Mitochondrial super complexes in the colorless chlorophycean alga Polytomella parva

Abstract

Chlorophycean algae of the genera Chlamydomonas and Polytomella share a common photosynthetic ancestor. However, members of the Polytomella lineage have adopted a heterotrophic lifestyle, having lost the photosynthetic apparatus and relying instead on acetate or ethanol as carbon sources, with energy production centered on oxidative phosphorylation (OXPHOS). In this study, we investigated the composition of the mitochondrial supercomplexes of the colorless alga Polytomella parva. Mitochondrial membranes were solubilized using mild detergents such as glycol-diosgenin and digitonin, followed by separation of OXPHOS complexes supramolecular assemblies via Blue Native electrophoresis and Fast Protein Liquid Chromatography (FPLC). Additionally, complexome profiling of solubilized algal mitochondria resolved by Blue Native Gel Electrophoresis was carried out (data are available via ProteomeXchange with identifier PXD075371). The resulting data indicate that the OXPHOS supercomplexes of Polytomella closely resemble those observed in situ in the mitochondria of its green relative Chlamydomonas reinhardtii, as revealed by electron cryo-tomography and subtomogram averaging.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10863-026-10103-3.

Introduction

Mitochondria are essential organelles in most eukaryotic cells, responsible for ATP production through oxidative phosphorylation (OXPHOS). Complexes I, III and IV oxidize substrates derived from the citric acid cycle, transferring electrons from NADH and succinate to oxygen while simultaneously generating a proton gradient across the inner mitochondrial membrane. This electrochemical gradient is then used by the F₁Fo-ATP synthase to make ATP (Mitchell 1961). OXPHOS complexes often associate into higher-order oligomeric assemblies known as super complexes, which were first isolated by solubilization of mitochondria with mild detergents such as digitonin followed by Blue Native Electrophoresis (BN-PAGE) (Schägger 2001). Across species, the catalytic cores of OXPHOS complexes share many conserved structural features, whereas variability is commonly found in peripheral regions, including N- or C-terminal extensions, subunit insertions, lineage-specific polypeptides, or even additional structural domains (Cardol et al. 2009; He et al. 2024; Klusch et al. 2021; Maldonado et al. 2021; Mühleip et al. 2023; Parey et al. 2021). Consequently, the organization of OXPHOS supercomplexes exhibits substantial evolutionary diversity. To date, supercomplexes have been isolated and structurally characterized in representatives of four of the 13 eukaryotic supergroups (Guo et al. 2017; He et al. 2024; MacLean et al. 2025; Mühleip et al. 2023), while in situ characterization has been achieved for only two species (Waltz et al. 2025; Zheng et al. 2024). In all mitochondrial supercomplexes described so far, complex III serves as the central scaffold for the association of complexes I, II, or IV (Protasoni et al. 2020; Mühleip et al. 2023; Miranda-Astudillo and Rico-Luna 2025). These supramolecular assemblies are thought to facilitate electron transfer between respiratory complexes, thereby reducing the production of reactive oxygen species (Berndtsson et al. 2020; Kohler et al. 2023; Chan et al. 2024). Moreover, they contribute to the efficient packing of OXPHOS components within the inner mitochondrial membrane and play a role in shaping cristae morphology—forming lamellar cristae in Opisthonkonta and Archaeplastida (Blum et al. 2019; Davies et al. 2011; Miranda-Astudillo et al. 2022), discoid cristae in Discoba (Mühleip et al. 2017), tubular cristae in ciliates (Mühleip et al. 2016), and bulbous cristae in Apicomplexa (Mühleip et al. 2021).

Previous work carried out with P. parva mitochondria solubilized with mild detergents, n-dodecyl-β-D-maltoside (here denoted as lauryl maltoside or LM) or digitonin, followed by separation on BN-PAGE, allowed the identification of three different associations, I-IV₆, I-III₄, and I-IV (Miranda-Astudillo et al. 2018). Here, we revisited the algal OXPHOS complexes using glyco-diosgenin (GDN)-solubilized mitochondria, BN-PAGE, chromatography and mass spectrometry-based complexome profiling.

