Functional Coupling of Presequence Processing and Degradation In

Functional coupling of presequence processing and degradation in human mitochondria Cansu Kuc€ ukk€ ose€ 1,2, Asli Aras Taskin1,3, Adinarayana Marada1, Tilman Brummer4,5,6,7, Sven Dennerlein8 and Friederike-Nora Vogtle€ 1,3

1 Faculty of Medicine, Institute of Biochemistry and Molecular Biology, ZBMZ,University of Freiburg, Germany 2 Faculty of Biology, University of Freiburg, Germany 3 CIBSS – Centre for Integrative Biological Signalling Studies, University of Freiburg, Germany 4 Faculty of Medicine, Institute of Molecular Medicine and Cell Research, University of Freiburg, Germany 5 Centre for Biological Signalling Studies BIOSS, University of Freiburg, Germany 6 Comprehensive Cancer Centre Freiburg, University of Freiburg, Germany 7 DKTK Partner Site Freiburg and DKFZ, Heidelberg, Germany 8 Department of Cellular Biochemistry, University Medical Center Gottingen,€ Germany

Keywords The mitochondrial proteome is built and maintained mainly by import of Alzheimer’s disease; integrated stress nuclear-encoded precursor proteins. Most of these precursors use N-termi- response; mitochondrial proteostasis; nal presequences as targeting signals that are removed by mitochondrial precursor protein import; presequence matrix proteases. The essential mitochondrial processing protease MPP degradation cleaves presequences after import into the organelle thereby enabling pro- Correspondence tein folding and functionality. The cleaved presequences are subsequently F.-N. Vogtle,€ Institute of Biochemistry and degraded by peptidases. While most of these processes have been discov- Molecular Biology, Stefan-Meier-Str. 17, ered in yeast, characterization of the human enzymes is still scarce. As 79104 Freiburg, Germany the matrix presequence peptidase PreP has been reported to play a role Tel: +49 761 2035270 in Alzheimer’s disease, analysis of impaired peptide turnover in human E-mail: nora.voegtle@biochemie. cells is of huge interest. Here, we report the characterization of uni-freiburg.de HEK293T PreP knockout cells. Loss of PreP causes severe defects in (Received 1 December 2019, revised 31 oxidative phosphorylation and changes in nuclear expression of stress March 2020, accepted 4 May 2020) response marker genes. The mitochondrial defects upon lack of PreP result from the accumulation of presequence peptides that trigger feed- doi:10.1111/febs.15358 back inhibition of MPP and accumulation of nonprocessed precursor proteins. Also, the mitochondrial intermediate peptidase MIP that cleaves eight residues from a subset of precursors after MPP processing is compromised upon loss of PreP suggesting that PreP also degrades MIP generated octapeptides. Investigation of the PrePR183Q patient mutation associated with neurological disorders revealed that the mutation destabilizes the protein making it susceptible to enhanced degradation and aggregation upon heat shock. Taken together, our data reveal a functional coupling between precursor processing by MPP and MIP and presequence degradation by PreP in human mitochondria that is crucial to maintain a functional organellar proteome.

Abbreviations PreP, presequence peptidase; MPP, mitochondrial processing protease.

600 The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. C. Kuc€ ukk€ ose€ et al. Coupling of presequence processing and degradation

Introduction The successive processing by MPP and Oct1 or Icp55 is therefore required to obtain stable and functional Mitochondrial proteostasis is essential for cellular sur- mitochondrial proteins. vival and mitochondrial dysfunctions are connected to As a result of presequence processing, not only a multitude of severe human diseases, for example, matured proteins are generated but also cleaved prese- neurological disorders, metabolic disorders, or car- quences are released into the mitochondrial matrix. diomyopathies [1-3]. To maintain a functional mito- Several studies mainly performed in the model organ- chondrial proteome, quality control mechanisms exist ism Saccharomyces cerevisiae have identified peptidases on several levels including protein biogenesis and pro- dedicated to degrade these free presequences. In yeast, tein turnover [4,5]. Mitochondrial proteases play a the concerted action of Cym1, Ste23, and Prd1, all decisive role in these processes, for example, by cleav- localized in the matrix, is securing efficient peptide ing targeting signals upon import for protein matura- degradation [20-22]. Deletion of these peptidases tion and functionality or by the degradation of results in growth defects that point into the direction damaged, superfluous, or misfolded proteins [4,5]. Sev- that accumulating presequences are toxic in vivo.In eral mutations in mitochondrial proteases have been human cells, two mitochondrial peptidases have been identified that result in severe human diseases, often identified in the matrix, the Cym1 homologue prese- affecting tissues with increased energy demands, like quence peptidase (PreP) and the Prd1 homologue, neu- brain or heart, but that have been also linked to global rolysin (NLN) [23,24]. While deletion strains in yeast human metabolic disorders [6-8]. However, analyses of enabled investigation of peptidase functions in vivo, the underlying pathomechanisms triggered by these analysis of the human enzymes so far mainly relied on mutations, the determinants of tissue specificity or dis- in vitro assays using recombinant proteins [23,24].As ease onset and progression are often lacking [9]. decreased proteolytic activity of PreP has been identi- The mitochondrial proteome is built mainly by fied in brain mitochondria of Alzheimer’s disease import of precursor proteins from the cytosol [10-12]. patients, mechanistic analysis of the function of PreP For targeting most of these precursors possess cleav- in human cells is of immense interest to understand able signals that are localized at the proteins’ N ter- the disease pathology [25]. Furthermore, recent studies mini [12,13]. These presequences are directing import have identified several patients with point mutations in into the mitochondrial matrix, where they are prote- PITRM1, encoding PreP, that present with a neurolog- olytically removed by the essential mitochondrial pro- ical disorder characterized by progressive spinocerebel- cessing protease MPP that is composed of the two lar ataxia, cognitive decline, and psychotic episodes subunits PMPCA and PMPCB [5,14]. Dysfunctional and in one family with severe progressive cerebellar MPP processing results in the accumulation of unpro- atrophy [26-28]. Conflicting reports on the effect of cessed precursor proteins in the mitochondrial matrix these mutations on PreP activity exist, suggesting [15]. These unprocessed precursors are prone to rapid either an enhanced turnover or a decreased enzymatic aggregation and are therefore not functional [15].As activity of the mutant protein as underlying cause of approximately 70% of all mitochondrial precursors mitochondrial dysfunction. In addition, these differing use presequences as targeting signals, proteolytic cleav- conclusions were based on in vitro degradation assays age of these presequences by MPP is indispensable to using recombinant PreP and on assays modeling the build up the mitochondrial proteome [13]. One-step disease mutation in the yeast homologous Cym1 pro- MPP processing of presequence precursors is often suf- tein [26-28]. However, missing knowledge of the func- ficient to generate functional proteins. However, sev- tional role of PreP has not permitted analysis of the eral precursors require a second maturation step pathophysiological consequences of the mutations in performed by the mitochondrial intermediate peptidase human cells in vivo so that clarification of these oppos- (MIP, Oct1 in yeast), which removes an octapeptide or ing findings has not yet been possible. the intermediate cleaving peptidase Icp55 (identified in Here, we report the generation of a PreP HEK293T yeast) that cleaves a single amino acid [5,13,16,17]. knockout cell line (PrePÀ/À) that allowed functional This two-step processing is required to convert MPP characterization of human PreP in vivo. Loss of PreP generated unstable processing intermediates into stable results in severe mitochondrial dysfunction character- and mature proteins. Basis for the instability of the ized by defects in the respiratory chain complexes and processing intermediates and the stability of their a decreased mitochondrial membrane potential. Cells mature counterparts are the identity of the N-terminal lacking PreP also displayed changes in nuclear expres- amino acid that correlates with the half-life of the pro- sion of genes associated with mitochondrial stress tein and follows a mitochondrial N-end rule [13,18,19]. responses. Mechanistic analysis of PrePÀ/À cells

