medium and subjected to Co-IP using anti-FLAG antibody prebound to protein G-conjugated Sepharose beads. The subunits Tim22, Tim54, and Sdh3 were efficiently immunoprecipitated with Tim18–FLAG in WT (Fig. 3A). However, the levels of Tim54, Sdh3, and Tim22 were significantly reduced in co-precipitates from IMS and TM4 mutants (Fig. 3A). Furthermore, the TM3 mutant showed impairment in Sdh3 interaction, whereas Tim22 and Tim54 were efficiently co-purified along with Tim18–FLAG (Fig. 3A). To ascertain specificity, Tim44 – a subunit of the TIM23 complex – was used as a negative control (Fig. 3A). To determine the interaction defects of TM1 and TM2 mutants, Co-IP was performed using mitochondria isolated from cells grown at 30°C in lactate medium. Interestingly, the TM2 mutant did not show any significant defect in the association of different subunits. Conversely, the amounts of Tim22 co-purified with Tim18–FLAG were significantly lesser in the TM1 mutant, though the reduction in amounts of co-precipitated Tim54 and Sdh3 were not substantial (Fig. 3B). This reduction in Sdh3 and Tim54 could be a consequence of Tim22 and Tim18 interaction defects. Furthermore, the TM2 mutant showed compromised interaction with Tim18 (along with slight reductions in Tim54 and Sdh3 pulldown) when subjected to heat shock at 37°C, suggesting that the TM2 region of Tim22 might be important for TIM22 complex stability at higher temperatures (Fig. 3C). To further explore the genetic interactions, we overexpressed different subunits of the TIM22 complex and tested for growth rescue of tim22 mutants upon overexpression of other subunits. Strikingly, overexpression of Tim18 rescued the growth defects of IMS and TM4 mutants on non-fermentable medium at 37°C (Fig. S3A), and the deletion of Tim18 was lethal in IMS, TM3, and TM4 mutants (Fig. S3B). Taken together, these results suggest that mutations in the IMS and TM regions of Tim22 severely influence the interaction of Tim18 and Sdh3 leading to destabilization of the TIM22 complex, in agreement with previous findings (Koehler et al., 2000; Gebert et al., 2011).

To test the regions of Tim22 required for Tim54 interaction, pulldown analysis was performed using Ni-NTA agarose beads in WT and mutant mitochondria (expressing C-terminally His-tagged Tim54) isolated from cells grown at 30°C in either dextrose (WT, IMS, TM3, and TM4) or lactate media (WT, TM1, and TM2). Upon analysis, Tim54–His from WT mitochondrial lysate efficiently bound to the Ni-NTA agarose beads and pulled down Tim22, Tim18, and Tim12 (Fig. 4A). In contrast, the IMS and TM4 mutants showed a drastic reduction in the association of Tim22, Tim18, Sdh3, and Tim12 (Fig. 4A). However, TM1, TM2, and TM3 mutants did not exhibit any interaction defect between Tim54 and Tim22, although the TM1 mutant showed mild impairment in Tim18 interaction, in agreement with the results of Tim18–FLAG immunoprecipitation (Fig. 4B). These observations highlight that Tim54 specifically interacts with the IMS and TM4 regions of Tim22. Furthermore, according to a previous report and bioinformatics analysis, the predicted topology of Tim54 contains one transmembrane region and a large IMS-exposed soluble domain (Mühlenbein et al., 2004). Based on our findings, it is reasonable to believe that the IMS regions of Tim54 and Tim22 are in close orientation for a stable interaction, while the TM region of Tim54 interacts with the TM4 region of Tim22.

