Mitochondrial-derived vesicles retain membrane potential and contain a functional ATP synthase

Abstract

Vesicular transport is a means of communication. While cells can communicate with each other via secretion of extracellular vesicles, less is known regarding organelle-to organelle communication, particularly in the case of mitochondria. Mitochondria are responsible for the production of energy and for essential metabolic pathways in the cell, as well as fundamental processes such as apoptosis and aging. Here, we show that functional mitochondria isolated from Saccharomyces cerevisiae release vesicles, independent of the fission machinery. We isolate these mitochondrial-derived vesicles (MDVs) and find that they are relatively uniform in size, of about 100 nm, and carry selective protein cargo enriched for ATP synthase subunits. Remarkably, we further find that these MDVs harbor a functional ATP synthase complex. We demonstrate that these vesicles have a membrane potential, produce ATP, and seem to fuse with naive mitochondria. Our findings reveal a possible delivery mechanism of ATP-producing vesicles, which can potentially regenerate ATP-deficient mitochondria and may participate in organelle-to-organelle communication.

Keywords: ATP synthase; membrane potential; mitochondria; mitochondrial-derived vesicles; protein distribution.

Figure 1. Isolation and characterization of mitochondria from Saccharomyces cerevisiae

1. Western blot analysis of purified mitochondria. Mitochondria were isolated from yeast BY4741 wild‐type strain, and their purity was assessed by western blot analysis using Tim44 (49 kDa), Tim23 (23 kDa), Nfs1 (54 kDa), and Atp2 (55 kDa) antisera as mitochondrial markers, Hxk1 (54 kDa) and Tdh1 (36 kDa) antisera as cytosolic markers and Pex13 (42.7 kDa) as a peroxisomal marker.

2. Percentage of the summed intensity of mitochondrial proteins in the mitochondrial fractions according to LC/MS/MS analysis. Each value represents the mean ± SD for n = 3 biological repeats, P‐value = 1.05e9.

3. In vitro import of Aco1. Aco1 mRNA was translated in rabbit reticulocyte lysate in the presence of [³⁵S] methionine. The labeled protein products were added to purified mitochondria for the indicated times, in order to measure the mitochondrial import efficiency. Samples were analyzed by SDS–PAGE and autoradiography. Processing of the Precursor, p form to a Mature, m form, indicated a functional import machinery (Vitrobot mark IV, FEI).

4. Representation of the vesicle isolation method. Isolated and purified mitochondria were incubated in an isotonic buffer (1) and separated from the supernatant by centrifugation (2). The supernatant was freed from debris by viva‐cell centrifugation (through vertical filtration membranes; Sisquella et al, ; 3), which was then ultracentrifuged at high speed to sediment MDVs (4). The figure was generated by the BioRender software (https://app.biorender.com/).

Source data are available online for this figure.

Figure 2. Characterization of MDVs from wild‐type mitochondria

1. Nanoparticle tracking analysis. Vesicle size distribution and concentration were performed using nanoparticle tracking analysis (NTA; Malvern Instruments, Nanosight NS300). Sample size distributions were calibrated in a liquid suspension by the analysis of Brownian motion via light scattering. Nanosight provides single particle size and concentration measurements.

2. Atomic force microscopy. Representative AFM image and a 3D AFM image of one representative vesicle (WT), adsorbed on a mica modified with Mg²⁺ and imaged under PBS.

3. TEM images of MDVs. Samples were stained with 2.0% uranyl acetate or 2.0% phosphotungstic acid. 5 μl of vesicles was placed on Formvar/carbon‐coated copper 200 mesh grids (EMS), mixed with 5 μl of PTA for 10–20 s, while excess stain was blotted off and grids were dried. Samples were examined with Jeol (Jem‐1400 Plus) transmission electron microscope. Scale bar—200 nm.

4. Cryo‐TEM analysis of mitochondrial‐derived vesicles, which can be detected to contain single and potentially double membranes. Samples were prepared using a vitrification robot system. Scale bar—100 nm.

Source data are available online for this figure.

Figure EV1. Characterization of MDVs derived from Dnm1 KO mitochondria

1. Nanoparticle tracking analysis. Vesicle size distribution and concentration were performed using nanoparticle tracking analysis (NTA; Malvern Instruments, Nanosight NS300). Sample size distributions were calibrated in a liquid suspension by the analysis of Brownian motion via light scattering. Nanosight provides single particle size and concentration measurements.

