DJ-1 acts in parallel to the PINK1/parkin pathway to control mitochondrial function and autophagy

Abstract

Mutations in DJ-1, PINK1 (PTEN-induced putative kinase 1) and parkin all cause recessive parkinsonism in humans, but the relationships between these genes are not clearly defined. One event associated with loss of any of these genes is altered mitochondrial function. Recent evidence suggests that turnover of damaged mitochondria by autophagy might be central to the process of recessive parkinsonism. Here, we show that loss of DJ-1 leads to loss of mitochondrial polarization, fragmentation of mitochondria and accumulation of markers of autophagy (LC3 punctae and lipidation) around mitochondria in human dopaminergic cells. These effects are due to endogenous oxidative stress, as antioxidants will reverse all of them. Similar to PINK1 and parkin, DJ-1 also limits mitochondrial fragmentation in response to the mitochondrial toxin rotenone. Furthermore, overexpressed parkin will protect against loss of DJ-1 and, although DJ-1 does not alter PINK1 mitochondrial phenotypes, DJ-1 is still active against rotenone-induced damage in the absence of PINK1. None of the three proteins complex together using size exclusion chromatography. These data suggest that DJ-1 works in parallel to the PINK1/parkin pathway to maintain mitochondrial function in the presence of an oxidative environment.

Figure 1.

Loss of DJ-1 results in multiple mitochondrial abnormalities. (A) M17 human dopaminergic neuroblastoma cells were stably transduced with control shRNA (lane 1) or DJ-1 shRNA (lane 2) and cell lysates blotted for DJ-1 (arrow). Markers on the right of the blot are in kilodaltons. (B) Δψ_m estimated in living cells using flow cytometry. Control cells (upper panel) or DJ-1 shRNA (lower panel) cell lines were stained with TMRE (x-axis) and DAPI for nuclei (y-axis). Flow cytometry was performed, and intensity of each signal for 10 000 cells was plotted and proportions of cells in each quadrant were calculated. (C) Δψ_m was estimated using live cell confocal imaging. After staining for TMRE, live cells were imaged and fluorescence pseudocolored to show differences in intensity. (D) Quantification of fluorescence in arbitrary units (a.u.) shows that DJ-1-deficient cells have significantly lower TMRE signal (P < 0.001 by t-test with Welsh's correction for unequal variances, n = 43–50 measurements per line). (E) Mitochondrial morphology was imaged in living cells transfected with mito-YFP plasmid. Mitochondria were elongated in control shRNA cells (upper panels) and at higher power (inset and right) showed a connected morphology. In DJ-1-deficient cells (lower panels), there were many small, disconnected mito-YFP-positive structures (inset and right). Scale bar is 2 μm. (F) Recovery of fluorescence was measured over a 12 s period after photobleaching a small region of mitochondria. Fluorescence recovery over time is plotted after normalization to both background fluorescence and non-photobleached mitochondrial fluorescence for 30 cells, showing decreased recovery of DJ-1-deficient cells (red symbols) compared with control lines (black symbols). Error bars show the SEM. (G) Mobile fraction of mito-YFP was lower in DJ-1-deficient mitochondria compared with controls. DJ-1-deficient cells had significantly lower FRAP for mito-YFP (P < 0.05 by t-test, n = 60 cells from duplicate experiments). (H) Control shRNA cells (upper panels) or DJ-1 shRNA cells (lower panels) either without treatment (left panels) or after treatment with 10 μm CCCP for 3 h (right panels) were transfected with LC3-GFP (green) and stained for the mitochondrial marker TOM20 (red). Arrows show the accumulation of LC3-GFP-positive punctatae near mitochondria in the cells. Scale bar is 10 μm. (I) Counts of LC3-GFP-positive punctatae per cell in n > 50 cells in three independent experiments showing the proportion of cells with less than 6 (open bars), between 6 and 15 (stippled bars), between 16 and 30 (striped bars) or more than 30 (filled bars) punctatae per cell, with error bars showing the SEM between experiments. Proportions of cells in each category were different by χ²-test comparing control and DJ-1 shRNA cells in each condition, untreated or with CCCP; ***P < 0.001; ns, not significant.

Figure 2.