Tim54 facilitates the docking of small Tim proteins to the core TIM22 complex by directly interacting with Tim10 (Hwang et al., 2007; Wagner et al., 2008). Intriguingly, the IMS and TM4 mutants showed a defect in Tim12 interaction (Fig. 4A), suggesting that impairment in these regions could influence the binding of small Tim proteins directly or indirectly. To further confirm this, Co-IP was performed at a permissive temperature using anti-Tim12 pre-bound to protein A-conjugated Sepharose beads. WT and the TM3 mutant did not display any impairment in the interaction of different subunits with Tim12 (Fig. 4C). However, IMS and TM4 mutants showed compromised co-precipitation of Tim22 and Tim18 (Fig. 4C). The defect in Tim18 pulldown could be a consequence of a defect in the interaction between Tim22 and Tim18. Strikingly, the interaction of Tim54 with Tim12 was also compromised in both of these mutants, whereas Tim10 was efficiently immunoprecipitated with Tim12 (Fig. 4C). These results indicate that the stable interaction of partner subunits with Tim22 is crucial for the association of Tim12, and any impairment in the Tim22–Tim54 interaction could result in defective docking of small Tim proteins.

The deletion or impairment of associations between different subunits of the TIM22 complex affects the import of polytopic IM proteins (Kerscher et al., 1997; Koehler et al., 1998, 2000; Sirrenberg et al., 1998; Adam et al., 1999). To determine whether the import of TIM22 pathway substrates is affected in tim22 mutants, in vitro import was performed using purified Tim23 tagged with 6× histidine (His) as a model substrate of the TIM22 pathway (Fig. 5A) (Kaldi et al., 1998; Davis et al., 2000; Koehler et al., 2000). Interestingly, the TM2 mutant showed mild impairment in import efficiency compared to the WT, in agreement with its interaction defects (Fig. 5B,C). In contrast, IMS and TM4 mutants exhibited a drastic reduction in the import efficiency of recombinant Tim23 (Fig. 5B,C). Furthermore, TM1 and TM3 mutants showed compromised import efficiency, correlating with their growth defects (Fig. 5B,C).

We further tested whether the growth defect of tim22 mutants is a consequence of impaired import of TIM22 pathway substrates then overexpression of substrates should exacerbate the growth defects of tim22 mutants. To validate this, we overexpressed metabolite carrier proteins such as PIC (also known as Pic2), AAC2 (also known as Pet9), and DIC1, as well as Tim23 (TIM22 pathway substrates) and tested for growth sensitivity. Strikingly, overexpression of PIC exacerbated the growth defects of TM1, TM2, and TM4 mutants, and resulted in a lethal phenotype in IMS and TM3 mutants (Fig. 5D; Fig. S4A). Furthermore, overexpression of AAC2 aggravated the growth defects of most of the tim22 mutants and was lethal in the TM3 mutant (Fig. 5E; Fig. S4B). Similarly, overexpression of DIC1 and Tim23 significantly enhanced the growth defects of most of the mutants (Fig. S4C,D). At the same time, as a control, the overexpression of an OM protein, Tom6, or a matrix protein, Abf2, did not affect the growth defects of tim22 mutants under the same experimental conditions (Fig. 5F,G). These results suggest that any impairment in the IMS and TM regions of Tim22 affects the import of complex substrate proteins, thus further emphasizing the significance of the conserved regions of Tim22 for its translocase activity and in vivo function. Intriguingly, overexpression of carrier proteins had a drastic effect on cellular viability and growth of TM3 mutants, indicating that this transmembrane region could play a critical role in the translocation of metabolite carrier proteins through the TIM22 complex.

The architecture of the mitochondrial IM mainly depends upon the accuracy in the import and precise integration of multiple transporters and translocation channels. The TIM22 translocase plays a crucial role in this process, thus overall influencing mitochondrial biogenesis. To test the role of Tim22 in modulating mitochondrial integrity, morphological differences between WT and tim22 mutants were analyzed using microscopy. The WT, TM1, TM2, and TM4 mutants showed normal tubular mitochondrial morphology at the permissive temperature (Fig. 6A). Conversely, IMS and TM3 mutants exhibited increased reticularity in their mitochondrial structures (Fig. 6A). To check whether the increased mitochondrial network could be a consequence of an increase in mitochondrial mass, flow cytometry was performed using a fluorescent dye, 10-N-Nonyl Acridine Orange (NAO), which specifically binds to cardiolipin in the mitochondrial membrane, thus providing an assessment of total mitochondrial mass (Gallet et al., 1995; Goswami et al., 2012). In agreement with our microscopic analysis, IMS and TM3 mutants displayed a significant increase in total mitochondrial mass as compared to that of WT (Fig. 6B). The increase in the mitochondrial mass in the IMS and TM3 mutants was further ascertained by testing for an enhancement in mitochondrial protein levels using western blotting, with Yme1 used as a mitochondrial marker and PGK1 serving as a cytosolic control (Fig. 6C,D). To test the functionality of the mitochondria, WT and tim22 mutants were stained with two membrane potential-dependent dyes, Tetramethylrhodamine ethyl ester (TMRE) and MitoTracker^® Deep Red (MTDR), followed by flow cytometric analysis. Strikingly, IMS and TM3 mutants exhibited increased fluorescence of both TMRE and MTDR, as compared to WT, suggesting that both mutants have an increase in the accumulation of functional mitochondrial mass (Fig. S5A,B). Remarkably, overexpression of carrier proteins drastically affected the growth of IMS and TM3 mutants; hence, the observed increase in mitochondrial mass could be an adaptive survival strategy for compensating the functional defect.