The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of 601 Federation of European Biochemical Societies Coupling of presequence processing and degradation C. Kuc€ ukk€ ose€ et al. demonstrated impaired presequence peptide degrada- immunodecoration of mitochondria isolated from tion and as a consequence compromised presequence wild-type and PreP knockout (PrePÀ/À) cells. PreP- cleavage by MPP with accumulation of nonprocessed specific antisera confirmed the absence of PreP precursor proteins revealing a functional coupling of (Fig. 1B). Deletion of PreP resulted in a significant presequence degradation and precursor processing. growth defect on glucose medium, which was further Similarly, the activity of MIP was strongly compro- enhanced upon growth on medium containing galac- mised in in organello import assays suggesting that tose as carbon source (Fig. 1C). PreP also degrades cleaved octapeptides and that their To study changes in mitochondrial physiology upon accumulation in the absence of PreP induces feedback loss of PreP, we analyzed the integrity of respiratory inhibition of MIP. The identification of the close link chain complexes by Blue-native-PAGE (BN-PAGE). of PreP and MPP in human cells also enabled analysis Complex III, complex IV, and respiratory chain super- of the functional consequences of a PrePR183Q patient complexes (SC) were reduced in PrePÀ/À mitochondria mutation. Expression of PrePR183Q in PrePÀ/À cells when compared to wild-type, while complex II was not revealed an MPP defect that was caused by decreased altered (Fig. 2A). To assess whether this decrease in PrePR183Q protein levels rather than defective prote- respiratory chain complexes impacts also functionally olytic activity of the mutant protein. Further analyses on oxidative phosphorylation, we measured oxygen of the PrePÀ/À R183Q cells revealed that the severely consumption by real-time respirometry (Fig. 2B). reduced protein levels of mutant PreP observed in the Already basal respiration was significantly decreased in patients are likely caused by an increased protein turn- PrePÀ/À cells compared to control. Total respiratory over and destabilization of the protein upon heat capacity can be determined by addition of FCCP shock. uncoupling the respiratory chain complexes from the ATP synthase and revealed a severe reduction of À/À Results OXPHOS capacity in PreP cells. Measurement of the mitochondrial membrane potential (Dw) indicated a reduction in PrePÀ/À cells upon growth on glucose- Loss of human presequence peptidase PreP containing medium, which was further exacerbated results in severe mitochondrial dysfunctions when galactose was used as carbon source (Fig. 2C). To elucidate the function of PreP in human cells, we The mitochondrial proteome is mainly built by generated a knockout of PreP using CRISPR/Cas9- import of precursor proteins that are translated in the mediated disruption of its alleles in HEK293T cells. cytosol and imported into mitochondria post-transla- For this purpose, we targeted exon 5 of PITRM1 tionally [10-12]. Targeting to the organelle is in the (Fig. 1A). Complete loss of PreP was controlled by majority of proteins encoded in N-terminal signals, sequencing (see Materials and methods) and termed presequences that are cleaved upon import

A PITRM1 Fig. 1. Generation of human PrePÀ/À cells. (A) Schematic of the strategy to target Exon 5 human PITRM1 encoding PreP by CRISPR- Cas9. The guide RNA was directed against exon 5 of PITRM1. Successful targeting of 5’...TTTCAGAATCTCCTCTCGGTGTATTTGGATGCCACCTTTTTCCCATGTTTACGCGAGCTGGATTTCTGG...3’ the gene was conﬁrmed by sequencing. (B) F Q N L L S V Y L D A T F F P C L R E L D F W Mitochondria isolated from HEK293T wild- GGT 3’...AAAGTCTTAGAGAGCCACATAAACCTAC GGAAAAAGGGTACAAATGCGCTCGACCTAAAGACC...5’ type (WT) and PrePÀ/À cells were analyzed guide RNA on SDS/PAGE followed by western blotting. Data are representative of one experiment. ) ) The experiment was performed at least in 5 BCHEK293T 5

50 8 WT triplicates. (C) Proliferation assay of WT and

WT À/À –/–

40 –/– PreP cells grown on medium 6 –/– supplemented with glucose (left bars) or WT PreP 30 kDa PreP 4 PreP galactose (right bars) as carbon source. 20 ** - PreP *** Dashed line reﬂects starting cell number. 100 - 2 10 Values represent means Æ SEM, n = 3. 70 - - TOM70 0 0 Statistical analysis was performed using Number of cells (×10 Number of cells Number of cells (×10 Glucose Galactose 12 Student’s t-test (**P < 0.01, ***P < 0.001).

602 The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies C. Kuc€ ukk€ ose€ et al. Coupling of presequence processing and degradation

ABα-COX1 α-RIESKE α-SDHA ) WT PreP–/– –/– –/– –1 –/– 1000 ·OD WT WT PreP PreP WT PreP Oligo- FCCP Rotenone –1 800 mycin Antimycin kDa ]SC ]SC 600 669 - - CIII 400 440 - - CIV 200 232 - - CII OCR (pmol·min 0 140 - 0204060 Time (min) 12 34 56

C D *** *** 100 WT [35S]Hsp10 100 –/– –/– 80 PreP 80 WT PreP Mito. Hsp10

Δψ S] 60 60 +–++– +++ 35 5 15 30 30 5 15 30 30 Time [ Prec. 40 40 kDa (min) (% of WT)(% of –/– –/– 20 10 - 20 WT WT PreP PreP 0 123456789 Imported

Membrane pot. (% of WT)Membrane pot. (% of Glu Gal 1020 0 3 Time (min)

Fig. 2. Analysis of mitochondrial functions in PrePÀ/À cells. (A) Mitochondria isolated from wild-type (WT) and PrePÀ/À cells were solubilized in digitonin and analyzed on Blue-native (BN-) PAGE using the indicated antibodies. SC, respiratory chain supercomplexes. Data are representative of one experiment. The experiment was performed four times. (B) Oxygen consumption rates (OCR) of WT and PrePÀ/À cells were measured at basal conditions and after addition of indicated compounds. Values represent means Æ SEM, n = 6. (C) Membrane potential measurement (Membrane pot.) of WT and PrePÀ/À cells after growth on glucose (Glu, left bars) or galactose (Gal, right bars) as carbon source. Values represent means Æ SEM, n = 3. Statistical analysis was performed using Student’s t-test (***P < 0.001). (D) Radiolabeled Hsp10 precursor protein was incubated with isolated WT and PrePÀ/À mitochondria for the indicated period of time. Where indicated, the membrane potential (Dw) was dissipated prior to the import reaction. Samples were treated with Proteinase K to digest nonimported precursor proteins and analyzed by SDS/PAGE and autoradiography. Quantiﬁcations represent means Æ SEM, n = 3.

[13]. Import of these presequence precursors into the Taken together, lack of the presequence peptidase matrix requires the membrane potential as driving PreP in human cells results in a growth defect and force for translocation across the inner membrane. We strong reduction in respiratory chain capacity. As a wondered if also protein import is affected in the consequence, the mitochondrial membrane potential PrePÀ/À cells due to the reduced membrane potential decreases, compromising the mitochondrial capacity to caused by defective respiratory chain complexes. We import newly synthesized precursor proteins from the assessed protein import by in organello assays using cytosol. Hsp10 as a model substrate. Hsp10 uses the presequence import pathway and depends therefore on the PreP activity is required for the maturation of membrane potential. However, in contrast to most newly imported precursor proteins presequence precursors Hsp10 is not processed by MPP upon import [15,29]. Hsp10 is therefore an ideal In vitro assays using recombinant PreP have shown substrate to test the presequence protein import path- that PreP can degrade peptides and cleaved prese- way independent of presequence processing. Import of quences that are generated upon precursor processing Hsp10 into mitochondria isolated from PrePÀ/À cells by MPP [23]. To assess if lack of PreP results in ham- grown on glucose-containing medium revealed a mildly pered peptide clearance in mitochondria, we analyzed reduced import pointing to a decreased import activity presequence peptide degradation in soluble mitochon- in the absence of PreP (Fig. 2D), which is most likely drial extracts [21,29]. For this, isolated mitochondria caused by the decreased membrane potential (Fig. 2C). from control and PrePÀ/À cells were solubilized in

