Report

Molecular architecture of the active mitochondrial protein gate

See allHide authors and affiliations

Science  25 Sep 2015:
Vol. 349, Issue 6255, pp. 1544-1548
DOI: 10.1126/science.aac6428

Dissecting the mitochondrial entry portal

Mitochondria, the powerhouses of the cell, are mainly composed of proteins made in the cytosol. These newly synthesized proteins need to be imported across the organelle's membrane through dedicated protein import machinery. Shiota et al. have worked out the architecture and mechanism of the mitochondrial protein import channel.

Science, this issue p. 1544

Abstract

Mitochondria fulfill central functions in cellular energetics, metabolism, and signaling. The outer membrane translocator complex (the TOM complex) imports most mitochondrial proteins, but its architecture is unknown. Using a cross-linking approach, we mapped the active translocator down to single amino acid residues, revealing different transport paths for preproteins through the Tom40 channel. An N-terminal segment of Tom40 passes from the cytosol through the channel to recruit chaperones from the intermembrane space that guide the transfer of hydrophobic preproteins. The translocator contains three Tom40 β-barrel channels sandwiched between a central α-helical Tom22 receptor cluster and external regulatory Tom proteins. The preprotein-translocating trimeric complex exchanges with a dimeric isoform to assemble new TOM complexes. Dynamic coupling of α-helical receptors, β-barrel channels, and chaperones generates a versatile machinery that transports about 1000 different proteins.

Mitochondria are pivotal for cellular adenosine triphosphate (ATP) production, numerous metabolic pathways and regulatory processes, and programmed cell death. Most mitochondrial proteins are synthesized as preproteins in the cytosol and are imported into mitochondria. Preproteins either contain N-terminal targeting sequences (presequences) or internal targeting information in the mature part (13). The protein translocator of the outer membrane (the TOM complex) functions as the main entry gate of mitochondria (13). Over 90% of all mitochondrial proteins are imported by the TOM complex, followed by transfer to distinct translocators for individual classes of preproteins. Whereas all structurally known membrane protein complexes consist of either α-helical or β-barrel proteins, the TOM complex is composed of both α-helical and β-barrel integral membrane proteins. The complex consists of the channel-forming β-barrel protein Tom40 and six other subunits, each containing single α-helical transmembrane (TM) segments: the receptor proteins Tom20, Tom22, and Tom70 and the regulatory small Tom proteins (13). The molecular architecture of the complex has not been elucidated. How α-helical and β-barrel membrane proteins can be combined into a functional complex and how diverse classes of preproteins can be transported by the same TM channel is unclear.

To define the architecture of the functional TOM complex, we mapped the interactions of Tom40 with preproteins in transit and α-helical subunits by in vivo and in organello site-specific cross-linking. Photoactivatable p-benzoylphenylalanine (BPA) was introduced at 108 different positions in the 387-residue Tom40 in yeast cells (fig. S1) (46). A structural model based on homology and cysteine scanning shows 19 antiparallel β strands with the first and last β strands annealing in parallel arrangement (79) (fig. S2A). In TM β strands, every second residue faces the pore lumen, and alternate residues face outward. BPA cross-linking to Tom22 revealed outward orientation of side chains (Fig. 1A and fig. S1). Whether preproteins transit through the lumen of the Tom40 β barrel or via the interstitial space between multiple β barrels that make up the TOM complex has been controversial (10, 11). To resolve this, we accumulated the model preproteins presequence of subunit 9-dihydrofolate reductase (pSu9-DHFR) and ADP-ATP carrier (AAC)–DHFR without a presequence, in the TOM complex (10, 12, 13) and irradiated it with ultraviolet (UV) light. Residues cross-linked to pSu9 and AAC were only found at positions facing the pore interior (Fig. 1, B to D; red residues face the pore). Similar results were obtained by sulfhydryl group (SH)–directed chemical cross-linking (fig. S2B). Thus, preproteins in transit are located inside the β-barrel pore of Tom40 and not in the interstitial space between Tom40 molecules.

Fig. 1 The pore interior of the β barrel of Tom40 is the protein import channel for preproteins.