The deletion or mutation of different subunits of TIM22 complex results in a petite-negative phenotype indicating their importance in mitochondrial DNA maintenance (Dunn and Jensen, 2003; Senapin et al., 2003; Hwang et al., 2007). Because the IMS and TM mutants of Tim22 showed compromised interaction with different subunits, we examined whether mutations in Tim22 lead to a petite-negative phenotype. Intriguingly, when WT and tim22 mutants were grown in YPD medium containing ethidium bromide (EtBr), most of the tim22 mutants showed a lethal phenotype (Fig. 6E). However, the TM2 mutant produced viable colonies on YPD supplemented with EtBr, but displayed a mild growth defect compared to WT (Fig. 6E). Furthermore, the overexpression of Tim18 did not rescue the growth defects of mutants upon EtBr treatment (Fig. S5C). Taken together, these results suggest that the functionality of the TIM22 complex is crucial to maintain the normal mitochondrial network as well as the viability of cells lacking mitochondrial DNA.

The TIM22 complex is one of the critical protein translocases required for the import and integration of complex polytopic IM proteins. Tim22, the central channel-forming module, is the only membrane component of the TIM22 pathway that is evolutionarily conserved along with the small Tim proteins Tim9, Tim13, Tim8 (TIM8A and TIM8B in humans), and Tim10 (TIM10A and TIM10B in humans; TIM10B acts as a functional homolog of yeast Tim12) (Bauer et al., 1999; Jin et al., 1999; Mühlenbein et al., 2004; Pacheu-Grau et al., 2018). Interestingly, the human TIM22 and yeast Tim22 share similar topological organizations, suggesting that the conserved regions of Tim22 might play a crucial role in maintaining the architecture, as well as the in vivo function, of the TIM22 complex. In this study, we have dissected the significance of conserved regions of Tim22 in regulation of the structural and functional organization of the TIM22 complex by selectively interacting with partner subunits (Fig. 7A). Our genetic and biochemical analyses provide compelling evidence that impairment in the interaction of the partner subunits with the IMS and TM regions of Tim22 structurally destabilizes the TIM22 core complex leading to compromised translocase function, particularly during import of complex inner membrane substrate proteins, which is reflected in the exacerbation of cellular growth defects upon overexpression of metabolite carrier proteins. Furthermore, we provide evidence to highlight that the functional TIM22 complex is required for maintenance of the mitochondrial network and viability of cells lacking mitochondrial DNA.

Our findings support a model in which the IMS and TM4 regions of Tim22 are required for the association of Tim54, Tim18, and Sdh3 (Fig. 7B). Additionally, mutation of the IMS and TM4 regions of Tim22 resulted in compromised Tim12 binding, probably due to the destabilization of the core complex. However, our findings also raise the possibility that Tim12 has dual interaction with the IMS region of Tim22 and Tim54, which requires further experimental validation. Furthermore, the TM1 and TM2 regions of Tim22 are involved in the association with Tim18, whereas the TM3 region is essential for interaction with Sdh3 (Fig. 7C,D). Based on our analysis, we propose that the IMS and TM4 regions of Tim22 are critical for TIM22 complex assembly. This further supports the recent finding where a point mutation within the IMS region of human TIM22 was shown to cause mitochondrial myopathy by affecting the stability of the TIM22 complex, leading to compromised translocase activity (Pacheu-Grau et al., 2018).