2. Atomic force microscopy. Representative AFM image and a 3D AFM image of one representative vesicle (Dnm1 KO), adsorbed on a mica modified with Mg²⁺ and imaged under PBS.

3. TEM images of MDVs. Samples were stained with 2.0% uranyl acetate or 2.0% phosphotungstic acid. 5 μl of vesicles was placed on Formvar/carbon‐coated copper 200 mesh grids (EMS), mixed with 5 μl of PTA for 10–20 s, while excess stain was blotted off and grids were dried. Samples were examined with Jeol (Jem‐1400 Plus) transmission electron microscope. Scale bar—2 μm.

Source data are available online for this figure.

Figure EV2. Vesicle characterization

1. Vesicle concentrations according to nanoparticle tracking analysis over time. Vesicle concentration was measured at the indicated time points, n = two repeats.

2. Mitochondria ATP levels with time. Isolated mitochondria from wild‐type cells were incubated for 0, 4 h, and 24 h at 30°C prior to measuring ATP production, in the absence or presence of succinate, by luciferin‐luciferase luminometry as described in the methods section. Each value represents the mean ± SD for n = 3 technical repeats. Significant differences were detected using t‐test, P‐value = 1.9558E‐05/1.03993E‐06 for 24 and 4 h mitochondria with succinate, respectively.

3. Western blot analysis of isolated MDVs. MDVs were purified from wild‐type or KO‐isolated mitochondria as mentioned, and the presence of Mdh1p and Por1p was assessed by western blot analysis using Mdh1 antisera.

4. Vesicles count in WT and ΔDnm1 strains. MDVs isolated from Wild‐type and Dnm1 KO strains were analyzed using nanoparticle tracking analysis (NTA; Malvern Instruments, Nanosight NS300). n = two repeats.

Source data are available online for this figure.

Figure 3. Selectivity and specificity of MDV protein cargo as shown by LC/MS/MS analysis

1. OptiPrep gradient purification of MDVs. Media components were fractionated by centrifugation (100,000 g, 18 h, 4°C) through a continuous 10–30% OptiPrep (Axis‐Shield) gradient. Fractions (1 ml) were collected from the top of the gradient and run on acrylamide gel for further analysis.

2. Venn diagram of three different repeats of LC/MS/MS analysis. 444 mitochondrial proteins were observed in mitochondrial fractions from three different repeats, and out of this group, 168 mitochondrial proteins were observed in MDVs.

3. Representation of sub‐compartmental localization of the MDV proteome. OMM—outer mitochondrial membrane; IMM—inner mitochondrial membrane; IMS—intermembrane space.

4. There is a selectivity in protein cargo of MDVs versus mitochondrial fractions. Percentage of representative mitochondrial proteins from the total intensity of mitochondrial proteins from three different repeats was calculated. EV = MDVs. Mito = mitochondria. Each value represents the mean ± SD for n = 3 biological repeats. Mitochondrial inner membrane P‐value = 0.0022. Mitochondrial outer membrane P‐value = n.s. Mitochondria ribosomal proteins P‐value = 0.0082. ATPases P‐value = 0.0078. Respiratory chain complexes P‐value = 0.0070. Enzymes P‐value = 0.0079.

5. Compromised mitochondria produce less MDVs. MDVs isolated from Wild‐type and KO strains (Δom45 and Δmdh1) were analyzed using nanoparticle tracking analysis (NTA; Malvern Instruments, Nanosight NS300). Each value represents the mean ± SD for n = 3 biological repeats. Significant differences were detected using t‐test, P‐value = 0.033/0.023 for ΔOM45 and ΔMdh1, respectively.

Source data are available online for this figure.

Figure EV3. Data for MDVs from yeast wild‐type mitochondria

1. Principal component analysis (PCA) visualizes the projection of the dataset defined by PCA in 2‐dimensional viewers, transforming the large set of protein information into a protein profile of each sample. The analysis was done by Perseus software.

2. The average number of gold particles of Atp2 is significantly higher compared with the controls. The number of gold particles was measured using ImageJ software (ImageJ.nih.gov). Each value represents the mean ± SD for n = 3 biological repeats. Significant differences were detected using t‐test, P‐value = 1.379E‐09.

Source data are available online for this figure.

Figure 4. F1Fo‐ATP synthase subunits are present in MDVs

1. GO annotation of biological processes of the MDV proteome. A list of 168 MDV proteins (based on three different repeats) was analyzed by the “STRING” tool. The results were filtered by strength > 0.5 and FDR < 0.05. Here, we represent the top 10 high significant values.

