Abstract Respiratory complex I plays a crucial role in the mitochondrial electron transport chain and shows promise as a therapeutic target for various human diseases. While most studies focus on inhibiting complex I at the Q-site, little is known about inhibitors targeting other sites within the complex. In this study, we demonstrate that diphenyleneiodonium (DPI), a N-site inhibitor, uniquely affects the stability of complex I by reacting with its flavin cofactor FMN. Treatment with DPI blocks the final stage of complex I assembly, leading to the complete and reversible degradation of complex I in different cellular models. Growing cells in medium lacking the FMN precursor riboflavin or knocking out the mitochondrial flavin carrier gene SLC25A32 results in a similar complex I degradation. Overall, our findings establish a direct connection between mitochondrial flavin homeostasis and complex I stability and assembly, paving the way for novel pharmacological strategies to regulate respiratory complex I. Keywords: Respiratory complex I, FMN, OXPHOS, DPI 1. Introduction In recent years, the involvement of respiratory complex I in the development of various common human diseases has become evident. Extensive research is currently underway to investigate drugs that target complex I as potential treatments for conditions such as ischemia reperfusion injury [[39]1] and cancer [[40]2]. Most drugs under study are active at the Q-site, which is the binding site for quinone molecules in the enzyme [[41][3], [42][4], [43][5]]. However, an alternative druggable site within complex I is the flavin mononucleotide (FMN) prosthetic group, which is known as the N-site due to its NADH-oxidizing activity. The reactivity of the N-site FMN has been extensively studied [[44]6], and its potential role as a source of reactive oxygen species (ROS) in both normal physiological processes and disease has attracted considerable attention in recent years [[45]7,[46]8]. Nevertheless, due to the limited availability of N-site inhibitors, there are still crucial aspects of complex I FMN biology that remain unexplored. Flavin cofactors are essential partners in numerous biological redox reactions [[47]9], and compounds derived from riboflavin are present in all living organisms. Besides complex I, several other mitochondrial enzymes also rely on flavin cofactors for their activity [[48]10]. However, the specific enzymes and transporters responsible for the uptake and conversion of riboflavin into FMN and FAD within mitochondria are still a subject of debate [[49]11]. Additionally, despite significant progress in understanding the intricated assembly pathway of complex I, the mechanism by which FMN is inserted into the enzyme remains unknown [[50]12]. In yeast, even a single amino acid substitution within the FMN binding pocket of the NDUFV1 subunit of complex I is enough to cause the complete loss of complex I [[51]13]. In humans, the notion that riboflavin may play a role in regulating and enhancing complex I activity has led to the use of riboflavin supplementation as a treatment for complex I deficiency. This approach has shown varying degrees of symptom improvement in some patients [[52][14], [53][15], [54][16]]. Whether the stability of mammalian complex I is regulated by its flavin cofactor, as it is the case for many other flavoenzymes [[55][17], [56][18], [57][19]], is yet unexplored. To address these unresolved questions, we conducted a study focusing on the biology of complex I inhibition using diphenyleneiodonium (DPI), a well-known N-site inhibitor. DPI functions by covalently reacting with the reduced FMN cofactor, thereby irreversibly blocking electron input into the enzyme [[58]20]. Our investigation revealed that sequestration of FMN by DPI has a profound effect on the stability of complex I. In the absence of a functional FMN, complex I assembly is disrupted, and the individual subunits of complex I undergo degradation. 