Materials and methods

1. Algal strain, growth conditions and mitochondria isolation.

P. parva (strain number 198.80 from the Culture Collection of Algae at the University of Göttingen) was grown in liquid mineral Tris-phosphate medium supplemented with sodium acetate (30 mM), and vitamins (biotin 10%, B₁₂ vitamin 10%, and B₁ vitamin 2 × 10–5% (w/v) (pH 7.0). Cells were harvested by centrifugation at 7,000 x g for 10 min and stored at -70 °C until use. Sedimented cells were resuspended in SPT buffer (0.3 M sucrose, 4 mM potassium-EDTA, and 20 mM Tris-HCl pH 7.2) and disrupted with a Potter homogenizer. An enriched mitochondrial membrane fraction was obtained by a two-step differential centrifugation as earlier described (Miranda-Astudillo et al. 2018).

• 2)Blue Native-Polyacrylamide Gel Electrophoresis (BN-PAGE).

Mitochondrial proteins were solubilized with 2.0 g lauryl maltoside (LM)/g protein (2.0%), or 4.0 g GDN/g protein (4.0%) in solubilization buffer (SB) containing 50 mM Tris-HCl, 1.5 mM MgSO₄, 100 mM NaCl, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 50 µg/ml tosyl-lysyl-chloromethylketone (TLCK) (pH 8.4). The mixture was incubated at 4 °C with gentle stirring for 30 min, and centrifuged at 30,000 × g for 30 min. The supernatants were subjected to BN-PAGE (Schägger 2001) in 3%–10% acrylamide gradient gels. For 2D-BN-PAGE, a lane was excised and loaded to a second acrylamide gradient with 0.03% LM in the cathode buffer (Wittig and Schägger 2005).

• 3)In-gel enzymatic activity staining.

In-gel staining of NADH/NBT oxidoreductase, cytochrome c oxidase, and ATPase activities were performed as previously described (Miranda-Astudillo et al. 2018).

• 4)OXPHOS complexes separation by size exclusion chromatography.

GDN-solubilized mitochondria were used in two independent approaches: anion-exchange chromatography and size exclusion chromatography.

For anion-exchange chromatography, 200 mg of algal mitochondria were solubilized in SB and centrifuged at 35,000 x g for 30 min. The supernatant was diluted in SB without NaCl (1:3), loaded on a Source 15Q 10/100 column and eluted with a continuous NaCl gradient (from 0 to 500 mM).

For size-exclusion chromatography, solubilized mitochondria were concentrated with an Amicon Ultra-15 Centrifugal Filter (EMD Millipore) to a final volume of 500 µL and injected to a couple of size exclusion Superose 6 10/300 columns connected in tandem (GE Healthcare Life Sciences) previously equilibrated with 50 mM Tris-HCl, 150 mM NaCl, 1mM MgSO₄, 10% glycerol, 1 mM PMSF, 50 µg/ml TLCK, and 0.01% GDN (pH 8.4). The elution was carried out at 0.25 mL/min. Fractions of 0.5 mL were collected and resolved by BN-PAGE.

• 5)Complexome profiling

• 5.1) In-gel digestion and peptide recovery

Isolated mitochondria (100 µg) were solubilized with either LM, digitonin or GDN at the following detergent/protein ratios (2 g of detergent per g of protein) and then resolved by BN-PAGE on a 3–16% acrylamide gradient gel (Wittig et al. 2006). After electrophoresis, the gel was fixed overnight in 50% methanol, 10% acetic acid and 100 mM ammonium acetate. It was then stained for 45 min with 0.025% Coomassie G-250 in 10% acetic acid, destained in 10% acetic acid, and stored in deionized water allowing the gel to re-swell to its original dimensions. The gels were scanned, and a full real size image was used for the cutting procedure.

