APOPT1/COA8 assists COX assembly and is oppositely regulated by UPS and ROS

Abstract

Loss-of-function mutations in APOPT1, a gene exclusively found in higher eukaryotes, cause a characteristic type of cavitating leukoencephalopathy associated with mitochondrial cytochrome c oxidase (COX) deficiency. Although the genetic association of APOPT1 pathogenic variants with isolated COX defects is now clear, the biochemical link between APOPT1 function and COX has remained elusive. We investigated the molecular role of APOPT1 using different approaches. First, we generated an Apopt1 knockout mouse model which shows impaired motor skills, e.g., decreased motor coordination and endurance, associated with reduced COX activity and levels in multiple tissues. In addition, by achieving stable expression of wild-type APOPT1 in control and patient-derived cultured cells we ruled out a role of this protein in apoptosis and established instead that this protein is necessary for proper COX assembly and function. On the other hand, APOPT1 steady-state levels were shown to be controlled by the ubiquitination-proteasome system (UPS). Conversely, in conditions of increased oxidative stress, APOPT1 is stabilized, increasing its mature intramitochondrial form and thereby protecting COX from oxidatively induced degradation.

Keywords: APOPT1‐COA8; cytochrome c oxidase; mitochondrial encephalopathy; proteasome–ubiquitin system; reactive oxygen species.

Figure 1. Generation of the Apopt1 knockout mouse model

1. A

   Strategy employed for the targeted disruption of mouse chromosome 12 Apopt1 coding exon 2 by CRISPR/Cas9. The mutation selected for the characterization of the mouse model produced an indel (c.188delAinsTG) resulting in a frameshift and truncated protein (p.Asp55Valfs*20).

2. B

   Chromatograms generated by Sanger sequencing of Apopt1 ^−/− (KO) and Apopt1 ^+/+ (WT) genomic DNA highlighting the mutated position in comparison with the wild‐type sequence.

3. C

   Relative mRNA expression of Apopt1 normalized to the expression of GAPDH in skeletal muscle and liver of wild‐type, heterozygous, and knockout mice of 3 months of age. Data are presented as mean ± SEM (n = 6 mice per genotype). The asterisks represent the significance levels calculated by two‐way ANOVA with Sidak's multiple comparisons test: muscle—***P = 0.0001 (WT vs. KO), *P = 0.0310 (WT vs. het), *P = 0.0395 (het vs. KO), liver—**P = 0.0022 (WT vs. KO), *P = 0.0101 (WT vs. het).

Figure 2. Clinical phenotype characterization of the Apopt1 knockout mice

1. A

   Body weight measured in female and male animals at 3, 6, and 12 months of age. Data are presented as mean ± SEM (n = 5 per group).

2. B

   Distance run by the tested female and male mice on the treadmill at 3 and 12 months of age. Data are presented as mean ± SEM (n = 6 per group). The asterisks represent the significance levels calculated by two‐way ANOVA with Sidak's multiple comparisons test: females—****P < 0.0001 (3 months), ***P = 0.0004 (12 months), males—****P < 0.0001 (3 months), **P = 0.0019 (12 months).

3. C

   Time in seconds spent by the female and male mice on the Rotarod cylinders before falling at different ages. Data are presented as mean ± SEM (n = 5 per group). The asterisks represent the significance levels calculated by two‐way ANOVA with Sidak's multiple comparisons test: females—**P = 0.0059 (3 months), **P = 0.0083 (6 months), *P = 0.0391 (12 months), males—**P = 0.0069 (12 months).

4. D

   Number of entries in each arm of the Y‐maze performed at different ages. Data are presented as mean ± SEM (n = 10 per group). The asterisks represent the significance levels calculated by two‐way ANOVA with Sidak's multiple comparisons test: ****P < 0.0001.

5. E

   Total spontaneous horizontal and vertical movements of 12‐month‐old mice measured in an activity cage. Data are presented as mean ± SEM (n = 6 per group). The asterisks represent the significance levels calculated by two‐way ANOVA with Sidak's multiple comparisons test: ****P < 0.0001 (horizontal).

