Myosin VI-Dependent Actin Cages Encapsulate Parkin-Positive Damaged Mitochondria

Abstract

Mitochondrial quality control is essential to maintain cellular homeostasis and is achieved by removing damaged, ubiquitinated mitochondria via Parkin-mediated mitophagy. Here, we demonstrate that MYO6 (myosin VI), a unique myosin that moves toward the minus end of actin filaments, forms a complex with Parkin and is selectively recruited to damaged mitochondria via its ubiquitin-binding domain. This myosin motor initiates the assembly of F-actin cages to encapsulate damaged mitochondria by forming a physical barrier that prevents refusion with neighboring populations. Loss of MYO6 results in an accumulation of mitophagosomes and an increase in mitochondrial mass. In addition, we observe downstream mitochondrial dysfunction manifesting as reduced respiratory capacity and decreased ability to rely on oxidative phosphorylation for energy production. Our work uncovers a crucial step in mitochondrial quality control: the formation of MYO6-dependent actin cages that ensure isolation of damaged mitochondria from the network.

Keywords: MYO6; NDP52; OPTN; Parkin; TAX1BP1; actin; mitochondrial quality control; mitophagy; myosin VI.

Graphical abstract

Figure 1

Endogenous and GFP-Tagged MYO6 Are Recruited to Damaged Mitochondria and Form a Complex with Parkin (A) HA-Parkin-expressing HEK293 cells were treated for 2 h with 10 μM CCCP or left untreated. Images were acquired by superresolution structured illumination microscopy (SR-SIM) after staining for endogenous MYO6, HA to detect Parkin, and cytochrome c (Cyt c) to visualize mitochondria. (B) Quantitation of the percentage of cells with endogenous MYO6 on Cyt c-labeled mitochondria from (A) by widefield microscopy. Data are represented as mean ± SEM. Two-tailed unpaired Student's t test, ^∗∗∗p < 0.001, n = 3 (≥427 cells per condition). (C) Line profile of MYO6- and Parkin-positive mitochondrion along the white line indicated in (A). (D) HEK293 cells stably expressing HA-Parkin transiently transfected with full-length (FL) GFP-MYO6 were left untreated or incubated for 2 h with 10 μM CCCP. Images were acquired by SR-SIM after staining for the GFP tag on MYO6, HA to detect Parkin, and TOMM20 to label the outer mitochondrial membrane. (E) Line profile of MYO6- and Parkin-positive mitochondrion along the white line indicated in (D). (F) Parkin was immunoprecipitated using antibodies either against the HA tag or Parkin from HA-Parkin-expressing HEK293 cells incubated for 1 h with 10 μM CCCP or left untreated. The inputs, control immunoglobulin G (IgG) immunoprecipitation (IP), and HA/Parkin IPs were immunoblotted for Parkin as well as co-immunoprecipitation of endogenous MYO6. Actin is shown as a loading control. (G) Endogenous MYO6 was immunoprecipitated from HA-Parkin-expressing HEK293 cells incubated for 1 h with 10 μM CCCP or left untreated. The inputs, IgG control IP, and MYO6 IPs were immunoblotted for MYO6 as well as co-immunoprecipitation of Parkin. Actin is shown as a loading control. Images in (A), (D), (F), and (G) are representative of three independent experiments. See also Figure S1.

Figure 2

MYO6 Binding to Ubiquitin, but Not to the Autophagy Receptors, Is Required for Parkin-Dependent Recruitment to Damaged Mitochondria (A) Illustration of MYO6 domain organization: catalytic motor domain (red), unique insert “reverse gear” (gray), IQ calmodulin-binding motif (orange), and a cargo-binding domain (CBD) in the tail region (blue). Enlarged is the CBD containing two protein-protein interaction motifs (RRL, green; WWY, purple) and ubiquitin-binding domains (motif interacting with ubiquitin [MIU], yellow; MYO6 ubiquitin-binding domain [MyUb], green). Amino acid residues of single point mutations in different domains are highlighted. (B) Quantitation of the degree of colocalization between the GFP-MYO6 constructs (wild-type [WT] or the indicated mutants) and TOMM20-labeled mitochondria in HeLaM cells stably expressing HA-Parkin left untreated or incubated for 2 h with 10 μM CCCP by determining the Pearson's correlation coefficient of confocal microscopy images. Data are represented as mean ± SEM. One-way ANOVA with post-hoc Bonferroni correction, ^∗∗∗p < 0.001, n ≥ 3 (≥50 cells per condition). (C) HeLaM cells were treated with siRNA against TAX1BP1, NDP52, and OPTN (TNO), then transiently transfected with GFP-MYO6 FL as well as mCherry-Parkin and incubated for 2 h with 10 μM CCCP. Images were acquired by confocal microscopy after staining the GFP tag on MYO6, the mCherry tag on Parkin with a DsRed antibody, and Cyt c to visualize mitochondria. (D) Western blot analysis of lysates corresponding to (C) confirming depletion of TAX1BP1, NDP52, and OPTN, as well as overexpression of GFP-MYO6 and mCherry-Parkin. α-Tubulin is shown as a loading control. Images in (C) and (D) are representative of three independent experiments. (E) Quantitation of the degree of colocalization between GFP-MYO6 and Cyt c-labeled mitochondria by determining the Pearson's correlation coefficient of confocal microscopy images from (C). Data are represented as mean ± SEM. Two-tailed paired Student's t test, ^∗p < 0.05, n = 3 (≥100 cells per condition). (F) Competition assay where GST-tagged MYO6 CBD (blue) and K63 tetra-ubiquitin (Ub⁴) chains (yellow) were pre-incubated with glutathione (GSH) sepharose (purple) and then incubated with increasing amounts of His-tagged TAX1BP1 (C-terminal half, pink). The samples were analyzed by immunoblotting with antibodies against GST to detect MYO6, bound ubiquitin (P4D1), and bound TAX1BP1. As controls, GST-MYO6 CBD, K63 Ub⁴, and His-TAX1BP1 were individually incubated with GSH sepharose to determine if there was any non-specific binding. GST-MYO6 CBD and His-TAX1BP1 were also incubated together to demonstrate direct binding. Inputs of K63 Ub⁴ and His-TAX1BP1 (TAX1BP1 input) are also shown. Images are representative of four independent experiments. See also Figures S2–S4.

