Phospho-proteome profiling in human neurons reveals targets of TBK1 in ALS/FTD-associated autophagy networks

SUMMARY

Loss-of-function variants in TBK1, encoding a protein kinase, are strongly associated with familial amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD). However, how haploinsufficiency for TBK1 leads to age-related neurodegeneration remains unresolved. Here, we utilize sets of isogenic induced pluripotent stem cells (iPSCs) with loss of TBK1 or loss of optineurin (OPTN) for quantitative global proteomics and phospho-proteomics in both stem cells and excitatory neurons. We found that TBK1 sustains the abundance and phosphorylation of its interacting adapter proteins, AZI2/NAP1, TANK, and TBKBP1/SINTBAD. Moreover, TBK1 regulates the phosphorylation of endo-lysosomal proteins, such as GABARAPL2, the late-endosome GTPase RAB7A, and selective autophagy cargo receptor proteins—including novel phospho-sites in p62/SQSTM1—in neurons. Finally, we provide a census of the phospho-proteome in nascent human neurons for further studies. Overall, TBK1 serves as a point of convergence in ALS/FTD-linked endo-lysosomal networks that act in a cell-autonomous manner to maintain protein homeostasis in neurons.

Graphical Abstract

In brief

Loss-of-function variants in TBK1 are strongly linked to ALS and dementia, but how decreased TBK1 kinase activity leads to neurodegeneration remains unclear. Smeyers et al. integrate isogenic stem cell lines and global phospho-proteomics to demonstrate that TBK1 regulates selective autophagy and endo-lysosomal pathways in excitatory neurons.

INTRODUCTION

TBK1 (TANK-binding kinase 1) is a member of the IKK (IkappaB kinase) family within the 538 protein kinases of the human kinome.¹ TBK1 is involved in the regulation of a diverse set of cellular pathways, including innate immunity, autophagy, and cell proliferation, and has been implicated in cancer, autoimmune diseases, glaucoma, and neurodegeneration.^2-8

Whole-exome sequencing of thousands of individuals has identified loss-of-function (LoF) variants in TBK1 strongly associated with familial amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD).^9,10 Subsequent studies confirmed that TBK1 variants comprise up to the third most common cause of ALS (following C9ORF72 and SOD1) and FTD associated with TDP-43 pathology (following C9ORF72 and GRN) and account for ~4%–10% of combined ALS/FTD in certain studied patient cohorts.^11-13 Disease-associated variants include damaging missense coding variants, occurring both within and outside its kinase domain, as well as rare nonsense variants.^2,14 How these variants affect TBK1 function in neurons remains unclear. Prior biochemical studies, predominantly in cancer cell lines, demonstrate that TBK1 phosphorylates additional ALS/FTD-associated autophagy proteins, including p62/sequestosome-1 (SQSTM1) and optineurin (OPTN).^15-17 Both heterozygous and homozygous LoF variants in OPTN are also sufficient to cause familial ALS/FTD.^12,18 Overall, this suggests that TBK1-dependent autophagy is critical to maintain protein homeostasis. However, given that TBK1 may be involved in many cellular processes, it remains unresolved as to how haploinsufficiency of this kinase contributes to age-related neurodegeneration.

Complete deletion of Tbk1 in mice is embryonic lethal due to severe liver inflammation.¹⁹ Heterozygous Tbk1 knockout (KO) mice do not exhibit any overt functional phenotypes that resemble ALS or FTD, aside from modest impairment of autophagy.²⁰ Paradoxically, partial deletion of Tbk1 extended survival in the commonly studied ALS mutant SOD1 overexpression mouse model.²¹ Hence, animal models to study the role of TBK1 in neurological disease remain limited.