Each lane was cut into 48 equal slices (upward from the bottom until the top of the gel), diced and transferred to 96-well filter plates (Millipore, MABVN1250) taped over waste collectors (Nunc MaxiSorp plates). Gel pieces were washed repeatedly in 50% methanol, 50 mM ammonium bicarbonate (ABC) until all blue dye had disappeared. Excess liquid was removed with brief centrifugations (1,000 x g, 20 s) between washes. Cysteines were reduced in 10 mM dithiothreitol, 50 mM ABC for 45 min, and further alkylated with 30 mM chloroacetamide, 50 mM ABC for 30 min in the dark. After a 15 min dehydration step in 50% methanol, 50 mM ABC and air-drying for 30–45 min at room temperature, 20 µl of a sequencing-grade trypsin solution at 5 ng µL⁻¹) in 50 mM ABC, 1 mM CaCl₂ were added to each well. The gel pieces were incubated at 4 °C for 20 min and covered with 50 µL of fresh ABC. The proteins were digested overnight at 37 °C. The resultant peptides were collected into clean 96-well PCR microplates after centrifugation at 1,000 x g for 60 s, followed by a 20 min incubation in 30% acetonitrile, 3% formic acid and finally eluted as in the previous step. The combined eluates were vacuum-dried (~ 3 h) in a Concentrator Plus (Eppendorf) and the peptides were resuspended in 20 µL of 5.0% acetonitrile, 0.5% formic acid. Samples were stored at − 20 °C until used.

• 5.2) Nano LC–MS/MS analysis

Thawed peptides were thoroughly agitated for ~ 20 min, and the plates were loaded onto an Ultimate 3000 UHPLC system coupled to an Orbitrap Eclipse Tribrid mass spectrometer (Thermo Fisher Scientific). Prior to electrospraying, 5 µL of peptides from each fraction were concentrated and desalted in a PepMap Neo Trap column (Thermo Fisher Scientific), followed by separation on an Elite Aurora column (1.7 μm C18, 15 cm × 75 μm ID (IonOpticks, Australia) maintained at 60 °C. Peptide elution was performed over 55 min using a linear gradient of solvent A (0.1% formic acid in water) and solvent B (0.1% formic acid in acetonitrile) at a flow rate of 0.3 µL/min, with the gradient programmed as follows: sample loading 0–4 min (2% B); separation 4–5 min (2–5% B), 5–35 min (5–40%B), and 35–40 min (40–90% B). The column was washed for 8 min at 90% B, then adjusted back to 2% B in 1 min followed by a 6 min re-equilibration at 2% B.

MS analysis was performed in positive mode. A NanoFlex source was used for electrospray ionization, applying 2.4 kV, with a source temperature of 275 °C. MS data were acquired in data-dependent acquisition (DDA) mode. Full MS scans were recorded from 375 to 1500 m/z at a resolution of 120,000, with an RF lens setting of 30%. MS1 data were collected in profile mode. The 20 most abundant precursors (charge states 2–7) were selected for MS/MS analysis. Fragmentation was performed using collision-induced dissociation (CID) at 35% collision energy. All other instrument parameters (e.g., AGC, injection times, dynamic exclusion) were set to default. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (Perez-Riverol et al. 2025) partner repository with the dataset identifier PXD075371.

• 5.3) Data processing and Profile Generation

Raw spectra were processed with MaxQuant v2.6.5.0 against two databases: the transcriptome of P.parva SAG 63 − 3 (sample MMETSP0052 from the Marine Microbial Eukaryote Transcriptome Sequencing Project) (Mallet and Lee 2006; Smith et al. 2010) and the mass spectrometry proteomic data obtained from Polytomella sp. SAG 198.80 (with identifier PXD035155 deposited at the ProteomeXchange Consortium) (Lacroux et al. 2022), with a false discovery rate (FDR) threshold of < 1%. The following modifications were specified: N-terminal acetylation (+ 42.01), deamidation of glutamine and asparagine (+ 0.98), and oxidation of methionine (+ 15.99) as variable modifications, while carbamidomethylation of cysteines (+ 57.02) was set as a fixed modification. were removed from the list. A total of 2,921 proteins were identified and annotated using BLAST in Phytozome, the Plant Comparative Genomics portal of the Department of Energy’s Joint Genome Institute (https://phytozome-next.jgi.doe.gov/) (Goodstein et al. 2012). Keratins, trypsin, and other common contaminants were not considered. Intensity-based absolute quantification (iBAQ) values were used to reflect protein abundances. Each profile was scaled to its own maximum, sorted by average-linkage hierarchical clustering with centered Pearson distance and displayed as black, yellow, and red heat maps in Microsoft Excel 2019. Further analysis was performed by manual inspection of the obtained profiles, and the iBAQ values were re-normalized within specific apparent molecular mass ranges. For OXPHOS complexes composed of multiple subunits, a single abundance profile was generated by averaging the iBAQ values of all quantified components. Specifically, we calculated the arithmetic mean of the subunits’ iBAQ values and, when indicated, further normalized these averaged values to facilitate comparisons of abundance patterns across complexes and detergent treatments. The detailed calculations used to generate these averaged profiles are provided in Supplementary File 1. Bovine heart mitochondria (BHM) solubilized under the same conditions as P. parva mitochondria were used for mass calibration of the 3–16% BN-PAGE.