(A) In vivo photo–cross-linking of BPA at the indicated positions of Tom40 with Tom22 analyzed by SDS-PAGE followed by immunoblotting (top) and quantification (bottom). (B and C) [35S]pSu9-DHFR (Su9 presequence fused to dihydrofolate reductase) (B) or [35S]AAC-DHFR (ADP-ATP carrier fused to DHFR) (C) was incubated with mitochondria [pretreated with valinomycin for (B)] with BPA-bearing Tom40 at 4°C (B) or 25°C in the presence of 1 μM methotrexate and 1 mM NADPH (C), for 10 min, and UV-irradiated. Affinity-purified cross-linked products (asterisk) were subjected to SDS-PAGE and radioimaging. Variation in the apparent molecular weight of the cross-linked products may reflect different configurations. β-strand numbers and BPA positions (red and green, side chain facing the pore interior or membrane, respectively) are indicated. The cross-links of the neighboring residues 356 (weak) and 357 (strong) with Tom22 may suggest imperfections in the last β strand 19 around this position. p, precursor form; i, processing intermediate form; m, processed mature form. (D) Quantification of the cross-linked products in (B) and (C). (E) The side chains of the Tom40 residues cross-linked to pSu9-DHFR or AAC-DHFR are shown with color reflecting the amount of cross-linked products detected. (F) The acidic patches (red) and hydrophobic patches (green) are shown in the pore interior of Tom40, in relation to the position of Tom22TM (Fig. 4, inset).

Do presequence-containing and carrier-family preproteins use the same path through the Tom40 channel? The cross-linking results revealed a non-identical pattern for pSu9-DHFR and AAC-DHFR (Fig. 1, D and E, and fig. S2C). Negatively charged residues are aligned in the pore from the cytosolic side to the intermembrane space (IMS) side, forming acidic patches near the cross-linked sites for pSu9-DHFR (red in Fig. 1F and fig. S3), whereas hydrophobic patches (green) are near the cross-linked sites for AAC-DHFR and partly for pSu9-DHFR [presequences form positively charged amphiphilic helices (13)]. Homology models for animal and plant Tom40 indicate similar acidic and hydrophobic patches inside the channel pore (fig. S4). We conclude that positively charged presequences follow an acidic path on the inner wall of the Tom40 pore, whereas carrier proteins interact with mostly hydrophobic residues. Thus, Tom40 can handle and chaperone (14) diverse classes of preproteins by providing distinct translocation paths.

Systematic analysis of the cross-linking partners of BPA-bearing Tom40 revealed that α-helical Tom proteins interact with the outside of the β barrel or loops of Tom40 (Fig. 2A and fig. S1). Unexpectedly, the IMS protein Tim10 was cross-linked to the N-terminal segment of Tom40 (Fig. 2, A and B), suggesting that this segment extends from the cytosolic side through the β-barrel pore of Tom40 to the IMS (8). We generated yeast mutants with N-terminal truncations of Tom40 and found that, whereas a 62-residue deletion (tom40Δ62) inhibited import of both presequence-containing and presequence-less preproteins, a 57-residue deletion selectively inhibited import of presequence-less preproteins (Fig. 2, C and D, and fig. S5). The tom40Δ57 yeast strain became sensitive to overexpression of carrier proteins, not of a presequence-containing preprotein, being unable to cope with the increased load of hydrophobic preproteins (Fig. 2E). Tim10 is a subunit of the hexameric Tim9-Tim10 (small TIM) chaperone in the IMS, functioning to guide presequence-less preproteins through the aqueous IMS (13). We thus conclude that the recruitment of these chaperones to the TOM channel exit promotes an efficient transfer of hydrophobic preproteins.

Fig. 2 The N-terminal segment of Tom40 interacts with elements of the carrier protein import pathway.

(A) Topology model and BPA cross-linking results of Tom40. Circles, pentagons, and squares indicate residues in the loop, α helix, or β strand, respectively. BPA-incorporated residues are labeled in bold. Cross-linked partners are indicated by different colors. (B) UV-dependent in vivo photo–cross-linking of BPA at the indicated position of Tom40 detected by immunoblotting (top) and quantification (bottom). (C and D) In vitro import of [35S]labeled presequence-containing (C) and presequence-less (D) preproteins into the tom40 mutant mitochondria at 25°C. After proteinase K treatment, imported proteins were analyzed by SDS-PAGE and radioimaging. The imported amount at the longest incubation time was set to 100% (control). Values are mean ± SEM (n = 3). (E) Serial dilutions of the Tom40 N-terminal truncation mutant cells with overexpression of the indicated proteins were spotted on SCD (-Ura, -Trp) (glucose), SCGal (-Ura, -Trp) (galactose), and SCGly + 0.05% d-glucose (-Ura, -Trp) (glycerol + galactose) media and grown at 30°C for 3 days (glucose and galactose) or 4 days (glycerol + galactose). WT, corresponding TOM40-(His)10 strain; PiC, phosphate carrier; DiC, dicarboxylate carrier.

