Vulnerable Parkin Loss-of-Function Drosophila Dopaminergic Neurons Have Advanced Mitochondrial Aging, Mitochondrial Network Loss and Transiently Reduced Autophagosome Recruitment

Abstract

Selective degeneration of substantia nigra dopaminergic (DA) neurons is a hallmark pathology of familial Parkinson's disease (PD). While the mechanism of degeneration is elusive, abnormalities in mitochondrial function and turnover are strongly implicated. An Autosomal Recessive-Juvenile Parkinsonism (AR-JP) Drosophila melanogaster model exhibits DA neurodegeneration as well as aberrant mitochondrial dynamics and function. Disruptions in mitophagy have been observed in parkin loss-of-function models, and changes in mitochondrial respiration have been reported in patient fibroblasts. Whether loss of parkin causes selective DA neurodegeneration in vivo as a result of lost or decreased mitophagy is unknown. This study employs the use of fluorescent constructs expressed in Drosophila DA neurons that are functionally homologous to those of the mammalian substantia nigra. We provide evidence that degenerating DA neurons in parkin loss-of-function mutant flies have advanced mitochondrial aging, and that mitochondrial networks are fragmented and contain swollen organelles. We also found that mitophagy initiation is decreased in park (Drosophila parkin/PARK2 ortholog) homozygous mutants, but autophagosome formation is unaffected, and mitochondrial network volumes are decreased. As the fly ages, autophagosome recruitment becomes similar to control, while mitochondria continue to show signs of damage, and climbing deficits persist. Interestingly, aberrant mitochondrial morphology, aging and mitophagy initiation were not observed in DA neurons that do not degenerate. Our results suggest that parkin is important for mitochondrial homeostasis in vulnerable Drosophila DA neurons, and that loss of parkin-mediated mitophagy may play a role in degeneration of relevant DA neurons or motor deficits in this model.

Keywords: MitoTimer; Parkinson’s disease; autophagy; dopaminergic neurons; mitochondria; mitochondrial dynamics; neurodegeneration; parkin.

Figure 1

Parkin loss-of-function flies have increased mitochondrial fragmentation and swelling in protocerebral posterior lateral region 1 (PPL1) DA neurons. Brains of park mutant flies expressing the mitochondrially targeted green fluorescent protein (mitoGFP) construct in tyrosine hydroxylase (TH)-producing cells were dissected, stained for TH (blue), and fixed on days 5, 10 or 20 post eclosion (PE). Images in (A) are representative intensity sums of Z-stacks used as 3D projections for the park control (park^+/+), heterozygote (park^+/−) and homozygote (park^−/−) conditions on day 20. Green (mitoGFP) labeling in the bottom row of (A) represents 3D isosurfaces that are generated based on signal above an automatic bright threshold. Digital image enhancement steps were standardized for green fluorophores. After using standardized image capture parameters, we used Image Pro Premier 3D image analysis software to measure and categorize the mitochondrial object size (B) and inverse of sphericity ([6V]/DS) (D) of mitoGFP objects within the TH-labeled regions for one PPL1 region per brain. Mitochondrion size frequency distribution histograms were generated by placing objects into small, medium and large bins based on volume (C). Bin size was determined by taking the difference of the largest and smallest data point and dividing it by three. The angled numbers along the x-axis indicate the bin center, or the value that falls halfway between the smallest and largest values of its bin. We also measured the inverse of sphericity for the largest five percent of objects (E). Green arrows indicate large, interconnected (not swollen) mitochondria. White and yellow arrows indicate fragmented and large, swollen mitochondria, respectively. Numbers in histogram bars indicate the number of PPL1 regions analyzed. For *P < 0.05; for **P < 0.01; for ***P < 0.001; for ****P < 0.0001. Error bars represent standard error of the mean; scale bar represents five microns.

Figure 2

Non-degenerating DA neurons in parkin loss-of-function flies do not have fragmented or swollen mitochondria networks. Brains of park mutant flies expressing the mitoGFP construct in TH-producing cells were dissected, stained for TH (blue), and fixed on days 5, 10 or 20 PE. Images in (A) are representative intensity sums of Z-stacks used as 3D projections for the park^+/+ and park^−/− conditions on day 20. Green (mitoGFP) labeling in the bottom row of (A) represents 3D isosurfaces that are generated based on signal above an automatic bright threshold. Digital image enhancement steps were standardized for green fluorophores. After using standardized image capture parameters, we used Image Pro Premier 3D image analysis software to measure and categorize the mitochondrial object size (B) and inverse of sphericity ([6V]/DS) (C) of mitoGFP objects within the TH-labeled regions for one PPM3 region per brain. Numbers in histogram bars indicate the number of regions analyzed. Error bars represent standard error of the mean; scale bar represents five microns.

Figure 3

Homozygous parkin loss-of-function flies have increased aged MitoTimer protein levels in PPL1 DA neurons. Brains of park mutant flies expressing the MitoTimer age reporter construct in TH-producing cells were dissected, stained for TH (blue), and fixed on days 5, 10 or 20 PE. Using standardized image capture and digital image analysis parameters, we measured the total volume of aged MitoTimer (red; A), newly synthesized MitoTimer (green; B), and the ratio (C) of aged to new MitoTimer within the TH-labeled region for one PPL1 region per brain. Images in (D) are digital sums of representative z-stacks for the park^−/−, park^+/−, and park^+/+ conditions on day 20. Digital image enhancement steps were standardized for red and green fluorophores in (D). Numbers in histogram bars indicate the number of PPL1 regions analyzed. For *P < 0.05; for **P < 0.01; for ****P < 0.0001. Error bars represent standard error of the mean; scale bar represents five microns.