2. Materials and methods Animals. All animal procedures conformed to the EU Directive 86/609/EEC and Recommendation 2007/526/EC regarding the protection of animals used for experimental and other scientific purposes, enforced by Spanish law under Real Decreto 53/2013. The mice were housed in pathogen-free animal facility at CNIC, Madrid with an artificial 12 h light cycle; temperature was kept at 22 °C ± 2, relative humidity oscillated between 45 % and 65 %. Mice were fed standard chow diet (LASQCdiet® and from December 2021 D184 SAFE®). Cell lines: all cell lines were grown in complete DMEM medium (D5796, containing 4500 mg/L glucose, 2 mM l-glutamine) supplemented with 10 % fetal bovine serum (FBS, Sigma F7524), 1 % penicillin-streptomycin (PenStrep, Lonza) and 1 mM sodium pyruvate (Pyr, Sigma), at 37 °C in an atmosphere of 5 % CO[2]/95 % air ([59]Table 1). Mouse adult fibroblasts (MAFs) were isolated from mouse ear of at least 2 different female mice: mouse ears were cut, incubated with collagenase (1 %, 10 min at 37 °C) and seeded in 24-well plates with the same medium as stated before, supplemented with 1 μg/ml Amphotericin B. After the first three passages, cells were immortalized with a lentiviral vector expressing the SV40 large T antigen (pLOX-Ttag-iresTK, Addgene). Table 1. Cell lines used in this study. Cell line Comments FC57 MAFs L929 nuclear background and C57BL/6 mtDNA [[60]21] 143B Human osteosarcoma cell line [[61]22] Clpp −/− MAFs Gift from Alexandra Trifunovic lab [[62]23] Ndufs4 −/− MAFs Generated in this study 143B SLC25A32^KO Generated in this study [63]Open in a new tab Estimation of cellular proliferation: For growth curves the CYQUANT™ NF Cell Proliferation Assay Kit (ThemoFisher Scientific [64]C35006) was used following manufacturer instructions. Cells were grown in DMEM medium supplemented with 5 % dialyzed FBS, 1 mM Pyr and either glucose 5 mM or galactose 5 mM. 2000 cells were seeded in 96-well plates and were grown for 3 days. Each day, fresh medium with or without inhibitors were added. Cell number was estimated by fluorescence (Excitation: 488 nm; Emission: 535 nm) after incubation with CYQUANT™ 30 min at 37 °C in the dark. Cell proliferation was calculated as the ratio between (estimated) number of cells at day 72 h and at the moment of the beginning of the experiment. Riboflavin free medium: Riboflavin free medium was prepared from RPMI 1640 Medium w/o l-Alanine, l-Glutamine, Folic acid, Riboflavin Culture Media powder (MyBioSource). Following manufacturer instructions. Briefly, 10.092 g of powder were dissolved in 900 ml of ddH[2]O[2] by 3 h stirring at room temperature. Then, 2 mM l-glutamine, 10 % dialyzed fetal bovine serum (Sigma), 1 % penicillin-streptomycin (Sigma), 1 mM sodium pyruvate (Sigma), 8 mg l-Alanine (sigma), 2 mg Folate (Sigma) were added, the pH was adjusted to 7.1 with HCl, the volume was adjusted to 1 L with ddH[2]O[2] and the mix was kept stirring for additional 30 min at room temperature. After that, pH was re-checked and 2 g of sodium bicarbonate were added and stirring was kept for additional 5 min. Control cells were grown in the same media with the addition of 1 μM riboflavin. Without riboflavin, cell proliferation decreased gradually and stopped after ten days; after five days, passaging cells already become difficult. For these reasons, it was important to plate a sufficient number of cells before the withdrawal of the vitamin. Mitochondria isolation and functional assays. Isolation of mitochondria from cell cultures was performed from 4 to 10 150 mm plates, according to the differential centrifugation method. Briefly, cells were lysed by hypotonic shock with 7 vol of hypotonic medium (83 mM sucrose, 10 mM MOPS, pH 7.2) and homogenized in a Teflon potter-type tissue homogenizer. Then, 7 vol of hypertonic medium (sucrose 510 mM, MOPS 30 mM, pH 7.2) were added. The homogenate was centrifuged at 1000 g, 5 min, the supernatant obtained was centrifuged at 10,000 g, 10 min. The pellet obtained was resuspended in medium A (0.32 M sucrose, 1 mM EDTA, 10 mM Tris-HCl, pH 7.4) and stored at −80 °C. Isolation from mouse kidney was similar, medium A was used throughout all homogenization steps. In case mitochondria were used fresh for functional assays (ex. respirometry), medium A supplemented with bovine serum albumin 0.1 % was used throughout all homogenization steps. Spectrophotometric activities. Enzyme activities were calculated according to the Beer-Lambert law by kinetic analysis of the absorbance variation measured by a UV–visible spectrophotometer. All assays were performed at 32 °C in 1 ml cuvettes, using frozen-thawed mitochondria resuspended in activity buffers. Complex I activity was initiated by addition of 130 μM NADH and monitored at 340 nm (ε = 6.22 mM^−1cm^−1). Rotenone-insensitive rates were measured in parallel and subtracted from the measured rates. For NADH:O[2] activity, 10 μg of mitochondria were resuspended in sucrose buffer (250 mM, HEPES 2 mM, EGTA 0.1 mM) and pre-incubated for 10 min with the inhibitor or DMSO plus 130 μM NADH. For NADH:decylubiquinone activity 10 μg of mitochondria were resuspended in sucrose buffer plus 1 μg/ml antimycin A and 130 μM decylubiquinone. For complex II activity, 30 μg of mitochondria were resuspended in sucrose buffer plus 100 μM dichlorophenol-indophenol (DCPIP), 1 μg/ml antimycin A, 130 μM decylubiquinone and 1 μM rotenone. The reaction was initiated by addition of 10 mM succinate and monitored at 600 nm (ε = 19.2 mM^−1cm^−1). For citrate synthase activity, 5 μg of mitochondria were resuspended in 10 mM Tris-HCl pH 8, 0,1 % Triton X-100, 46 μg/ml acetyl-CoA, 200 μM 5,5-Dithio-bis-2-nitrobenzoic acid (DTNB). The reaction was initiated by addition of 0.5 mM oxaloacetate and monitored at 412 nm (ε = 13.6 mM^−1cm^−1). Respirometry. Oxygen consumption was assayed using a Clark type polarographic oxygen sensor (Oroboros instruments) in a 2 ml isolated chamber, in agitation with a magnetic stirrer, at 37 °C, in Mir 05 respiration medium (Oroboros instruments). Either 50 μg of mitochondria or 2 million of cells permeabilized with 20 μg of digitonin were sequentially incubated with the indicated substrates and inhibitors. Respiration following antimycin A incubation was set as baseline; complex IV respiration was calculated as ascorbate/TMPD minus azide oxygen consumption rate. Flavin fluorescence assay. Flavin fluorescence in mitochondria and in cells was measured following the protocol of [[65]24]. Briefly, either 2 million of cells or 50 μg of mitochondria were resuspended in Tris-HCl 10 mM pH = 7.5 and incubated on ice for 10 min to allow lysis by osmotic shock. Flavin fluorescence was measured at 435/550 nm (em/ex) in a plate fluorescence reader. A regression line was built in parallel using known concentration of FAD. The flavin independent background fluorescence was subtracted by adding in each well a reducing solution of sodium dithionite, because reduced flavin have no fluorescent emission. Flavin fluorescence in CN gels were revealed with an iBright scan (Invitrogen) after 2 min incubation of the gel in a solution with 2 % SDS, 1 % β-mercaptoethanol. During the preparation of this manuscript a similar protocol was published elsewhere [[66]25]. Quantitative analysis of flavins. Mitochondrial and cytosolic samples purified from cells DPI-treated and untreated were boiled for 5 min, and then precipitated proteins and cellular remains were removed by centrifugation. Supernatant was then analyzed using an Alliance HPLC system (Waters) equipped with a 2707 autosampler and a HSST3 column (4.6 × 50 mm, 3.5 mm; Waters) preceded by a pre-column of the same material (4.6 × 20 mm, 3.5 mm; Waters). An aliquot of 50 μl of the solution was applied and the chromatography was developed at 1 ml/min with a 6 min isocratic run of methanol 40 % (vol/vol) in 5 mM ammonium acetate pH 6.0 [[67]26]. RF, FMN and FAD standard curves acquired under the same conditions were used to quantify the flavin content present at each assayed condition. Mitochondria solubilization. Solubilization of mitochondrial membranes was carried out with digitonin, in order to visualize both respiratory SCs and free form complexes. The ratio of grams of detergent:grams of mitochondrial protein used was 4:1.100 μg of purified mitochondria were incubated with digitonin for 5 min in 50 mM