Human pluripotent stem cell models have provided many insights into familial ALS and related diseases.^22-24 However, several technical sources of variability among induced pluripotent stem cell (iPSC) lines that include reprogramming methods, quality control, and rate and efficiency of neuron differentiations can present potential challenges for defining phenotypes from disease-associated variants, especially those that may act with incomplete penetrance.²⁵ Hence, studies of isogenic iPSCs may help to increase rigor and define disease mechanisms, such as for SOD1-ALS.²⁶ Recently, the well-characterized KOLF2.1J cell line was carefully selected as a standard iPSC line for the field,²⁷ potentially equivalent to the widely used C57BL/6 mouse genetic background, to model a wide range of neurodegenerative diseases,²⁸ though some possible caveats have been recently acknowledged.²⁹

Here, we integrate isogenic iPSC lines and unbiased multi-omic approaches to define the cellular consequences of TBK1 LoF variants, complete loss of TBK1 function, and complete loss of OPTN function. Through quantitative proteomics across thousands of phospho-proteins in both human stem cells and excitatory neurons, we delineate the role of TBK1 in targeting factors involved in selective autophagy and lysosomal pathways. Additionally, our study provides a detailed phospho-proteome reference atlas for further understanding the dysregulation of protein phosphorylation in neurons across major neurodegenerative diseases.

RESULTS

TBK1 interacts with its adapters AZI2, TANK, and TBKBP1 in neurons

To investigate the impact of TBK1 and OPTN LoF in human neurons, we validated isogenic iPSCs in the reference KOLF2.1J background, including those with near-complete deletion of the TBK1 locus and complete deletion of OPTN (Figure 1A). We also examined two ALS/FTD-associated TBK1 coding variants: E696K, located in the C-terminal coiled-coil domain, which may affect interactions with adapter proteins and OPTN, and G217R, residing in the N-terminal serine/threonine kinase domain, which may impair kinase activity (Figure 1B).³⁰ All lines were comprehensively characterized using Sanger sequencing, transcriptomic profiling, quantitative PCR (qPCR), and immunoblotting, as well as immunostaining to confirm pluripotency (Figures 1C-1F; S1A and S1B). Interestingly, in the TBK1 KO CRISPR-edited lines, there was upregulation of the few residual exons both upstream and downstream of the CRISPR-targeted site, suggesting a compensatory transcriptional response to loss of TBK1 (Figures 1C; S2A). We confirmed that the TBK1 G217R variant was associated with reduced TBK1 levels in a gene dosage-dependent manner compared to the isogenic “revertant” control in iPSCs, consistent with this variant contributing to haploinsufficiency for TBK1 function in affected carriers (Figures S4A-S4C). Finally, we validated that the CRISPR-mediated OPTN KO resulted in a complete loss of OPTN expression (Figures S2B-S2D).

Figure 1. Validation of a human iPSC model to study TBK1 and OPTN loss of function.

(A) Experimental approach in isogenic induced pluripotent stem cell (iPSC) lines and differentiations into induced neurons (iNeurons) for multi-omic analysis.

(B) Schematic of TBK1 protein domains, highlighting its kinase domain and the ALS/FTD-associated point mutations examined in this study.

(C) Genomic location of the CRISPR-Cas9 deletion within the TBK1 locus and the resulting disruption in TBK1 expression via RNA-seq. See Figure S2 for more detailed views at both cut sites.

(D) Sanger sequencing analysis confirming the presence of the TBK1 c.2086G>A; p.E696K variant in the TBK1 E696K isogenic iPSC lines: homozygous (A/A) or heterozygous (G/A).

(E and F) Quantitative PCR (qPCR) (E) and immunoblotting (F) analyses for TBK1 transcript and protein levels normalized to GAPDH in the isogenic iPSC lines. Each dot for quantification of immunoblots represents a biological replicate (n = 3–5) from three individual experiments, and data are presented as mean ± s.e.m.; one-way ANOVA with Kruskal-Wallis test and Dunn’s multiple comparison correction. TBK1 KO vs. control (** adj. p val. = 0.0020). No significant changes were detected in the TBK1 E696K SNV/WT or SNV/SNV lines compared to control.

(G) Longitudinal tracking of isogenic iPSC-derived neurons over a 5-day time course with quantification of neurite length and neurite branch points from live microscopy (****p < 0.0001), along with a survival curve from neuron counts over time for each (ns: not significant). A representative experiment is shown, and two independent experiments gave similar results.

(H) Representative brightfield images of neurons from each genotype at day 5 (scale bar, 400 μm).

