Translocation of proteins across a biological membrane represents a crucial step in the biosynthesis of many proteins. Most proteins that are exported are initially synthesized on cytosolic ribosomes as precursors with an N-terminal signal peptide that directs it to the translocation site. The left panel of figure 1 (adapted from 1) shows a generalized model for protein translocation, revealing two main modes of protein export, namely: post- and cotranslational translocation. Both converge at the membrane-embedded translocation site that is made up of receptors as well as a protein-conducting channel (1). In posttranslational translocation, the nascent polypeptide chain is bound by cytosolic chaperones and following translation, the newly synthesized polypeptide chain is delivered at the translocation site. In cotranslational translocation, the polypeptide chain is bound by targeting factors as soon as it emerges from the ribosomal exit tunnel that guide it also to the translocation site. Following capture by receptors, substrate proteins are transferred to the protein-conducting channel (translocon) that comprises a small pore that spans the membrane and facilitates transport of substrate proteins across the membrane as well as lateral release of sufficiently hydrophobic substrate proteins into the lipid bilayer (1).
Examples of protein-conducting channels include the Sec61 complex of the endoplasmic membrane and the mitochondrial TOM and TIM23 complexes. These are able to translocate a wide variety of substrate proteins typically in an unfolded, extended state (2,3). Although most proteins are translocated linearly, a subset is transported in a fully folded form. This typically occurs posttranslationally and concerns substrate proteins that maturate (e.g. fold and assemble cofactors) before translocation. The bacterial Tat system represents (figure 1 right panel) the best characterized system for the export of folded proteins (4). However, the molecular mechanism employed by the Tat machinery is not fully understood. The Tat system is universally present in bacteria and has also been identified in archaea as well as chloroplasts, while most mitochondria have lost it during the evolution from their bacterial ancestor. Nevertheless, mitochondria have retained a substrate protein, RISP (Rieske iron-sulfur protein), that is in bacteria and chloroplasts exported in a folded form Tat dependently. Mitochondrial RISP is a crucial electron-relaying subunit of respiratory complex III (cytochrome c reductase or bc1 complex) that is anchored to the inner membrane through its transmembrane domain (TMD) with its C-terminal domain localized to the intermembrane space and its N-terminus in the matrix. The C-terminal domain is, moreover, equipped with a 2Fe-2S redox cofactor. RISP is synthesized as a precursor in the cytosol and imported into mitochondrial matrix through the TOM and TIM23 complexes. Here, the 2Fe-2S cluster is added to the C-terminal domain and the complete protein is subsequently translocated across the inner membrane in a folded state. It has been shown that the mitochondrial AAA ATPase Bsc1 is responsible for the export of folded RISP across the inner membrane to the intermembrane space, indicating that Bsc1 has replaced the Tat system as an export apparatus for folded RISP in mitochondria (5,6). Detailed structures of Bsc1 were presented recently, providing profound insight into its molecular mechanism (7,8). Here, I will discuss these structures as well as the current functional understanding.
Biosynthesis of Rieske iron-sulfur protein, RISP
Rieske iron-sulfur proteins were first described by John S Rieske and co-workers in 1964 (9) and it is now well established that they are present in the genomes of most organisms. Rieske proteins comprise a large protein family made up of members that possess a 2Fe-2S redox cofactor and mediate electron transfer. The Rieske protein family consists of two main subgroups, namely: so-called high potential and low potential Rieske proteins (10). Members of the latter group occur as individual proteins such as ferredoxins or as domains of hydroxylases or oxygenases. Whereas proteins of the first group are typically membrane-bound and comprise subunits of respiratory or photosynthetic complexes (10). For example, the yeast mitochondrial Rieske protein, RISP, is a crucial electron-transducing subunit of respiratory complex III (cytochrome c reductase or bc1 complex) with a mass of about 23 kDa. Its structure is shown in the left panel of figure 2 with coloring according to secondary structure (11), revealing that it contains a single α-helical TMD that anchors the protein into the inner membrane. The TMD is connected to the globular C-terminal domain, which is exposed to the intermembrane space and contains the 2Fe-2S redox cluster (in red spheres). This domain is made up of three β-sheets and a single α-helix. The N-terminus of RISP resides into the matrix. The mitochondrial import pathway and subsequent assembly of RISP into the bc1 complex were recently assessed (5). This is presented in figure 2 (adapted from 6), showing that RISP is nuclear-encoded and translated by cytosolic ribosomes initially as a precursor with a matrix targeting signal that is imported into the mitochondrial matrix via the TOM and TIM23 complexes. Here, the targeting signal is proteolytically removed in two consecutive steps by mitochondrial processing peptidase (MPP) and intermediary peptidase (MIP), yielding the mature-sized protein. Moreover, the 2Fe-2S redox cluster is loaded onto the apoprotein in the matrix after which the C-terminal domain is folded (5).
