MTCH2 is a mitochondrial outer membrane protein insertase

1Mitochondria are organelles of endosymbiotic origin that have evolved to play a central role in eukaryotic cell metabolism and signaling. Mitochondrial function and their ability to communicate with the cytosol depend on proteins embedded in the outer mitochondrial membrane. As a result, dysregulation of outer membrane protein function is associated with ageing and the pathogenesis of a variety of human diseases including Alzheimer’s, Parkinson’s, and many cancers. In mammals, the insertion of α-helical proteins into the outer membrane, a function that would not have been required in the ancestral endosymbiont, remains poorly understood. In yeast and trypanosomes the mitochondrial import protein 1 (Mim1) and pATOM36, respectively, have been implicated in this process, but no clear homologs exist in mammalian mitochondria. One important class of α-helical outer membrane proteins are tail-anchored proteins (TAs), which are characterized by a single C-terminal transmembrane domain (TMD) and mediate diverse functions including apoptosis, innate immunity, and mitochondrial turnover and dynamics. Therefore, we set out to systematically identify and characterize the factors required for mitochondrial TA biogenesis in human cells.

2Using an in vitro competition assay we first showed that TA insertion does not strictly require the TOM complex, the major outer membrane translocase (Fig. 1A, fig. S1–2;). Therefore, to enable CRISPR-based screens, we adapted and validated (Fig. 1B, fig. S3) a split-GFP reporter to measure insertion of the model TA, OMP25, into mitochondria. Amongst hits that increased mitochondrial integration of OMP25 were the ER membrane protein complex (EMC) and the ubiquilin (UBQLN) chaperone family (Fig. 1C; fig. S4A). These results are consistent with the EMC serving as the major insertase for mislocalized mitochondrial TAs into the ER (fig. S4;), and the UBQLNs’ role in degrading mislocalized mitochondrial TAs, leading to their accumulation in the cytosol (fig. S5).

3Conversely, depletion of the outer membrane resident mitochondrial carrier homologue 2 (MTCH2) resulted in the most pronounced loss of OMP25 integration (Fig. 1D, fig. S6A). MTCH2 is a member of the solute carrier 25 (SLC25) family, integral membrane proteins best known for their role in transporting metabolites into the mitochondrial matrix, but its localization and sequence suggests its function has potentially diverged, and it has no known substrates or transporter activity. Further, loss of MTCH2 is associated with a variety of pleotropic phenotypes including defects in mitochondrial fusion, lipid homeostasis, and apoptosis. However, the underlying biochemical activity of MTCH2 is not known.

4Because of the diverse phenotypes attributed to MTCH2, we excluded the possibility that dysregulation of lipogenesis (fig. S6B,C;), the outer membrane, or general mitochondrial protein biogenesis (Fig. 1E) could explain the observed biogenesis defect on OMP25. We next sought to determine if MTCH2 could be playing a more general role in the biogenesis of other mitochondrial outer membrane proteins. Using a quantitative proteomics strategy, we compared the steady-state levels of endogenous proteins in mitochondria isolated from wildtype or MTCH2 depleted cells (Fig. 2A, fig. S7A, Tables S2–3). We identified several outer membrane α-helical TA, signal anchored, and multipass proteins, that were reproducibly decreased upon loss of MTCH2 (Fig. 2B). Because MTCH2 levels do not appreciably alter the mRNA levels for these proteins (fig. S7C;), we concluded that the effects of MTCH2 on the mitochondrial outer membrane proteome must be occurring post-transcriptionally. To determine if MTCH2 exerts these effects specifically on biogenesis of nascent substrates, we tested a panel of mitochondrial proteins using our fluorescent reporter strategy (Fig. 1B). Consistent with the proteomics, MTCH2 affected the biogenesis of a functionally and biophysically diverse set of TA, signal anchored, and multipass proteins (Fig. 2C, D, fig. S8).

5Based on these experiments, we reasoned that MTCH2 may have evolved the ability to insert α-helical proteins into the outer membrane. To test this hypothesis, we focused on TA proteins, because they are the largest class of α-helical outer membrane proteins and adopt a uniform topology. Using an in vitro insertion assay with purified mitochondria (Fig. 3A, fig. S1C, fig. S9), we found that loss of MTCH2 affected the insertion of several mitochondrial TA proteins, but not unrelated intermembrane- or matrix-targeted controls (Fig. 3B, C and fig. S10, fig S11). Further, using site-specific crosslinking (Fig. 3D, Table S4;), we demonstrated that MTCH2 physically associated with nascent substrates during their insertion (Fig. 3E, fig. S12).

