Respiratory supercomplexes enhance electron transport by decreasing cytochrome c diffusion distance

Abstract

Respiratory chains are crucial for cellular energy conversion and consist of multi-subunit complexes that can assemble into supercomplexes. These structures have been intensively characterized in various organisms, but their physiological roles remain unclear. Here, we elucidate their function by leveraging a high-resolution structural model of yeast respiratory supercomplexes that allowed us to inhibit supercomplex formation by mutation of key residues in the interaction interface. Analyses of a mutant defective in supercomplex formation, which still contains fully functional individual complexes, show that the lack of supercomplex assembly delays the diffusion of cytochrome c between the separated complexes, thus reducing electron transfer efficiency. Consequently, competitive cellular fitness is severely reduced in the absence of supercomplex formation and can be restored by overexpression of cytochrome c. In sum, our results establish how respiratory supercomplexes increase the efficiency of cellular energy conversion, thereby providing an evolutionary advantage for aerobic organisms.

Keywords: bioenergetics; competitive fitness; cryo-EM; mitochondria; respiratory chain supercomplexes.

Figure EV1. Workflow of cryo‐EM processing

1. A

   Schematic illustration of Workflow of cryo‐EM processing.

2. B

   Gold‐standard Fourier shell correlation (FSC) of the CIII₂/CIV supercomplex and the CIV density maps after final refinements. FSC = 0.143 is presented as dashed line.

3. C, D

   Local resolution maps of the CIII₂/CIV supercomplex viewed along the mitochondrial inner membrane (IM) and from the mitochondrial intermembrane space (IMS) side (C), and the CIV density after particle subtraction of CIII₂ and the digitonin micelle viewed from two rotations along the mitochondrial inner membrane (D).

Figure 1. A detailed structure of the CIII‐CIV interface allows to design supercomplex‐disrupting mutations

1. A

   Side view of the overall structure of the Saccharomyces cerevisiae CIII₂/CIV supercomplex and its position in the mitochondrial inner membrane (IM). IMS: intermembrane space.

2. B, C

   Zoom‐in of the CIII₂/CIV interaction sites with residues and distances annotated. The inner membrane protein–lipid–protein interactions (Rip1‐CL-Cox5a), where Rip1 is marked in gold, cardiolipin (CL) in gray, and Cox5a in red (B), and the mitochondrial matrix, protein–protein interaction of Cor1 (blue), and Cox5a (red) (C) are shown.

3. D

   Blue‐native gel electrophoresis of S. cerevisiae strains with indicated mutations in Cor1.

Figure EV2. Example densities of CIV subunits Cox12, Cox13, and Cox26

1. A

   Ribbon diagram of CIV with Cox12, Cox13, and Cox26 annotated. IM: inner membrane; IMS: intermembrane space.

2. B–D

   Example densities of Cox12 (B), Cox13 (C), and Cox26 (D) demonstrating parts of the subunits that improved due to the re‐processing of the data set.

Figure 2. Characterization of strains lacking respiratory supercomplexes

1. A

   Blue‐native gel electrophoresis and immunoblots of cells expressing the wild‐type form of Cor1 (Cor1^WT), as well as the mutants Cor1^{N63A, N187A, D192A} (Cor1*) and Cor1^{N63A, N187A, D192A, V189A, Y65A, L238A, K240A} (Cor1**) in wild‐type (WT) background as well as in cells lacking CRD1 (crd1Δ). Blots were probed with antibodies against Cor1 to visualize Complex III (CIII), as well as Cox1, to monitor Complex IV (CIV). CV: Complex V; Cor1Cor2: multimer consisting of Cor1 and Cor2.

2. B

   Reduced‐minus‐oxidized difference spectra of strains described in (A). Heme peaks are indicated for Cor1^WT in WT background.

3. C–E

   Steady‐state protein levels of strains as described in (A) cultivated on YP media containing glycerol (C) or CM media with galactose as carbon source (D, E).

Figure 3. Abrogation of supercomplexes decreases competitive fitness

1. A

   Growth analysis of strains expressing the wild‐type form of Cor1 (Cor1^WT) or the mutant Cor1^{N63A, N187A, D192A, V189A, Y65A, L238A, K240A} (Cor1**). Growth curves and calculated doubling times from the exponential phase (insert) are presented. Cells were cultivated in media containing either glucose or glycerol as carbon source.

2. B

   Determination of chronological lifespan via propidium iodide (PI) staining of cells described above.

3. C

   Scheme of the experimental setup to evaluate competitive fitness.

4. D, E

   Analysis of competitive fitness as described in (C). Cor1^WT strains harbored a chromosomally integrated hygromycin selection cassette, whereas Cor1** contained a clonNAT selection cassette (D). Subsequently, selection markers were switched as a control (E).