Figure 3

Time Course of MYO6 and F-Actin Assembly around Damaged Mitochondria (A) HEK293 cells stably expressing HA-Parkin were left untreated or incubated for 2 h with 10 μM CCCP. Images were acquired by SR-SIM after staining for endogenous MYO6, F-actin was visualized with phalloidin, and HA to detect Parkin. (B) Line profile of MYO6- and Parkin-positive mitochondrion that is F-actin-positive along the white line indicated in (A). (C) HA-Parkin-expressing HEK293 cells were left untreated or incubated for 5 min and 2 h with 10 μM CCCP. Images were acquired by SR-SIM after staining for endogenous MYO6, F-actin was visualized with phalloidin, and Cyt c as a mitochondrial marker. (D and E) Line profiles of F-actin-positive mitochondria along the white lines indicated in (C) that are MYO6-negative after 5 min CCCP treatment (D) or MYO6-positive after 2 h CCCP incubation (E). (F) Quantitation of the percentage of cells with endogenous MYO6 or F-actin on Cyt c-labeled mitochondria from (C) at the indicated time points by widefield microscopy. Data are represented as mean ± SEM, n = 3 (≥299 cells per condition). (G and H) Quantitation of the percentage of HA-Parkin-expressing HeLaM cells transiently transfected with GFP-MYO6, either FL or tail, incubated for 5 min with 10 μM CCCP (Figure S5C) that have F-actin on mitochondria (G) and fragmented Cyt c-labeled mitochondria (H) by widefield microscopy. Data are represented as mean ± SEM. Two-tailed paired Student's t test, ns, not significant, n = 3 (≥213 cells per condition). Images in (A) and (C) are representative of three independent experiments. See also Figure S5.

Figure 4

MYO6 and Actin Regulators/Nucleators Mediate the Assembly of F-Actin Cages around Damaged Mitochondria (A and B) HEK293 cells stably expressing HA-Parkin were incubated for 2 h with 10 μM CCCP and inhibitors of actin regulators Rho (0.5 μg/mL Rho inhibitor I, RhoI, or 50 μM Rhosin), Rac1 (100 μM NSC23766, 100 μM W56, or 10 μM EHT 1864), or cdc42 (20 μM ML141), and actin nucleators Arp2/3 complex (100 μM CK666), formins (20 μM SMIFH2), or N-WASP (5 μM Wiskostatin) (Figure S6). Quantitation of the percentage of cells with F-actin (A) and endogenous MYO6 (B) on mitochondria by widefield microscopy. Data are represented as mean ± SEM. One-way ANOVA with post-hoc Bonferroni correction, ^∗∗∗p < 0.001, ns, not significant, n ≥ 3 (≥339 cells per condition). (C) HeLaM cells stably expressing HA-Parkin transiently transfected with GFP-MYO6, either FL or tail, were incubated for 2 h with 10 μM CCCP. Images were acquired by SR-SIM after staining for the GFP tag on MYO6, with phalloidin to visualize F-actin, and Cyt c as a mitochondrial marker. Images are representative of three independent experiments. (D and E) Line profiles of GFP-MYO6-positive mitochondria along the white lines indicated in (C) that are actin-positive in the case of GFP-MYO6 FL (D) or actin-negative for the tail (E). (F) Quantitation of the percentage of HA-Parkin-expressing HeLaM cells expressing GFP-MYO6, either FL WT, tail, or the indicated mutants, incubated for 2 h with 10 μM CCCP that have F-actin on mitochondria by widefield microscopy. Data are represented as mean ± SEM. One-way repeated measures ANOVA with post-hoc Bonferroni correction, ^∗∗∗p < 0.001, ns, not significant, n = 3 (≥241 cells per condition). See also Figure S6.