Figure 3. Morphological, biochemical, and structural analysis in mouse tissues

1. A

   Representative hematoxylin and eosin staining of skeletal muscle in 3‐month‐old individuals.

2. B

   Histochemical reaction specific to COX and SDH in skeletal muscle, cerebellar cortex, and kidney of the same individuals.

3. C

   COX (CIV) enzymatic activity normalized to the activity of citrate synthase (CS) in five animals per genotype as indicated at 3 and 12 months of age. Data are presented as mean ± SEM (n = 6 per group). The asterisks represent the significance levels calculated by two‐way ANOVA with Sidak's multiple comparisons test: Three months: Muscle (SM)—***P = 0.0001 (WT vs. KO), kidney (K)—***P = 0.0002 (WT vs. KO), heart (H)—***P = 0.0001 (WT vs. KO), brain (B) and liver (L)—****P < 0.0001 (WT vs. KO). Twelve months: SM and L—****P < 0.0001, B—***P = 0.0008 (WT vs. KO).

4. D

   Western blot analysis of SDS–PAGE of total lysates from liver, skeletal muscle, and brain from the indicated genotypes, each lane showing the results for one animal. The graph shows the densitometric quantification of the signals obtained in the liver. Data are presented as mean ± SEM (n = 2 WT; n = 2 het; n = 3 KO). The asterisks represent the significance levels calculated by two‐way ANOVA with Sidak's multiple comparisons: ***P = 0.0002 and ****P < 0.0001 (WT vs. KO).

5. E

   Western blot analysis of 1D‐BNGE of mitochondria from skeletal muscle from the indicated genotypes, each lane showing the results from one animal.

6. F

   Western blot analysis of 2D‐BNGE of mitochondria from skeletal muscle from one individual of the indicated genotype. Red arrows indicate the accumulation of COX assembly intermediates in Apopt1 ^−/− samples.

Source data are available online for this figure.

Figure 4. Mitochondrial sub‐localization of APOPT1

1. A

   Western blot of SDS–PAGE of different fractions from 143B cells expressing APOPT1^HA. Tot: total lysate. Cyto: post‐mitochondrial fraction (cytoplasm). Mt: isolated mitochondria. Mt sol: soluble mitochondrial fraction. Mt memb: mitochondrial membranes. ${\text{CO}}_3^{2-}$ pellet: pellet after carbonate extraction with 0.1 M Na₂CO₃, pH 10.5 for 30 min. ${\text{CO}}_3^{2-}$ sol: soluble fraction after the carbonate extraction.

2. B

   Western blot of SDS–PAGE of mitochondria used for protease protection assay after digitonin treatment. The experiment was carried out in isolated mitochondria from 143B cells expressing APOPT1^HA exposed to increasing amounts of digitonin (expressed in μg) and 50 μg/ml trypsin. Complete solubilization with 1% Triton X‐100 was used as a control of protease sensitivity. TOM20: Translocase of the outer membrane (OM) 20 kDa. ACO2: aconitase 2 (mitochondrial isoform). AIF: apoptosis‐inducing factor. AK2: adenylate kinase 2. OM: outer mitochondrial membrane. IM: inner mitochondrial membrane. IMS: intermembrane space.

3. C

   Western blot of SDS–PAGE of mitochondria used for protease protection assay after hypotonic shock. The experiment was carried out in isolated mitochondria from 143B cells expressing APOPT1^HA incubated either in isotonic (Iso) or in hypotonic buffers for 5 min (Hypo 5′) or 15 min (Hypo 15′) and 50 μg/ml trypsin. Complete solubilization with 1% Triton X‐100 was used as a control of protease sensitivity.