6Finally, to determine whether MTCH2 is sufficient for TA insertion, we purified MTCH2 (Fig. 3F) and optimized conditions for its reconstitution into liposomes (fig. S13). Using a panel of α-helical substrates, we show that purified MTCH2 specifically stimulated insertion of MTCH2 dependent, but not MTCH2 independent TAs and signal anchored proteins (Fig. 3G, fig. S14). To reconcile these results with earlier observations that trypsin-treated mitochondria remain competent for TA insertion, we found that in contrast to several subunits of the TOM complex, MTCH2 is largely trypsin resistant (Fig. 3H). Cumulatively, the requirement for MTCH2 in vivo and in vitro for TA insertion, together with its reconstituted insertase activity and physical association with substrates, rigorously establishes MTCH2 as an insertase for α-helical mitochondrial outer membrane proteins.

7Bioinformatic analysis reveals that in addition to MTCH2, other examples of SLC25 family members lacking canonical sequence motifs are found in both mitochondria and peroxisomes (Fig. 4A, fig. S15). Indeed, depletion of the close paralog MTCH1, which is also localized to the mitochondrial outer membrane, had an additive effect to loss of MTCH2 on biogenesis of many mitochondrial TAs (Fig. 4B, fig. S17). This result is consistent with our genome-wide screen (fig. S17C) and the synthetic lethal relationship between MTCH1 and 2. We therefore propose that MTCH1/2 are the founding members of a unique class of membrane protein insertases that exploit the SLC25 transporter fold (fig. S18).

8In contrast to other solute carrier family members, in which the transmembrane helices close to form a pore that allows charged species to cross the membrane, the AlphaFold2 predicted model of MTCH2 contains a prominent groove that is accessible to the membrane and lined with charged and polar residues (Fig. 4C, fig. S19). By introducing mutations at positions that altered the electrostatic potential of its intramembrane surfaces, we identified mutants that both diminish and enhance biogenesis of MTCH2-dependent but not independent substrates (Fig. 4C, fig. S19–20). We therefore concluded that MTCH2’s role in TA insertion relies on a hydrophilic surface within the bilayer.

9Given MTCH2’s central role in mitochondrial TA biogenesis, we asked whether it may broadly affect cellular proteostasis. We find that indeed, depletion of MTCH2 leads to an increase in ER insertion of mitochondrial TAs, while MTCH2 overexpression leads to a commensurate decrease in their mistargeting to the ER (Fig. 4D, fig. S21–22). This effect was enhanced by further depleting ATP13A1, an ER dislocase for mislocalized mitochondrial TAs (fig. S21C). These data suggest that MTCH2 is a central ‘gatekeeper’ for the mitochondrial outer membrane: MTCH2 levels and activity dictate the cytosolic reservoir of mitochondrial TAs, which then can be re-routed to the ER if unable to successful integrate into mitochondria (Fig. 4D).

10Finally, considering that insertion of several MTCH2-dependent TAs play a central role in apoptosis, we reasoned that MTCH2 activity may affect cellular sensitivity to apoptotic stimuli. To test this, we overexpressed MTCH2 in human K562 cells, which are derived from a myelogenous leukemia cell line known to upregulate the anti-apoptotic TA, BCL2L1. We treated cells with imatinib, a leukemia treatment which targets the BCR-ABL oncogene, and measured apoptosis. We found that while knockout of MTCH2 did not appreciably alter apoptosis propensity in this system, overexpression of wild type MTCH2 markedly sensitizes K562 cells to imatinib treatment (Fig. 4E). Critically, by expressing a series of MTCH2 mutants we found that this sensitization depends on MTCH2’s insertase activity.

11In summary, we have demonstrated that MTCH2 is a defining member of a family of membrane protein insertases that are necessary and sufficient for insertion of TAs into human mitochondria. MTCH2’s insertase activity relies on a hydrophilic groove within the bilayer, an apparent example of convergent evolution of many membrane protein translocases including the EMC, Hrd1, and YidC. A significant number of mitochondrial TAs are enriched in basic residues immediately C terminal to their TMDs, and may be particularly reliant on charged surfaces along their route into the membrane. MTCH2’s role also appears to extend to the integration of a broader class of α-helical proteins into the outer membrane, including signal anchored and multipass proteins. Homologs of MTCH2 are present in metazoan peroxisomes and its orthologs are found throughout holozoa, suggesting that the MTCH2 family has co-opted the SLC25 transporter fold to function in diverse biological membranes. The transition from a solute carrier, which mediates transport of small molecules across the membrane, to an insertase appears to have been enabled by the evolution of a membrane accessible hydrophilic groove absent from MTCH2’s SLC25 ancestors.

12Previously, loss of MTCH2 has been reported to lead to a diverse range of phenotypes including dysregulation of mitophagy, mitochondrial fragmentation, recruitment of tBID, and altered lipid homeostasis, and was also identified in a recent genome-wide association study for Alzheimer’s disease. The identification of MTCH2 as a key gatekeeper for α-helical outer membrane proteins now provides a molecular explanation for its pleotropic phenotypes, many of which can be directly ascribed to defects in biogenesis of MTCH2 substrates.