• 6)In silico modelling

The 3D structural models of P. parva complexes III and IV were built using the SWISS-MODEL tool (Waterhouse et al. 2018), based on the structures of C. reinhardtii complex III (PDB 9F5Z) and complex IV (PDB 9F60) respectively (Waltz et al. 2025). The structure of P. parva complex I (PDB 7ARD) and the models of complex III and complex IV were assembled into the EMD 50,210 electron density map and the PDB 9F62 structure of the C. reinhardtii respiratory supercomplexes (Waltz et al. 2025) using ChimeraX 1.9 (Meng et al. 2023).

Results

To investigate the supramolecular organization of the OXPHOS complexes in P. parva, isolated mitochondria solubilized with LM or GDN were subjected to BN-PAGE. The previously characterized pattern of algal OXPHOS complexes solubilized with LM and resolved in 3–10% acrylamide gradient gels stained with Coomassie blue exhibited three main bands corresponding to dimeric complex V (V₂), complex I (I), dimeric complex III (III₂). Four weaker bands were also noticed, two from oligomeric forms of complex V (V₄ and V₆) and two representing monomeric and dimeric complex IV (IV and IV₂) (Fig. 1A lane 1). GDN solubilized mitochondria exhibited less V₂ but higher proportions of the oligomeric forms V₄ and V₆. This was evident in both Coomassie Blue and ATPase activity staining, indicating that the milder detergent GDN preserves more complex-complex interactions than LM (Fig. 1A, lanes 2 and 5). While both GDN and LM led to a similar band pattern of the complex V forms, the respiratory complexes displayed greater variation in migration. In-gel activity staining of complexes I and IV (Fig. 1A, lanes 3 and 4) evidenced their presence in several supramolecular associations. Further resolution of GDN-solubilized samples by 2D-BN-PAGE in the presence of LM, revealed that V₆ dissociated into V₄, V₂, and V; V₄ was partially conserved, but also dissociated into V₂ and V; and lastly, V₂ fully dissociated to V. Complex I was found to be present in three larger super-complexes while complex IV appeared in four larger oligomeric forms. Furthermore, complex III₂ was found in three supramolecular associations. In summary, these results indicate four main oligomeric associations of the respiratory complexes: III₂-IV, I-III₂, I₂-III_W-IV_Z and I₂-III_X-IV_Y (where W, X, Y and Z represent uncertain stoichiometries, since the exact ones are difficult to define with this approach) (Fig. 1B).

Fig. 1.

ATPase oligomers and respiratory supercomplexes in P. parva mitochondria. Isolated mitochondria were solubilized with the indicated detergent: lauryl maltoside (LM) at 2.0 g/g protein or glyco-diosgenin (GDN) at 4.0 g/g protein. (A) soluble fractions resolved by BN-PAGE in a 3–10% polyacrylamide gradient gel showing Coomassie Blue staining (lanes 1 and 2); in-gel NADH-dehydrogenase activity staining (lane 3); in-gel cytochrome c oxidase (COX) activity staining (lane 4); and in-gel ATPase activity staining (lane 5). (B) The GDN-solubilized samples resolved by BN-PAGE (Lanes 2 to 5 from Fig. 1A) were incubated in the presence of LM and resolved in 2D-BN-PAGE gels, allowing the rupture of higher-order associations but maintaining intact the composition of individual complexes I, III and IV. The four 2D gels were subject to Coomassie Blue and activity staining, revealing the composition of the different supercomplexes.

