Protein import into mitochondria

D. Mokranjac, W. Neupert


Mitochondria comprise approx. 1000–3000 different proteins, almost all of which must be imported from the cytosol into the organelle. So far, six complex molecular machines, protein translocases, were identified that mediate this process. The TIM23 complex is a major translocase in the inner mitochondrial membrane. It uses two energy sources, namely membrane potential and ATP, to facilitate preprotein translocation across the inner membrane and insertion into the inner membrane. Recent research has led to the discovery of a number of new constituents of the TIM23 complex and to the unravelling of the mechanisms of preprotein translocation.

  • chaperone
  • import motor
  • mitochondria
  • mitochondrial translocase
  • protein translocation
  • TIM23 complex


Biogenesis of mitochondria is under the control of two genetic systems, their own genome and that of the cell nucleus. However, contribution of the mitochondrial genome to the mitochondrial proteome is very limited and does not exceed 1% of the total number of proteins. The biogenesis of mitochondria, therefore, depends on the import of several hundreds of proteins from the cytosol. The highly complex processes of recognition and sorting of mitochondrial preproteins are mediated by protein translocases present in both the outer and inner mitochondrial membranes (Figure 1) [14]. Some of these translocases were identified only during the last few years and a growing number of new components are being added to the previously known translocases.

Figure 1 An overview of mitochondrial translocases

Most of the mitochondrial proteins are synthesized in the cytosol and are subsequently imported into the organelle. So far, six molecular machines have been identified that mediate a highly complex process of recognition and sorting of mitochondrial proteins. For details, see the text.

Mitochondrial precursor proteins are delivered to the organelle by virtue of specific targeting signals. These signals are highly divergent in nature but are all recognized by the specialized receptor proteins, which are located in the mitochondrial outer membrane and expose their receptor domains to the cytosol. The receptors deliver the precursor proteins to the translocation channel of the TOM (translocase of the outer mitochondrial membrane) complex through which the preproteins cross the outer membrane. The TOM complex itself is sufficient for translocation of a small subset of outer membrane proteins and some IMS (intermembrane space) proteins. For translocation of all other mitochondrial preproteins, the TOM complex co-operates with other mitochondrial translocases. The TOB [topogenesis of outer membrane, or SAM (sorting and assembly machinery)] complex, together with the TOM complex, mediates import, integration and assembly of outer membrane β-barrel proteins [5,6]. Small, metal-co-ordinating proteins of the IMS require the co-operation of the TOM complex with a recently identified import machinery whose first component is Mia40 (Tim40) [79]. Precursors of a subclass of hydrophobic inner membrane proteins, in particular the members of the mitochondrial carrier family, use the TOM complex to cross the outer membrane. They then traverse the IMS in association with soluble small Tim complexes and finally interact with the TIM22 (translocase of the inner mitochondrial membrane) complex to insert into the inner membrane. The last step of the process, membrane integration, is dependent on the membrane potential across the inner membrane.

A major translocase of the inner membrane is the TIM23 complex. It is used by virtually all matrix proteins and the majority of inner membrane proteins. The TIM23 translocase needs two energy supplies, namely the membrane potential and ATP, for mediating translocation across and insertion into the inner membrane. Although the TIM23 complex can recognize various types of targeting signals, most of its known substrates contain N-terminal, positively charged presequences. Once in the matrix, these presequences are usually proteolytically removed by the matrix-processing peptidase. Some inner membrane proteins follow a conservative import pathway. They are first fully imported into the matrix via TOM and TIM23 complexes and are then exported from the matrix into the inner membrane in a process that involves the OXA (oxidase assembly) complex. Research in the last couple of years has identified a number of new components of the TIM23 complex. Thereby, our understanding of the function of this molecular machine has been greatly improved. Our current knowledge of the structure and function of the TIM23 translocase will be summarized in this review.