Next, we performed transcriptional profiling of the corresponding isogenic induced neurons and assessed for signs of TDP-43 pathology, the main neuropathological marker in ALS/ FTD.³¹ In isogenic iPSCs and iPSC-derived neurons, the expression of TARDBP (encoding TDP-43) was constant. There was no detectable induction of the STMN2 cryptic exon, which is a sensitive readout for loss of TDP-43 nuclear function in neurodegeneration,^24,32,33 while STMN2 levels were unchanged across genotypes (Figure S3A). Additionally, TDP-43 localization was predominantly nuclear in both control and TBK1 KO neurons (Figure S3B). This suggests that even complete loss of TBK1 or OPTN function is not sufficient to perturb TDP-43 in iPSC-derived neurons in the absence of additional disease-associated stimuli or stressors.

To further assess neuron-specific phenotypes, we longitudinally tracked both the survival and neurite outgrowth of isogenic neurons with TBK1 deletion compared to wild-type control. Using automated image analysis, we found that loss of TBK1 significantly impaired neurite outgrowth and neurite branching over time (Figures 1G and 1H). On the other hand, there was no significant difference in neuron survival in vitro, hence the observed decrease in neurite complexity is not simply due to changes in the number of soma (Figures 1G and H). This suggests that the kinase TBK1 has cell-autonomous effects that sustain excitatory neuron growth and function.

Next, to determine how loss of these ALS/FTD-associated proteins affects protein homeostasis, we focused on an unbiased, comprehensive assessment of the proteome in these isogenic iPSCs through tandem mass tag (TMT)-based quantitative proteomics (Figure 2A). Following protein processing and extensive fractionation, mass spectrometry (LC-MS/MS) captured 155,307 peptides, yielding the relative quantification of over 8,000 proteins across 18 TMT-channels (Figure 2A; Table S1). In OPTN KO cells, there was notable enrichment for increased lysosomal proteins (adj. p = 1.265 × 10⁻²⁷), including cathepsin A, B, and H (Figure 2B), compared to isogenic control cells; however, there was no significant enrichment for this pathway in cells with loss of TBK1 function. In TBK1 KO iPSCs, the second least abundant protein was the TBK1 adaptor protein AZI2 (NAP1), which may mediate TBK1 localization and facilitate interactions with its substrates.³⁴ TANK, another TBK1 adapter, was trending toward reduced abundance in TBK1 KO iPSCs (Figure 2C), and TANK was significantly decreased in isogenic TBK1 G217R^SNV/WT iPSCs compared to G217R^REV/WT iPSCs (Figures S4D and S4E). CYLD (cylindromatosis lysine 63 deubiquitinase), which was recently implicated in ALS/FTD,³⁵ was also decreased in abundance in TBK1 KO iPSCs (Figure 2B).

Figure 2. Isogenic global proteomics reveals TBK1 complexes in cells and excitatory neurons.

(A) Schematic of isogenic iPS cell collection for global proteomics analysis with each replicate (n = 3–4) containing ~40 million cells. Proteins were extracted, digested into peptides, and peptides from each replicate were labeled with tandem mass tags (TMT-18plex). Then, all samples were combined and fractionated for quantitative LC-MS/MS. Peptides were stringently assigned to proteins for analysis with bioinformatic tools (see methods). Based on 155,307 peptides detected, a total of 8,196 proteins were quantified across 18 channels. Note that the detected residual TBK1 is likely due to ratio compression, a phenomenon commonly observed with this method, resulting from the co-elution of peptides with similar sizes.

(B and C) Volcano plots (B) for differential protein abundance in TBK1 isogenic lines and OPTN KO iPSC lines compared to control iPSC. AZI2 (blue) is the least abundant protein in TBK1 KO cells aside from TBK1 itself. The OPTN KO cells exhibit increased abundance of proteins that were enriched for lysosomal components (GO:0005764: lysosome; adj. p value of 1.265 × 10⁻²⁷), including lysosomal proteases (colored yellow): cathepsins A, B, and H (CTSA, CTSB, and CTSH). Mass spectrometry-based proteomic analysis of TBK1 adapter proteins across isogenic iPSCs (C) reveals reduced abundance of AZI2 in TBK1 KO vs. control (****p = 3.15E-05), TBK1 E696K^SNV/WT vs. control (*p = 0.04255), TBK1 E696K ^SNV/SNV vs. control (*p = 0.01489), and OPTN KO vs. control (**p = 0.00415). Note, TBKBP1 was not detected in the iPSC proteome.