The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of 603 Federation of European Biochemical Societies Coupling of presequence processing and degradation C. Kuc€ ukk€ ose€ et al. digitonin followed by separation of soluble and mem- indicating that human PreP is required for degradation brane fractions via centrifugation [21,29]. Authentic of presequence peptides of different lengths. presequence peptides of the dually processed MPP We next investigated the functional consequences of substrate Frataxin (FXN1À41 and FXN42À80) or prese- impaired presequence peptide degradation on human quence peptides of mitochondrial malate dehydroge- mitochondria. Deletion of the yeast PreP homologue nase (MDH21À19) were added to the soluble extract, Cym1 and the matrix peptidase Ste23 was previously incubated for different time points and peptide degra- shown to result in feedback inhibition of MPP pro- dation monitored on Nu-PAGE. While the prese- cessing revealing a link between presequence degrada- quence peptides were efficiently degraded in wild-type tion and presequence processing in yeast [21,29]. extract over the time of incubation, this process was Presequence peptides accumulating upon lack of Cym1 strongly delayed in the absence of PreP (Fig. 3A–C) and Ste23 probably compete with incoming precursors for binding to MPP. This results in defective MPP processing, and as a consequence of this MPP inhibition, nonprocessed precursor proteins accumulate A WT PreP-/- Mito. extract [21,29]. Furthermore, PreP can also degrade amyloid- beta peptides and the activity of PreP has been 015305 015305 Time (min) kDa reported to decrease in brain mitochondria of Alzhei- 10 - - FXN1–41 mer’s disease patients and animal models of Alzhei- 100 - 70 - mer’s disease (AD) [23,25]. A potential functional 55- coupling of PreP and MPP could therefore also play a 40- role in disease pathogenesis, since accumulation of 35- control Coomassie nonprocessed precursor proteins was also identified in 12345678 mitochondria isolated from the temporal cortex of AD patients [29]. B WT PreP-/- Mito. extract To assess whether presequence processing is com- kDa 0203010 0203010 Time (min) promised in the absence of PreP, we analyzed imports 10 - of the presequence containing precursors TFAM (mi- - FXN42–80 tochondrial transcription factor A), which is processed by MPP, and OTC (mitochondrial ornithine car- 70 - - GRP75 bamoyltransferase), which is sequentially processed by 12345678 MPP and MIP, into isolated mitochondria from wild- type and PrePÀ/À cells. While TFAM was imported C into mitochondria in a membrane potential-dependent WT PreP-/- Mito. extract manner and processed by MPP in the control mito- kDa 05102 02 510Time (min) chondria, the mature form was strongly reduced in 10 - À/À - MDH21–19 PreP mitochondria (Fig. 4A). Furthermore, nonprocessed TFAM precursor accumulated in PrePÀ/À 70 - - GRP75 mitochondria, which was resistant to externally added 12 345678 Proteinase K. Similar to TFAM, import of OTC into isolated mitochondria was also dependent on the mem- Fig. 3. Functional analysis of PreP in presequence peptide brane potential and a two-step cleavage was observed À/À degradation. (A, B) Isolated wild-type (WT) and PreP in wild-type mitochondria. In contrast, processing by mitochondria were solubilized in digitonin and soluble proteins extracted via centrifugation. Authentic presequence peptides (first MPP from the precursor to intermediate OTC was À/À (FXN1À41) and second (FXN42À80) part of the presequence of compromised in PreP mitochondria and the second Frataxin) were added to the soluble mitochondrial extracts and processing by MIP generating the mature protein was their degradation monitored over time. Presequence peptides were strongly impaired (Fig. 4C). À detected by immunoblotting (a-FXN42 80) or with Coomassie We hypothesized that TFAM and OTC were still 1À41 staining (FXN ). GRP75, loading control. Data are imported into PrePÀ/À mitochondria, but MPP and representatives of one experiment. Experiments were performed MIP processing in the matrix was impaired. Steady- five times. (C) Same analysis as in (A and B) using the presequence peptide of MDH2 (MDH21À19). Samples were state protein levels of both subunits of MPP, PMPCA analyzed by SDS/PAGE and immunoblotting. GRP75, loading and PMPCB, and also of MIP were not changed in À/À control. Data are representative of one experiment. The PreP compared to wild-type mitochondria (Fig. 4B, experiment was performed five times. D) excluding that the observed processing defects were

604 The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies C. Kuc€ ukk€ ose€ et al. Coupling of presequence processing and degradation caused by a lack of the proteolytic enzymes. As MPP PrePÀ/À mitochondria resulting in accumulation of is essential for cell survival and the majority of mito- intermediate FXN and reduced levels of the mature chondrial precursors require MPP activity for their form (Fig. 4G). maturation, we further assessed MPP activity in Defects in mitochondrial protein biogenesis can ulti- PrePÀ/À mitochondria by using soluble mitochondrial mately result in imbalances of the mitochondrial pro- extracts. Soluble extracts have the advantage to uncou- teome and compromised mitochondrial proteostasis ple protein maturation by MPP processing from mem- has been shown to cause mitochondria-to-nucleus sig- brane potential dependent import across both naling that triggers changes in the expression of mitochondrial membranes and therefore allow direct nuclear genes aiming to ameliorate mitochondrial assessment of MPP activity [21,29]. We generated sol- function [31-33]. We wondered if loss of PreP with its uble extract from wild-type and PrePÀ/À mitochondria accumulation of presequence peptides and unprocessed and added radiolabeled precursors of Frataxin (FXN). precursor proteins could also trigger mitochondrial Frataxin is processed twice by MPP, making it a sensi- stress signaling and therefore assessed nuclear expres- tive substrate for MPP dysfunction [6]. While 35S- sion of stress markers by quantitative reverse tran- FXN was processed first to the intermediate and then scription-PCR (qRT-PCR). Indeed, transcripts of mature form in the wild-type extract, the second MPP genes associated with mitochondrial stress and here processing was not detectable in the PrePÀ/À sample especially with the integrated stress response (ATF4, and more unprocessed precursor was observed CHAC1, ASNS, PCK2) were increased upon loss of (Fig. 4E). Intriguingly, we observed appearance of a PreP (Fig. 4H) [34,35]. radiolabeled band below the molecular weight marker In summary, PreP is required for efficient peptide of 10 kDa only in the PrePÀ/À sample. We speculate turnover in the mitochondrial matrix of human cells. In that this band is the Frataxin presequence (FXN1À41), the absence of PreP, presequence peptides accumulate which is still partially cleaved off by MPP and then and trigger feedback inhibition of MPP and MIP. Dys- accumulates due to the lack of PreP, which would nor- functional MPP processing in turn leads to the accumu- mally degrade cleaved presequences. Analysis of the lation of nonprocessed precursor proteins. Changes in kinetics of the two-step Frataxin processing revealed mitochondrial proteostasis induced by loss of PreP fur- that the first MPP processing (precursor to intermedi- ther trigger changes in the expression of nuclear genes ate) is kinetically faster than the second processing (in- associated with mitochondrial stress responses likely in termediate to mature) reaction [30]. This difference in order to compensate for mitochondrial dysfunctions. kinetics could explain why only the second more sensi- Efficient presequence degradation by PreP therefore tive MPP processing is affected in the absence of PreP. appears to be crucially required to build and maintain a If accumulating presequence peptides in the absence functional mitochondrial proteome. of PreP result in feedback inhibition of MPP, then an overloading of PreP by exogenous addition of prese- PreP patient mutation is proteolytically active quence peptides to wild-type extract should mimic this feedback inhibition and also result in inhibition of The identification of the functional consequences of MPP processing. To assess this, we used mitochondrial PreP loss and its impact on mitochondrial physiology extract from wild-type cells and added increasing by inhibition of MPP processing enabled us for the amounts of presequence peptides (FXN42À80). MPP first time to investigate the pathological mechanism activity was monitored by processing of radiolabeled triggered by mutations in PITRM1 that have been Frataxin (Fig. 4F). While Frataxin was processed to identified in several patients in vivo [26,28]. Functional its mature form in the absence of peptides, increasing analyses of these PreP mutations have so far only been concentrations of FXN42À80 resulted in gradual accu- performed in vitro or in yeast models and yielded con- mulation of the processing intermediate and also pre- tradictory results [26-28]. While one study claimed a cursor form of radiolabeled Frataxin. Concomitantly destabilization of PreP by the PrePR183Q mutation [26], we observed accumulation of the cleaved presequence another study proposed that the same mutation affects in the low molecular weight range that could not be the catalytic activity of PreP [27]. However, PreP sta- degraded in the wild-type extract due to the saturation bility was not directly tested, but deduced from analy- of PreP activity by the exogenous peptides. sis of protein steady-state levels of the yeast We wondered if accumulation of peptides upon loss homologue Cym1, that did not include assays to deter- of PreP could result in an MPP processing defect also mine the half-life of the protein [26]. Furthermore, in vivo and analyzed FXN protein steady-state levels. PreP activity of wild-type and mutant was only mea- Indeed, FXN processing was impaired in vivo in sured in vitro using recombinant PreP and fluorogenic