To define the subunit organization within the TOM complex, we probed the interactions of the Tom40 molecule with Tom40 itself and with the core receptor Tom22. We asked whether Tom40 molecules are close enough to make direct contact with each other, like bacterial β-barrel proteins (15), by chemical cross-linking with the SH-directed homobifunctional cross-linkers BMB or M2M. We introduced Cys at membrane-facing positions as well as pore-facing positions of a Cys-free Tom40 variant (8). Cross-linked products were evident for the pairs of membrane-facing Cys residues in endogenous Tom40 (Fig. 3A) or imported Tom40 (fig. S6A). The distances between the Sγ atoms of cross-linked Cys are ~6.9 and ~12.0 Å for M2M and BMB cross-linking, respectively (16), indicating that two Tom40 molecules are located within a distance of ~6.9 Å (Fig. 3B).

Fig. 3 The native TOM complex dynamically exchanges with the Tom40 dimeric complex.

(A) Mitochondria with the indicated Tom40 Cys residues were treated with (+) or without (–) the cross-linker (1,4-bismaleimidobutane) [BMB (XL)]. Proteins were analyzed by SDS-PAGE and immunoblotting with antibodies to Tom40. Cross-linked products are indicated by the asterisk and arrowheads. (B) Schematic models for (A). (C) (Top) The Tom40 β barrel showing the residues cross-linked with Tom22. (Botom) Tom22-interacting regions of Tom40 (blue and yellow) with the Tom22-cross-linked residues (pink circles). (D) Tom40-interacting regions of Tom22TM (orange and green) (18). (E) In vivo cross-linking of BPA in Tom40 and Tom22 was detected by immunoblotting. Cross-linked products are indicated. (F) Schematic models for (E). (G) Mitochondria with the Tom40 Cys mutations were treated with (+) or without (–) the cross-linker BMOE [bis(maleimido)ethane]. Proteins were analyzed by SDS-PAGE (lanes 1 to 4), BN-PAGE (lanes 5 and 6), or 2D-PAGE (first- BN-PAGE and second-dimensional SDS-PAGE) and immunoblotting. Cross-linked products are indicated by the asterisk and arrowheads.

Because the TM helix of Tom22 (Tom22TM) interacts with two Tom40 molecules (17), we analyzed the geometrical arrangement of Tom22 and Tom40. We introduced BPA into Tom40 at two of the positions 86, 309, 350, and 357 in the narrow vertical Tom22-interacting regions along the β-barrel axis (Fig. 3C and fig. S6B) simultaneously. UV irradiation generated cross-linked Tom40:[Tom22]2 oligomers for BPA positions 86 and 309 (fig. S6B). Simultaneous introduction of BPA into Tom40 and Tom22 in their interacting regions (Fig. 3, C and D), generated cross-linked Tom40:[Tom22]2 and [Tom40]2:Tom22 oligomers only for a specific combination of introduced BPA (Fig. 3E). These results pose a geometrical constraint that is consistent only with a threefold rotational symmetric arrangement of three molecules each of Tom40 and Tom22 (Fig. 3F), in which the distance between the Tom40 molecules bridged by Tom22 must be larger than 22 Å, which is incompatible with a distance of ~6.9 Å between the Tom40 molecules (fig. S6A).