NaCl buffer, 50 mM imidazole, 5 mM aminocaproic acid at a concentration of 10 μg/μl. The insoluble fraction was removed by centrifugation at 13,000g in microfuge for 30 min at 4 °C. The pellet obtained was discarded and the supernatant was mixed with 4× loading buffer (5 % Coomassie Blue-G250 in 1 M aminocaproic acid) for gel loading. Preparation of polyacrylamide gels. Gradient polyacrylamide gels were prepared in house with a gradient former, composed of two chambers, where two solutions of different percentage of acrylamide were added. Mini Protean III system (Biorad) 1.5 mm was used normally, for mass spectrometry experiments, larger gels (23 × 16.5 cm) were used. The gradient used was different depending on the kind of high molecular weight SCs that wanted to be visualized, in general, for the resolution of SCs and complexes (DIG) gels were 5–13 % [[68]27]. CN gels were essentially the same, but 0.01 % of digitonin was added to all gel solutions and Ponceau red buffer (Ponceau red, glycerol) was used instead of the normal loading buffer. BN-PAGE and CN-PAGE. The amount of sample loaded in each well was that obtained from solubilization of 100 μg of mitochondria, except in the case of samples intended for analysis by mass spectrometry, for which samples were of 300 μg. Cathode buffer A (tricine 50 mM, bis-tris 15 mM, pH 7.0, Coomassie Blue G-250 0.02 %), cathode buffer B (tricine 50 mM, bis-tris 15 mM, pH 7.0, Coomassie Blue G-250 0.002 %) and anode buffer (bis-tris 50 mM, pH 7.0) were used for electrophoresis. Electrophoresis was performed in cold chamber. The run was developed for half hour at 90 V with cathode buffer A. Then, the cathode buffer was exchanged for cathode buffer B and electrophoresis continued for approximately one more hour at 300 V. In CN gels the whole electrophoresis was performed in cathode buffer B, with no changes. From the electrophoresis it was possible to reveal the activity of respiratory complexes in gel, to perform western blotting with antibodies against proteins of the respiratory complexes, to stain the gel to later analyze the gel composition by mass spectrometry, or reveal flavin fluorescence. SDS-PAGE. Electrophoresis for protein separation was performed in denaturing polyacrylamide gels. The protein samples were incubated for 1 min at 95 °C with loading buffer (Tris-HCl 50 mM pH 6.8, 2 % SDS, 10 % glycerol, 1 % β-mercaptoethanol, 0.02 % bromophenol blue). Subsequently, 30 μg of each protein sample was loaded on a 12.5 % acrylamide gel and electrophoresis was developed in Tris-glycine solution at 10 mA per gel during migration on the stacking gel and at 20 mA per gel once the sample passed to the resolving gel. Immunoblotting. Immunodetection by Western blot was performed on either type of electrophoresis previously described. Briefly, proteins were transferred to PVDF membrane (Immobilon-FL, 0.45 μm) by transfer in Bio Rad Mini Trans-Blot Cell or Trans-Blot Cell systems, in 48 mM Tris, 39 mM glycine, 20 % methanol transfer solution, 1 h at 100 V or o/n at 30 V. Once the membrane was obtained, it was blocked in 0.1 % PBS-tween, 5 % BSA solution for 1 h and incubated with primary antibody overnight in agitation at 4 °C ([69]Table 2). After three washes with 0.1 % PBS-tween, it was incubated with secondary antibody for 1 h, after which another three washes were performed before development. Membrane revealing was performed by using fluorescent secondary antibody and revealed with the Odyssey imaging system (LI-COR biosciences). Table 2. Antibodies used in this study. Target Origin/conjugate Clonality Company ACAD9 Mouse Poli- Abcam Actin Rabbit Poli- Sigma CS Rabbit Mono- Abcam COX1 Mouse Mono- Invitrogen NDUFA9 Mouse Mono- Abcam NDUFS1 Rabbit Poli- Abcam NDUFS2 Rabbit Poli- Abcam NDUFS3 Mouse Mono- Abcam NDUFS4 Mouse Mono- Abcam NDUFS5 Rabbit Poli- Proteintech NDUFV1 Rabbit Mono- Sigma NDUFV2 Rabbit Mono- Abcam SDHA Mouse Mono- ThermoFisher TOM20 Rb Poli- Santa Cruz UQCRC2 Mouse Mono- Abcam UQCRFS1 Mouse Mono- Abcam VDAC Mouse Mono- Abcam Anti-Mouse DyLight 800™ Poli- Rockland Anti-Rabbit DyLight 800™ Poli- Rockland