Figure 4

Homozygous parkin loss-of-function non-degenerating DA neurons show signs of normal mitochondrial aging. Brains of park mutant flies expressing the MitoTimer age reporter construct in TH-producing cells were dissected, stained for TH (blue), and fixed on days 5, 10 or 20 PE. Using standardized image capture and digital image analysis parameters, we measured the total volume of aged MitoTimer (red) within the TH-labeled region for one PPM3 region per brain (A,B). Images in (A) are digital sums of representative z-stacks for the park^−/−, park^+/−, and park^+/+ conditions on day 20. Digital image enhancement steps were standardized for red fluorophores. Numbers in histogram bars indicate the number of PPM3 regions analyzed. Error bars represent standard error of the mean; scale bar represents five microns.

Figure 5

Levels of active and inactive autophagy 8a (atg8a) are unchanged in park mutant fly heads. Ten microgram of homogenized park^−/−, park^+/−, or park^+/+ fly brain fractions per lane was transferred to a nitrocellulose membrane, which was incubated in REVERT^™ protein stain and imaged using an automatic bright threshold. The REVERT™ stain was washed, and the membrane was probed with a rabbit anti-gamma amino butyric acid receptor associated protein (GABARAP) primary antibody (recognizing Drosophila atg8a) followed by an anti-rabbit secondary antibody with an infrared fluorescent conjugate (IRDye^® 800CW, LICOR) (A). Total protein was determined by measuring REVERT™ fluorescence between 75 kDa and 37 k Da for each lane (B, black boxes). Fluorescence signal for each lane was divided by the signal for the lane that had the highest intensity. For atg8a measurements, the signals from the inactive 14 kDa and the active 12 kDa bands were divided by the corresponding signal for the band that had the highest intensity. To control for loading, these values were divided by the relative REVERT™ total protein value for each lane (Eaton et al., 2013). There was no effect of the park mutation on total or activated atg8a levels (C). Each data point is an average of three technical replicates for one collection day; lysates were generated on four different collection days (indicated in histogram bars).

Figure 6

Exogenous expression of atg8a in TH-producing neurons does not ameliorate DA degeneration or climbing deficits in park mutant flies. Brains of flies expressing the mitoGFP and mCherry-atg8a constructs in TH-producing cells were dissected, stained for TH (blue), and fixed on days 5, 10 or 20 PE. Images in (A) are digital sums of representative z-stacks for day 20 park^+/+, park^+/−, park^−/− PPL1; digital image enhancement steps were standardized for the mCherry (red) fluorophore. Using standardized image capture and digital image analysis parameters, we identified and counted the number of atg8a puncta (white arrows) within TH-labeled regions and divided that value by the number of TH-positive cells for one PPL1 region per brain (B). We also counted the number of TH-positive cells per cluster (C). (D) Climbing assays were performed by placing individual flies expressing mitoGFP and mCherry-atg8a into polycarbonate tubes; the fly’s position in the tube was recorded each second for 20 min. The total distance climbed during the 20-min recording period was measured, and divided by number of climbs to calculate the average height climbed. A receiver operating characteristic (ROC) curve was generated from mitoGFP, mCherry-atg8a control fly climbing data to distinguish “climbing” from “non-climbing” flies. Numbers in histogram bars indicate the number of PPL1 regions analyzed for (B,C); for (D), number in histogram bars represent the number of flies tested. For ** and ^bbP < 0.01; for ^###P < 0.001; ^####P < 0.0001; for **** and ^aaaaP < 0.0001. Error bars represent standard error of the mean; scale bar represents five microns.

Figure 7

Autophagosome recruitment to mitochondria is selectively and transiently reduced in park mutant PPL1. Brains of park mutant flies expressing the mitoGFP and mCherry-atg8a constructs in TH-producing cells were dissected, stained for TH (blue), and fixed on days 5, 10 or 20 PE. Red and green objects in (A) represent 3D isosurfaces that are generated based on signal above background for representative z-stacks of the park^−/− and park^+/+ conditions on day 10. TH signal is represented as a summary of the z-stack. Insets highlight mitochondria that are colocalized with atg8a (A). Using standardized image capture and digital image processing analysis, we identified and counted the number of atg8a puncta that colocalized with mitoGFP according to a positive Pearson’s correlation coefficient within the TH-labeled regions for one PPL1 and one PPM3 region per brain (B,D, respectively). Because there was a decrease in PPL1 neurons in park^−/− flies and no change in colocalizations per cell on day 20, we divided the number of colocalizations by the number of TH-positive (DA) neurons for the corresponding PPL1 cluster in day 20 park^+/+ and park^−/− (C). Numbers in histogram bars indicate the number of DA regions analyzed. For **P < 0.01. Error bars represent standard error of the mean; scale bars represent five microns.

Figure 8

Mitochondrial network volume is decreased in park mutant fly DA cells. Brains of park mutant flies expressing only the mitoGFP construct in TH-producing cells were dissected, stained for TH (blue), and fixed on days 5, 10 or 20 PE. Using standardized image capture and digital image analysis, we took the sum of the volume of all mitoGFP objects fluorescing above an automatically generated brightness threshold that were within the TH-labeled region for one cell, and divided that value by the volume of the corresponding TH-labeled soma and proximal projections for one region per brain (B,C). Images in (A) are sums of representative z-stacks for park^+/+, park^+/− and park^−/− day 5, 10 and 20 PPL1 and day 10 PPM3. Digital image enhancement steps were standardized for green fluorophores. Numbers in histogram bars indicate the number of regions analyzed. For *P < 0.05; for **P < 0.01; for ***P < 0.001. Error bars represent standard error of the mean; scale bar represents five microns.