Fast Protein Liquid Chromatography (FPLC) using either anion exchange or size-exclusion columns allowed us to identify different super complexes in GDN-solubilized Polytomella mitochondria. In the first case, protein complexes were eluted with a continuous gradient from 0 to 500 mM NaCl, giving rise to the elution profile shown in Fig. 2A. Collected fractions were subjected to BN-PAGE and in-gel activity staining for both complexes I and IV. The ATP synthase oligomeric forms V₂, V₄ and V₆ eluted earlier than respiratory complexes I, III and IV (Fig. 2A) and their supramolecular associations, here identified as III₂-IV, I-IV, and I-III₂. In the second case, our tandem size exclusion chromatography setup allowed separation of proteins ranging 5–5000 kDa. The elution pattern shows the sequential separation of V₆, V₄ and V₂ oligomers followed by the supramolecular associations of respiratory complexes I-IV and III₂-IV (Fig. 2B). The molecular masses of the identified oligomeric states were estimated as shown in Supplementary Figure S1.

Fig. 2.

OXPHOS supercomplexes are resolved by ion exchange chromatography and size exclusion chromatography. Two hundred milligrams of mitochondria were solubilized with GDN and loaded into an anion exchange column, then eluted with continuous NaCl gradient. (A) BN-PAGE of elution fractions from ion exchange chromatography, the upper gray triangle represents the NaCl gradient. (B) BN-PAGE of fractions derived from size-exclusion chromatography. The gels were stained for complex I and IV activity

To further explore the effect of the detergents GDN, digitonin, and LM on mitochondria solubilization and super-complex stability, composition, and abundance, complexome profiling (CP) was performed. This approach involves separating mitochondrial proteins by native electrophoresis, fractionating entire gel lanes, and identifying proteins in each individual fraction by tandem mass spectrometry followed by abundance pattern–based hierarchical clustering (Heide et al. 2012; Cabrera-Orefice et al. 2022). CP enables unbiased and comprehensive identification of potential protein–protein interactions.

Complexome profiling analysis allowed us to identify around 2,900 proteins, among which we found most of the subunits that form part of the mitochondrial respiratory complexes. For complex I, 38 out of the 51 subunits were identified; for complex III, 7 out of 10; and for complex IV, 9 out of 12 (Fig. 3, Supplementary Table S2). It should be noted that the subunits that cannot be identified, such as cytochrome b of complex III, or some subunits belonging to the P distal module of complex I, are highly hydrophobic membrane proteins, which makes them difficult to digest with trypsin and subsequently to be identified by MS. On the other hand, all the subunits of complex II and complex V could be identified (4 for complex II and 18 for complex V) (Supplementary Figure S2). In addition, this experimental approach allowed us to identify 5 isoforms of previously characterized proteins (Supplementary Table S1 and Supplementary Figure S3). For complex I, two isoforms were identified: one of NUOP4 belonging to the P distal module, and another for gamma carbonic anhydrase (CA3). In complex IV, the presence of an isoform of the Cox6b subunit was found (Fig. 3, highlighted in red). In the case of complex II and complex V, which are not constituents of the supercomplexes, an isoform of subunit D and an isoform of subunit Asa9 were identified, respectively (Supplementary Figure S2). Under the growth conditions for the colorless alga, using acetate as carbon source, all these proteins are produced in different proportions.

Fig. 3.

Complexome profiling of mitochondria isolated from P. parva solubilized with different detergents. P. parva mitochondria were solubilized with the detergents GDN, LM, and digitonin at a ratio of 2:1 mg of detergent per mg of protein. The samples were separated by BN-PAGE followed by quantitative mass spectrometry analysis. The iBAQ abundance of the subunits composing each complex was visualized using a heat map. For complex I, we identified at least 42 of the 51 known subunits; for complex III, 8 of 10; and for complex IV, 9 of 12. Complex I migrated with apparent molecular masses of approximately 950 kDa in GDN, 1100 kDa in LM, and 930 kDa in digitonin. Complex III showed apparent molecular masses of about 470, 510, and 430 kDa in GDN, LM, and digitonin, respectively. Finally, complex IV displayed apparent molecular masses of roughly 230, 150, and 200 kDa in GDN, LM, and digitonin

With GDN-solubilized mitochondria, the migration of complexes I, III and IV and its supramolecular associations were followed by mass spectrometry (Fig. 4). Normalization of the main peaks representing the supercomplexes allowed the identification of three different organizations: 1-IV (~ 1,188 kDa), I-III₄-IV₂ (~ 2,320 kDa) and I₂-III₄-IV₂ (~ 3,130 kDa) (Fig. 4, inset).