Subsequently, the C-terminal domain of RISP is translocated across the inner membrane into the intermembrane space, while its TMD is inserted into the inner membrane proper. These two steps require the mitochondrial AAA ATPase Bsc1, which is embedded into the inner membrane with its catalytic AAA domain exposed to the matrix. Bsc1 belongs to the AAA+ protein family with members ubiquitously present in all organisms that typically play a role in protein degradation and unfolding. In the biosynthesis of RISP, however, Bsc1 functions as a translocase by facilitating the export of the Fe-S domain. Following export of the C-terminal domain, RISP is assembled into the bc1 complex and to this end, RISP is released from Bsc1 and inserted into the bc1 precomplex in an ATP-dependent fashion (5). The precomplex is made up of cytochrome c1, cytochrome b1 and the core proteins 1 and 2, while RISP and Qcr10 are absent. After incorporation of RISP, the bc1 complex dimerizes and assembles into supercomplexes with complex IV (cytochrome c oxidase). In bacteria and chloroplasts, related Rieske proteins are exported in a folded conformation through the Tat pathway. Intriguingly, Bsc1 recognizes fully folded RISP and it is therefore conceivable that Bsc1 has replaced the Tat system as an export apparatus for folded proteins in mitochondria.
Biochemical and structural features of Bcs1
Yeast Bcs1 is a protein of about 51 kDa with ATPase activity and resides in the mitochondrial inner membrane. Homologous are present in most eukaryotic proteomes and, moreover, Bcs1 belongs to the AAA+ family (ATPases associated with diverse cellular functions). AAA proteins are universally conserved in all organisms and are typically involved in protein degradation and unfolding. Structurally, most AAA proteins form a ring-shaped hexameric complex. Each subunit contains one or two ATP-binding domains and together they form a pore-like channel through which substrate proteins are transported at the expense of ATP hydrolysis. Phylogenetically, AAA proteins can be divided into six major clades, namely: (i) proteasome subunits, (ii) metalloproteases, (iii) members with one AAA domain (D1), (iv) members with two AAA domains (D2), (v) meiotic AAA proteins and (vi) Bcs1 variants. Thus, Bcs1 and its homologous comprise a separate clade of the AAA+ family (10). As described above, Bcs1 facilitates the final steps in the assembly of the bc1 complex. The first indications that Bcs1 serves as an assembly factor came from yeast cells lacking the bcs1 gene. These mutants lacked the bc1 complex and contained reduced levels of RISP and therefore the protein was named ubiquinol-cytochrome c reductase (bc1) synthesis, or Bcs1 (12). Additionally, more than 25 pathological mutations in BCS1L, the human Bcs1 gene, are known that are linked to various clinical manifestations such as neurological and metabolic disorders with different tissue involvement and disease progression (13). These mutations typically result in bc1 complex deficiency, increase in the production of reactive oxygen species and iron overload.
As shown in figure 3 (adopted from 6), yeast Bcs1 is made up of different structural regions that include a short N-terminal sequence that is localized to the intermembrane space and contains an internal targeting signal required for mitochondrial import, a single TMD inserted into the inner membrane and AAA domain that is connected to the membrane-embedded part through a short N-linker. In contrast to other AAA proteins, Bsc1 contains only one AAA domain that is situated in the matrix. This domain includes the characteristic Walker A and B motifs of the nucleotide-binding site but lacks the so-called pore loop region. This is a structural motif commonly found in AAA proteins that probably plays a key role in the binding, unfolding and translocation of substrate proteins. Due to the absence of the pore loop region it is not immediately clear how Bcs1 interacts with substrate proteins. Moreover, Bcs1 contains two arginine residues (Arg376 and Arg379) that are important for structural integrity as well as ATP hydrolysis. Blue native gel electrophoresis of detergent-solubilized mitochondrial inner membranes from yeast showed that Bcs1 is present in an oligomeric complex with a mass of about 500 kDa (6).