The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of 605 Federation of European Biochemical Societies Coupling of presequence processing and degradation C. Kuc€ ukk€ ose€ et al. substrates but not authentic presequence peptides [27]. downstream consequences of the patient mutation The identiﬁcation of a functional coupling of PreP and in vivo using MPP activity as read-out for changes in MPP in human cells allowed now investigation of the PreP activity. – A [35S]TFAM 100 WT B –/ –/– WT WT PreP Mito. 80 kDa PreP +–++– +++ Δψ - PreP 60 100 - Prec. Time kDa 5 15 30 30 5 15 30 30 –/– (min) PreP 55 - - PMPCA -p 40

25 - WT)(% of 55 - -m - PMPCB

S] TFAM processing processing S] TFAM 20

123456789 35 [ 70 - - TOM70 1020 0 3 Time (min) 12 C D 35 100 WT – [ S]OTC –/ –/– 80 PreP

WT PreP Mito. WT Δψ kDa + ++–– +++ 60 - PreP Time 100 - Prec. 13 991399

kDa (min) - MIP

- 40 - p WT)(% of 70 - 40 - processing OTC - i ] –/– S 70 - - m 20 PreP - TOM70 35 [ 123456789 12 5 10 Time (min)

E [35S]FXN F 35 WT PreP–/– Mito. extract [ S]FXN WT Mito. extract 15 30 60 15 30 60 Time (min) kDa Prec. 42–80

kDa Prec. 0 5 10 20 50 FXN (µM)

25 - - - p - 25 - p - i - i - m - m FXN 10 - 10 - - preseq. FXN - preseq. peptide peptide 123456 1234567

G H *** 8 –/–

/WT) 7 –/– WT PreP kDa 6 WT - PreP –/– 100 - PreP 5 *** -i 15 - 4 FXN *** -m 3 *** *** ** 2 * 70 - - GRP75 * ** n.s. 12 1

Gene expression (fold PreP (fold Gene expression 0 ATF5HSPA9HSPD1 LONP CLPP ATF4 CHAC1 ASNS PCK2 TOMM70

606 The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies C. Kuc€ ukk€ ose€ et al. Coupling of presequence processing and degradation

We re-expressed wild-type (PreP) and the mutant Taken together, the patient mutation PrePR183Q does version (PrePR183Q) in PrePÀ/À cells to assess impact not impact on peptidase activity in vivo. Instead, of the mutation on MPP function. Re-expression of rather the low protein levels detected in the patients PreP in the PrePÀ/À cells (PrePÀ/À Resc.) partially res- and eventually degradation capacity of PreP seem to cued the MPP defect restoring FXN processing to its be triggering mitochondrial dysfunction by feedback mature form (Fig. 5A, lane 3). Interestingly, re-expres- inhibition of MPP. sion of PreP did not yield similar protein steady-state levels as detected in the wild-type cells and also MPP PreP patient mutation results in a destabilized processing of Frataxin was only partially restored and heat-shock-sensitive protein (compare lanes 1 and 3 in Fig. 5A). MPP activity, assessed by the ratio of intermediate to mature Fra- Based on our results that the severe reduction of taxin, therefore seems to directly correlate with PreP mutant PreP protein levels and not a compromised protein levels. We speculated that the amount of catalytic activity is likely responsible for the patho- PrePR183Q protein levels might be the underlying cause physiological consequences in the patients, we aimed for the pathophysiological consequences in the patients to identify the underlying cause of this low protein and not a compromised proteolytic activity [26,27]. abundance as it might be opening up a new avenue for Therefore, we re-expressed wild-type (PrePÀ/À Resc.) therapeutic intervention. We first assessed protein and mutant PreP (PrePR183Q) in PrePÀ/À cells and turnover in organello and compared degradation of compared cells that had equal protein steady-state PreP in mitochondria isolated from wild-type and levels of mitochondrial PreP (Fig. 5A, lanes 5 and 6), mutant PreP cells. Indeed, PrePR183Q was degraded to cells with strongly reduced PrePR183Q protein levels faster than its wild-type form (Fig. 5B) indicating that (Fig. 5A, lane 4), which reflects the PrePR183Q levels in the mutation destabilizes PreP and triggers degrada- patient fibroblasts [26]. When the expression levels of tion of the otherwise functional protein. However, PrePR183Q reached almost wild-type levels, no signifi- reduced protease levels can often be tolerated due to a cant difference in FXN processing was detectable fast turnover of substrates. We wondered why the (Fig. 5A, lanes 5 and 6). In contrast, if PrePR183Q was residual amounts of mutant PreP still result in severe expressed to a much lower extent, accumulation of MPP inhibition and speculated that the mutation elic- intermediate and reduced mature FXN protein levels its additional restraints on the enzyme that render the was observed (Fig. 5A, lanes 3 and 4). This demon- PrePR183Q mutant nonfunctional under certain condi- strates that the patient mutation is fully functional tions in vivo. We therefore assessed the sensitivity of in vivo, implicating that the strong reduction of mutant PrePR183Q upon stress conditions, for example, heat PreP protein levels in the patients is the underlying shock. For this, isolated mitochondria were incubated cause for mitochondrial dysfunction, and not loss of for 30 min at 39 °C followed by solubilization with enzymatic activity. TX-100 and nonsoluble and soluble fractions were

Fig. 4. Mechanistic analysis of PreP knockout on MPP activity. (A) 35S-TFAM precursor proteins were incubated with isolated wild-type (WT) and PrePÀ/À mitochondria for the indicated time. Where indicated, the membrane potential (Dw) was dissipated prior to the import reaction. Samples were treated with Proteinase K and analyzed by SDS/PAGE and autoradiography. Prec. and p, precursor; m, mature. Quantifications represent means Æ SEM, n = 3. (B) Immunoblot analysis of mitochondria isolated from WT and PrePÀ/À cells. TOM70, loading control. Data are representative of one experiment. The experiment was performed four times. (C) Same analysis as in (A) using the radiolabeled precursor of OTC. i, processing intermediate. Quantifications of the MIP processing step represent means Æ SEM, n = 3. (D) Immunoblot analysis of isolated mitochondria from WT and PrePÀ/À cells. TOM70, loading control. Data are representative of one experiment. The experiment was performed four times. (E) Processing of radiolabeled FXN precursor protein in soluble mitochondrial extract from WT and PrePÀ/À mitochondria. Samples were incubated over the indicated time and analyzed by SDS/PAGE and autoradiography. preseq., presequence. Shown is the result of a single experiment. The experiment was performed three times. (F) Processing of 35S-FXN in soluble mitochondrial extract from WT mitochondria with addition of increasing amounts of FXN42À80 presequence peptides. Samples were incubated for 15 min and analyzed by SDS/PAGE and autoradiography. Shown is the result of a single experiment. The experiment was performed three times. (G) Analysis of protein levels in mitochondria isolated from WT and PrePÀ/À cells by SDS/PAGE and immunodecoration. GRP75, loading control. Data are representative of one experiment. The experiment was performed two times. (H) Analysis of representative transcripts of the mitochondrial unfolded protein response and the integrated stress response encoded by nuclear DNA by qRT-PCR. TOMM70, control. n = 3, data represent mean Æ SEM. Statistical analysis was performed using Student’s t-test (*P < 0.05, **P < 0.01, ***P < 0.001, n.s. not significant).