(D) Abundance of TBK1 adapters across isogenic iNeurons. AZI2 was decreased in TBK1 KO vs. control iNeurons (**p = 0.00260) and increased in TBK1 E696K^SNV/WT vs. control (**p = 0.00368); TANK was decreased in TBK1 KO vs. control (**p = 0.00864) and TBK1 E696K^SNV/SNV vs. control iNeurons (**p = 0.04582). Decreased abundance of TBKBP1 was also observed in TBK1 KO vs. control iNeurons (**p = 0.00781). Each dot represents a biological replicate (n = 3–4) from three independent neuron differentiations. Data are presented as mean ± s.e.m. Significance was assessed by comparing each isogenic group to control using a two-tailed t test (see STAR Methods).

(E) Schematic of the affinity-purification mass spectrometry (AP-MS) methodology to identify TBK1 interacting proteins in excitatory neurons.

(F) Venn diagram illustrating the results of AP-MS experiments using two monoclonal antibodies against endogenous TBK1 compared to IgG control, which was used to help define non-specific interactions. Upon bioinformatic filtering, six proteins were consistently detected, including TBK1 and all three of its adapters (AZI2, TANK, and TBKBP1).

(G) Relative abundance of TBK1 and selected candidate interactors expressed in arbitrary units (AU) (N-term mAb = anti-TBK1 N; C-term mAb = anti-TBK1 C). Each dot represents a binding reaction (n = 3).

(H) Reverse co-immunoprecipitation (coIP) where neuron lysates where incubated with the indicated antibodies and then assessed for TBK1 via immunoblotting, confirming biochemical interactions between TBK1 and AZI2 and between TBK1 and TANK in neurons.

Subsequently, isogenic iPSCs were differentiated into glutamatergic cortical neurons by inducing the expression of the neuralizing transcription factor NGN2 coupled with small-molecule patterning factors and then collected at day 28 across independent differentiations for TMT-based proteomics.^36,37 All TBK1 adapter proteins, AZI2 (NAP1) along with TANK and TBKBP1/SINTBAD, were significantly decreased in abundance in TBK1 KO compared to control neurons (Figure 2D). Transcriptomic analysis confirmed that the corresponding transcripts of these adapters were unchanged (Figure S2E), suggesting that TBK1 promotes the stability of these three proteins in human neurons.

To validate TBK1 protein-protein interactions in neurons, we performed affinity purification-MS (AP-MS) using two monoclonal antibodies to capture endogenous TBK1 (Figure 2E; Table S2). This bait and five interacting proteins, AZI2, TANK, TBKBP1, FTH1 (ferritin heavy chain), and JPH1 (Figure 2F), were consistently detected. The relative abundances of AZI2, TANK, and TBKBP1 were greater than those of other detected proteins (Figure 2G), suggesting a central role for these adapters in mediating TBK1 functions. The interactions between TBK1 and its adapters AZI2 and TBKBP1 were further validated by reverse co-immunoprecipitation (coIP) of endogenous proteins in neurons (Figure 2H). Notably, TAX1BP1 was captured with the anti-TBK1 C-terminal antibody, but not the N-terminal antibody, which may sterically block its kinase domain, suggesting that we detected a transient interaction between this kinase and one of its key substrates (Figure 2G, and see below). Collectively, these results demonstrate that TBK1 interacts with the adapters AZI2, TANK, and TBKBP1 in the disease-relevant cellular context of excitatory neurons, forming complexes that are disrupted by loss of TBK1.