4. D

   N‐SIM super‐resolution micrographs showing 0.8 μm Maximum Intensity Projection (0.15 μm for each Z‐stack) of 143B cells expressing APOPT1^GFP shown in green, specific markers for each mitochondrial compartment: TOMM20 (OM), COX8A (IM), MitoTracker (IM + M), and matrix‐target (mScarlet) shown in red. Scale bar: 5 μm. The graph shows Pearson's coefficient in co‐localized volume of different sub‐compartment combinations with APOPT1^GFP. Data show mean ± SEM (n = 5). One‐way ANOVA (*P = 0.0298 and **P = 0.0076) was employed for the statistical analysis.

Source data are available online for this figure.

Figure 5. Defective COX assembly and accumulation of assembly intermediates in APOPT1‐less human cells

1. A

   APOPT1^HA was expressed in S6 and S2 immortalized fibroblasts as shown by the Western blot immunovisualization. The expression levels were tested at different days after transduction. The graph on the right shows the activities of CIV normalized to CS in three biological replicates per cell line. Data are presented as mean ± SEM (n = 3). The asterisks represent the significance levels calculated by two‐way ANOVA with Tukey's multiple comparisons test: **P = 0.0030 (S6 naïve vs. S6 APOPT1^HA), *P = 0.0165 (S6 EV vs. S6 APOPT1^HA).

2. B

   APOPT1^GFP was expressed in S6 and S2 immortalized fibroblasts as shown by the Western blot immunovisualization. The graph on the right shows the activities of CIV normalized to CS in four biological replicates per cell line. Data are presented as mean ± SEM (n = 4). The asterisks represent the significance levels calculated by two‐way ANOVA with Tukey's multiple comparisons test: **P = 0.0066 (S6 naïve vs. S6 APOPT1^GFP), ***P = 0.0006 (S6 GFP vs. S6 APOPT1^GFP), *P = 0.0169 (S2 naïve vs. S2 APOPT1^GFP), ***P = 0.0008 (S2 GFP vs. S2 APOPT1^GFP).

3. C

   Western blot analysis of 1D‐BNGE of mitoplasts extracted from the indicated cell lines. Fully assembled COX migrating at 200 kDa was immunovisualized with antibodies recognizing MT‐CO1 and MT‐CO2. The presence of the assembly intermediates “MITRAC” and “S3” was also detected (arrows).

4. D

   Western blot analysis of 2D‐BNGE of mitoplasts extracted from the indicated cell lines. Fully assembled COX was immunovisualized with antibodies recognizing MT‐CO1 and MT‐CO2. The presence of the assembly intermediates “MITRAC” and “S3” was also detected (arrows). Anti‐GFP immunodetection revealed the presence of the low molecular weight APOPT1^GFP complex in the complemented cells expressing the tagged protein.

5. E

   L‐[³⁵S]‐Methionine pulse‐chase labeling of mtDNA‐encoded proteins. After a 2‐h exposure with the radioactive label (pulse), cells were cultured in cold medium for the indicated chase times. The graphs show the densitometric quantification of the bands corresponding to MT‐CO1 (left graph) and MT‐CO2 + MT‐CO3 (right graph) normalized to the ATP6 band over the indicated time points. Graphs represent the values of three biological replicas for each cell line. Data are presented as mean ± SEM (n = 4 for controls, n = 2 for S2/S6/S2 APOPT1^GFP/S6 APOPT1^GFP). The symbols represent the significance levels calculated by two‐way ANOVA with Tukey's multiple comparisons test: MT‐CO1—3.5 h: *P = 0.0202 (controls vs. S2), ^†† P = 0.0039 (controls vs. S6); 6.5 h: *P = 0.0118 (controls vs. S2), ^†† P = 0.0011 (controls vs. S6), ^§ P = 0.0163 (S6 vs. S6 APOPT1^GFP); 20 h: ***P = 0.0002 (controls vs. S2), ^†††† P < 0.0001 (controls vs. S6), ^# P = 0.0343 (S2 vs. S2 APOPT1^GFP), ^§§§ P = 0.0007 (S6 vs. S6 APOPT1^GFP). MT‐CO2/MT‐CO3—3.5 h: ^††† P = 0.0003 (controls vs. S6), ^§ P = 0.0194 (S6 vs. S6 APOPT1^GFP); 6.5 h: ^§ P = 0.0375 (S6 vs. S6 APOPT1^GFP); 20 h: **P = 0.0013 (controls vs. S2), ^†††† P < 0.0001 (controls vs. S6), ^# P = 0.0234 (S2 vs. S2 APOPT1^GFP), ^§§§§ P < 0.0001 (S6 vs. S6 APOPT1^GFP).