Important insight into the structure of this complex was recently provided by detailed cryo-EM reconstructions with or without ADP, using affinity purified yeast Bcs1 complex from a crude mitochondrial fraction (7). The structure of the ADP-bound Bcs1 complex at 3.4 Å resolution is presented in figure 4 in surface (upper panel) or ribbon presentation (bottom panel). This reveals that Bcs1 forms a heptameric oligomeric assembly with a central pore unlike other AAA proteins that typically assemble into a hexameric complex. Overall, the Bcs1 heptamer resembles an upside-down mushroom with an iris-shaped AAA ring, a central stalk made up of the middle domains and a TMD region. The middle domains connect the AAA ring and TMD region. The upper right and bottom left panels show a view from the matrix side towards the AAA ring with its central pore and bound ADP in red spheres. Clearly, ADP is present in the AAA domains of all seven subunits. The structure of an individual subunit is presented in the bottom right panel with the single TMD in yellow, the middle domain in magenta and the subdomains of the AAA region in green and orange, respectively, and bound ADP as red spheres. The TMDs of all subunits contact each other in the membrane and are bundled into a basket-like arrangement. Structurally, the middle domain comprises a β-sheet with four antiparallel strands that are stabilized by two helices, joining the middle domain to the AAA domain. Moreover, this domain contains a loop, the so-called seal loop, that protrudes into the central pore of the heptamer. The AAA region is made up of two subdomains, the large and small domain, that form a typical nucleotide binding pocket. Moreover, the AAA domain includes two loops that resemble the commonly observed pore loops of other AAA proteins, which grab and pull substrate polypeptides. However, the loops of Bcs1 don’t protrude into the central pore as in other AAA protein but, rather, point away from the central pore. Moreover, these loops also lack the conserved aromatic residues that are important for substrate interaction.
Figure 5 shows that the heptameric assembly of Bcs1 allows the formation of two cavities within the matrix and inner membrane, respectively. All seven AAA domains as well as the middle domains form a large chamber that is accessible from the matrix. The TMDs, moreover, form a second chamber in the membrane. Both chambers are large enough to house the folded Fe-S domain of RISP. The matrix and inner membrane chamber are separated by the middle domains through a small seal-forming pore (right panel) at the center. This pores probably seals the central channel of the heptamer between the chambers during protein translocation and comprises small loops made up by hydrophobic interactions by Met160, Ile161 and Ile163 (shown as red spheres in each subunit). These interactions result in a highly hydrophobic character of the seal pore at the matrix side, while at the membrane side, a charged lysine of each subunit creates a hydrophilic ring at the seal pore. The membrane chamber represents an aqueous environment that almost spans the entire inner membrane. It is as wide as the matrix chamber at the matrix side but narrows towards the top at the inter membrane space. Here, the TMDs contact each other and form an iris-like helical bundle with a gap between each other that allows direct access from the membrane chamber to the lipid phase of the inner membrane (right panel).
The structure of the Bcs1 complex was also assessed in the absence of nucleotides. In fact, two different conformations were obtained under these conditions, termed Apo state 1 and Apo state 2 at 4.4 and 4.6 Å resolution, respectively. These are shown in figure 6 (upper panel), revealing two heptameric structures with empty nucleotide-binding sites. Conformationally, both structures differ from the ADP bound one but also between each other. The bottom panel of figure 6 provides a view from the matrix towards the opening of the AAA ring showing that in both apo structures the opening of the matrix chamber is smaller when compared to the ADP state due to an inward movement of the AAA domain in the absence of nucleotides.
The arginine finger comprises a conserved loop in the AAA domain of AAA proteins, which is important for ATP hydrolysis. This loop contains two specific arginine residues of which one projects into the ATP-binding site of an adjacent subunit. Likewise, in the ADP bound state two arginine residues of the AAA domain of Bsc1, Arg376 and Arg379, rotated towards the nucleotide-binding site of the neighboring subunit when compared to apo state 1 (left panel of figure 7). In the ADP and apo 1 states, the middle domain adopts a similar conformation, while the main structural difference between the apo states occurs in this region as shown in figure 7 (right panel). In apo state 2, the middle domain shifted away from the membrane plane, thereby separating the pore loops and creating an opening between both chambers. This opening is large enough for the passage of RISP from the matrix to the membrane chamber. Conceivably, opening of the seal pore could be triggered by the ATPase activity of Bcs1.
Molecular mechanism of Bcs1-facilitated translocation
The hexameric structure of Bcs1 allows the formation of two aqueous chambers in the matrix and inner membrane that are large enough to accommodate fully folded RISP. Both chambers are separated by a seal-forming middle domain. An outward movement of these domains results in opening of the seal, thereby allowing RISP to move from the matrix to the inner membrane chamber, while, concomitantly, maintaining the permeability barrier of the inner membrane. The targeting of RISP from the matrix to the Bsc1 complex is not fully understood. But once this has been accomplished, translocation probably occurs in three steps as shown in figure 8 (adapted from 7). Specifically: (1) loading of RISP into the matrix chamber of Bcs1, (2) opening of the seal and translocation of RISP into the inner membrane chamber through the widened seal pore and (3) release of RISP to the intermembrane space and release of its TMD into the lipid bilayer of the membrane. Subsequently, RISP is transferred and incorporated into the bc1 complex. Though speculative but despite the absence of classical pore loops that bind and translocate a substrate polypeptide, the pore loops of Bcs1 point away from central channel of the heptameric assembly but could still transiently interact with RISP. Subsequent ATP hydrolysis could power substrate translocation as in AAA proteins.