TBK1 phosphorylates protein cargo receptors and autophagy factors

To determine the role of the kinase TBK1 in the regulation of the phospho-proteome, additional TMT-based proteomics was performed to provide the relative phosphorylation status of tens of thousands of residues across thousands of proteins (Figure 3A). Following rigorous normalization to protein abundance in our global proteomics, a total of 50,872 phospho-peptides were quantified across all genotypes (Table S3). In TBK1 KO cells, Gene Ontology (GO) analysis of the top decreased phospho-peptides revealed enrichment for autophagy-related processes (GO:0006914, adj. p = 6.01 × 10⁻³), especially selective autophagy cargo receptors (SACRs; Figure 3B), in contrast to modest-to-absent changes in their protein abundance (Figure S5). On the other hand, only modest changes were observed in cells with either one or two TBK1 E696K alleles, consistent with this variant not impairing TBK1 kinase activity (Figure 3B). By focusing on amino-acid sequences matching TBK1 putative consensus motifs, a serine residue, and a hydrophobic leucine or phenylalanine (SL/SF; Table S4), we found significant decreases in the phosphorylation of SACRs at specific sites, including SQSTM1 (pS361; p = 2.5 × 10⁻ , NCOA4 (pS412; p = 0.00096), and CALCOCO2 (pS355; p = 0.02979) in TBK1 KO iPSCs, as well as NBR1 (pS622) across several genotypes, TBK1 KO (p = 0.00170), TBK1 E696K^SNV/SNV (p = 0.03972), and OPTN KO cells (p = 0.03671; Figure 3C). RAB7A, a key regulator of late endosomes, was previously characterized as a target of TBK1³⁸ and LRRK1.³⁹ We observed decreased phosphorylation of RAB7A (pS72; p = 0.00050), as well as altered phosphorylation of additional autophagy and endo-lysosomal proteins, including RB1CC1/FIP200 (pS1286; p = 0.00029); GABARAPL2 (pS10; p = 0.00018), an Atg8 ortholog; and RAB11B (pS42; p = 0.00972), a mediator of retromer trafficking, in isogenic TBK1 KO cells (Figure 3D). In general, the effect of complete loss of TBK1 on the phospho-proteome was greater than the TBK1 E696K variant (Figure 3E).

Figure 3. Phospho-proteomics defines the impact of TBK1 function on endo-lysosomal pathways.

(A) Schematic of global phospho-proteome-wide profiling. Briefly, phospho-peptides were enriched using titanium dioxide (TiO2) before being fractionated into 50 fractions for subsequent mass spectrometry analysis.

(B) Volcano plots representing differential abundance of phospho-peptides of isogenic cell lines, following normalization to the global-proteomics data. A total of i = 50,872 normalized phospho-peptides were quantified. Phospho-peptides of selective autophagy cargo receptors (SACRs; in red) and autophagy factors (in yellow) were decreased in abundance in TBK1 KO cells compared to control, and Gene Ontology analysis demonstrated enrichment for the autophagy pathway. No significant pathway enrichment was found in either the TBK1 E696K or OPTN KO lines.

(C and D) Specific phospho-peptides with phosphoserine residues consistent with TBK1 kinase consensus sites in autophagy proteins, including several SACRs (C) and key regulators of autophagy and endo-lysosomal pathways (D). Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

(E) Heatmap representing a selection of phospho-peptides from autophagy factors, ALS-associated proteins, SACRs, and novel targets of interest.

(F) Volcano plots representing the differential abundance of phospho-peptides normalized to the global proteomics data from TBK1 G217R hiPSC lines SNV/WT and SNV/SNV compared to REV/WT (isogenic CRISPR-revertant control). A total of i = 30,810 normalized phospho-peptides were quantified. Phosphorylation of the SARC p62/SQSTM1 (highlighted in red) was decreased in TBK1 G217R^SNV/SNV compared to REV/WT.

(G) Quantification of selected phospho-peptides with phosphoserine residues consistent with TBK1 kinase consensus sites of p62/SQSTM1, OPTN, GABARAPL2, and RAB7A across genotypes, demonstrating a variant-dosage effect on the phosphorylation of specific residues of key TBK1 substrates. Data are presented as mean ± s.e.m.; each dot represents an independent biological replicate (n = 4). Pairwise comparisons among genotypes were performed using a one-way ANOVA with Tukey’s multiple comparisons post hoc test. Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Next, we validated our findings in an orthogonal phospho-proteome set that included isogenic TBK1 G217R^SNV/WT , G217R^SNV/SNV, and G217R^REV/WT iPSCs, resulting in the relative quantification of 30,180 phospho-peptides, normalized to the abundance of 7,229 proteins (Table S5; Table S6). Although the heterozygous genotype (SNV/WT) matches ALS/FTD risk factor carriers, the homozygous (SNV/SNV) cells exhibited greater alterations in the phosphorylation status of key selective autophagy proteins. For instance, a clear variant-dosage effect was identified for p62/SQSTM1, where the homozygous cells exhibited decreased phosphorylation compared to the heterozygous cells (SNV/SNV vs. SNV/WT, p < 0.0001), which had reduced phosphorylation compared to wild-type (SNV/WT vs. REV/WT, p < 0.0001). Hence, this suggests that the G217R variant acts as a trans-phospho-protein quantitative trait locus (ppQTL) for the phosphorylation status of p62/SQSTM1, another protein strongly linked to ALS/FTD. Additionally, OPTN (pS177), GABARAPL2 (pS10), and RAB7A (pS72) exhibited significantly decreased phosphorylation due to the effect of a single nucleotide variant in isogenic cell lines (Figures 3F and 3G) and were overwhelmingly consistent with the effects of complete loss of TBK1 function on these autophagy proteins (Figures 3F and 3G).