Understanding the structural organization of the TIM22 complex is of vital importance because it is required for the biogenesis of metabolite carriers and the TIM23 complex, which are essential for mitochondrial integrity and cell survival. Remarkably, carrier proteins, the classical substrates of the TIM22 pathway, are present in high abundance as compared to Tim22. Hence, any impairment in TIM22 assembly is detrimental to cells due to the susceptibility to aggregation of complex hydrophobic proteins in the cytosol, leading to mitochondrial precursor over-accumulation stress (mPOS), activation of the mitochondrial unfolded protein response (UPRmt), and organelle dysfunction (Wang and Chen, 2015; Rolland et al., 2019). Such pathological implications are well known to associate with aging and neurodegeneration in humans (Lin and Beal, 2006; Tatsuta and Langer, 2008). This is further supported by the observations that Tim22 is essential across the phylogeny of eukaryotes, and that mutations in several subunits of the TIM22 pathway are implicated in the development of neurodegenerative diseases, such as Mohr–Tranebjaerg syndrome and mitochondrial myopathy (Sirrenberg et al., 1996; Koehler et al., 1999; Roesch et al., 2002; Kang et al., 2018; Pacheu-Grau et al., 2018). Moreover, multiple sequence alignment of Tim22 across species reveals that most of the amino acid positions mutated in the mutants isolated in our genetic screen are evolutionarily conserved, which further highlights the functional significance of these regions. Most notably, mutations isolated in the TM1, TM3, and TM4 regions of Tim22 reside in conserved glycine residues that are critical for maintaining the architecture of the TIM23 complex (Pareek et al., 2013; Demishtein-Zohary et al., 2017; Matta et al., 2017).

In summary, regardless of the recent identification of new interacting subunits of the TIM22 complex, our understanding of the function and pathophysiology associated with the TIM22 complex remains poorly understood due to the lack of structural characterization of the complex. In this context, our genetic and biochemical analyses provide valuable information towards understanding the functionally distinct roles of different regions of Tim22 involved in structural organization of the TIM22 complex. Therefore, this study provides critical findings that represent an important advancement towards a better understanding of the structural organization of the carrier translocase, in order to gain mechanistic insights into pathological conditions related to the TIM22 pathway.

The details of yeast strains and primers used in this study are listed in Table S1. The WT TIM22 gene from positions −700 to +905 was PCR amplified from W303 yeast genomic DNA using suitable forward and reverse primers. The PCR-amplified products were cloned into NotI/SalI-digested yeast centromeric vector pRS315 harboring LEU2 as an auxotrophic marker (gift from Prof. Elizabeth A. Craig (University of Wisconsin-Madison, WI). The constructed plasmid was labeled as pRS315-Tim22. A library of random mutations in pRS315-Tim22 was generated through error-prone PCR in the coding region of TIM22 using low-fidelity Taq DNA polymerase (Cirino et al., 2003; McCullum et al., 2010). The randomly mutagenized PCR fragments were cloned into the XbaI/SalI-digested pRS315 vector and the generated library was transformed into a tim22Δ strain carrying pRS316-Tim22 using the lithium acetate method (Gietz and Wood, 2002). The strains expressing only Tim22 encoded by pRS315 were selected by patching transformants onto 5-fluoroorotic acid (5-FOA) plates. The strains were tested for growth phenotype by streaking on synthetic complete (SC) medium omitted for leucine (LEU) and supplemented with either 2% dextrose or 2% lactate, followed by incubation at different temperatures. The colonies exhibiting growth sensitivity were isolated, and the positions of the mutated nucleotides were identified by the sequencing of plasmids rescued from mutants, according to a previously published protocol (Robzyk and Kassir, 1992).