Anti-Mouse AlexaFluor 680 Poli- LIfe technologies Anti-Rabbit AlexaFluor 680 Poli- LIfe technologies [70]Open in a new tab Complex I In-gel activity. Measurement of NADH dehydrogenase activity of complex I was determined on the same gel after BN-PAGE electrophoresis. The gel was incubated in 0.1 M Tris-HCl, pH 7.4, 0.14 mM NADH and 1 mg/ml NitroBlue tetrazolium solution at room temperature. Visualization was achieved by the purple precipitate produced after reduction of NitroBlue tetrazolium by the NADH dehydrogenase activity of CI. CRISPR/Cas9 genome editing. KO cell lines were generated by CRISPR/Cas9 mediated genome editing with the TLCV2 viral vector (Addgene) [[71]28]. A two-target sequence strategy was employed. Target sequences were generated using the CHOP/CHOP tool [[72]29]. The two-best gRNA, in terms of low off-target and high on-target activity, were chosen against the first exon of the SLC25A32 gene ([73]Table 3). The sequences were cloned in two TLCV2 vectors and used to transduce cells. A non-targeting gRNA was used to generate a control cell line. Lentivirus production was carried out by the Viral Vector unit at CNIC. After two days of transduction, puromycin 1 μg/ml was added to select transduced cells. After one day of selection doxycycline was added to induce the expression of Cas9. The following day GFP positive cells were seeded in a 96 wells plate containing conditioned DMEM of the parental line, one cell per well. For the following weeks the growth of clones was monitored. Clone screening was performed by PCR, using primers flanking the PAM sequences targeted by the two sgRNAs. The clones which presented abnormal PCR amplicons were grown. Since no good anti SLC25A32 antibody was found, the knockouts were validated by qPCR. Table 3. Primers used in this study. Name Primer hSLC25A32_g1 BsbI Fwd CACCGAATAGGGGTCGCAAGCACGGGGG hSLC25A32_g1 BsbI Rev AAACCCCCCGTGCTTGCGACCCCTATTC hSLC25A32_g2 BsbI Fwd CACCGTGGGAACACGTTTATTCCAGAGG hSLC25A32_g2 BsbI Rev AAACCCTCTGGAATAAACGTGTTCCCAC hSLC25A32_ex1 Fwd GCATAAGAGTCCTCTCGTTGGT hSLC25A32_ex1 Rev AAAAGACGGAGGAGATCCAGTT [74]Open in a new tab Proteomics analysis. High-Throughput protein profiling was performed in collaboration with the CNIC-proteomics Unit. Total cellular pellet of 143 B cells was digested with trypsin and labelled with tandem mass tags TMT (10 plex), according to manufacturer's instructions. Resulting tryptic peptides were injected onto a C-18 reversed phase (RP) nano-column (75 μm I.D. and 50 cm, Acclaim PepMap, Thermo Fisher, San José, CA, USA), connected to a nEasy LC-1000 chromatography system (Thermo Fisher, San José, CA, USA) and analyzed in a continuous acetonitrile gradient consisting of 8–31 % B-solution (B = 0.5 % formic acid in acetonitrile) for 240 min, 50–90 % B for 1 min. Subsequently, peptides were eluted from the RP nanocolumn at a flow rate of ∼200 nL/min to an emitter nanospray needle for real-time ionization and MS analysis in a Q-Exactive HF mass spectrometer (Thermo Fisher). Mass spectra were acquired in a data-dependent manner, with an automatic switch between MS and MS/MS using a top 20 method. The raw files were analyzed with Proteome Discoverer (version 2.1, Thermo Fisher Scientific), using Uniprot database. The parameters selected for database searching were: trypsin digestion with 2 maximum missed cleavage sites, precursor mass tolerance of 800 ppm, fragment mass tolerance of 0.03 atomic mass unit (amu). MS/MS spectra were also queried against inverted databases constructed from the same target databases. Peptide identification from MS/MS data was performed using the probability ratio method [[75]30]. False discovery rates (FDR) of 1 % was used as a threshold for peptide identification. Quantitative information were extracted from TMT reporters ions in the MS/MS spectra, and protein abundance changes were analyzed using the WSPP model [[76]31] by applying the Generic Integration Algorithm [[77]32]. Transcriptomic analysis. RNA sequencing was performed by the team of CNIC Genomics Unit. Downstream analysis was performed by CNIC Bioinformatics Unit. RNA