The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of 607 Federation of European Biochemical Societies Coupling of presequence processing and degradation C. Kuc€ ukk€ ose€ et al.

ABPreP–/– Resc. 0 2 4 8 12 Time (h) –/– Resc. –/– R183Q –/– –/– R183Q –/– Resc. kDa - PreP 100 - PreP PreP WT PreP PreP kDa PreP 70 - - GRP75 - PreP(short 100 - exp.) 12345 - PreP(long 100 - exp.) PreP–/– R183Q - i 15 - FXN kDa 0 2 4 8 12 Time (h) - PreP - m 100 - 1234 56 70 - - GRP75 678910

C - heat shock D + heat shock –/– R183Q –/– Resc. PreP PreP PreP–/– Resc. PreP–/– R183Q SN P SN P kDa kDa SN P SN P - PreP - PreP 100 - 100 - 70 - - TOM70 70 - - TOM70

12 34 12 34

E Functional coupling of presequence processing and degradation in human mitochondria

Presequence MPP Mature Cleaved PreP Presequence precursor protein presequence degradation

Octa- MIP peptide

Feedback inhibition of presequence processing upon loss of PreP

Presequence Cleaved precursor Presequence presequence degradation PreP MPP re protein

Octa- MIP peptide

Fig. 5. Functional analysis of PreP patient mutation. (A) Analysis of protein levels in mitochondria isolated from wild-type (WT), PrePÀ/À, and PrePÀ/À cells with re-expression of either WT PreP (PrePÀ/À Resc.) or the R183Q PreP mutant (PrePÀ/À R183Q). Shown are two PrePÀ/À R183Q cell lines with different steady-state protein levels of PrePR183Q. Samples were analyzed by SDS/PAGE and immunodecoration. Shown is the result of a single experiment. The experiment was performed two times. (B) Mitochondria isolated from PrePÀ/À Resc. and PrePÀ/À R183Q cells were incubated at 37 °C and analyzed by SDS/PAGE and immunoblotting. Shown is the result of a single experiment. The experiment was performed two times. (C, D) Isolated mitochondria from PrePÀ/À Resc. and PrePÀ/À R183Q cells were solubilized in TX-100 and separated into soluble (supernatant, SN) and aggregated (pellet, P) fraction. + heat shock, mitochondria were incubated for 30 min at 39 °C prior to solubilization. Data are representatives of one experiment. Experiments were performed three times. (E) Model of the functional coupling of presequence processing by MPP and MIP and presequence and octapeptide degradation by PreP in human mitochondria (top panel). Loss of PreP results in the accumulation of cleaved presequences and octapeptides that trigger feedback inhibition of MPP and MIP. As a consequence, unprocessed precursor proteins accumulate that rapidly aggregate and are not functional (lower panel). The degradation of presequence peptides by PreP is therefore essential to maintain mitochondrial proteostasis. separated via centrifugation. Wild-type and mutant major part of PrePR183Q was found in the pellet frac- PreP were found in the soluble fraction (supernatant) tion containing nonsoluble, for example, aggregating if mitochondria had not been exposed to the heat proteins, while the wild-type protein was still predomi- shock (Fig. 5C). However, when isolated mitochondria nantly recovered in the soluble fraction (Fig. 5D). were subjected to the mild in organello heat shock, a These results indicate that PrePR183Q is less stable than

608 The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies C. Kuc€ ukk€ ose€ et al. Coupling of presequence processing and degradation the wild-type protein and additionally prone to mis- recombinant IDE and a synthetic presequence peptide folding and aggregation under heat-shock conditions. implicated degradation of the cleaved targeting signal The low protein amounts still detectable in the patient by IDE [39]. Furthermore, the Prd1 homologue mito- would therefore be labile under, for example, fever, chondrial NLN has also been shown to cooperate with which might trigger the detrimental circuit of disease PreP in degradation of long presequence and also onset and progression by further compromising pep- amyloid-beta peptides in vitro [24]. While the role of tide clearance in the mitochondrial matrix. these two peptidases in presequence turnover and the consequences of their lack in vivo await analysis, the À/À Discussion characterization of PreP cells revealed that loss of PreP is sufficient to trigger MPP dysfunction. Most mitochondrial precursor proteins have N-termi- We observed the accumulation of presequence pep- nal presequences for targeting and translocation into tides, nonprocessed precursors, and processing inter- mitochondria that are proteolytically removed upon mediates that all can cause disturbances in import. This presequence processing by MPP is mitochondrial proteostasis in PrePÀ/À mitochondria. It required to generate functional proteins as nonpro- has been suggested that mammalian cells sense and cessed precursors are rapidly aggregating [15]. Precur- respond to mitochondrial dysfunctions by activation sor processing is therefore essential to build and of protective transcriptional responses that promote, maintain a functional mitochondrial proteome. The for example, synthesis of nuclear-encoded mitochon- importance of precursor maturation is also underlined drial chaperones and proteases [31-35]. We found a by identification of several patients with mutations in strong transcriptional increase of components of the one of the two MPP subunits that suffer from neu- integrated stress response involving the transcription rodegeneration [6,36-38]. As a side product of MPP factor ATF4 and its targets CHAC1, ASNS, and processing free presequence peptides are generated that PCK2, while other previously reported stress marker are degraded by matrix-localized peptidases. Studies in genes (ATF5, HSPA9, CLPP, LONP, HSPD1) yeast revealed that this presequence degradation is showed no or only very mild changes. Therefore, loss required to maintain MPP activity [21,29]. Here, we of PreP elicits a mitochondria-to-nuclear signaling identify a functional coupling between the presequence response adapting nuclear gene expression to support peptidase PreP and the presequence proteases MPP mitochondrial function. The discovery of the changed and MIP in human mitochondria. Loss of PreP results nuclear expression upon loss of PreP further suggests in accumulation of presequence peptides that in turn that PreP is playing a dominant role for peptide degra- induce feedback inhibition of MPP and MIP dation and mitochondrial proteostasis in human cells. (Fig. 5E). The identification of the functional coupling of pre- Analysis of yeast mitochondria revealed that three sequence degradation and presequence processing in peptidases, Cym1, Ste23, and Prd1, degrade cleaved human cells also enabled characterization of a PreP presequences. Intriguingly, the studies that analyzed mutation (PrePR183Q), which was identified in patients the function of the yeast PreP homologue Cym1 used suffering from a neurological disorder characterized by in vitro assays with addition of exogenous presequence slowly progressive spinocerebellar ataxia, mental retar- peptides to elicit feedback inhibition of MPP [21,29]. dation, and psychosis [26]. While analysis in yeast and This suggests that the three yeast presequence pepti- in vitro assays using the recombinant enzyme resulted dases have overlapping substrate spectra and can par- in contradictory results regarding the effect of the tially compensate each other. In contrast, lack of PreP patient mutation, we were now able to analyze the in human cells resulted in a severe impairment of MPP consequences of impaired PreP activity in vivo by processing already in the absence of additional pep- assessing feedback inhibition of MPP activity. Expres- tides. Therefore, PreP seems to have a predominant sion of PrePR183Q at the low protein levels detected in role in peptide degradation in human cells as the dele- patient cells was resulting in a strong impairment of tion of PreP alone already results in the accumulation MPP activity in vivo. In contrast, expression of the of a significant amount of peptides that induce MPP PrePR183Q mutation in PrePÀ/À cells at wild-type levels dysfunction in organello and in vivo. restored MPP processing. Therefore, low PreP protein Also in human mitochondria, two further peptidases levels are likely responsible for the mitochondrial have been reported to be capable of presequence defects observed in the patient and not a reduced pro- degradation besides PreP. An isoform of human IDE, teolytic activity. Using HEK293T cells, we could also homologous to yeast Ste23, has been suggested to investigate the underlying nature of the decreased pro- localize to mitochondria [39]. In vitro analysis using tein levels in the PreP patients and found that

The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of 609 Federation of European Biochemical Societies Coupling of presequence processing and degradation C. Kuc€ ukk€ ose€ et al.