Figure 5

F-Actin Cages around Mitochondria Restrict Fragment Size and Lead to a Reduced Refusion Rate (A) HEK293 cells with endogenous Parkin were left untreated or incubated for 6 h with 10 μM CCCP. Images were acquired by SR-SIM after staining F-actin with phalloidin and Cyt c to label mitochondria. Images are representative of three independent experiments. (B) Line profile of F-actin-positive mitochondrion along the white line in (A). (C) Quantitation of the area of mitochondria that are actin positive (+) or negative (–) in regions of interest from cells imaged in (A). Data are represented as mean ± SEM. Two-tailed paired Student's t test, ^∗∗p < 0.01, n = 17 from three independent experiments. (D) Quantitation of the area of mitochondria in regions of interest with GFP-MYO6 FL (with actin) or tail (without actin) recruitment in HA-Parkin-expressing HeLaM cells treated for 2 h with 10 μM CCCP and imaged in Figure 4C. Data are represented as mean ± SEM. Two-tailed unpaired Student's t test, ^∗p < 0.05, n = 18 from three independent experiments. (E) HA-Parkin-expressing HEK293 cells were treated for 2 h with 10 μM CCCP alone or with the addition of 100 μM CK666. After washout, cells were fixed every hour for 3 h, stained with Cyt c to visualize mitochondria and DAPI to label nuclei, and imaged by widefield microscopy. Images were scored according to three categories of mitochondrial morphology: clustered, refusing, and network (representative images taken by confocal microscopy). Data are represented as means ± SEM. Two-way repeated measures ANOVA with post-hoc Bonferroni correction, ^∗p < 0.05, ^∗∗p < 0.01, n = 3 (≥531 cells per time point). (F) Schematic model on the role of MYO6 and F-actin: after a mitochondrial insult, the network fragments and Parkin (orange) is selectively recruited to damaged mitochondria, where it ubiquitinates outer mitochondrial membrane proteins. Subsequently, MYO6 (black) is recruited to damaged mitochondria by binding to ubiquitin chains (yellow), and damaged mitochondria are isolated in an F-actin cage (red) from the neighboring population. F-actin cages around damaged mitochondria serve as a barrier to prevent refusion with the mitochondrial network and require MYO6 as well as several actin regulators including cdc42, Arp2/3 complex, formins, and N-WASP.

Figure 6

Loss of MYO6 Leads to an Accumulation of Damaged Mitochondria in Autophagosomes after Mitophagy Induction (A) Immortalized wild-type (Myo6^+/+) and Snell's waltzer (Myo6^sv/sv) mouse embryonic fibroblasts (MEFs) were transiently transfected with HA-Parkin and treated for 12 or 24 h with 20 μM CCCP. Cells were imaged by confocal microscopy, processed for correlative light electron microscopy, and imaged by transmission electron microscopy. The star indicates ruptured mitochondria and the arrow shows the phagophore. Scale bars, 500 nm in subpanels (a, b, e, and f), 1 μm in subpanels (c and d). (B) Quantitation of the number of autophagosomes containing mitochondria (white) or endoplasmic reticulum and cytoplasmic material (black) in (A). At least 19 randomly selected areas were counted per cell type. (C) Western blot analysis confirming the absence of MYO6 from Snell's waltzer immortalized MEFs compared with wild-type lysates. Actin is shown as a loading control. See also Figure S7.

Figure 7

Cells Lacking MYO6 Accumulate Dysfunctional Mitochondria (A) The mitochondrial mass of wild-type (Myo6^+/+) and Snell's waltzer (Myo6^sv/sv) immortalized MEFs was measured using the MitoTracker Green FM dye by flow cytometry. Data are represented as mean ± SEM. Two-tailed paired Student's t test, ^∗p < 0.05, n = 5. (B) The oxygen consumption rate (OCR), an indicator of mitochondrial respiration, of wild-type and Snell's waltzer immortalized MEFs was measured using the XF^e24 Extracellular Flux Analyser. Oligomycin (1 μM), FCCP (0.75 μM), and rotenone (1 μM) were injected at the indicated times to determine the oxygen consumption rate at regular intervals. Normalized to rotenone as baseline. Data are represented as mean ± SEM, n = 3. (C) Respiratory control ratio ([RCR]; OCR after FCCP addition divided by oligomycin) in wild-type and Snell's waltzer immortalized MEFs. Data are represented as mean ± SEM. Two-tailed paired Students t test, ^∗p < 0.05, n = 3. (D–I) HEK293 cells stably expressing HA-Parkin (D and E) or HEK293 cells (G and H) were depleted of MYO6 by siRNA transfection and growth curves in media containing (D and G) glucose (GLU) or (E and H) galactose (GAL) were obtained by quantitative live-cell phase contrast imaging. Data are represented as mean ± SEM, n = 3. (F) Representative western blot analysis of lysates from (D) and (E) confirming depletion of MYO6 and similar expression levels of Parkin. Actin is shown as a loading control. (I) Representative western blot analysis of lysates from (G) and (H) confirming depletion of MYO6 using actin as a loading control.