The tim18Δ strains were generated by PCR-based homologous recombination to substitute the entire ORF with a KanMx selection cassette (Longtine et al., 1998). The TIM54 gene was C-terminally tagged with six histidines (His) using PCR-mediated gene replacement to exchange the endogenous stop codon with a cassette containing the His-tag and Hph selection marker. For overexpression of Tim18, phosphate carrier (PIC), ADP-ATP carrier (AAC), dicarboxylate carrier (DIC), Tim23, Tom6, and Abf2, the corresponding ORFs were PCR amplified from yeast genomic DNA and cloned into the pRS416_TEF vector (gift from Prof. Elizabeth A. Craig (University of Wisconsin-Madison, WI).).

Yeast cells were grown in SCD–LEU (0.67% yeast nitrogen base without amino acids, 0.069% Leu dropout supplement and 2% glucose), SCD–URA (0.67% yeast nitrogen base without amino acids, 0.072% Ura dropout supplement and 2% glucose), SCLac (0.67% yeast nitrogen base without amino acids, 0.079% complete supplement mixture and 2% lactate, pH 5.6) and YPD (1% yeast extract, 2% peptone and 2% glucose). Strains containing the KanMX6 or Hph genes were selected on YPD supplemented with either 400 µg/ml G418 sulfate or 250 µg/ml hygromycin. To eliminate the URA3-containing plasmid, yeast cells were cultured on SCD containing 1 mg/ml 5-FOA.

BN-PAGE was performed essentially according to a previously published protocol (Wittig et al., 2006). Briefly, mitochondria (1 mg/ml) were solubilized in 100 µl of digitonin buffer (1% digitonin, 50 mM NaCl, 50 mM imidazole, 2 mM 6-aminohexanoic acid and 1 mM EDTA, pH 7.0) for 30 min at 4°C. The insoluble fraction was removed by centrifugation (16,900 g for 20 min at 4°C) and 15 µl of sample buffer (2.5 µl 5% Coomassie Brilliant Blue-G and 12.5 µl 50% glycerol) was added to 100 µl of supernatant. Samples were loaded on a 6–16% gradient native imidazole PAGE gel followed by western blotting and immunodetection using Tim22, Tim54 and Tim18–FLAG specific antibodies.

Mitochondria (1 mg/ml) were solubilized in 250 µl of digitonin buffer (1% digitonin, 25 mM Tris-HCl pH 7.4, 50 mM KCl, 5 mM EDTA, 10% glycerol and 1 mM PMSF). The solubilized fraction was separated by centrifugation and diluted with 750 µl of buffer (25 mM Tris-HCl pH 7.4, 50 mM KCl, 5 mM EDTA, 10% glycerol and 1 mM PMSF) followed by incubation with 20 µl of protein G-conjugated Sepharose beads prebound to anti-FLAG or anti-Tim12 antibodies. Samples were slowly rotated at 4°C for 1 h followed by washing (three times) with 1 ml of buffer (0.05% digitonin, 25 mM Tris-HCl, 50 mM KCl, 5 mM EDTA, 10% glycerol and 1 mM PMSF, pH 7.4). The immunoprecipitated proteins were eluted by boiling in SDS-PAGE sample buffer followed by separation by SDS-PAGE and analysis by immunoblotting using the indicated antibodies. Immunoprecipitated protein levels were quantified by ImageJ software (National Institutes of Health) and plotted as percentages by setting signals from WT mitochondria samples as 100%.

Mitochondria (1.5 mg/ml) were lysed in 250 µl of solubilization buffer (1% digitonin, 25 mM Tris-HCl pH 7.4, 50 mM KCl, 20 mM imidazole, 10% glycerol and 1 mM PMSF) for 30 min at 4°C. The insoluble fraction was separated by centrifugation and the soluble supernatant was incubated with 25 µl of prewashed and equilibrated Ni-NTA agarose beads. Samples were gently rotated at 4°C for 1 h and, after three washing steps, bound materials were eluted with SDS-PAGE sample buffer containing 300 mM imidazole. Samples were analyzed by SDS-PAGE followed by immunoblotting with specific antibodies.