Data information: C1 and C2: control immortalized fibroblasts. EV: Cells transduced with the empty vector. Naïve: Non‐transduced immortalized fibroblasts. GFP: Cells transduced with a recombinant GFP expression vector. Tubulin, SDHB, and SDHA were used as loading controls in the different experiments as indicated.Source data are available online for this figure.

Figure 6. Effects of proteasome inhibition and ROS overproduction on APOPT1 stability

1. A

   Western blot analysis of SDS–PAGE of MGM132‐treated 143B APOPT1^HA cells. The graph represents the densitometric quantification of the signals for the precursor and mature protein.

2. B

   The upper part of the panel (Input) shows Western blot analyses of APOPT1^HA in the MGM132‐treated cells. Note the appearance of higher molecular weight bands upon longer exposure in the samples treated with the proteasome inhibitor. The bottom part of the panel (Purified fractions) shows the analysis of fractions from the same cells immunoprecipitated with anti‐HA. Note that the higher molecular weight species are cross‐reacting with both anti‐HA and anti‐ubiquitin.

3. C

   Western blot analysis of SDS–PAGE of total lysates from 143B cells overexpressing tagged APOPT1 (as indicated) and exposed to 100 μM H₂O₂, as illustrated by the scheme (H₂O₂ treatment), for the indicated times. The graphs represent the densitometric quantification of the tagged APOPT1 signal at each time point. The graph inset shows that the increase of APOPT1 occurs in the first minutes after the exposure to H₂O₂.

4. D

   Western blot analysis of SDS–PAGE of total lysates from 143B cells overexpressing tagged APOPT1 (as indicated) and exposed to 5 μM MitoParaquat (MitoPQ), as illustrated by the scheme (MitoPQ treatment), for the indicated times. The graphs represent the densitometric quantification of the tagged APOPT1 signal at each time point. The graph inset shows that the increase of APOPT1 occurs in the first minutes after the exposure to MitoPQ.

Source data are available online for this figure.

Figure 7. Effects of MitoPQ treatment in the absence or presence of APOPT1

1. A

   Western blot analysis of SDS–PAGE of different mitochondrial proteins in total lysates from the indicated cell lines treated with 5 μM MitoPQ at the indicated times. UT: untreated cells.

2. B

   Densitometric quantification of APOPT1^GFP signal during the treatment in two biological replicas. Data are presented as mean ± SEM (n = 2).

3. C

   Densitometric quantification of MT‐CO1 signal in the non‐complemented APOPT1‐less cells (S6 GFP) vs. the complemented cells (S6 APOPT1^GFP). Three biological replicas and two technical replicates were carried out for each cell line. The signals in UT S6 APOPT1 were considered 100%. Data are presented as mean ± SEM (n = 3). The levels of MT‐CO1 were significantly lower in the S6 GFP patient cells after 20 h of MitoPQ treatment compared to the untreated cells (*P = 0.0202, two‐tailed unpaired Student's t‐test).

4. D

   Densitometric quantification of MT‐CO2 signal in the non‐complemented APOPT1‐less cells (S6 GFP) vs. the complemented cells (S6 APOPT1^GFP). Three biological replicas and two technical replicates were carried out for each cell line. The signals in UT S6 APOPT1 were considered 100%. Data are presented as mean ± SEM (n = 3).

Source data are available online for this figure.