Fig. 4.

Respiratory complexes and supercomplexes in GDN-solubilized mitochondria. Relative abundance graph and heat maps of the migration profiles of the averaged subunits of the algal respiratory complexes in their individual forms (OXPHOS-GDN): IV (232 kDa, green line), III₂ (468 kDa, orange line), and I (943 kDa, blue line). Inset: amplified region showing the migration zone of the supercomplexes (SC-GDN) (1,200-5,000 KDa), where the associations I/IV (1,200 kDa), I-III₄-IV₂ (2,300 kDa) and I₂-III₄-IV₂ (3,100 kDa) could be identified

In contrast, digitonin-solubilized mitochondria subjected to BN-PAGE in the same conditions seem to partially disrupt supramolecular associations, since higher levels of free complexes I, III and IV were present as compared with GDN-solubilized mitochondria (Fig. 5). In these conditions, the three following associations were identified: I-IV (~ 1,188 kDa), I-III₂-IV (~ 1,445 kDa) and I₂-III₂-IV₂ (~ 2,532 kDa) (Fig. 5, inset).

Fig. 5.

Respiratory complexes and supercomplexes in digitonin-solubilized mitochondria. Relative abundance graph and heat maps of the migration profiles of the averaged subunits of the algal respiratory complexes in their individual forms (OXPHOS-digitonin): IV (192 kDa, green line), III₂ (421 kDa, orange line), and I (922 kDa, blue line). Inset: amplified region showing the migration zone of the supercomplexes (SC-digitonin) (1,200-5,000 KDa), where the associations I/IV (1,200 kDa), I-III₂-IV₄ (1,500 kDa) and I₂-III₂-IV₂ (2,500 kDa) could be identified

Finally, and as expected, LM-solubilized mitochondria displayed a polypeptide pattern exhibiting mainly free respiratory complexes and negligible amounts of supercomplexes and (Fig. 6).

Fig. 6.

LM destabilizes supramolecular associations in P. parva mitochondria, preventing higher-order assemblies. Relative abundance graph and heat maps of the migration profiles of the subunits of the complexes in their individual forms (OXPHOS-LM): IV (200 kDa, green line), III₂ (508 kDa, orange line), and I (1,086 kDa, blue line). Inset: amplified region of the migration zone of supercomplexes (SC-LM) showing negligible amounts of high-order associations (No SC´s)

Discussion

Earlier studies recognized the mitochondrial ATP synthase of the colorless alga P. parva as a stable dimer resistant to dissociation by detergents such as LM (Vázquez-Acevedo et al. 2006; van Lis et al. 2007). The enzyme contains conserved subunits within its rotary and catalytic domains but also incorporates several atypical subunits known as ASA subunits, which form the peripheral arms of the complex and participate in the dimerization of the enzyme. In addition to its dimeric form, higher-order oligomers—including tetramers and hexamers—were observed by BN-PAGE analysis (Miranda-Astudillo et al. 2018). Low-resolution imaging further revealed the overall silhouette of the dimer, characterized by robust peripheral stalks (Dudkina et al. 2005). More recently, high-resolution 3D structures have provided detailed insights into subunit interactions and the mechanism of rotational catalysis (Allegretti et al. 2015; Murphy et al. 2019).

