Scaf1 promotes respiratory supercomplexes and metabolic efficiency in zebrafish

Abstract

The oxidative phosphorylation (OXPHOS) system is a dynamic system in which the respiratory complexes coexist with super-assembled quaternary structures called supercomplexes (SCs). The physiological role of SCs is still disputed. Here, we used zebrafish to study the relevance of respiratory SCs. We combined immunodetection analysis and deep data-independent proteomics to characterize these structures and found similar SCs to those described in mice, as well as novel SCs including III₂ + IV₂ , I + IV, and I + III₂ + IV₂ . To study the physiological role of SCs, we generated two null allele zebrafish lines for supercomplex assembly factor 1 (scaf1). scaf1^-/- fish displayed altered OXPHOS activity due to the disrupted interaction of complexes III and IV. scaf1^-/- fish were smaller in size and showed abnormal fat deposition and decreased female fertility. These physiological phenotypes were rescued by doubling the food supply, which correlated with improved bioenergetics and alterations in the metabolic gene expression program. These results reveal that SC assembly by Scaf1 modulates OXPHOS efficiency and allows the optimization of metabolic resources.

Keywords: OXPHOS super-assembly; SCAF1/COX7A2L; metabolism; mitochondria; zebrafish.

Figure 1. OXPHOS super‐assembly in zebrafish

1. A–D

   Blue native gel electrophoresis (BNGE) of mouse (M) C57BL/6J (111), CD1 (113), and zebrafish (ZF) skeletal muscle digitonin‐solubilized mitochondria. (A, B) Immunodetection of the indicated proteins after BNGE, (C) in‐gel activity for CI and (D) CIV (shown is a representative gel from two technical and two biological replicates).

2. E–H

   BNGE of whole‐body zebrafish digitonin‐solubilized mitochondria of scaf1 ^Δ1/Δ1 (−/−) and its respective scaf1 ^+/+ counterpart. (E, F) Immunodetection of the indicated proteins, (G) in‐gel activity of CI and (H) CIV (representative gel from two technical and three biological replicates).

3. I

   2D BNGE/SDS electrophoresis: 1^st dimension with digitonin (Dig) and 2^nd dimension with SDS, followed by immunoblotting with the indicated antibodies to identify the proteins detected by the commercial anti‐SCAF1 antibody. Asterisks indicate missing bands in scaf1 ^Δ1/Δ1.

Figure EV1. Characterization of OXPHOS super‐assembly in zebrafish

1. A–C

   Split channels of BNGE of mouse (M) C57BL/6J (111), CD1(113), and zebrafish (ZF) muscle mitochondria shown in Fig 1A and B.

2. D, E

   Immunodetection of the indicated proteins after 2D BNGE/DDM electrophoresis of whole‐body zebrafish mitochondria (representative of n = 3). Merged (D) and (E) split channels.

Figure EV2. OXPHOS super‐assembly in whole zebrafish and mouse liver in homeostasis and low‐protein/low‐fat diet

1. A–E

   BNGE of mouse (M) C57BL/6J (111) and CD1(113) liver mitochondria and zebrafish (ZF) whole‐body mitochondria, digitonin‐solubilized. (A–C) Immunodetection of the indicated proteins, (D) CI and (E) CIV in‐gel activity (representative of two technical and three biological replicates).

2. F

   Immunodetection of the indicated proteins of digitonin‐solubilized whole‐body zebrafish mitochondria with different concentrations of digitonin.

3. G–I

   BNGE of zebrafish fed with low‐protein/low‐fat diet (LP/LF) and control diet (CD). (G) variation of BMI of fish after 6 weeks in LP/LF and CD (H) Representative images. (I) Representative BNGE of whole fish mitochondria of fish fed during 6 weeks in the indicated diet (experimental replicates n = 2 are composed by a pool of n = 2 biological replicates).

Figure 2. Blue‐DiS proteomics of scaf1 ^+/+ and scaf1 ^−/− isolated mitochondria

1. Quantitative data‐independent scanning (DiS) mass spectrometry protein profiles for CI, CIII, and CIV. Vertical numbers indicate the BNGE gel slices. Left and right profiles correspond to scaf1 ^+/+ and scaf1 ^Δ1/Δ1 animals, respectively. Red heatmap corresponds to the E‐score from two proteotypic Scaf1‐derived tryptic peptides spanning sequences ascribed to the CIII‐interacting site (in green) and to the CIV‐interacting site (in yellow) in scaf1 ^+/+ fish. Thick blue line, marked with an asterisk, indicates the putative proteolytic site in Scaf1.

2. Sequence alignment of Scaf1 protein in mouse and zebrafish. Structural and functional regions previously described in mouse are indicated in shaded gray boxes. Thick blue line indicates the proteolytic processing site in the mouse sequence.