Of note, given the described function of TBK1 in regulating the interferon pathway in immune cells, we found no overt signs of dysregulation of this pathway at the protein level in TBK1 KO or E696K isogenic hiPSCs. However, our approach was sensitive enough to detect a lower level of IRF3 pS173 in TBK1 G217R variant cells, despite the absence of applying inflammatory stimuli (Figure S4F), suggesting that TBK1 may act in additional roles in certain cell types.

Proteins associated with neurodegenerative diseases are detected in iNeurons

iPSC-derived neurons provide a valuable model to study neurological diseases, yet their phospho-proteome has not been well defined. Therefore, we compared the number of phospho-peptides detected in the phospho-proteome of control iPSCs and iNeurons, in the absence of any stressors, focusing on 29 proteins implicated in various neurodegenerative diseases, including Alzheimer’s disease (AD), frontotemporal lobar degeneration (FTLD), ALS, Huntington’s disease (HD), Parkinson’s disease/dementia with Lewy bodies (PD/DLB), and neurofilaments, which are generally used as disease biomarkers (Figure 4A; Tables S6; S7 and S8). As expected, a higher number of phospho-peptides for neural-enriched proteins were detected in iNeurons, including MAPT, KCNQ2, STMN2, NEFL, and NEFH. Interestingly, an increase in phospho-peptide coverage was observed in iNeurons for PSEN1, ATXN2, FUS, NEK1, OPTN, PFN1, TBK1, TMEM106B, VAPB, and VSP35, highlighting the relevance of iNeurons in the study of neurodegeneration (Figure 4A). We then mapped the numerous phosphorylation sites of tau detected in our dataset, highlighting the exact phospho-residues recognized by commonly used phospho-specific antibodies for tau neuropathology, including AT8 (pS202/pT205) and PHF1 (pS396/pS404), suggesting that tau is regulated by complex levels of phosphorylation in nascent neurons (Figure 4B). Additionally, we observed the phosphorylation of alpha-synuclein at pS129, the critical site for identifying pathological inclusions in synucleinopathies (Figure 4C). Interestingly, TDP-43 pS409/S410, a phospho-site used as a common marker for TDP-43 pathology in ALS/FTLD and additional TDP-43 proteinopathies, was detected in both iPSCs and iNeurons (Figure S6). Hence, phosphorylated forms of key proteins that comprise pathological inclusions in age-related neurodegenerative diseases could be detected in iNeurons, underscoring the importance of this disease model for investigating protein phosphorylation. For example, in iNeurons but not iPSCs, we identified a novel phospho-site for TBK1, located in a stretch of three consecutive serine residues within one of its coiled-coil domains (Figures 4D and 4E).

Figure 4. A phospho-proteome census of neurodegenerative markers in iNeurons.

(A) Comparison of the number of phospho-peptides detected in the phospho-proteome of hiPSCs and iNeurons for 29 selected proteins implicated in neurodegenerative diseases: AD (Alzheimer’s disease), FTLD-tau (frontotemporal lobar degeneration with tau pathology), ALS (amyotrophic lateral sclerosis)/FTLD-TDP (FTLD associated with TDP-43 pathology), HD (Huntington’s disease), PD/DLB (Parkinson’s disease/dementia with Lewy bodies), and neurofilaments, which are used as general biomarkers of neurodegeneration.