PrePR183Q was faster degraded compared to the wild- of 8 lgÁmLÀ1 polybrene. Positive clones were selected using À type protein. Furthermore, PrePR183Q was also prone 0.5 lgÁmL 1 puromycin. to aggregation upon heat shock, which could render the residual amounts identified in the patients non- Cell proliferation assay functional upon stresses like fever that might be rele- vant in triggering disease onset or further exacerbate Growth analysis was performed by seeding cells at a den- disease progression. sity of 200 000 cells per well into a 6-well plate. Cells were Taken together, we identified a functional coupling counted after 3 days by Countess II FL-automated cell of presequence processing by MPP and MIP and pre- counter (Thermo Fisher, Waltham, MA, USA). sequence degradation by PreP in human mitochondria that is likely playing a role in the pathogenesis of sev- Isolation of mitochondria eral diseases. The identification of the crucial role of PreP for the activity of the essential mitochondrial Mitochondria were isolated as previously described [41]. processing protease MPP opens up new avenues for Cells were harvested and resuspended in solution A treatment of dysfunctions not only caused by muta- (220 mM mannitol, 70 mM sucrose, 20 mM HEPES-KOH Á À1 tions in PreP but also for MPP dysfunctions and Alz- (pH 7.6), 1 mM EDTA, 0.5 mM PMSF, and 2 mg mL heimer’s disease, in which a decreased PreP activity BSA). Subsequently, cells were homogenized using a glass potter. Samples were centrifuged at 800 g for 5 min at 4 °C has been reported [6,25,26]. to remove cellular debris, and the supernatant was subjected to centrifugation at 10 000 g for 15 min at 4 °C. Mitochondrial pellet was resuspended in solution B (solu- Materials and methods tion A without BSA).

Cell culture MPP activity assay Human embryonic kidney cell lines HEK293T were cul- tured in DMEM (Gibco, Carlsbad, CA, USA) containing Isolated mitochondria were solubilized in reaction buffer À1 4.5 gÁL glucose or galactose and supplemented with 10% (10 mM HEPES-KOH (pH 8), 1 mM MnCl2, and 1 mM (v/v) fetal bovine serum (FBS) (Sigma, F7524, St. Louis, DTT) containing 1% (w/v) digitonin, incubated on ice for MO, USA) and 2 mM L-glutamine (Sigma, G7513) at 15 min, and centrifuged at 10 000 g for 10 min at 4 °C.

37 °C in a humidiﬁed incubator with 5% CO2. All func- Supernatant was incubated with radiolabeled Frataxin pre- tional analyses were performed with cells grown on glucose cursor at 37 °C for different time points and analyzed by as carbon source. SDS/PAGE followed by autoradiography. The PageRuler prestained marker (Thermo Fisher Scientiﬁc, #26616) was used as molecular weight marker. Generation of knockout and stable cell lines

CRISPR/Cas9 genome editing was used to generate PreP Peptide degradation assay knockout HEK293T cell line as described previously and Isolated mitochondria were solubilized in reaction buffer the pSpCas9(BB)-2A-GFP (PX458) was a gift from Feng (20 mM HEPES-KOH (pH 8), 10 mM MgCl ,1mM Zhang (Addgene plasmid, #48138) [40]. Guide RNA 2 MnCl ) containing 1% (w/v) digitonin, incubated on ice for sequence (GCGTAAACATGGGAAAAAGGTGG) target- 2 15 min, and centrifuged at 10 000 g for 10 min at 4 °C. ing human PITRM1 exon 5 was cloned into pSpCas9 Obtained supernatant was incubated with 30 lM Frataxin (BB)-2A-GFP vector. HEK293T cells were transfected by presequence peptides 1-41 (MWTLGRRAVAGLLASP Lipofectamine 2000 (Invitrogen, #11668-027, Waltham, MA, SPAQAQTLTRVPRPAELAPLCGRRG), 20 lM 42-80 (LRTD USA). Five hours after transfection, GFP-positive cells were IDATCTPRRASSNQRGLNQIWNVKKQSVYLMNLRK), or single-cell sorted into a 96-well plate by FACS. Clones were 80 lM MDH2 presequence peptide 1-19 (MLSALARPA- expanded and screened for PreP expression by western blot- SAALRRSFST) at 37 °C for different time points sepa- ting and indel mutations were veriﬁed by sequencing. The rately. Samples were analyzed by Nu-PAGE (Invitrogen) obtained PrePÀ/À cells were complemented by retroviral followed by immunoblotting. transduction. PITRM1 and PITRM1-R183Q cDNA were cloned into pBABE-puromycin vector. NIH293T cells were transfected by retroviral constructs using Lipofectamine Blue-native-PAGE LTX (Invitrogen, #15338-100). After 48 h, viral supernatant À/À Respiratory chain complexes were analyzed by Blue-native was collected and used to infect PreP cells in the presence polyacrylamide gel electrophoresis. Mitochondria were

610 The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies C. Kuc€ ukk€ ose€ et al. Coupling of presequence processing and degradation solubilized in solubilization buffer containing 1% (w/v) RNA isolation and qRT-PCR digitonin, 20 mM Tris/HCl (pH 7.4), 0.5 mM EDTA, 10% Total RNA was isolated using the RNeasy Mini Kit (QIA- glycerol, 50 mM NaCl and incubated on ice for 15 min. Samples were centrifuged at 20 000 g for 5 min at 4 °C GEN, Venlo, Netherlands) and treated with DNase to pre- and supernatant was loaded on 4–10% gradient Blue-na- vent DNA contamination. cDNA was synthesized with tive-PAGE followed by immunoblotting. High-Capacity cDNA kit (Applied Biosystems, Foster City, CA, USA) from 1 lg RNA. PCR amplification and detection were done by CFX384 Real-time PCR detection sys- Mitochondrial respiration tem (Bio-Rad, Hercules, CA, USA) using SsoAdvanced Mitochondrial respiration was measured by XF96 Extracel- Universal SYBR Green Supermix (Bio-Rad). Glyceralde- lular Flux analyzer (Seahorse Bioscience, Billerica, MA, hyde-3-phosphate dehydrogenase (GAPDH) was used as a USA). Cells were seeded on a poly-D-lysine-coated plate at housekeeping gene for normalization and relative mRNA a density of 20 000 cells per well the day before the mea- levels were calculated using the delta-delta Ct method. surement. Basal levels of oxygen consumption rate (OCR) Each experiment was performed in triplicates. The follow- and OCR in the presence of electron transport chain inhibi- ing primer pairs were used: ATF4-F, CAGCAAGGAGGA tors and uncouplers (2 lM oligomycin, 0.3 lM carbonyl TGCCTTCT; ATF4-R, CCAACAGGGCATCCAAGTC; cyanide-4-(trifluoromethoxy) phenylhydrazone, 1 lM anti- CHAC1-F, GTGGTGACGCTCCTTGAAGA; CHAC1-R, mycin, and 1 lM rotenone) were measured. Bradford assay TTCAGGGCCTTGCTTACCTG; ASNS-F, GATGAACT was performed after the assay and OCR results were nor- TACGCAGGGTTACA; ASNS-R, CACTCTCCTCCTCG malized to protein content. GCTTT; PCK2-F, AAACCCTGGAAACCTGGTG; PCK2-R, CAATGGGGACACCCTCTG; ATF5-F, GAGCCCCTGG CAGGTGAT; ATF5-R, CAGAGGGAGGAGAGCTGT Membrane potential measurement GAA; CLPP-F, AAGCACACCAAACAGAGCCT; CLPP-R, Membrane potential was measured by TMRE mitochon- AAGATGCCAAACTCCTGGG; HSPA9-F, TGGTGAGCG drial membrane potential assay kit (Abcam, ab113852, ACTTGTTGGAAT; HSPA9-R, ATTGGAGGCACGGACA Cambridge, UK). Tetramethylrhodamine, ethyl ester ATTTT; HSPD1-F, ACTCGGAGGCGGAAGAAA; HSP (TMRE) was added to cells at a final concentration of D1-R, TGTGGGTAACCGAAGCATTT; LONP-F, CCCG 50 nM and incubated 15 min at 37 °C. Cells were analyzed CGCTTTATCAAGATT; LONP-R, AGAAAGACGCCGA by BD LSR II flow cytometer (488/575 nm ex/em). Data CATAAGG GAPDH-F, AGGGTCATCATCTCTGCCCC analysis was performed by FLOWJO software (BD, Becton CTC; and GAPDH-R, TGTGGTCATGAGTCCTTCCAC Dickinson Company, Ashland, OR, USA). GAT; TOMM70-F, TTTTGCATTGTACCGCCAGG and TOMM70-R, ATAGCCTTCGGCACACCTTG.