Figure EV3. Analysis of mitochondrial complexes by Blue‐DiS proteomics

1. A, B

   Correlation between the abundance of proteins (expressed as sum of E‐scores of the corresponding peptides) detected in the analysis of scaf1 ^+/+ and scaf1 ^−/− animals. Proteins were considered mitochondrial according to the classification in the mouse MitoCarta 2.0 database. Non‐mitochondrial proteins include true non‐mitochondrial proteins and potential mitochondrial proteins that failed to be identified as such. In (B), only the mitochondrial proteins are represented, indicating the proteins from the indicated groups.

2. C, D

   Heatmaps showing the summed absolute abundance of selected protein groups across BNGE gel slices. For a better comparison, absolute abundances were normalized using the values of slice 9 as a reference. Qualitative migration of the added E‐score value for each indicated complex, subcomplex, or protein. For each line, data were normalized within a 100–0 range, with 100 being the value of slice 9. The color scale is established as a linear increase from black (being 0) to the green in slice 9 (being 100). Any value over 100 is white.

3. E–G

   Analysis of the quantitative differences between Blue‐DiS profiles of scaf1 ^+/+ and scaf1 ^−/− animals. Differences in quantity profiles of CI‐, CIII‐, and CIV‐related complexes and SCs in (E) control and (F) scaf1 ^−/− samples. The insets focus on the differences at very high molecular weights (slices 1–15). (G) Comparative analysis of quantitative profiles of complexes and SCs in control or scaf1 ^−/− samples. Arrows indicate increase, decrease, or shifts of complexes observed between scaf1 ^−/− and control samples.

Figure 3. Phenotype consequences of Scaf1 loss of function

1. A, B

   Representative images from scaf1 ^+/+ and scaf1 ^−/− (A) female and (B) male adult zebrafish.

2. C–F

   Size of scaf1 ^Δ1/Δ1 and scaf1 ^Δ2/Δ2 (scaf1 ^−/−) fish in comparison with their respective scaf1 ^+/+ wild type (WT) lines, (C) length and (E) weight of females (Δ1 +/+ n = 10, Δ1 −/− n = 12, Δ2 +/+ n = 24, Δ2 −/− n = 18); (D) length and (F) weight of males (Δ1 +/+ n = 16, Δ1 −/− n = 13, Δ2 +/+ n = 13, Δ2 −/− n = 23).

3. G–K

   Adipose tissue measurements on hematoxylin–eosin (H&E)‐stained adult zebrafish sagittal sections. (G, I) Adipose tissue area per total section area (average of 3 sections/biological replicate) and (H, J) adipocyte size (average of 20–30 adipocytes of ventral adipose tissue per biological replicate) of females (G, H) (Δ1 n = 5, Δ2 n = 8, same number of animals for homozygous mutants and controls) and males (I, J) (Δ1 n = 8, Δ2 n = 7, same number of animals for homozygous mutants and controls). (K) Representative images of ventral fat deposits in females (dotted lines).

4. L–N

   Effect of Scaf1 loss of function on female fertility. (L) Number of eggs per clutch (Δ1 +/+ n = 12, Δ1 −/− n = 13, Δ2 +/+ n = 13, Δ2 −/− n = 10). (M) Quantification of mature ovary follicles per ovary section (average of three sections/biological replicate; Δ1 n = 5, Δ2 n = 8; same number of animals for homozygous scaf1 ^+/+ and scaf1 ^−/−). (N) Representative images of H&E‐stained ovaries. Dotted lines delineate adipose tissue.

Data information: One‐way ANOVA. Outliers are shown in gray and were not considered for the statistical analysis. Data are represented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Scale bars = 500 μm.

Figure EV4. Transgenic expression of scaf1 recovers CIII and CIV super‐assembly and fish size of scaf1 ^−/− fish

1. A

   Scheme of the transgenic construct Tg(ubi:scaf1). ubi, ubiquitin promoter. Βg‐intron, beta‐globin intron, pA, poly A, crys, crystalline promoter (used as selection marker).

2. B

   Immunodetection of the indicated proteins of BNGE from whole fish mitochondria scaf1 ^Δ1/Δ1 (−/−) expressing the transgenic construct Tg(ubi:scaf1) in heterozygosis, their scaf1 ^Δ1/Δ1 siblings with no transgenic expression, and scaf1 ^+/+ zebrafish (representative BNGE of n = 5). Asterisks mark bands’ absence in scaf1 ^Δ1/Δ1 (−/−) recovered by the transgenic expression of scaf1.

3. C, D

   Representative images from scaf1 ^+/+, scaf1 ^Δ1/Δ1 (scaf1 ^−/−), and scaf1^−/−, Tg/− (C) female and (D) male zebrafish (3 mpf).