TIM23 translocase: an overview

The TIM23 complex can be structurally and functionally subdivided into the membrane-integrated translocation channel and the import motor, which is located at the matrix face of the channel (Figure 2). Although the membrane-integrated part of the translocase is sufficient for insertion of some preproteins into the inner membrane, complete translocation into the matrix strictly requires the motor section of the complex. Known components of the membrane-integrated part of the complex are Tim17, Tim21, Tim23 and Tim50. The motor comprises at least five proteins, Tim14, Tim16, Tim44, Mge1 and mtHsp70 (mitochondrial heat-shock protein 70). The membrane-integrated part of the TIM23 complex contains receptors that recognize the precursor proteins as they appear at the outlet of the TOM complex and which deliver them to the translocation channel. The initial translocation of the presequence across the inner membrane is a strictly membrane potential-dependent step. As the presequence appears at the matrix face of the translocation channel, it is caught by the motor part of the complex. Even though the exact mechanism of the import motor is still under debate, it is clear that several rounds of ATP-dependent binding to and release from the mtHsp70 lead to complete translocation of polypeptides into the matrix. All components of the TIM23 translocase are highly conserved throughout the eukaryotic kingdom. Except Tim21, they are all essential for cell viability.

Figure 2 Mitochondrial TIM23 preprotein translocase

The TIM23 complex is a major translocase of the inner membrane. It needs two energy sources, the membrane potential and ATP, to mediate translocation across and insertion into the inner membrane. The TIM23 complex can be structurally and functionally subdivided into the membrane-embedded translocation unit (light grey) and the import motor (dark grey) located at the matrix face of the channel. OM, outer mitochondrial membrane; IM, inner mitochondrial membrane.

Transfer of preproteins from TOM to TIM23

After crossing the outer membrane through the TOM complex, presequence-containing preproteins are directed to the TIM23 complex. The TOM complex appears to play an active role in this process as a single-point mutation in Tom40, the central component of the TOM complex, specifically affects transfer of precursors to the TIM23 complex, leaving other functions of the TOM complex intact [10]. Tim50 is the first component of the TIM23 complex that interacts with precursors after they have been partially translocated across the outer membrane [1113]. It is anchored into the mitochondrial inner membrane with a single transmembrane domain exposing its large C-terminal domain into the IMS. Besides binding to the preproteins, this C-terminal domain also interacts with the N-terminal domain (amino acid residues 50–100) of Tim23. It is still not clear whether the receptor function of Tim50 is independent of its binding to Tim23.

TOM and TIM23 complexes do not appear to form a stable supercomplex in the absence of the precursor. Several factors, however, appear to bring the two complexes together. First, Tim23 was shown to have a rather unusual topology. While its C-terminal domain spans the inner membrane with four transmembrane helices, its first 50 amino acid residues cross the outer membrane [14]. This two-membrane spanning topology brings two mitochondrial membranes in close contact, thereby enhancing the efficiency of protein translocation. The domain of Tim23 associated with the outer membrane does not appear to interact stably with the TOM complex. A direct interaction between TOM and TIM23 complexes was shown only recently [15,16]. The recombinantly expressed C-terminal, IMS-exposed domain of Tim21 specifically binds to the TOM complex or more precisely to the C-terminal, IMS-exposed domain of Tom22. This interaction was, however, only shown in vitro and its significance awaits further in vivo and in organello studies. This is particularly so, as the deletion of either Tim21 or the C-terminal domain of Tom22 does not lead to a significant, if any, growth or import defect. On the other hand, depletion of, or point mutations in, Tim50 prevent stable accumu-lation of presequence-containing precursors in the TOM complex [15]. Based on these findings and the fact that Tim50 interacts with precursors as soon as they appear at the trans side of the TOM complex, it is reasonable to assume that Tim50 is in close proximity to the TOM complex, even though no direct interaction has been demonstrated so far.

Translocation channel of the TIM23 complex

The molecular nature of the translocation channel of the TIM23 complex is still an enigma. Two components are discussed as possible constituents, Tim17 and Tim23. Both proteins are anchored in the inner mitochondrial membrane with four predicted transmembrane helices. As mentioned above, Tim23 has an additional N-terminal domain that is not present in Tim17. The hydrophobic domains of Tim17 and Tim23 are homologous but they cannot substitute for each other. Studies with recombinantly expressed Tim23, after refolding and reconstitution into black lipid bilayers, showed that it can form cation-selective and voltage-activated channels [17]. These channels specifically respond to presequence-like peptides. The N-terminal domain of Tim23 (amino acid residues 1–100) is responsible for presequence sensitivity of the channel, while the C-terminal domain forms the actual channel. These observations confirm a previous study with intact mitochondria that showed that the N-terminal domain of Tim23 has a receptor function and also responds to the membrane potential [18]. The exact role of Tim17 in the formation of the translocation channel remains unclear. Depletion of, or mutations in, Tim17 completely block translocation of the presequence across the inner membrane [19,20]. This suggests that Tim17 is absolutely necessary for the proper functioning of the translocation channel, probably as an integral component. Experiments with reconstituted systems using natively purified proteins seem to be required in future studies.

