Ubiquitination and receptor-mediated mitophagy converge to eliminate oxidation-damaged mitochondria during hypoxia

Abstract

The contribution of the Ubiquitin-Proteasome System (UPS) to mitophagy has been largely attributed to the E3 ubiquitin ligase Parkin. Here we show that in response to the oxidative stress associated with hypoxia or the hypoxia mimic CoCl₂, the damaged and fragmented mitochondria are removed by Parkin-independent mitophagy. Mitochondria isolated from hypoxia or CoCl₂-treated cells exhibited extensive ubiquitination, predominantly Lysine 48-linked and involves the degradation of key mitochondrial proteins such as the mitofusins MFN1/2, or the import channel component TOM20. Reflecting the critical role of mitochondrial protein degradation, proteasome inhibition blocked CoCl₂-induced mitophagy. The five conserved ubiquitin-binding autophagy receptors (p62, NDP52, Optineurin, NBR1, TAX1BP1) were dispensable for the ensuing mitophagy, suggesting that the mitophagy step itself was independent of ubiquitination. Instead, the expression of two ubiquitin-independent mitophagy receptor proteins BNIP3 and NIX was induced by hypoxia or CoCl₂-treatment followed by their recruitment to the oxidation-damaged mitochondria. By employing BNIP3/NIX double knockout and DRP1-null cell lines, we confirmed that mitochondrial clearance relies on DRP1-dependent mitochondrial fragmentation and BNIP3/NIX-mediated mitophagy. General antioxidants such as N-Acetyl Cysteine (NAC) or the mitochondria-specific Mitoquinone prevented HIF-1α stabilization, ameliorated hypoxia-related mitochondrial oxidative stress, and suppressed mitophagy. We conclude that the UPS and receptor-mediated autophagy converge to eliminate oxidation-damaged mitochondria.

Keywords: HIF-1α; Hypoxia; Mitochondria; Mitophagy; Oxidative stress; Proteasome; Ubiquitin.

Graphical abstract

Fig. 1

The hypoxia mimic CoCl₂ induces mitochondrial oxidative stress and mitophagy in HeLa cells. The cells were treated with 500 μM CoCl₂ in combination with 25 μM Z-VAD-FMK for 24 h unless otherwise indicated. Additionally, the cells were pretreated with 10 mM N-Acetyl Cysteine for 2 h before treatment with CoCl₂ wherever indicated. A. HeLa cells were treated with 500 μM CoCl₂ alone for the indicated time points. The whole cell lysates were subjected to immunoblotting to analyze the expression of HIF-1α, LC-3B, cleaved caspase-3 and GAPDH (as a loading control). B. The mitochondrial respiration measurements were performed on Seahorse Extracellular Flux analyzer. Changes in the oxygen consumption rate (OCR) were measured by sequential injections of Oligomycin, FCCP and equimolar concentration of Rotenone + Antimycin A. The OCR values were normalized by protein concentration per well determined by Bradford assay at the end of experiment. C. The extent of mitochondrial fragmentation and mitochondrial ROS levels were determined by live cell imaging dyes Mitotracker Green and MitoSOX red using a confocal microscope and represented as fold change relative to control. D. Evaluation of functional mitophagy by confocal microscopy-based Mitokeima assay. The extent of mitophagy was represented as mitophagy index. E. Alternatively, the extent of mitophagy was determined by mitokeima-based FACS assay and expressed as percent cells showing the red mitokeima fluorescence. F. The control and CoCl₂ treated whole cell extracts were analyzed for the expression of the indicated mitochondrial proteins with GAPDH as a loading control. The blots are representative of three independent experiments. The data is expressed as mean + SEM of three independent experiments. Scale bar: 10 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 2

Genuine hypoxia causes mitochondrial oxidative stress leading to their ubiquitination. HeLa cells were exposed to 1% Oxygen for 24 h (Hypoxia) or 21% Oxygen (Normoxia). A. The cell lysates corresponding to the indicated time points following hypoxia were analyzed for detection of HIF-1α and GAPDH by western blotting. B. The cells were immunostained with HSP60 antibody followed by image acquisition on a confocal microscope and the percent mitochondrial fragmentation was determined by evaluating a minimum of 100 cells across 10 different microscopic fields. C. The Oxygen Consumption Rate (OCR) was determined by Seahorse Extracellular Flux analyzer using the MitoStress test kit. D. The cells were stained with live cell imaging dyes Mitotracker Green and MitoSOX red and observed under a confocal microscope. The MitoSOX Red staining intensity was expressed in terms of fold change relative to the control. At least 100 cells across 10 different microscopic fields per condition were evaluated. E. The cells were exposed to normoxia and hypoxia alone or in combination with 1 μM Mitoquinone Q (MitoQ) or vehicle control (VC) and whole cell extracts were subjected to immunoblotting for the detection of HIF-1α, PHD2 and Actin (loading control). F. The cells were stained with MitoTracker Red followed by immunostaining with LC-3B antibody. G. The whole cell lysates of HeLa cells exposed to Normoxia (N) and Hypoxia (H) were subjected to immunoblotting for detection of the indicated mitochondrial proteins. H. The isolated mitochondrial fractions from control and CoCl₂ treated cells (left) and normoxic and hypoxic cells (right) were subjected to western blotting for evaluating the level of total ubiquitin and ATP5a (loading control). I. Wild type HeLa cells were stained with Mitotracker Red followed by immunostaining with Ubiquitin and LC-3B antibodies. The numbered circles mark their corresponding position in the individual channels and in the merged image to indicate the extent of their overlap. Scale bar for microscopy images: 10 μm. The blots are representative of three experiments. The data is expressed as mean + SEM of three experiments. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