4. E, F

   Size of scaf1 ^+/+, scaf1 ^Δ1/Δ1 (scaf1 ^−/−), and scaf1^{−/−, Tg/−} fish (E) length of females (scaf1 ^+/+ n = 23, scaf1 ^−/− n = 39 and scaf1 ^{−/−, Tg/−} n = 25); and (F) males (scaf1 ^+/+ n = 15, scaf1 ^−/− n = 31 and scaf1^{−/−, Tg/−} n = 44).

Data information: One‐way ANOVA. ns P > 0.05, **P < 0.01, ****P < 0.0001. Data are represented as mean ± SD.

Figure 4. Scaf1 loss of function leads to alterations in mitochondrial structure and performance

1. A–C

   Transmission electron microscopy image of cardiac muscle from scaf1 ^Δ1/Δ1 (n = 3) and scaf1 ^+/+ fish (n = 3). (A) Representative images showing mitochondria. (B) Mitochondria size (100–150 mitochondria per biological sample). (C) Cristae lumen width (average of three cristae per mitochondria, 20 mitochondria per biological sample). Different biological replicates are represented with different color tones.

2. D, E

   Mitochondrial DNA copy number per nuclear copy number in muscle in females (D) and males (E) (Δ1 n = 6, Δ2 n = 6 and same number for their respective controls).

3. F

   Survival curve of 4 days post‐fertilization embryos treated with different concentrations of the indicated OXPHOS inhibitors (three experimental replicates per biological replicate and three biological replicates).

4. G, H

   Oxygen consumption of 48 h post‐fertilization embryos using the XFe24 Seahorse analyzer, (G) representative oxygen consumption rate (OCR) profile along time, and (H) maximum OCR (Δ1 n = 11, Δ2 n = 11, n = 10, n = 11, respectively, for their controls).

5. I

   Maximum uncoupled (FCCP) OCR in isolated mitochondria from adult fish (male and females Δ1 n = 4 and Δ2 n = 4, and same number for their respective controls) with the indicated site I [pyruvate (Pyr), glutamate (Glu), malate (Mal)] or site II [succinate (Succ)] substrates.

6. J, K

   Respiratory control ratio (RCR; State 3/State 4) (J) and P/O ratio (K) in isolated mitochondria from adult fish (male and females Δ1 n = 4, and same number for their respective controls) with the indicated substrates.

Data information: (B–E, G, H, J, K) Unpaired t‐test, (F) two‐way ANOVA, and Sidak's multiple comparison test. (I) Two‐way ANOVA, post hoc Fisher's LSD test. ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Data are represented as mean ± SD, or (G) as ± SEM. Scale bars = large image 1 μm, small image 50 nm.

Figure 5. Diet‐induced recovery of scaf1 ^−/− phenotypes

Data from females.

1. A

   Representative images of scaf1 ^−/− and scaf1 ^+/+ fish fed with the indicated diets.

2. B, C

   Changes in (B) length and (C) weight over time (Δ1 +/+ n = 10, Δ1 −/− n = 10, Δ2 +/+ n = 10, Δ2 −/− n = 12–13).

3. D–F

   Adipose tissue measurements on hematoxylin–eosin (H&E)‐stained adult zebrafish sagittal sections. (D) Adipose tissue area per total section area (average of three sections/biological replicate) and (E) adipocyte size (average of 20–30 adipocytes of ventral adipose tissue per biological replicate; standard diet Δ1 +/+ n = 3, Δ1 −/− n = 3, Δ2 +/+ n = 3, Δ2 −/− n = 3, double diet Δ1 +/+ n = 3, Δ2 +/+ n = 3, Δ1 −/−n = 4, Δ2 −/− n = 4). scaf1 ^Δ1 and scaf1 ^Δ2 are represented with circles and squares, respectively. (F) Representative images of ventral fat deposits (dotted lines).

4. G

   Number of eggs per clutch (standard diet Δ1 +/+ n = 8, Δ1 −/− n = 8, Δ2 +/+ n = 7, Δ2 −/− n = 12, double diet Δ1 +/+ n = 7, Δ1 −/− n = 8, Δ2 +/+ n = 8, Δ2 −/− n = 9).

5. H

   Representative images of H&E‐stained ovaries. Black dotted lines outline the ovaries, and blue dotted line indicates a mature follicle.

6. I

   Quantification of mature ovary follicles per ovary section (average of three sections/biological sample; Δ1 n = 2–3, Δ2 n = 6).

Data information: (B, C) Two‐way ANOVA, (D, I) unpaired t‐test, and (E, G) one‐way ANOVA. Data are represented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Scale bars = 500 μm.

