ALS- and FTD-associated missense mutations in TBK1 differentially disrupt mitophagy

Significance Missense mutations in TANK-binding kinase 1 (TBK1) have diverse biophysical and biochemical effects on the molecule and are associated with the neurodegenerative diseases amyotrophic lateral sclerosis (ALS) and fronto-temporal dementia. TBK1 plays an essential role in clearing damaged mitochondria. Here, we investigate the impact of 10 ALS-linked TBK1 mutations on the critical early stage of mitophagy. We find that both TBK1 recruitment and kinase activity contribute to the clearance of the damaged mitochondria. Furthermore, in neurons, expression of TBK1 mutants alone affects mitochondrial network health. Our investigation utilizes disease-linked mutations to further refine the current model of mitophagy, identifying regulatory crosstalk between the kinases TBK1 and ULK1, and providing insights into the roles of TBK1 in neurodegenerative pathogenesis.


Primary hippocampal culture and transfection: A suspension of embryonic day 18
Sprague Dawley rat hippocampal neurons were provided from the Neurons R Us Culture Service Center at the University of Pennsylvania. Cells were plated on 35 mm glassbottom dishes (MatTek) at a density of 250,000 cells/dish; dishes were precoated with 0.5 mg/ml poly-L-lysine (Sigma Aldrich). Cells were initially plated in MEM supplemented with 10% horse serum, 33 mM D-glucose, and 1 mM sodium pyruvate and left for 2-5 hours. The media was then replaced with Neurobasal (Gibco) supplemented with 33 mM D-glucose, 2 mM GlutaMAX (Invitrogen), 100 units/ml penicillin, 100 g/ml streptomycin, and 2% B-27 (ThermoFisher) (Maintenance Media; MM) and cells were maintained at 37 C in a 5% CO2 incubator. AraC (5 M) was added the day after plating to prevent glia cell proliferation. Neurons were transfected at 5 DIV with DNA (0.8-1.2 g of total plasmid) and siRNA (45 pmol) mixtures using Lipofectamine 2000 Transfection Reagent (ThermoFisher) and incubated 36-48 hours. To induce mitophagy, media was fully replaced with MM containing 3 nM Antimycin A for 2 hours; in control conditions media was replaced with standard MM.
Labeling, treatment, and fixation. HeLa-M and HEK cells: To tag Halo-tagged or SNAP-tagged proteins, cells were incubated with the respective Halo or SNAP ligands.
For Halo, cells were incubated with 190 nM Halo ligand for at least 20 min. Cells were then washed two times with conditioned media and allowed to rest for at least 20 min in conditioned media. For SNAP, cells were incubated with 1.25 μM SNAP ligand for at least 1 hour. Cells were then washed two times with conditioned media and allowed to rest for at least 30 min in conditioned media. When both SNAP and Halo were used, cells were incubated with Halo tag first, then SNAP tag, with their respective protocols. Cells were washed two more times with conditioned media. When applicable, cells were treated with 5 μM ULK-101 for 2.5 hr. Cells were then treated with 20 μM CCCP or a combination of 10 μM Oligomycin A and 10 μM Antimycin A in conditioned media for 1.5 hours.
Immediately afterward, cells were washed with warmed PBS then fixed with warmed 4% paraformaldehyde for 10-12 min. For experiments with antibody tagging, cells were permeabilized with 0.5% Triton X at room temperature for 5 min, then blocked with 3% BSA, 0.2% Triton X for one hour. Cells were incubated with primary antibodies overnight at 4 ºC. Afterward, cells were washed 4x 5 min in PBS and incubated with secondary antibodies for one hour. Cells were then washed 4x 5 min in PBS and imaged. For the transfection levels test (Supplemental Figure 1C), after fixation cells were labeled with Hoechst 33342 for 5 min, then again washed 4x 6 min before imaging. Hippocampal neurons: Prior to imaging, Halo-tag (JaneliaFluor 646, 100 nM) and SNAP-tag (SNAP-Cell 430, 2 M) ligands were added for 30 min, followed by two quick washes and a 30 min washout. Mitochondrial membrane potential was assessed by loading mitochondria with 2.5 nM TMRE for 30 min.  Figure 1C), cells were imaged with widefield microscopy. Images were taken from three fields per dish, and the only requirement for each field was that the nuclei (Hoechst staining) appeared healthy and regularly spaced in an area that was close to fully confluent. For all other experiments, samples were imaged with a Nikon Eclipse Ti Microscope with a 100X objective (Apochromat, 1.49-N.A. oil immersion) and an UltraView Vox spinning disk confocal system (PerkinElmer). Zstacks at 0.15 nm/step or timelapse confocal images at 30 seconds/frame were collected with Volocity acquisition software (PerkinElmer). Fields of view were chosen to maximize the number of cells that expressed detectable components of interest. In fixed samples, z-stacks were collected through the majority of cells' midsections.