libraries were produced using TrueSeq RNASeq kit from Illumina and sequenced in the HiSeq 2500 Illumina Sequencer. Adaptors from RNAseq were removed from reads using the trimgalore v-0.6.6 and cutadapt v-1.18 software. Then, trimmed reads were mapped and quantified on the transcriptome GRCm38 gene-build 91 using RSEM v-1.3.1 [[78]33]. Data were then normalized by using trimmed mean of M-values method [[79]34] and differential expression analyzed using the function Voom from the bioconductor package Limma v-3.50.3 [[80]35]. False discovery rate was corrected using Benjamini–Hochberg method, considering differentially expressed genes for an adjusted p-value ≤0.05. Pathway enrichment analysis was performed GSEA software [[81]36,[82]37] through REACTOME pathways database [[83]38] using log[2]FC values as the gene ranks. Transcription factor activity was inferred by virtual inference of protein activity by enriched regulon analysis (VIPER) [[84]39] using DoRothEA regulons [[85]40]. We filter for regulons with level of confidence (“A” and “B”) and we filtered out genes with an adjusted p value of the log[2]FC greater than 0.0001. Generation of heatmaps and Figures was done in R using Complex Heatmap package [[86]41]. 3. Results 3.1. Diphenyleneiodonium induces respiratory complex I degradation In isolated mitochondria, the NADH:ubiquinone oxidoreductase activity of respiratory complex I is sensitive to the N-site inhibitor diphenyleneiodonium (DPI) [[87]20]. Compared with piericidin A and rotenone, two commonly used Q-site inhibitors, DPI is at least one hundred times less potent ([88]Fig. 1A) [[89]42]. However, DPI treatment completely suppresses cell division of immortalized mouse adult fibroblasts (MAFs) in galactose medium, showing that it interferes with cellular respiration as rotenone does ([90]Fig. 1B). We have previously demonstrated that deficiency of complex III leads to the degradation of complex I, and that the use of rotenone effectively inhibits this degradation process [[91]43]. In this study, we aimed to investigate whether the site of complex I inhibition has any effect on the structural organization of complex I within the mitochondrial electron transport chain. Fig. 1. [92]Fig. 1 [93]Open in a new tab Diphenyleneiodonium induces respiratory complex I degradation. (A) Inhibition of complex I NADH:O[2] activity of frozen-thawed mouse kidney mitochondria by rotenone, piericidin A and DPI. Samples were pre-incubated for 10′ in the presence of NADH and the indicated inhibitor, n = 3 replicates per point. (B) Effect of 72 h DPI and rotenone treatment on MAFs cell proliferation in glucose or galactose media. Data are mean ± S.E.M., ordinary two-way ANOVA, Dunnett's multiple comparisons test. (C, D) Immunoblot analysis of the effect of 72 h, 1 μM DPI, rotenone or piericidin A treatment on the supramolecular organization of complex I (C), complex III and complex IV (D), Veh = DMSO. (E) Immunoblot analysis of the effect of 72 h, 400 nM DPI and rotenone on complex I and complex II subunits abundance relative to citrate synthase. The quantification is displayed in [94]Fig. S1A. (F) Immunoblot analysis of the short-term effect of 400 nM DPI treatment on the supramolecular organization of complex I. “CI#” and “CI + III[2]#” refer to complex I and supercomplex CI + III[2]sub-complex, respectively. (G) In-gel activity analysis of the short-term effect of DPI on complex I. (H–J) In-gel activity (H) and immunoblot analysis (I and J) of the effect of 2 μg/g DPI i. p. Injection on kidney mitochondria complex I. To explore this, we conducted experiments using mitochondria isolated from MAFs cultivated in glucose medium and treated with different complex I inhibitors (DPI, rotenone, and piericidin) for a duration of 72 h. Isolated mitochondria were then solubilized with digitonin and analyzed by blue-native gel electrophoresis (BN-page). Surprisingly, we observed that treatment with DPI resulted in a complete and selective loss of respiratory complex I and its associated supercomplexes, I + III[2] and I + III[2]+IV ([95]Fig. 1C), with no effect on other respiratory