Data information: For (A, B, D, and E), line graphs indicate mean ± standard error of the mean (s.e.m.). Two‐way ANOVA mixed design followed by Bonferroni post hoc test was used to analyze data from 4 individual yeast clones (n = 4 biological replicates). Significances of main effects are visualized as: n.s.: not significant (P ≥ 0.05) and ^### P < 0.001. For insert in (A), data are presented as mean (square) ± s.e.m., median (center line), and single data points (n = 4 biological replicates). Thereby, a two‐tailed independent sample t‐test was performed and significances are given as: n.s.: not significant. A detailed description of statistical analyses performed is given in Table EV6.

Figure EV3. Disruption of supercomplexes does not modulate oxidative stress or cell death

1. A, B

   Scheme of 2′,7′‐dichlorodihydrofluorescein diacetate (DCFH₂‐DA), and propidium iodide (PI) staining and gating strategy for flow cytometry. PI is taken up by cells upon loss of membrane integrity, thus indicating dead cells. Non‐fluorescent DCFH₂‐DA is oxidized to fluorescent 2′,7′‐dichlorofluorescein (DCF) upon oxidative stress (A). Hence, the DCF mean fluorescence intensity of PI‐negative cells was applied as a measure for oxidative stress. The threshold used to separate PI‐negative and PI‐positive cells is shown as black line in (B).

2. C, D

   Flow cytometric quantification of oxidative stress indicated via DCFH₂‐DA/PI counterstaining of cells expressing the wild‐type form of Cor1 (Cor1^WT), as well as the mutant Cor1^{N63A, N187A, D192A, V189A, Y65A, L238A, K240A} (Cor1**). Cells were analyzed 16 h (C) and 24 h (D) after inoculation in media either containing glucose or galactose as indicated. To visualize fold values, normalization was performed to Cor1^WT cells of the respective medium used.

3. E, F

   Analysis of cell death via flow cytometric quantification of PI stained cells as described above. Experiments were performed 16 h (E) and 24 h (F) after inoculation, and the percentages of PI‐positive cells are presented.

Data information: Mean (square) ± s.e.m., median (center line), and single data points are depicted (for C and D: n = 8 biological replicates; for E and F: n = 4 biological replicates). Outliers were detected using the 2.2‐fold interquartile range labeling rule and are indicated in turquoise. Two‐tailed independent sample t‐tests were used for statistical analysis, except for (C glycerol) and (D glycerol), where two‐tailed Mann–Whitney U tests were applied. Significances are visualized as: n.s.: not significant (P ≥ 0.05), *P < 0.05. A detailed description of statistical analyses performed is given in Table EV6.

Figure EV4. Absence of supercomplex formation does not impact mitochondrial function in living cells

1. A, B

   Scheme of Mitotracker CMXRos staining and gating strategy for flow cytometric analysis. The fluorescence probe is taken up by mitochondria depending on the transmembrane potential (Δψ_m,). Dead cells accumulate the dye due to a lack of membrane integrity (A) and are excluded from the analysis. The threshold used to separate dead and living cells is shown as black line in (B). The mean fluorescence intensity of cells alive is quantified as a readout for Δψ_m.

2. C–E

   Analysis of Δψ_m via Mitotracker CMXRos staining of cells expressing the wild‐type form of Cor1 (Cor1^WT), as well as the mutant Cor1^{N63A, N187A, D192A, V189A, Y65A, L238A, K240A} (Cor1**). Cells were cultivated in CM media containing indicated carbon sources. Flow cytometric quantification 16 h (C) and 24 h (D) after inoculation are shown, as well as representative confocal micrographs (Z‐projections) after 24 h (E). Normalization in (C) and (D) was performed to Cor1^WT cells of the respective medium used.

3. F, G

   Oxygen consumption quantified in intact cells. Strains described in (C) were analyzed 16 h (F) and 24 h (G) after inoculation. Normalization of data was performed as stated in (C) to present fold values.

4. H

   Cellular ATP content measured from cells described in (C). The measurement was performed with cells cultivated for 24 h in CM media containing the indicated carbon sources. Normalization was performed as stated in (C).

Data information: Mean (square) ± s.e.m., median (center line), and single data points (n = 4 biological replicates) are depicted. Two‐tailed independent sample t‐tests were used for statistical analysis. For (C glycerol), Welch correction was performed. Significances are presented as: n.s.: not significant (P ≥ 0.05). A detailed description of statistical analyses performed is given in Table EV6. Scale bar in (E) represents 5 μm.

Figure 4. Lack of supercomplexes impairs electron transport from CIII to CIV

1. A, B

   Spectrophotometric measurement of Cyt c oxidase (CIV; A) and NADH Cyt c reductase (CIII; B) activities in solubilized mitochondrial extracts. Mitochondria were isolated from cells expressing the wild‐type form of Cor1 (Cor1^WT) or the mutant Cor1^{N63A, N187A, D192A, V189A, Y65A, L238A, K240A} (Cor1**).

2. C, D

   Polarographic measurement of KCN‐sensitive oxygen consumption, driven by NADH (C) or succinate + glycerol‐3-phosphate (G3P; D) in isolated coupled mitochondria of indicated strains. Measurements were performed in the absence (basal respiration) or presence (phosphorylating condition) of ADP.