2. Western blot analysis of gradient purified MDVs. Media components were fractionated by centrifugation (100,000 g, 18 h, 4°C) through a continuous 10–30% OptiPrep (Axis‐Shield) gradient. Fractions (1 ml) were collected from the top of the gradient, ran on acrylamide gels, and assessed by western blot analysis using Atp2 antisera.

3. Immunogold labeling of MDVs. Isolated mitochondrial‐derived vesicles were incubated with specific primary antibodies (anti‐Atp1 or anti‐Atp2) and secondary antibodies conjugated to 12 nm gold (goat anti‐rabbit). Samples were viewed by Tecnai 12 TEM 100 kV (Phillips, Eindhoven, the Netherlands). AUc—uranyl acetate. Scale bar—500 nm. Arrows indicate MDVs.

Source data are available online for this figure.

Figure 5. MDVs produce ATP and can rejuvenate damaged mitochondria

1. MDVs produce ATP in an enzymatic reaction dependent on membrane potential. Mitochondrial‐derived vesicles from wild‐type (wtMDVs) or devoid of malate dehydrogenase (dMDH MDVs), were incubated in the presence or absence of ADP, CCCP, or oligomycin for 15 min at 30°C or 4°C. Subsequently, ATP was measured by luciferin‐luciferase luminometry. Each value represents the mean ± SD for n = 3 biological repeats. Significant differences were detected using t‐test, P‐value = 0.01/0.0004 for CCCP and oligomycin, respectively.

2. Membrane potential of MDVs as visualized by imaging flow cytometry. Atp2‐GFP MDVs stained with TMRE (tetramethylrhodamine, ethyl ester) were treated with CCCP and imaged by ImageStreamX mark II (Amnis, Part of Luminex, Au. TX). Each value represents the mean ± SD for n = 3 biological repeats. Significant differences were detected using t‐test, P‐value = 0.0039.

3. MDVs can be taken up by mitochondria and rejuvenate damaged mitochondria. Isolated mitochondria from wild‐type (WT mito) and respiratory deficient mitochondrial (dATP2 mito) were preincubated in the presence or absence of mitochondrial‐derived vesicles from wild‐type (WT MDVs), washed and incubated for 15 min at 30°C in the presence or absence of ADP. Subsequently, ATP formation was measured by luciferin‐luciferase luminometry. Each value represents the mean ± SD for n = 3 biological repeats. Significant differences were detected using t‐test, P‐value = 2.38872E‐13 (for WT mito+ADP)/ 5.44794E‐06 (for WT mito + ADP + MDVs) 2.38872E‐13 (for Δatp2 mito + ADP)/ 4.84404E‐13 (for Δatp2 mito + ADP + MDVs).

Source data are available online for this figure.

Figure EV4. Co‐localization of mitochondria and MDVs as visualized by imaging flow cytometry

mCherry‐labeled mitochondria were preincubated in the presence or absence of GFP‐labeled MDVs, washed, and imaged by ImageStreamX mark II (Amnis, Part of Luminex, Au. TX). Each value represents the mean ± SD for n = 3 biological repeats. Significant differences were detected using t‐test, P‐value = 0.018. Mito, mitochondria; Ves, MDVs.Source data are available online for this figure.

Figure EV5. MDVs derived from HEK293‐isolated mitochondria

1. Nanoparticle tracking analysis of HEK293‐derived MDVs. Vesicles derived from HEK293‐isolated mitochondria were subjected to size distribution and concentration using nanoparticle tracking analysis (NTA).

2. TEM images of HEK293‐derived MDVs. Samples were stained with 2.0% uranyl acetate or 2.0% phosphotungstic acid. 5 μl of vesicles was placed on Formvar/carbon‐coated copper 200 mesh grids (EMS), mixed with 5 μl of PTA for 10–20 s, while excess stain was blotted off and grids were dried. Samples were examined with Jeol (Jem‐1400 Plus) transmission electron microscope. Scale bar—500 nm.

3. HEK293‐derived MDVs, produce ATP in an enzymatic reaction dependent on membrane potential. Mitochondrial‐derived vesicles from HEK293‐isolated mitochondria were incubated in the presence or absence of ADP, CCCP, or oligomycin for 15 min at 37°C. Subsequently, ATP was measured by luciferin‐luciferase luminometry. Each value represents the mean ± SD for n = 3 biological repeats. Significant differences were detected using t‐test, P‐value = 6.69E‐06/5.40E‐07 for oligomycin and CCCP, respectively.

Source data are available online for this figure.