It is commonly accepted that mitochondria originate from a free-living bacterium that was taken up by a primordial eukaryotic cell about 2 billion years ago. During their subsequent evolution, most mitochondrial genes were transferred to the nucleus, which, in turn required the development of specialized protein translocation machineries to facilitate the import of nuclear-encoded mitochondrial proteins. This typically requires that the substrate protein is unfolded because most protein translocases, including the mitochondrial import machineries and the Sec61 complex of the endoplasmic membrane, are only able to handle linear polypeptides. However, a subset of proteins is translocated in a fully folded form and this typically concerns substrate proteins that maturate (e.g. fold and assemble cofactors) before translocation. The bacterial Tat system represents the best characterized system for the export of folded proteins (4). The Tat system is universally present in bacteria and has also been identified in archaea as well as chloroplasts, while most mitochondria have lost it during the evolution from their bacterial ancestor. Nevertheless, mitochondria have retained a substrate protein, RISP (Rieske iron-sulfur protein), that is in bacteria and chloroplasts exported in a folded form Tat dependently. Rieske proteins comprise a large protein family made up of members that possess a 2Fe-2S redox cofactor and mediate electron transfer. RISP is a crucial electron-transducing subunit of respiratory complex III (cytochrome c reductase or bc1 complex) that is anchored to the inner membrane through its single TMD with its C-terminal domain localized to the intermembrane space and its N-terminus in the matrix. Moreover, the C-terminal domain is equipped with a 2Fe-2S redox cofactor. Following import into the matrix through the TOM and TIM23 complex, RISP is translocated from the matrix across the inner membrane with its C-terminal domain exposed to the intermembrane space and its TMD inserted into the lipid bilayer of the inner membrane. These steps require the mitochondrial AAA ATPase Bsc1, which is embedded into the inner membrane. Bsc1 belongs to the AAA+ protein family with members ubiquitously present in all organisms that typically play a role in protein degradation and unfolding. In the biosynthesis of RISP, however, Bsc1 functions as an assembly factor. Intriguingly, Bsc1 recognizes fully folded RISP and it is therefore conceivable that Bsc1 has replaced the Tat system as an export apparatus for folded proteins in mitochondria. The detailed structures of Bsc1 that were presented recently provide profound insight into its molecular mechanism that is perhaps similar to the way substrate proteins are handled by the Tat system.
1. Wickner W, Schekman R. 2005. Protein translocation across biological membranes. Science. 310: 1452-1456.
2. Rapoport TA, Li L, Park E. 2017. Structural and Mechanistic Insights into Protein Translocation. Annu Rev Cell Dev Biol. 33: 369-390.
3. Wiedemann N, Pfanner N. 2017. Mitochondrial Machineries for Protein Import and Assembly. Annu Rev Biochem. 86: 685-714.
4. Palmer T, Stansfeld PJ. 2020. Targeting of proteins to the twin-arginine translocation pathway. Mol Microbiol. Doi: 10.1111/mmi.14461.
5. Wagener N, Ackermann M, Funes S, Neupert W. 2011. A pathway of protein translocation in mitochondria mediated by the AAA-ATPase Bcs1. Mol Cell. 44: 191-202.
6. Wagener N, Neupert W. 2012. Bcs1, a AAA protein of the mitochondria with a role in the biogenesis of the respiratory chain. J Struct Biol. 179: 121-125
7. Kater L, Wagener N, Berninghausen O, et al. 2020. Structure of the Bcs1 AAA-ATPase suggests an airlock-like translocation mechanism for folded proteins. Nat Struct Mol Biol. 2: 142-149.
8. Tang WK, Borgnia MJ, Hsu AL, et al. 2020. Structures of AAA protein translocase Bcs1 suggest translocation mechanism of a folded protein. Nat Struct Mol Biol. 2: 202-209.
9. Rieske JS, Hansen RE, Zaugg WS. 1964. Studies on the electron transfer system. J Biol Chem. 239: 3017-322.
10. Schneider D, Schmidt CL.2005. Multiple Rieske proteins in prokaryotes: where and why? Biochim Biophys Acta. 1710: 1-12.
11. Lange C, Nett JH, Trumpower BL, Hunte C. 2001. Specific roles of protein-phospholipid interactions in the yeast cytochrome bc1 complex structure. EMBO J. 20: 6591-600.
12. Conte L, Trumpower BL, Zara V. 2011. Bcs1p can rescue a large and productive cytochrome bc (1) complex assembly intermediate in the inner membrane of yeast mitochondria. Biochim Biophys Acta. 1813: 91-101.
13. Fernández-Vizarra E, Zeviani M. 2015. Nuclear gene mutations as the cause of mitochondrial complex III deficiency. 6:134.