Electron microscopy and cryo-electron tomography revealed the presence of densely packed cristae in Polytomella mitochondria (Dudkina et al. 2006). Using electron cryo-tomography and subtomogram averaging at resolutions of up to 4.2 Å, ATP synthase could be visualized in vivo following rapid freezing of actively growing algal cells. The 3D reconstruction of a mitochondrion showed series of ATP synthase dimers adopting a near-helical arrangement, with multiple parallel rows organized within the disk-shaped cristae (Dietrich et al. 2024). Additionally, individual ATP synthases were observed in proximity to the cristae junctions. As shown here, some of these higher V₂ oligomeric states can be preserved after mild extraction (Fig. 1), and these associations remain stable after size-exclusion or anion-exchange chromatography (Fig. 2).

The existence of respiratory super complexes in Polytomella was suggested by the identification in LM-solubilized mitochondria of I-IV₆, I-III₄, and I-IV associations by BN-PAGE and the in vitro reconstitution of supercomplexes III₂-IV and I-III₂-IV₂ by ion exchange chromatography (Miranda-Astudillo et al. 2018). Here, we demonstrate the presence of the super complex III₂-IV in GDN-solubilized mitochondrial membranes (Fig. 1A) as well as the existence of dimeric complex IV and of at least one respirasome.

Recent advances in in situ cryo-electron microscopy have provided direct images of mitochondria, enabling the determination of the 3D structures of respiratory supercomplexes in their native states from both porcine (Zheng et al. 2024) and reinhardtii mitochondria (Waltz et al. 2025). Therefore, there is no doubt about the existence of dynamical super-structures in the mitochondrial inner membrane that exhibit a huge diversity of associations and stoichiometries among different species (Eldeeb et al. 2024; Guan et al. 2022). In mammalian mitochondria, four distinct super complex organizations have been identified: I-III₂-IV (~ 1,980 kDa), I₂-III₂-IV₂ (~ 3,380 kDa), I-III₂-IV₂ (~ 2,128 kDa), and I₂-III₄-IV₂ (~ 3,906 kDa). Notably, the latter two arrangements had not previously been detected in vitro using mild non-ionic detergents for extraction (Vercellino and Sazanov 2024), highlighting the unique ability of in situ cryo-electron microscopy to reveal otherwise elusive associations. In contrast, the C. reinhardtii respirasome displayed an in situ I₂-III₄-IV₆ arrangement (Waltz et al. 2025), which differs from supercomplexes reported in other organisms, including Sus scrofa [I-III₂-IV₂ (PDB 8UGI), I-III₂-IV₂ (PDB 8UGJ), I₂-III₂-IV₂ (PDB 8UGN), I₂-III₄-IV₂ (PDB 8UGR)] (Zheng et al. 2024); Mus musculus [I-III₂-IV (PDB 8PW7), I-III₂-IV₂ (PDB 8PW5)] (Vercellino and Sazanov 2024); Euglena gracilis [I-III₂-IV] (He et al. 2024); and Tetrahymena thermophila [I₂-II₂-III₄-IV₂ (PDB 8GYM)] (Han et al. 2023). Among them, the T. thermophila super complex is the only respirasome characterized so far to include complex II, a feature attributed to lineage-specific subunits in both complexes II and IV, that promote an unusually large supramolecular assembly exceeding in size the I₂-III₄-IV₆ arrangement of the C. reinhardtii respirasome.

These different supramolecular organizations may influence local membrane curvature, rendering it more convex or concave depending on their organization and stoichiometry (Zheng et al. 2024). Putative respiratory strings have been proposed to be present in mammalian, chlorophycean and plant mitochondria, where I₂-III₂-IV₂ and I₂-III₄-IV₂ associations should work as building blocks for larger circular or linear organizations, respectively (Bultema et al. 2009; Guo et al. 2018; Letts et al. 2016; Miranda-Astudillo et al. 2018). To date, no direct evidence of these large associations has been confirmed using in-situ studies (Davies et al. 2011, 2012; Dietrich et al. 2024; Mühleip et al. 2016, 2017; Zheng et al. 2024). Electron cryotomography of C. reinhardtii mitochondria further revealed a spatial segregation within cristae of respiratory complexes and rows of ATP synthase dimers, as has been reported in other systems. Biochemical isolation and characterization of the C. reinhardtii respirasome yielded a super complex with an I-III₂-IV₂ composition, whose structure was resolved at 2.8 Å using single-particle cryo-EM (Waltz et al. 2025). Apart from the absence of a dimeric complex IV, this isolated form appears to represent one half of the intact respirasome observed in situ.