Recombinant Tim23 was purified following previously published protocols (Truscott et al., 2001; Pareek et al., 2013). The TIM23 ORF was PCR amplified and cloned into a pRSFDuet-1 expression vector (gift from Prof. Elizabeth A. Craig (University of Wisconsin-Madison, WI). Escherichia coli BL21-CodonPlus-RIL cells (gift from Prof. Elizabeth A. Craig (University of Wisconsin-Madison, WI) containing the Tim23 construct were grown to mid-log phase at 37°C. 1 mM IPTG was added to the culture for the induction of Tim23 protein expression, followed by incubation for 4 h at 37°C. Cells were harvested by centrifugation and lysed in lysis buffer (20 mM Na-phosphate pH 7.5, 20 mM imidazole, 100 mM NaCl, 10% glycerol and 1 mM PMSF) by adding 0.2 mg/ml lysozyme followed by incubation at 4°C for 45 min with gentle rotation. The lysates were further subjected to 0.2% deoxycholate and 10 μg/ml DNase I treatment for 15 min at 4°C. Pellet and supernatant fractions were separated by centrifugation (31,000 g for 45 min at 4°C). Tim23 protein was recovered from the inclusion bodies and subjected to purification using Ni-NTA chromatography.

The in vitro import assay was performed using isolated mitochondria. WT or mutant mitochondria were incubated at 37°C for 10 min followed by suspension in import buffer (250 mM sucrose, 10 mM MOPS-KOH pH 7.4, 80 mM KCl, 2 mM ATP, 20 mM NADH, 5 mM MgCl2, 5 mM DTT and 1% BSA). Purified Tim23–His was added to the mitochondrial extract for different time intervals and incubated at 25°C. The import reaction was terminated by disrupting membrane potential using 10 µg/ml valinomycin (Sigma) and excess substrate was removed by digesting with 50 µg/ml proteinase K treatment for 20 min at 4°C. The proteinase K was inactivated by treatment with 1 mM PMSF and the samples were analyzed by SDS-PAGE followed by immunoblotting using His-specific antibodies. ImageJ software was used for quantification of imported protein bands and the signal obtained from the longest time point for WT was set to 100%.

The visualization of mitochondrial morphology was performed as described previously (Bankapalli et al., 2015). To observe mitochondrial morphology, yeast cells expressing the mitochondrial targeting sequence (MTS)–mCherry were generated by transforming the pRS416 MTS–mCherry construct (Bankapalli et al., 2015). MTS–mCherry decorates the mitochondria due to the presence of the MTS from subunit 9 of ATP synthase (pSU9), thereby permitting visualization of mitochondrial morphology. The yeast cells were grown to mid-log phase (A₆₀₀ of 0.5–0.6) and cells were harvested by centrifugation followed by two washes with 1× phosphate-buffered saline (PBS). The cells were placed on 2% agarose pads and images were captured with a Delta Vision Elite fluorescence microscope (GE Healthcare) using a 100× objective lens. The excitation wavelength (587 nm) and the emission wavelength (610 nm) were used for mCherry visualization. The images were later deconvolved and analyzed using SoftWoRx 6.1.3 software.

For the estimation of total mitochondrial mass, WT and mutant strains were grown to early log phase, harvested and washed with 1× PBS followed by treatment with 10 µM NAO in the dark for 15 min at 30°C. The cells were washed once with 1× PBS, subjected to fluorescence-activated cell sorting (FACS) analysis using an excitation wavelength of 488 nm and an emission wavelength of 520 nm on a BD FACSVerse™ flow cytometer. For functional mitochondrial mass assessment, cells in early log phase were harvested and incubated with either 8.75 µM TMRE or 1 µM MTDR in the dark for 20 min at room temperature. Subsequently, cells were washed once with 1× PBS and subjected to FACS analysis using an excitation/emission wavelength of 549/575 nm for TMRE and 644/665 nm for MTDR. 10,000 events were recorded for each sample, and the data were analyzed using BD FACSuite software. Relative fluorescence signals were quantified by calculating the fold change in the median fluorescence intensity of mutants in comparison to WT.

The loss of mitochondrial DNA was examined as described in previous protocols (Dunn et al., 2006). WT and tim22 mutants were grown to early log phase, followed by serial dilution and spotting on YPD, or YPD supplemented with 25 μg/ml EtBr. The plates were incubated at the permissive temperature and images acquired after 2 days.