Oxidative stress causes extensive mitochondrial ubiquitination but mitophagy occurs independent of ubiquitin-binding autophagy receptors. HeLa cells were treated with 500 μM CoCl₂ along with 25 μM Z-VAD-FMK for 24 h. A. The cells were either treated with CoCl₂ alone or in combination with NAC. The isolated mitochondrial fractions were analyzed for the extent of total ubiquitin, K48 and K63-linked polyubiquitin by immunoblotting. B. The cells were stained with Mitotracker Red followed by immunostaining with LC-3B and Ubiquitin antibodies. Panels on the right display co-localization between mitochondria, ubiquitin and LC-3B by fluorescence intensity line measurement (yellow line). Scale bar: 10 μm. C. The whole cell extracts of CoCl₂ treated or control HeLa WT and 5KO cells were analyzed for HIF-1α, MFN2 and tubulin (loading control) by immunoblotting. Alternatively, isolated mitochondrial fractions from these cells were subjected to western blotting for detection of total ubiquitination and HSP60 (loading control). D. Control or CoCl₂ treated WT and 5KO HeLa cells stably expressing Mitokeima were analyzed by FACS for mitophagy quantification (ns: not significant). Data is shown as mean + SEM of three experiments. E. Control, CoCl₂ and CCCP (10 μM for 6 h) treated HeLa cells overexpressing Parkin were immunostained with TOM20 and OPTN or NDP52 antibodies. Scale bar: 10 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4

BNIP3 and NIX proteins drive mitophagy of oxidatively damaged mitochondria. A. Whole cell extracts of CoCl₂ (500 μM CoCl₂ for 6 h) and hypoxia-treated (1% O₂ for 24 h) WT HeLa cells were analyzed for the expression of HIF-1α, BNIP3, NIX, GAPDH and Tubulin proteins by immunoblotting. B. The expression of BNIP3 and NIX transcripts in response to CoCl₂ treatment (500 μM CoCl₂ for 6 h) was analyzed by SYBR Green real time PCR relative to GAPDH and expressed as 2^−ΔCt. WT HeLa cells treated with 500 μM CoCl₂ plus 25 μM Z-VAD-FMK for 16 h were stained with Mitotracker Red followed by immunostaining with BNIP3 (C) and NIX (D). E. Following the indicated treatments for 24 h, the whole cell extracts of WT HeLa cells were analyzed for HIF-1α, BNIP3, NIX and Tubulin proteins by western blotting. F. The mitochondrial bioenergetics were studied by Seahorse flux analyzer. G. The whole cell extracts of control and CoCl₂-treated WT and DRP1 KO HeLa cells were analyzed for DRP1, HIF-1α, MFN2, TOM20, BNIP3, NIX, total ubiquitin and GAPDH proteins (loading control) by immunoblotting. H. The control and CoCl₂-treated WT and DRP1 KO HeLa cells were stained with MitoSOX red to evaluate the extent of ROS generated in response to CoCl₂-induced oxidative stress and the MitoSOX red staining intensity was expressed as fold change relative to the control. Scale bar for microscopy images: 10 μm. Data represented as Mean + SEM of three experiments. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

Mitochondrial ROS generation and ubiquitination occurs independent of DRP1-dependent fragmentation. HeLa WT and DRP1 KO cells were either treated with 500 μM CoCl₂ plus 25 μM Z-VAD-FMK for 24 h or left untreated (control). A. The cells were stained with Mitotracker Red dye followed by immunostaining with ubiquitin and LC-3B antibodies. Right panels indicate semiquantitative co-localization analysis between mitochondria, ubiquitin and LC-3B by fluorescence intensity line measurement (blue line). Scale bar: 10 μm. B. The whole cell extracts of WT and BNIP3-NIX double knockout (DKO) HeLa cells were analyzed for HIF-1α, NIX, BNIP3, MFN2, TOM20 and Tubulin proteins by immunoblotting. C. The mitophagy analysis was performed by mitokeima FACS analysis of WT, DRP1 KO and DKO cells. D. The enriched mitochondrial fractions from WT, DRP1 KO and DKO cells were analyzed for total ubiquitin levels by immunoblotting with HSP60 being the loading control. The data is representative of three experiments and expressed as mean + SEM. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6

Proteasome inhibition blocks CoCl₂-induced mitophagy. HeLa WT cells were treated with 500 μM CoCl₂ plus 25 μM Z-VAD-FMK either alone or in combination with NAC (10 mM; pretreatment for 2 h), mDivi-1 (20 μM; throughout 24 h), Velcade (500 nM; for the last 12 h) and Chloroquine (50 μM; for the last 6 h). A. The whole cell extracts were analyzed by immunoblotting for the expression of HIF-1α, TOM20, MFN2, DRP1, LC-3B and GAPDH (loading control). The normalized expression of TOM20 and MFN2 relative to GAPDH is indicated. Comparing the ratio of LC-3BII/LC-3BI indicates that CoCl₂ exhibits an active flux of autophagy with reference to the lysosomal (autophagy) inhibitor Chloroquine. B. The analysis of functional mitophagy was performed by Mitokeima FACS assay. C. Mitokeima analysis by microscopy and the extent of mitophagy is indicated as mitophagy index. The data is represented as mean + SEM of three experiments. D. The proposed model of oxidative stress-induced mitophagy depicts ROS-mediated stabilization of HIF-1α, mitochondrial ubiquitination followed by fragmentation and subsequent expression of ubiquitin-independent mitophagy receptors BNIP3 and NIX.