(B) Protein domain schematic of tau with detected phospho-residues (serines above and threonines and tyrosines below), with residues detected by commonly used phospho-specific antibodies of tau pathology underlined (Note: phospho-residues are plotted on 4R tau based on the common numbering convention, though 3R tau is more highly expressed in induced neurons).

(C) Protein databank structure of SNCA/α-synuclein protein (1XQ8) with arrow indicating the only detected phosphoserine, pS129, which is the same site detected by antibodies that mark pathological inclusions in synucleinopathies.

(D) MS-MS spectra belonging to the tryptic peptide spanning amino acids L508-R525 of human TBK1, obtained by HCD fragmentation of a precursor ion 774.3785⁺³. An excess mass of 79.9663 compared to the unmodified sequence, corresponding to the phosphate group, is observed in the precursor and in sequence ions that contained serine 510. Experimental masses of the most representative sequence ion peaks are labeled in the spectrum, also indicating the fragment type (Roepstorff-Fohlmann-Biemann ions nomenclature). The positions of the fragmentation events generating these ions are indicated in the sequences over the spectrum. A table indicating the theoretical masses of the sequence ions according to the proposed modified sequences is shown in the right side, with red font indicating observed fragments. S^Phospho, phosphorylated serine; TMT, tandem mass tag derivatized N terminus.

(E) Ribbon representation of the human TBK1 structure as a dimer (PDB: 4IWO).⁴⁰ The coiled-coil domain containing the novel phospho-site is a lighter tone with a stretch of three consecutive serine residues highlighted in yellow, including S510 (arrow).

(F) TBK1 targets multiple sites on SACR family proteins. Schematic of protein domains with phospho-sites identified in normalized phospho-proteomics in iPSCs (blue), iNeurons (green), or both (black).

(G) AlphaFold structural models of selected SACRs: p62/SQSTM1 (Q13501), NBR1 (Q14596), and NCOA4 (Q13772; for illustration purposes, scale varies). The arrows point to TBK1 phospho-sites in exposed flexible domains, highlighted in green.

Additional structural modeling analysis revealed that TBK1 targets multiple serine residues of SACRs that are located in accessible, flexible domains, which we speculate might mediate oligomerization into biomolecular condensates or protein-protein interactions with selective autophagy cargos (Figures 4F and 4G). For instance, p62 co-localizes with a variety of neuronal protein inclusions in neurodegenerative diseases, including tau in AD, HTT in HD, and DPRs in C9ORF72-ALS/FTD.^31,41-44 Phosphoserine 366 falls adjacent to its ubiquitin-binding region (UBA) within its FIP200-interacting region that promotes autophagy.⁴⁵ NDP52 functions in several forms of selective autophagy, including xenophagy, and phosphoserine 355, which strongly matched a TBK1/IKKE consensus site between its coiled-coil domain and C-terminal ubiquitin binding domain (UBZ), appears analogous to pS366 in p62 (Table S11).

TBK1 regulates the interaction between TAX1BP1 and NCOA4, which are both involved in the selective autophagy of ferritin and intracellular iron metabolism.⁴⁶ We identified a TBK1-dependent phosphoserine residue in TAX1BP1 within its coiled-coil region, which regulates its oligomerization, as well as a novel phosphoserine residue in NCOA4 within its iron-sulfur cluster domain that is critical for the sensing of intracellular iron and regulating ferritinophagy.⁴⁷ This provides a potential mechanism of how TBK1 might target SARCs to promote protein-protein interactions and the lysosomal turnover of ferritin that warrants further study. NBR1 (neighbor to BRCA1) is a highly conserved SARC implicated in cancer metastasis and found in Lewy bodies in PD.^48,49 NBR1 has been implicated in ESCRT-complex-dependent endosomal micro-autophagy (eMI).^50,51 TBK1 regulated a phosphoserine in NBR1 that falls between its LIR1 and LIR2 domains, which interact with LC3 and GABARAP family proteins to facilitate binding to autophagosome membranes.