The import motor of the TIM23 translocase

The motor part of the TIM23 translocase mediates the ATP-dependent completion of translocation into the matrix. Some preproteins, those that have the transmembrane-spanning hydrophobic segment directly after the presequence, can be laterally sorted into the inner membrane without the help of the import motor. In these cases, insertion of the transmembrane part seems to be the driving force. All other TIM23 substrates depend on the import motor for their translocation into mitochondria. Tim44 is the main organizer of the import motor. It recruits mtHsp70, a major chaperone of the mitochondrial matrix, to the translocation sites, along with its nucleotide-exchange factor, Mge1. On the other hand, Tim44 is also responsible for the recruitment of the Tim14–Tim16 subcomplex to the translocation channel. Tim44 is a peripheral inner membrane protein that can be released from the membrane by high-salt treatment [21]. The exact binding site for Tim44 on the translocation channel is not known. It appears that Tim44 binds stably to Tim17 and Tim23 only when these proteins are assembled with each other. Depletion of either of these two proteins precludes stable binding of Tim44 to the remaining one [22]. Purified Tim44 also binds to pure lipid vesicles containing cardiolipin, suggesting that the lipid molecules have a role in recruitment of Tim44 to the inner mitochondrial membrane [23]. Tim44 forms a subcomplex with mtHsp70 in an ATP-dependent manner. mtHsp70 is a typical member of the Hsp70 chaperone family and a close relative of the bacterial DnaK. It consists of an N-terminal ATP-binding (ATPase) domain followed by the C-terminal peptide-binding domain. Hsp70 chaperones use ATP hydrolysis-dependent conformational changes to bind and release unfolded polypeptides [24,25]. When ATP is bound to the ATPase domain, the peptide-binding pocket of Hsp70 is open and polypeptides can be not only easily bound but also released. Upon hydrolysis to ADP, the peptide-binding pocket closes so that bound polypeptides are bound tightly and cannot be easily released. Upon exchange of ADP by a new molecule of ATP, the peptide-binding pocket opens again and the bound polypeptide is released. Hsp70 chaperones depend on various co-chaperones for progression through their ATPase cycles. The rate-limiting step of the cycle, ATP hydrolysis, is accelerated by J-proteins, and the last step, nucleotide exchange, is helped by nucleotide-exchange factors. In mitochondria, the nucleotide-exchange factor of mtHsp70 is Mge1. Tim14 is the J-protein co-chaperone of mtHsp70 in the import motor [22,26,27]. It is anchored in the inner membrane by a single transmembrane domain exposing its C-terminal domain into the matrix. The J-domain of Tim14 is part of this conserved C-terminal domain. In yeast, Tim14 has a short N-terminal segment exposed into the IMS, which was recently proposed to bind to Tim17 [15]. A recombinantly expressed N-terminal segment of Tim14 was observed to be capable of binding to Tim17 and a mutant form of Tim17 (tim17-5) was found to be defective in recruitment of the import motor to the membrane-integrated part of the translocase. Together, these results lead to the conclusion that the N-terminal domain of Tim14 is the main binding site of the motor to the channel. However, this domain of Tim14 is present only in yeast, speaking against an evolutionarily conserved function. Furthermore, the tim17-5 mutation leads not only to lack of Tim14–Tim16 recruitment but also to a loss of binding of Tim44 to the Tim17–Tim23 subcomplex. This nicely confirms a previous finding that, in the absence of Tim44, neither Tim14 nor Tim16 is recruited to the translocation channel [28].