The quantification of protein bands was performed using ImageJ software. Subsequently, the data were processed in Excel (Microsoft) and all statistical analyses were performed using GraphPad Prism 6.0 software. Error bars in FACS analysis indicate ±s.d. and were obtained from a minimum of four individual experiments. Significance testing was performed using a one-way ANOVA and Dunnett's multiple-comparison test was used to compare mutant values against WT values. For Fig. 4D and Fig. S3A,B asterisks are used to represent significance: *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001. P<0.05 was considered significant.

The antibodies against Tim9 (1:3000 dilution), Tim12 (1:3000 dilution), and Tim18 (without presequence, residues 43–192; 1:2500 dilution) proteins were raised in rabbits commercially by Abgenex Pvt. Ltd. Antisera against Tim22 (1:250 dilution) were gifted by Prof. Agnieszka Chacinska (Warsaw University, Poland). Antisera against Tim54 (1:1000 dilution) were gifted by Prof. Toshiya Endo (Kyoto Sangyo University, Japan). Antisera specific for Sdh3 (1:250 dilution) and Yme1 (1:250 dilution) were gifted by Prof. Nikolaus Pfanner (University of Freiburg, Germany). Antisera against Tim10 (1:1000 dilution) were gifted by Prof. Carla M. Koehler (UCLA, USA). Tom70 (1:5000 dilution) and Tim44 (1:5000 dilution) antisera were gifted by Prof. Elizabeth A. Craig (University of Wisconsin–Madison, USA). Antisera against PGK1 (1:7000 dilution) were gifted by Prof. P. N. Rangarajan (Indian Institute of Science, Bangalore, India). Antibodies for Tim23 detection were raised in rabbits as reported previously (Pareek et al., 2013; Matta et al., 2017). The following antibodies were obtained from Sigma-Aldrich: FLAG tag antibody (1:1000 dilution) raised in mouse (F1804) and 6× His-tag antibody (1:5000 dilution) raised in mouse (SAB2702219).

Yeast extract, peptone, dextrose, and agar were purchased from BD Difco. Yeast nitrogen base without amino acids, 5-fluoroorotic acid, zymolyase, and PMSF were purchased from US Biologicals. G418, hygromycin, and digitonin were purchased from Calbiochem. NAO, TMRE, and MTDR dyes were obtained from Molecular Probes. Other chemicals or reagents were purchased from Sigma-Aldrich unless specified otherwise.

Mitochondria were isolated according to previously published protocols (Gambill et al., 1993). Tricine polyacrylamide gels were used for the detection of proteins smaller than 15 kDa wherever required. Immunoblots were performed using an enhanced chemiluminescence system from Bio-Rad following the manufacturer's instructions. Tim22 sequences from different species (Saccharomyces cerevisiae, Schizosaccharomyces pombe, Neurospora crassa, Caenorhabditis elegans, Xenopus laevis, Mus musculus, Homo sapiens and Arabidopsis thaliana) were retrieved from the NCBI database, and multiple sequence alignment was performed using Clustal Omega online tool. The primers used in this study were synthesized by Sigma-Aldrich, and sequencing reactions were performed by AgriGenome Labs Pvt. Ltd.

We thank Prof. Toshiya Endo (Kyoto Sangyo University, Japan) for providing yeast strains (WT Tim22-316 and Tim18_FLAG/WT Tim22-316) and Tim54 antibodies, Prof. Agnieszka Chacinska (Warsaw University, Poland) for Tim22 antibodies, Prof. Nikolaus Pfanner (University of Freiburg, Germany) for Sdh3 and Yme1 antibodies, Prof. Carla M. Koehler (UCLA, CA) for Tim10 antibodies, Prof. Elizabeth A. Craig (University of Wisconsin–Madison, WI) for Tom70 and Tim44 antibodies, and Prof. P. N. Rangarajan (Indian Institute of Science, Bangalore, India) for PGK1 antibodies. We thank Ena Dhaddha for her help with microscopy-related experiments. We acknowledge the Flow Cytometry facility of the Indian Institute of Science (Bangalore, India).

The peer review history is available online at https://jcs.biologists.org/lookup/doi/10.1242/jcs.244632.reviewer-comments.pdf