Multicellular chlorophycean algae such as Volvox diverged from their unicellular ancestors at least 200 million years ago (MYA) (Herron et al. 2009). It can be inferred that the Polytomella lineage separated from the Chlamydomonas less than 200 MYA. The loss of photosynthesis in Polytomella may have driven extensive changes in chloroplast composition and function, ultimately leading to the emergence of colorless plastids in this genus (Figueroa-Martínez et al. 2015). In contrast, mitochondrial evolution appears to have followed a more conservative trajectory, preserving many features of its green algal relatives. The stoichiometry of the P. parva respirasomes characterized in this study may reflect supramolecular associations derived from the dissociation of a larger, Chlamydomonas-like I₂-III₄-IV₆ respirasome. Nevertheless, we did not detect in vitro a complete respirasome comparable to the one observed in situ in the green algal mitochondria. This discrepancy may arise from several factors, including the disruptive effects of detergent extraction on complex organization or the loss of structural lipids such as cardiolipin (Waltz et al. 2025). Alternatively, it is possible that the various supra complex species identified by complexome profiling assemble in a context-dependent manner, reflecting differential associations that adapt to the metabolic requirements of the cell.

In the C. reinhardtii respirasome, four contact regions have been identified between complexes I and III, and two between complexes I and IV. It was proposed that the I/III super complex is the most stable, whereas complex IV is the least resistant to biochemical purification (Waltz et al. 2025). In the present work, we isolated an I₂-III₄-IV₂ super complex from P. parva, identifying each of the orthologous subunits that form the complexes taking as a reference the subunits that make up the same complexes in C. reinhardtii (Supplementary Table S3). However, it is likely that up to four complex IV units were lost during the isolation procedure. We propose that P. parva contains a higher-order I₂-III₄-IV₆ respirasome which, upon detergent solubilization of mitochondria, may dissociate into the various single complexes and supercomplexes detected in this study (Fig. 7). While differences between isolated supercomplexes and their counterparts observed in situ are well recognized, multiple supramolecular arrangements with varying stoichiometries can coexist within the same organism (Zheng et al. 2024). Accordingly, the I₂-III₄-IV₂ super complex characterized here may also represent a physiologically relevant organization.

Fig. 7.

The supramolecular associations of respiratory complexes in P. parva. The main, central figure shows a model for the putative P. parva respirasome modelled upon the C. reinhardtii mitochondrial respirasome I₂-III₄-IV₆, and its fate after solubilization with each of the three detergents used in this work: LM, digitonin and GDN. Solubilization with the detergent LM caused the dissociation of the supercomplexes into their individual complexes I (blue), III (orange) and IV (green). Digitonin preserved three interactions: the supramolecular associations I-IV, I-III₂-IV, and I₂-III₂-IV₂. By contrast, solubilization GDN yielded six different supramolecular species, two of which were respirasomes with different stoichiometries: I-III₄-IV₂ and I₂-III₄-IV₂. P. parva complexes III and IV were modeled upon the C. reinhardtii PDB structures PF5Z and 9F60 (Waltz et al. 2025) respectively, using Swiss-Model. For P. parva Complex I the structure 7ARD from PDB was used (Klusch et al. 2021). The P. parva respirasome was assembled in Chimera X using the density map EDM 50,210 together with the PDB structure 9F62 (Walz et al., 2025)

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Marcos Ostolga-Chavarría: formal analysis, investigation, validation, methodology, writing, review and editing.Anaiza Rico-Luna: investigation, validation, methodology, review and editing.Sergio Fuentes-Hernández: investigation, validation, methodology, review and editing.Héctor V. Miranda-Astudillo: conceptualization, formal analysis, funding acquisition, methodology, resources, supervision, review and editing.Alfredo Cabrera-Orefice: data curation, formal analysis, methodology, investigation, funding acquisition, methodology, resources and writing, review, and editing. Diego González-Halphen: conceptualization, funding acquisition, methodology, project administration, resources, supervision and writing original draft.

Data availability

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD075371.

Declarations

Competing interests

The authors declare no competing interests.