The mitochondrial import motor is unique among the known Hsp70 chaperone systems in so far as the chaperone and its J-protein partner are continuously held at the place of the action, the translocation sites. This was thought to create the danger of an unwanted stimulation of the mtHsp70's ATPase activity, idling of the motor and waste of ATP. Therefore it was speculated that Tim14 is prevented from stimulating the ATPase activity of mtHsp70 either by Tim44 or by another unknown protein in the inactive translocase, i.e. in the absence of the translocating chain [22]. Indeed, such a protein was identified and named Tim16 [28,29]. Tim16 forms a stable subcomplex with Tim14 and is predicted to have the same fold as Tim14. Both proteins are predicted to have the characteristic fold of the J-domain, namely three helices with the absolutely conserved HPD (His-Pro-Asp) motif present in the loop between helices two and three. This HPD motif is, however, missing in Tim16, suggesting that Tim16 is not a functional J-protein. Tim16 was therefore described as a J-like protein. In vitro ATPase assays using purified proteins confirmed that Tim14 specifically stimulates mtHsp70 and that the HPD motif is absolutely necessary for this stimulatory activity. On the other hand, Tim16 lacks such an activity. Even introduction of an HPD motif into Tim16 does not make Tim16 a functional J-protein, suggesting that additional motifs are required for this function [30]. Interestingly, addition of Tim16 prevents Tim14 from stimulating the ATPase activity of mtHsp70. The inhibitory effect of Tim16 is specific for Tim14 since addition of Tim16 does not affect the stimulatory activity of Mdj1, a J-protein of the mitochondrial matrix, which acts as a co-chaperone for mtHsp70 in protein folding. This suggests that the inhibitory activity of Tim16 is not due to its binding to the J-binding site on mtHsp70 but rather by its binding to Tim14 and thus preventing the activity of the latter. The exact mechanism of this inhibition will be revealed probably only by careful biochemical studies using purified components and the high-resolution structure of this complex.

The actual mechanism of the import motor is a matter of continuing debate [31,32]. Two models were put forward. According to the Brownian ratchet or trapping model, Tim44 binds to the peptide-binding domain of mtHsp70 and thereby recruits it to the outlet of the translocation pore. This allows mtHsp70 to bind to the incoming polypeptide chain as soon as it emerges from the channel. In this way, backsliding of even short segments is prevented. mtHsp70 represents a trap that blocks the retrograde movement, and spontaneous Brownian forward movement can be transduced into the vectorial transport by repeated cycles of mtHsp70 binding and release. According to the power stroke or pulling model, mtHsp70 is bound via its ATPase domain to Tim44 and actively pulls on the translocating chain. The force needed to pull in the polypeptide chain is generated by ATP hydrolysis-driven conformational changes of mtHsp70. A considerable amount of experimental evidence has been presented supporting a Brownian ratchet mechanism [33,34]. The kinetics of import correlates with the rate of spontaneous unfolding of a protein domain to be imported, but not with the force required to unfold a protein or the thermodynamic stability of a protein. Determination of the actual binding site for Tim44 on mtHsp70 might help to distinguish between the two models. Establishment of the exact sequence of events during one cycle reconstituted with purified components will bring new evidence for one or the other model. Such studies have been initiated [35], but it is of utmost importance that all the components are known as even if one component is missing, it may lead to major misconceptions. In the course of the enterprise of determining the structure and function of the TIM23 translocase, it has become apparent that this molecular machine holds for more mysteries than originally anticipated.


  • Mechanistic and Functional Studies of Proteins: A Focus Topic at BioScience2005, held at SECC Glasgow, U.K., 17–21 July 2005. Edited by S. Crosthwaite (Manchester, U.K.), M. Ginger (Oxford, U.K.), K. Gull (Oxford, U.K.), A. Lee (Southampton, U.K.), H. McWatters (Oxford, U.K.), J. Mottram (Glasgow, U.K.), P. Rich (University College London, U.K.), C. Robinson (Warwick, U.K.) and H. van Veen (Cambridge, U.K.).

Abbreviations: IMS, intermembrane space; mtHsp70, mitochondrial heat-shock protein 70; TIM, translocase of the inner mitochondrial membrane; TOM, translocase of the outer mitochondrial membrane


View Abstract