Mouse DJ-1 rescues effects of human DJ-1 deficiency. (A) Control (lanes 1–6) or DJ-1-deficient cells (lanes 7–12) were either untransfected (lanes 1–3 and 7–9) or transfected with HA-tagged mouse DJ-1 (lanes 4–6 and 10–12). Protein extracts blotted with antibodies to HA for the tagged mouse DJ-1 construct (filled arrowhead, upper panel) endogenous human DJ-1 (arrow, middle panel) and β-actin as a loading control (open arrowhead, lower panel). Because mouse DJ-1 has a different cDNA sequence from human DJ-1, the construct is resistant to the shRNA. Markers on the right of the blots are in kilodaltons. (B) FRAP for mito-YFP was used to measure mitochondrial connectivity in control shRNA (white) or DJ-1 shRNA lines (red) alone or after transfection with mouse DJ-1 (striped bars). Each bar shows the mean mobile fraction from n = 60 individual cells measured over two independent experiments and error bars indicate the SEM. Statistical significance was calculated using one-way ANOVA with Newman–Keuls post hoc test; *P < 0.05; **P < 0.01; ns, not significant. (C) Autophagy was monitored using LC3-GFP transfection, counting the proportion of cells with varying numbers of LC3 punctatae per cell as indicated in the legend and comparing control and DJ-1-deficient cells with or without mouse DJ-1 transfection. Each bar is the counts from n = 50 cells from two to four experiments per cell line and the error bars indicate SEM between experiments.

Figure 3.

Endogenous oxidative stress contributes to mitochondrial phenotypes in DJ-1-deficient M17 human neuroblastoma cells. (A) ROS were measured in living cells using the fluorescent dye DFFDA. Under basal conditions, DJ-1-deficient cells (red boxes) had higher DFFDA signal than control lines (black boxes). DFFDA signals were lower, and differences between the lines were not seen when the cells were pretreated for 24 h with 100 μm GSH-EE prior to imaging. Two-way ANOVA was used to compare cell lines and treatments as separate factors and then Bonferroni post hoc tests were used to compare cell lines for each treatment; ***P < 0.001; ns, not significant. (B) TMRE and DAPI were used to estimate Δψ_m and nuclear integrity, respectively, and measured using flow cytometry. Controls (upper panels) had higher TMRE staining than DJ-1-deficient cells (lower panels). Treatment with GSH-EE (right pair of panels) increased the proportion of DJ-1-deficient cells with higher TMRE staining. (C) Cells were treated with GSH-EE and Δψ_m estimated by imaging of live cells using TMRE. DJ-1 shRNA lines (red) bars had lower TMRE signals compared with control lines (open bars) and this difference was diminished after treatment with GSH-EE. Two-way ANOVA was used to compare cell lines and treatments as separate factors and then Bonferroni post hoc tests were used to compare cell lines for each treatment; ***P < 0.001; ns, not significant. (D) Mitochondrial morphology in controls (upper panels) or DJ-1-deficient cells (lower panels) was imaged after transfection with plasmids encoding mitochondrially targeted YFP. Addition of GSH-EE caused a change in the morphology of mitochondria in DJ-1-deficient cells toward a more elongated, connected phenotype. Scale bar is 2 μm. (E) Quantitation of mitochondrial connectivity using FRAP shows a recovery of signal in the DJ-1-deficient cells (red boxes) after treatment with the glutathione precursor. Statistical significance was calculated using two-way ANOVA with Bonferroni post hoc tests to compare cell lines; ***P < 0.001; ns, not significant (n = 60 cells per condition from duplicate experiments). (F) Counts of LC3-GFP-positive punctatae as a marker of autophagy show that the difference between controls and DJ-1 shRNA lines is diminished after GSH-EE treatment. ***P < 0.001; ns, not significant by χ²-test, n = 3 experiments with >50 cells counted per experiment.

Figure 4.

Endogenous oxidative stress contributes to mitochondrial phenotypes in DJ-1-deficient MEFs. (A) ROS were measured in living cells using the fluorescent dye DFFDA. Under basal conditions, DJ-1 knockout MEFs (red boxes) had higher DFFDA signals than wild-type lines (black boxes). DFFDA signals were lower, and differences between the lines were not seen when the cells were pretreated for 24 h with 100 μm GSH-EE prior to imaging. Two-way ANOVA was used to compare cell lines and treatments as separate factors and then Bonferroni post hoc tests were used to compare cell lines for each treatment; ***P < 0.001; ns, not significant. (B) TMRE and DAPI were used to estimate mitochondrial membrane potential and nuclear integrity in knockout MEFs (lower panels). Treatment with GSH-EE (right pair of panels) increased the proportion of DJ-1 knockout MEFs with higher TMRE staining compared with controls. (C) Cells were treated with GSH-EE and mitochondrial membrane potential imaged using TMRE. DJ-1 knockout cells (red) bars had lower TMRE signals compared with wild-type lines (open bars) and this difference was diminished after treatment with GSH-EE. Two-way ANOVA was used to compare cell lines and treatments as separate factors and then Bonferroni post hoc tests were used to compare cell lines for each treatment; ***P < 0.001; ns, not significant.

Figure 5.