How can the opposing cross-linking results be explained? We hypothesized that both dimeric and trimeric isoforms of TOM complexes, previously observed by single-particle electron microscopy analyses (1820), exist in organello. Tom40 Cys mutant mitochondria were subjected to cross-linking and analyzed by SDS–polyacrylamide gel electrophoresis (PAGE) and blue native (BN)–PAGE (Fig. 3G). SDS-PAGE demonstrated the cross-linked dimers, and BN-PAGE revealed that the cross-linked Tom40 dimers arose from a ~100-kD (100K) subcomplex, not from the mature (large) TOM complex. The two Tom40 molecules are thus close together, bridged by short chemical cross-linkers only in the 100K complex (Fig. 3B). The 100K complex contains Tom40 and small Tom proteins, but not Tom22; it functions as a late assembly intermediate of newly imported Tom40 on the pathway to the mature TOM complex (17, 21). Thus, the 100K complex containing the dimer exchanges with the mature trimeric TOM complex. A dynamic exchange between dimeric and trimeric forms provides the means for template-driven assembly of new subunits through their exchange for old subunits.

The Tom22TM has been conserved through evolution, with an invariant Pro (Pro112), flanked by basic residues on the cytosolic side and acidic residues on the IMS side (fig. S7). We generated yeast strains with mutant Tom22 and analyzed destabilization of the TOM complex and in vitro import of presequence-containing and presequence-less preproteins by amino acid replacements of those residues (fig. S8). These analyses suggest that the full-sized mature TOM complex tethered by Tom22 (Fig. 4, inset) is required for efficient preprotein import. We also determined the interactions of Tom40 with the receptor Tom20 and the small subunits Tom5, Tom6, and Tom7 (summarized in Fig. 2A and figs. S1 and S9). A complete model of the subunit arrangement in the TOM core complex is shown in Fig. 4 (middle panel).

Fig. 4 Subunit organization of the TOM complex.

Subunit arrangement of the Tom40 β barrel and TM α helices of Tom5, Tom6, Tom7, and Tom22, based on the BPA cross-linking results and the proposed model for the exchange of the Tom40-Tom22 trimeric complex with the Tom40 dimer. The inset shows possible interactions between Tom22TM and the Tom40 β barrel, with key residues suggested from fig. S8. Basic residues, acidic residues, and others are colored in blue, red, and pink, respectively. The Tom22TM α helix, possibly bent at Pro112, tethers two Tom40 molecules through the interactions of its N-terminal and C-terminal parts with conserved residues of adjacent Tom40 molecules.

We conclude that the trimeric mature TOM complex dynamically exchanges with a dimeric Tom22-free form that provides an assembly platform for the integration of new subunits (Fig. 4, right panel). The dynamic α/β organization of the TOM complex favors both assembly of the complex and cooperative preprotein transfer from receptors to the import channel and IMS chaperones, ensuring the efficient translocation of different classes of preproteins into mitochondria.

Supplementary Materials

www.sciencemag.org/content/349/6255/1544/suppl/DC1

Materials and Methods

Figs. S1 to S9

Tables S1 to S5

References (2239)

REFERENCES AND NOTES

  1. Acknowledgments: We thank C. Stubenrauch, M. Belousoff, A. Traven, and the members of the Endo lab for discussions and critical comments on the manuscript. We are grateful to P. G. Schultz for the materials for in vivo cross-linking. This work was supported by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (JSPS) and a CREST Grant from the Japan Science and Technology Agency (JST) (T.E.); the Platform for Drug Discovery, Informatics, and Structural Life Science from the Ministry of Education, Culture, Sports, Science and Technology and Japan Agency for Medical Research and Development (K.T., K.I., and Y.F.); Grants-in-Aid for Scientific Research on Innovative Areas (“Matryoshka-type evolution,” no. 3308) (K.I., K.T., and Y.F.); the Strategic Japanese-Swedish Cooperative Program on “Multidisciplinary BIO” (JST-Verket För Innovationssystem/Swedish Foundation for Strategic Research) (K.I., N.S., Y.F., S.H., A.E., K.T., P.H., and T.E.); the Deutsche Forschungsgemeinschaft (PF 202/8-1), Sonderforschungsbereiche 746 and 1140; and the Excellence Initiative of the German federal and state governments (EXC 294 BIOSS, GSC-4 Spemann Graduate School) (N.W., N.P., and J.Q.). T.S. is a Research Fellow of the JSPS and was supported by the Toyobo Bio Foundation, H.S.S. is an Australian Research Council (ARC) Super Science Fellow (FS110200015), and T.L. is an ARC Australian Laureate Fellow (FL130100038). The data presented in this paper are tabulated in the main paper and the supplementary materials.
View Abstract

Navigate This Article