Figure EV5. Diet‐induced recovery of scaf1 ^−/− phenotypes in males and diet effect in SCAF1‐deficient mice

1. A

   Representative images of scaf1 ^−/− and scaf1 ^+/+ males fed with the indicated diets.

2. B, C

   Size of males after the indicated diet. (B) Changes in length and (C) weight over time (Δ1 +/+ n = 10, Δ1 −/− n = 10, Δ2 +/+ n = 10, Δ2 −/− n = 7–8).

3. D, E

   Effect of SCAF1 loss of function on weight gain in mice after starvation. (D) Scheme of the food restriction experiment in mice. (E) Impact of food restriction in C57BL/6JOlaHsd mice with the functional version of SCAF1 113/113, with the spontaneous mutation in SCAF1 111/111 (natural C57BL/6JOlaHsd mice harbor a non‐functional version of SCAF1) and in C57BL/6JOlaHsd mice without SCAF1 (SCAF1 KO, −/−). Males 111/111 n = 3; males KO n = 6; males 113/113 n = 2; females 111/111 n = 5; females KO n = 8; females 113/113 n = 5.

Data information: (B, C) Two‐way ANOVA. Data are represented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001. (E) Data are represented as mean ± SEM.

Figure 6. Lack of recovery of Scaf1^−/− phenotypes after high‐fat diet

1. A, B

   Representative images of scaf1 ^−/− and scaf1 ^+/+ females (A) and males (B) fed with the indicated diets.

2. C, D

   Length of females (C) and males (D) after the indicated diets (Δ1 +/+ n = 5, Δ1 −/− n = 5, Δ2 +/+ n = 5, Δ2 −/− n = 5).

3. E, F

   Weight of females (E) and males (F) after the indicated diets (Δ1 +/+ n = 5, Δ1 −/− n = 5, Δ2 +/+ n = 5, Δ2 −/− n = 5).

4. G

   Number of eggs per clutch (control diet: Δ1 +/+ n = 4, Δ1 −/− n = 3, Δ2 +/+ n = 3, Δ2 −/− n = 2, high fat diet: Δ1 +/+ n = 5, Δ1 −/−, n = 5 Δ2 +/+ n = 1, Δ2 −/− n = 2).

5. H

   Quantification of mature ovary follicles per ovary section in hematoxylin–eosin (H&E) histological sections (average of three sections/biological sample; Δ1 n = 3, Δ2 n = 3).

6. I–L

   Adipose tissue quantification in H&E sections (Δ1 n = 3, Δ2 n = 3) in females (I, K) and males (J, L): adipose tissue area (I, J) and adipocyte size (K, L).

Data information: (C–F) Two‐way ANOVA. (G–L) Unpaired t‐test. Data are represented as mean ± SD. ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 7. Molecular basis for diet‐induced recovery of scaf1 ^−/− phenotypes

1. A

   Immunoblot of the indicated proteins of BNGE from female scaf1 ^+/+ and scaf1 ^−/− whole zebrafish mitochondria for the indicated diet (representative of n = 2).

2. B

   Maximum uncoupled (FCCP) oxygen consumption rate in whole zebrafish mitochondria (females Δ1 n = 4 and Δ2 n = 4, and same number for their respective controls) with glutamate (Glu), malate (Mal), and succinate (Succ).

3. C–G

   RNAseq data from scaf1 ^+/+ and scaf1 ^−/− skeletal muscle for the indicated diet (standard diet scaf1 ^+/+ and scaf1 ^−/− n = 4, double diet scaf1 ^+/+ n = 3, scaf1 ^−/− n = 4). (C‐F) Volcano plots of differentially expressed genes (DEGs). (C) Comparison between scaf1 ^−/− and scaf1 ^+/+ zebrafish in standard diet. (D) Comparison between scaf1 ^−/− and scaf1 ^+/+ zebrafish in double diet. (E) Comparison of scaf1 ^+/+ zebrafish in double diet and standard diet. (F) Comparison of scaf1 ^−/− zebrafish in double diet and standard diet. In blue, significant DEGs (Padj < 0.05, log₂FC > |1|); in gray, not significant DEGs; red circles represent non‐significant differentially regulated OXPHOS genes, green circles represent significant differentially regulated OXPHOS genes, and purple circle represents scaf1 (cox7a2l). (G) Heatmap of metabolic pathways differentially regulated according to gene set enrichment analysis (GSEA) in the indicated comparisons.

4. H

   Heatmap of differentially regulated growth hallmarks according to GSEA analysis in the indicated comparisons.

Data information: (G, H) White squares Padj > 0.05, colored squares Padj < 0.05. Color scales goes from blue (downregulated) to red (upregulated) gene sets. (B) T‐test analysis. Data are represented as mean ± SD. *P < 0.05, **P < 0.01.