DJ-1-deficient cells have a defect in mitochondrial fusion. (A) Control shRNA (upper panels) or DJ-1 shRNA (lower panels) cells were transfected with a mitochondrially directed, photoactivatable GFP which was then photoactivated within a small region of interest. Sequential images of the single cells (0, 15, 30, 45 and 60 min as indicated above the images) after photoactivation are shown. Note that the fluorescence intensity is equal across the cell in the control line in the upper panel by 30 min, whereas the DJ-1-deficient cell retains areas of higher intensity. Scale bar is 5 μm. (B) Quantification of experiments as in (A) (three experiments, with n = 9–10 cells measured per time point per experiment; error bars indicate SEM) shows the loss of fluorescence over time for control cells (black) and DJ-1 shRNA (red) cell lines. The difference between cell lines was significant by two-way ANOVA (***P < 0.001). (C) Mito-YFP plasmid was co-transfected into DJ-1 and control shRNA cell lines with plasmids containing Opa1, Mfn1 or myc-tagged K38A Drp1. Scale bar is 2 μm. (D) Co-transfecting fusion proteins resulted in rescue of DJ-1 shRNA mitochondrial connectivity back to control levels and a dominant negative Drp1 K38A resulted in mobile fraction values higher than basal control cells. Statistical significance (n = 60 cells from duplicate experiments) was calculated using one-way ANOVA with Newman–Kuels post hoc test; ***P < 0.001.

Figure 6.

DJ-1, PINK1 and parkin each protect against mitochondrial fragmentation induced by mitochondrial complex I inhibition. (A) M17 cells were transfected with mito-YFP without (left panels) or with (right panels) wild-type human DJ-1. Cells were either left untreated (upper panels) or exposed to 100 nm rotenone for 24 h (lower panels) and mitochondria imaged. For each image, high-power views of mitochondria are shown on the right. Rotenone induces mitochondrial fragmentation that is prevented by co-expression of DJ-1. (B) To quantify this, we measured recovery of fluorescence over time for vector (black) or DJ-1 transfected (red) cells either in the absence (open symbols) or presence (closed symbols) of rotenone. Each data point is the average of 60 cells from two replicate experiments and error bars indicate the SEM. (C) Mobile fraction values were calculated from each cell and averages plotted (error bars indicate SEM) for vector and DJ-1 expressing cells in the absence (open bars) or presence (filled bars) of rotenone. Two-way ANOVA was used to compare cell lines and treatments as separate factors and then Bonferroni post hoc tests were used to compare cell lines for each treatment; ***P < 0.001; ns, not significant. Experiment similar to (C) using transient transfection with wild-type human PINK1 (D) or wild-type human parkin (E) in the absence (open bars) or presence (filled bars) of rotenone. Statistical tests were as above; ***P < 0.001; ns, not significant.

Figure 7.

The PINK1/parkin pathway rescues mitochondrial defects in DJ-1-deficient cells. (A) Mobile fraction from FRAP experiments shows rescue of loss of mitochondrial connectivity by parkin or PINK1 in DJ-1-deficient cells (red bars) or control shRNA cells (open bars). Statistical significance was calculated using one-way ANOVA with Newman–Keuls post hoc test; *P < 0.05, n = 30 cells per condition, representative of duplicate experiments. (B) Similar experiment in DJ-1 wild-type (open bars) or knockout (red bars) MEFs. Statistical significance was calculated using one-way ANOVA with Newman–Keuls post hoc test; ***P < 0.001, n = 30 cells per condition, representative of duplicate experiment. (C) Control (open bars) or PINK1-deficient (purple bars) M17 cells were transfected with DJ-1 and treated with 100 nm rotenone for 24 h as indicated. Statistical significance was calculated using one-way ANOVA with Newman–Keuls post hoc test; ***P < 0.001; **P < 0.01, n = 30 cells per condition, representative of duplicate experiments.

Figure 8.

Lack of formation of co-complexes of DJ-1 and PINK1/parkin. (A and B) Size exclusion chromatography was performed on cell lysates from M17 lines stably transduced with V5-tagged PINK1 and transfected with myc-parkin either without (A) or after treatment with CCCP (B). Fractions (0.25 ml) were taken and run on SDS–PAGE gels then blotted for V5 for PINK1 (upper blots), myc-parkin (middle blots) or DJ-1 (lower blots). Markers on the right indicate sizes on SDS–PAGE in kilodaltons. (C and D) Quantification of proteins in distinct native complexes for blots as in (A) and (B) for the preprotein of PINK1 (purple), the mature PINK1 protein (green), parkin (blue) or DJ-1 (black). For each protein, the immunoreactivity in each fraction is plotted as a percentage of the total protein immunoreactivity in all fractions against fraction number. Error bars indicate SEM from n = 3 experiments from independent transfections on different occasions.