complexes ([96]Fig. 1D). In contrast, the Q-site inhibitors rotenone and piericidin A did not exhibit any noticeable changes in the BN-page profile of the mitochondria. To further validate these findings, we performed a similar experiment but employed denaturing conditions using SDS-page. Consistent with our previous results, MAFs treated with DPI showed a decrease in the abundance of complex I subunits, whereas rotenone treatment did not ([97]Fig. 1E and [98]Fig. S1A). This result highlights a unique feature of the N-site inhibitor DPI, which piqued our interest and prompted us to investigate deeper its mechanism of action. In subsequent experiments, we observed a correlation between the loss of complex I and the appearance of faster migrating bands below complex I and complex I-containing supercomplexes on the BN gel (marked as “CI#” and “CI + CIII[2]#“, [99]Fig. 1F). The accumulation of the “sub-complexes” was time-and dose-dependent: it started after 45 min, it peaked after 4–6 h, and by 72 h all holo- and sub-complexes had disappeared ([100]Fig. 1F, [101]Fig. S1, B and C; for 72 h see [102]Fig. 1, C and F). On the other hand, inhibition of in-gel NADH:NBT oxidoreductase activity occurred immediately without any lag time ([103]Fig. 1G). To validate the impact of DPI on complex I beyond cell culture, we administered DPI intraperitoneally (i.p.) to wild-type CD1 mice. DPI treatment resulted in the inhibition of complex I and the accumulation of complex I sub-complexes in vivo, as evidenced by [104]Fig. 1H and I. At the 24-h time point, the abundance of CIII[2] and CIII[2]–CIV complexes appeared to increase, likely due to their detachment from complex I-containing supercomplexes subsequent to DPI treatment ([105]Fig. 1J). 3.2. Molecular characterization of complex I “sub-complexes” To gain insights into the nature of the observed “sub-complexes,” we employed two complementary approaches. Firstly, we resolved respiratory complexes by two-dimensional BN/SDS-page. Subsequent immunoblot analysis showed the absence of NDUFS1, NDUFV1, NDUFV2 and NDUFS4 subunits from the bands corresponding to the “sub-complexes” ([106]Fig. 2A and B). These subunits belong to the N-module, which is the NADH-oxidizing region of complex I where the FMN cofactor binds. Importantly, DPI acts specifically at this site. We further confirmed the absence of the N-module from all subcomplexes by simultaneous incubation with antibodies against N-module subunit NDUFS1 and Q-module subunit NDUFA9 in one-dimensional BN gels ([107]Fig. 2C). Fig. 2. [108]Fig. 2 [109]Open in a new tab Molecular characterization of complex I sub-complexes. (A–B) Subunit composition of complex I sub-complexes revealed by immunoblot analysis of two-dimensional BN/SDS polyacrylamide gel (A) and localization of the subunits in the structure of complex I (B), (PDB accession number: [110]5LDW, N-module subunits are in red). In (A), MAFs were treated with 400 nM DPI for 24 h. (C) Immunoblot analysis of the subunit composition of complex I sub-complexes revealed by co-incubation of N-module and Q-module specific antibodies. Mitochondria isolated from MAFs treated with 400 nM DPI for 24 h. (D) Structural complex I models summarizing the results from LCMS analysis of in-gel digested BN-PAGE CI and CI# bands. The two eluted bands are the same as CI and CI# bands from [111]Fig. 2C. The color code for the modules was adopted from Ref. [[112]12]. Pink, NDUFAB1; shady gray, subunit not found in either band. (E) Immunoblot analysis of the accumulation of NDUFAF2 assembly factor upon DPI treatment. Mitochondria isolated from MAFs treated with 400 nM DPI for 24 h (F) Comparison between the short-term and long-term effect of 400 nM DPI treatment on complex I. (G–H) Immunoblot analysis of the effect of short-term and long-term 400 nM DPI treatment on N-module and Q-module subunits abundance (G) and quantification relative to β-actin (H). Data represented as mean ± S.E.M, ordinary two-way ANOVA, Dunnett's multiple comparisons test. (For interpretation of the references to color in this figure legend, the reader is referred to