For live cell imaging, conditioned media was replaced with Leibovitz's L-15 Medium (Gibco, 11415064) supplemented with 10% fetal bovine serum. Cells were then rested for at least 10 min in the 37 ºC imaging chamber of the microscope. For timelapse mitochondrial damage, a z position was chosen in the midsection of a healthy-appearing cell with a regularly shaped nucleus (nucleus characterized by absence of tagged TBK1). 5-10 frames were collected at basal conditions, then a volume of imaging media at least 50% of the initial volume was added, including CCCP to bring the total concentration to 20 μM as frame collection continued.
Hippocampal neurons: Neurons were imaged in HibernateE (Brain Bits) supplemented with 2% B27 and 33 mM D-glucose; Antimycin A was added to the imaging media for treated conditions. TMRE was added to the imaging media for TMRE experiments.
Mitochondrial enrichments and immunoblots. For standard cell lysis, cells were washed two times with warmed PBS, then lysed with ice cold RIPA buffer (50 mM Tris-HCl, 1 mM EDTA, 2 mM EGTA, 1% Triton X, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 150 mM NaCl) with added Halt Protease and Phosphatase Inhibitor Cocktail (ThermoFisher Scientific, 78444) and scraped into sample tubes and incubated with continuous gentle inversion at 4 ºC for 20 min. Samples were then spun at 4 ºC in a microcentrifuge at 17 g for 20 min, and the supernatant was transferred to a separate tube. Samples were assayed for protein concentration with Pierce BCA Protein Assay Kit (ThermoFisher Scientific, 23225). Mitochondrial enrichment was performed with ThermoScientific isolation kit for cultured cells (89874) and mitochondrial fractions were diluted in RIPA buffer with Halt Cocktail as above.
30 μg of each sample was loaded into a 10% gel (or 14% gel for TOM20 detection).
Image processing and analysis.
Images were deconvolved with Huygen's Professional version 17.10 software (Scientific Volume Imaging, The Netherlands, http://svi.nl) to remove background noise and increase resolution and signal-to noise ratio. The Classic Maximum Likelihood Estimation (CMLE) algorithm with theoretical PSF was performed for 50 iterations at most. The signal-to-noise ratio for all channels was set between 10 and 30, depending on the individual construct; all other settings were default.
Images were assembled in Illustrator (Adobe). Most images shown are deconvolved, with the exceptions of widefield images (Supplemental Figure 1C), all Supplemental wholefield images, the experiment to determine ULK1 dependence (Figure 6), and neuronal mitochondrial TMRE images ( Figure 8A). All intensity measurements were carried out on raw intensity images, not deconvolved images.
In Figures 1E,F Feature selection used color/intensity, edge, and texture up to σ = 5 pixels. Binary images were exported as .tifs using simple segmentation, and the Analyze Particles function of ImageJ was used to count mitochondria. Mitochondria were considered positive for LC3 if the fluorescent LC3 ring surrounded at least half of the organelle. Each data point represents the percentage of LC3-positive mitochondria for a single cell. For Figure 5, OPTN rings were identified as before. Then OPTN ring regions of interest (ROIs) were transferred to the TBK1 and phospho-OPTN channels of the same image, and mean fluorescence intensity was measured for each ROI on the non-deconvolved images.
Intensity data points were plotted for WT-TBK1-expressing cells, and all intensities above the 25 th percentile were considered "positive," while all intensities below were considered "negative" (see Supplemental Figure 6B,C Transfection level images included Hoechst and JF646 channels (Supplemental Figure 1C). Both channels were maximally projected, and the JF696 channel was background subtracted with ImageJ's rolling ball radius set to 25.0 pixels, with sliding paraboloid and disabled smoothing. Images were then imported to CellProfiler software (59) and nuclei were identified as primary objects; then cells were delineated by the propagation method in the JF646 channel. Thus, only cells expressing SNAP-TBK1 were identified and the mean intensities of their cytoplasmic signals were exported to Excel.
These values were displayed on a histogram to demonstrate the relative frequencies of mean intensities (GraphPad).
For immunoblots, ImageStudio Software (Version 5, Li-Cor) was used to scan bands to ensure no patches were overexposed. ImageStudio was used to subtract background and quantify band intensities, which were normalized to the total protein signal for their respective lanes with Excel (Microsoft). For mitochondrial enrichment, bands were normalized to TOM20. Those values were graphed in GraphPad (Version 9, Prism).

SUPPLEMENTAL FIGURES
WT R47H Y105C S151C S151F G217R    A. Representative whole-field images corresponding to images in Main Figure 3A. Cells were tagged with an antibody to total TBK1 (magenta). Scale bars, 10 μm.