In organello import of radiolabeled precursor proteins Protein aggregation assay

Radiolabeled precursor proteins (human TFAM, human Isolated mitochondria (40 lg) were lysed in PBS containing Frataxin, rat OTC, yeast Hsp10) were synthesized in vitro 0.5% Triton X-100 and incubated 5 min on ice. Samples with rabbit reticulocyte lysate system (Promega, Madison, were centrifuged at 16 000 g for 10 min at 4 °C. The pellet 35 WI, USA) in the presence of S-methionine. Radiolabeled was resuspended in PBS containing 0.5% Triton. All frac- precursor proteins and isolated mitochondria were incu- tions were subjected to TCA precipitation. Alternatively, bated in import buffer (250 mM sucrose, 5 mM magnesium mitochondria were subjected to an in organello heat shock acetate, 80 mM potassium acetate, 10 mM sodium succinate, (39 °C for 30 min) prior to the aggregation assay. Samples 20 mM HEPES-KOH (pH 7.4)) supplemented with 1 mM were analyzed by SDS/PAGE and immunoblotting. DTT and 5 mM ATP for the indicated time points at 37 °C. Membrane potential was disrupted prior to the import reaction by addition of AVO (8 lM antimycin, 1 lM In organello degradation assay valinomycin, 20 lM oligomycin). Samples were treated with 20 lgÁmLÀ1 Proteinase K to digest nonimported precursor Isolated mitochondria were incubated in sucrose buffer ° proteins. Mitochondria were re-isolated by centrifugation (10 mM HEPES-KOH (pH 7.6), 0.5 M sucrose) at 37 C. at 10 000 g for 10 min at 4 °C. Samples were analyzed by Samples were collected at indicated time points, and mito- SDS/PAGE and autoradiography. chondria were re-isolated by centrifugation at 10 000 g for 10 min at 4 °C. Samples were analyzed by SDS/PAGE and immunoblotting.

The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of 611 Federation of European Biochemical Societies Coupling of presequence processing and degradation C. Kuc€ ukk€ ose€ et al.

Statistical analysis machinery: cleavage, quality control and turnover. Cell Tissue Res 367,73–81. Æ Data shown represent means standard error of the mean 6Vogtle€ FN, Brandl€ B, Larson A, Pendziwiat M, (SEM). Statistical details of each experiment can be found Friederich MW, White SM, Basinger A, Kuc€ ukk€ ose€ C, in the figure legends. Student’s t-test was applied to com- Muhle H, Jahn€ JA et al. (2018) Mutations in PMPCB pare between two groups. Significances are indicated with encoding the catalytic subunit of the mitochondrial < < < asterisks: ***P 0.001, **P 0.01, *P 0.05, not signifi- presequence protease cause neurodegeneration in early > cant (n.s.) P 0.05. childhood. Am J Hum Genet 102, 557–573. 7 Eldomery MK, Akdemir ZC, Vogtle€ FN, Charng WL, Acknowledgements Mulica P, Rosenfeld JA, Gambin T, Gu S, Burrage LC, Al Shamsi A et al. (2016) MIPEP recessive variants We thank Dr. Ralf Zerbes for support with tissue cul- cause a syndrome of left ventricular non-compaction, ture, Dr. Mike Ryan and Dr. Feng Zhang for hypotonia, and infantile death. Genome Med 8, 106. NIH293T cells and plasmids, and Dr. Chris Meisinger 8 Ahola S, Langer T & MacVicar T (2019) Mitochondrial and Dr. Jan Riemer for discussion. This work was proteolysis and metabolic control. Cold Spring Harb supported by the Deutsche Forschungsgemeinschaft Perspect Biol 11, a033936. (DFG), under Germany’s Excellence Strategy (CIBSS- 9 Lightowlers RN, Taylor RW & Turnbull DM (2015) EXC-2189-Project ID 390939984), the SFB 1381 (Pro- Mutations causing mitochondrial disease: what is new ject ID 403222702; to FNV), and the Emmy-Noether and what challenges remain? Science 349, 1494–1499. Programme (to FNV). TB is supported by the DFG 10 Hansen KG & Herrmann JM (2019) Transport of by a Heisenberg-Professorship and AM by DST- proteins into mitochondria. Protein J 38, 330–342. SERB, India. 11 Dimmer KS & Rapaport D (2012) Unresolved mysteries in the biogenesis of mitochondrial membrane proteins. Biochim Biophys Acta 1818, 1085–1090. Conflict of interest 12 Schulz C, Schendzielorz A & Rehling P (2015) The authors declare no conflict of interest. Unlocking the presequence import pathway. Trends Cell Biol 25, 265–275. 13 Vogtle€ FN, Wortelkamp S, Zahedi RP, Becker D, Author contributions Leidhold C, Gevaert K, Kellermann J, Voos W, Sickmann A, Pfanner N et al.(2009)Globalanalysisof CK, AAT, and AM performed the experiments. CK, the mitochondrial N-proteome identifies a processing TB, SD, and FNV designed experiments and analyzed peptidase critical for protein stability. Cell 139,428–439. and interpreted the data. CK and FNV developed the 14 Teixeira PF & Glaser E (2013) Processing peptidases in project. FNV wrote the manuscript and coordinated mitochondria and chloroplasts. Biochim Biophys Acta and directed the project. All authors approved the 1833, 360–370. final version of the manuscript. 15 Poveda-Huertes D, Matic S, Marada A, Habernig L, Licheva M, Myketin L, Gilsbach R, Tosal-Castano S, Papinski D, Mulica P et al. (2020) An early mtUPR: References redistribution of the nuclear transcription factor Rox1 1 Nunnari J & Suomalainen A (2012) Mitochondria: in to mitochondria protects against intramitochondrial sickness and in health. Cell 148, 1145–1159. proteotoxic aggregates. Mol Cell 77, 180–188. 2 Gorman GS, Chinnery PF, DiMauro S, Hirano M, 16 Chew A, Buck EA, Peretz S, Sirugo G, Rinaldo P & Koga Y, McFarland R, Suomalainen A, Thorburn DR, Isaya G (1997) Cloning, expression, and chromosomal Zeviani M & Turnbull DM (2016) Mitochondrial assignment of the human mitochondrial intermediate diseases. Nat Rev Dis Primers 2, 16080. peptidase gene (MIPEP). Genomics 40, 493–496. 3 Dudek J, Hartmann M & Rehling P (2019) The role of 17 Gakh O, Cavadini P & Isaya G (2002) Mitochondrial mitochondrial cardiolipin in heart function and its processing peptidases. Biochim Biophys Acta 1592, implication in cardiac disease. Biochim Biophys Acta 63–77. Mol Basis Dis 1865, 810–821. 18 Vogtle€ FN, Prinz C, Kellermann J, Lottspeich F, 4 Quiros PM, Langer T & Lopez-Ot ın C (2015) New Pfanner N & Meisinger C (2011) Mitochondrial protein roles for mitochondrial proteases in health, ageing and turnover: role of the precursor intermediate peptidase disease. Nat Rev Mol Cell Biol 16, 345–359. Oct1 in protein stabilization. Mol Biol Cell 22, 5 Poveda-Huertes D, Mulica P & Vogtle€ FN (2017) The 2135–2143. versatility of the mitochondrial presequence processing 19 Varshavsky A (2011) The N-end rule pathway and regulation by proteolysis. Protein Sci 20, 1298–1345.