3. E

   Respiratory control ratio calculated from polarographic measurements of KCN‐sensitive oxygen consumption driven by NADH and succinate + glycerol‐3-phosphate (G3P) in isolated coupled mitochondria of indicated strains.

4. F, G

   Polarographic measurement of KCN‐sensitive oxygen consumption driven by NADH in mitoplasts in the absence (endog. Cyt c) or presence (+ exog. Cyt c) of exogenous oxidized Cyt c. NADH oxidation (E) and the calculated ratio +Cyt c/−Cyt c of substrate oxidation (F) are visualized.

Data information: Mean (square) ± s.e.m., median (center line) and single data points are depicted. Two‐tailed independent sample t‐tests were used to analyze data from at least three individual mitochondrial isolations (n ≥ 3 biological replicates), For (A, B, and F endog. Cyt c), Welch correction was performed, and for (E Succinate + G3P), a two‐tailed Mann–Whitney U test was applied. Significances are given as: n.s.: not significant (P ≥ 0.05), *P < 0.05, **P < 0.01, and ***P < 0.001. A detailed description of statistical analyses performed is given in Table EV6.

Figure EV5. The lack of supercomplexes impairs electron transport from CIII to CIV

1. Assays to verify the intactness of the outer mitochondrial membrane. Mitochondria were isolated from strains expressing the wild‐type form of Cor1 (Cor1^WT), as well as the mutants Cor1^{N63A, N187A, D192A} (Cor1*) and Cor1^{N63A, N187A, D192A, V189A, Y65A, L238A, K240A} (Cor1**) in wild‐type (WT) background as well as in cells lacking CRD1 (crd1Δ). Intactness of the outer mitochondrial membrane was tested by following the stability of the intermembrane protein Cmc2 1 h after treatment with 6 μg/ml proteinase K (pK). Negative control samples were created by membrane permeabilization with 0.5% digitonin (dig.).

2. Immunoblot analysis of isolated mitochondria from strains expressing the wild‐type form of Cor1 (Cor1^WT) or the mutant Cor1^{N63A, N187A, D192A, V189A, Y65A, L238A, K240A} (Cor1**). Cells either overexpressed cytochrome c (Cyt c ^OE) or contained the empty plasmid (Empty) as a control. Blots were probed with antibodies against Cyt c, as well as porin as loading control.

3. Immunoblot analysis as described in (B) from intact cells. Blots were probed with antibodies against Cyt c, as well as tubulin as loading control.

4. Densitometric quantification of immunoblots validating overexpression of cytochrome c (Cyt c ^OE) in strains expressing either the wild‐type form of Cor1 (Cor1^WT) or the mutant Cor1^{N63A, N187A, D192A, V189A, Y65A, L238A, K240A} (Cor1**). The Cyt c signal was normalized to tubulin intensities as a loading control, followed by normalization to Cor1^WT strains harboring the empty vector to present fold values.

5. Flow cytometric quantification of loss of membrane integrity as visualized via propidium iodide (PI) staining of cells described in (D). Cells were cultivated in CM media either containing glucose, galactose or glycerol as carbon source.

6. Doubling time of strains described in (E).

Data information: Mean (square) ± s.e.m., median (center line), and single data points (n = 4 biological replicates) are depicted one‐way ANOVA followed by Bonferroni post hoc test was applied for statistical analysis, and significances are presented as: n.s.: not significant (P ≥ 0.05); **P < 0.01. A detailed description of statistical analyses performed is given in Table EV6.

Figure 5. Cytochrome c overexpression restores competitive fitness of strains lacking supercomplexes

1. A

   Polarographic measurement of KCN‐sensitive oxygen consumption, driven by NADH in isolated coupled mitochondria. Mitochondria were isolated from cells expressing the wild‐type form of Cor1 (Cor1^WT) or the mutant Cor1^{N63A, N187A, D192A, V189A, Y65A, L238A, K240A} (Cor1**). Thereby, strains either overexpressed cytochrome c (Cyt c ^OE) or harbored the empty plasmid as a control (Empty). Measurements were performed in the absence (basal respiration) or presence (phosphorylating condition) of ADP.

2. B–D

   Analysis of competitive fitness of strains described in (A). Cor1^WT strains harbored a chromosomally integrated clonNAT selection cassette, whereas Cor1** contained a hygromycin selection cassette. Strains were cultivated in CM media containing glucose (B), galactose (C), or glycerol (D) as carbon source.

Data information: For (A), mean (square) ± s.e.m., median (center line), and single data points (n = 6 biological replicates) are depicted, and an one‐way ANOVA followed by a Bonferroni post hoc test was used for statistical analysis. For (B–D), mean ± s.e.m. (n = 4 biological replicates) are depicted, and data were statistically evaluated by two‐way ANOVA mixed design followed by a Bonferroni post hoc test. Significances are visualized as: n.s.: not significant (P ≥ 0.05), *P < 0.05, **P < 0.01, ***P < 0.001. A detailed description of statistical analyses performed is given in Table EV6.