612 The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of Federation of European Biochemical Societies C. Kuc€ ukk€ ose€ et al. Coupling of presequence processing and degradation

20 Alikhani N, Berglund AK, Engmann T, Spanning E, intermediate forms are cleaved at different rates. J Biol Vogtle€ FN, Pavlov P, Meisinger C, Langer T & Glaser Chem 275, 41469–41475. E (2011) Targeting capacity and conservation of PreP 31 Topf U, Wrobel L & Chacinska A (2016) Chatty homologues localization in mitochondria of different mitochondria: keeping balance in cellular protein species. J Mol Biol 410, 400–410. homeostasis. Trends Cell Biol 26, 577–586. 21 Taskin AA, Kuc€ ukk€ ose€ C, Burger N, Mossmann D, 32 Andreasson C, Ott M & Buttner€ S (2019) Mitochondria Meisinger C & Vogtle€ FN (2017) The novel orchestrate proteostatic and metabolic stress responses. mitochondrial matrix protease Ste23 is required for EMBO Rep 20, e47865. efﬁcient presequence degradation and processing. Mol 33 Suomalainen A & Battersby BJ (2018) Mitochondrial Biol Cell 28, 997–1002. diseases: the contribution of organelle stress responses 22 Vogtle€ FN, Burkhart JM, Gonczarowska-Jorge H, to pathology. Nat Rev Mol Cell Biol 19,77–92. Kuc€ ukk€ ose€ C, Taskin AA, Kopczynski D, Ahrends R, 34 Quiros PM, Prado MA, Zamboni N, D’Amico D, Mossmann D, Sickmann A, Zahedi RP et al. (2017) Williams RW, Finley D, Gygi SP & Auwerx J (2017) Landscape of submitochondrial protein distribution. Multi-omics analysis identiﬁes ATF4 as a key regulator Nat Commun 8, 290. of the mitochondrial stress response in mammals. J Cell 23 Falkevall A, Alikhani N, Bhushan S, Pavlov PF, Busch Biol 216, 2027–2045. K, Johnson KA, Eneqvist T, Tjernberg L, Ankarcrona 35 Fiorese CJ, Schulz AM, Lin YF, Rosin N, Pellegrino M & Glaser E (2006) Degradation of the amyloid beta- MW & Haynes CM (2016) The transcription factor protein by the novel mitochondrial peptidasome, PreP. ATF5 mediates a mammalian mitochondrial UPR. Curr J Biol Chem 281, 29096–29104. Biol 26, 2037–2043. 24 Teixeira PF, Masuyer G, Pinho CM, Branca RMM, 36 Jobling RK, Assoum M, Gakh O, Blaser S, Raiman Kmiec B, Wallin C, Warml€ ander€ SKTS, Berntsson RP, JA, Mignot C, Roze E, Durr€ A, Brice A, Levy N Ankarcrona M, Graslund€ A et al. (2018) Mechanism of et al. (2015) PMPCA mutations cause abnormal peptide binding and cleavage by the human mitochondrial protein processing in patients with mitochondrial peptidase neurolysin. J Mol Biol 430, non-progressive cerebellar ataxia. Brain 138, 348–362. 1505–1517. 25 Alikhani N, Guo L, Yan S, Du H, Pinho CM, Chen 37 Choquet K, Zurita-Rendon O, La Piana R, Yang S, JX, Glaser E & Yan SS (2011) Decreased proteolytic Dicaire MJ, Care4Rare Consortium, Boycott KM, activity of the mitochondrial amyloid-b degrading Majewski J, Shoubridge EA, Brais B & et al. enzyme, PreP peptidasome, in Alzheimer’s disease brain (2016) Autosomal recessive cerebellar ataxia caused mitochondria. J Alzheimers Dis 27,75–87. by a homozygous mutation in PMPCA. 26 Brunetti D, Torsvik J, Dallabona C, Teixeira P, Brain 139, e19. Sztromwasser P, Fernandez-Vizarra E, Cerutti R, Reyes 38 Joshi M, Anselm I, Shi J, Bale TA, Towne M, Schmitz- A, Preziuso C, D’Amati G et al. (2016) Defective Abe K, Crowley L, Giani FC, Kazerounian S, PITRM1 mitochondrial peptidase is associated with Ab Markianos K et al. (2016) Mutations in the substrate amyloidotic neurodegeneration. EMBO Mol Med 8, binding glycine-rich loop of the mitochondrial 176–190. processing peptidase-a protein (PMPCA) cause a severe 27 Smith-Carpenter JE & Alper BJ (2018) Functional mitochondrial disease. Cold Spring Harb Mol Case Stud requirement for human pitrilysin metallopeptidase 1 2, a000786. arginine 183, mutated in amyloidogenic neuropathy. 39 Leissring MA, Farris W, Wu X, Christodoulou DC, Protein Sci 27, 861–873. Haigis MC, Guarente L & Selkoe DJ (2004) Alternative 28 Langer Y, Aran A, Gulsuner S, Abu Libdeh B, translation initiation generates a novel isoform of Renbaum P, Brunetti D, Teixeira PF, Walsh T, insulin-degrading enzyme targeted to mitochondria. Zeligson S, Ruotolo R et al. (2018) Mitochondrial Biochem J 383, 439–446. PITRM1 peptidase loss-of-function in childhood 40 Ran FA, Hsu PD, Wright J, Agarwala V, Scott DA & cerebellar atrophy. J Med Genet 55, 599–606. Zhang F (2013) Genome engineering using 29 Mossmann D, Vogtle€ FN, Taskin AA, Teixeira PF, the CRISPR-Cas9 system. Nat Protoc 8, Ring J, Burkhart JM, Burger N, Pinho CM, Tadic J, 2281–2308. Loreth D et al. (2014) Amyloid-b peptide induces 41 Johnston AJ, Hoogenraad J, Dougan DA, Truscott mitochondrial dysfunction by inhibition of preprotein KN, Yano M, Mori M, Hoogenraad NJ & Ryan maturation. Cell Metab 20, 662–669. MT (2002) Insertion and assembly of human tom7 30 Cavadini P, Adamec J, Taroni F, Gakh O & Isaya G into the preprotein translocase complex of the outer (2000) Two-step processing of human frataxin by mitochondrial membrane. J Biol Chem 277, mitochondrial processing peptidase. Precursor and 42197–42204.

The FEBS Journal 288 (2021) 600–613 ª 2020 The Authors. The FEBS Journal published by John Wiley & Sons Ltd on behalf of 613 Federation of European Biochemical Societies