TBK1 regulates selective autophagy and endo-lysosomal proteins in neurons

Many studies have focused on the function of TBK1 in immune cells and immortalized cancer cell lines; however, the precise role of TBK1 in non-proliferating human neurons remains poorly understood. To better define the function of TBK1 in neurons, we generated isogenic neurons for phospho-proteome-wide profiling, where TMT-based analysis identified 66,278 normalized phospho-peptides, providing even greater coverage than for iPSCs (Figure 5A; Tables S7; S8). TBK1 KO iNeurons exhibited a robust signature with the most significantly decreased phospho-peptides mainly belonging to three major groups: selective SACRs, autophagy factors, and the TBK1 adapter family (Figure 5B). Several phospho-targets of TBK1 in iPSCs were also observed in neurons, including p62/SQSTM1 pS366, NCOA4 pS412, TAX1BP1 pS593, and OPTN pS177 (Figure 5C). Of note, perturbations in the phosphorylation status of OPTN exhibited a pattern different from the other SARCs, where a significant decrease in OPTN phosphorylation was found in isogenic excitatory neurons with the E696K ALS/FTD-associated variant. This suggests that disruption of the C-terminal OPTN-binding region in TBK1 is sufficient to decrease phosphorylation of this key substrate, even in the presence of an intact kinase domain. On the other hand, the complete loss of OPTN did not seem to affect the phosphorylation of other SARCs, suggesting that the diverse roles of TBK1 in the regulation of selective autophagy factors are mainly independent of OPTN function. Nonetheless, we detected an intriguing increase in the phosphorylation of TDP-43 (pS409/pS410), which is a neuropathological hallmark of ALS and FTD, in both TBK1 and OPTN variant excitatory neurons compared to controls (Figure S6).

Figure 5. Phospho-proteomics identifies TBK1 targets in iNeurons.

(A) Schematic overview of phospho-proteome profiling in iNeurons.

(B) Volcano plots representing differential abundance of phospho-peptides, normalized to global-proteomics data, across genotypes compared to control. A total of 66,278 normalized phospho-peptides were quantified. Phospho-peptides belonging to the selective autophagy cargo receptors (SACRs; red), TBK1 and its adapters (blue), and additional autophagy factors (yellow) were significantly decreased in TBK1 KO iNeurons.

(C–E) Selected phospho-residues from SACR family proteins that matched TBK1 consensus sites (C) and autophagy factors (D), and selected phospho-residues from TBK1 interacting proteins (E) are shown.

(F), Heatmap of 14 representative phospho-peptides corresponding to those containing a TBK1 SL/SSL consensus phosphorylation site, including some of those displayed in (C) and (E). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

For the most common genetic form of ALS/FTD (C9-ALS/FTD), haploinsufficiency for C9ORF72 may contribute to neural dysfunction.^52,53 C9ORF72 has been implicated in autophagy and forms a complex with WDR41 and SMCR8, which directly interacts with FIP200/RB1CC1, a key factor in autophagy initiation.^54-56 Interestingly, we found significantly decreased phosphorylation of the C9ORF72-interacting protein SMCR8, specifically at pS492, in both TBK1 KO (p = 0.007686) and TBK1 E696K^SNV/WT (p = 0.00024) neurons compared to control neurons. In TBK1 KO iNeurons, we also found decreased phosphorylation of FIP200/RB1CC1 (pS1285; p = 0.00024) as well as ATG9A (pS755; p = 0.00052), which regulates autophagosome formation and lipid trafficking and interacts with TBC1D15.⁵⁷ Additionally, RAB7A phosphorylation was significantly decreased across TBK1 variant E696K^SNV/WT, E696K^SNV/SNV , and TBK1 KO neurons compared to control neurons (Figure 5D). Finally, phosphorylation of TBK1 adapters on non-canonical serine sites, AZI2 pS353, TANK pS225, and TBKBP1 pS335, was also significantly downregulated in TBK1 KO neurons (Figure 5E). Hence, loss of TBK1 perturbs the phosphorylation across a broad swath of disease-relevant autophagy and endo-lysosomal proteins in excitatory neurons (Figure 5F).

TBK1 regulates NCOA7 and lysosomal function

Given that TBK1 regulates both autophagy and lysosomal proteins in neurons, we sought to investigate its role in lysosomal function. First, we assessed lysosome abundance using a fluorescent dye (LysoTracker) that labels acidic organelles in live cells under basal conditions (Figure 6A). We observed that both TBK1 KO compared to control cells and TBK1 G217R^SNV/SNV compared to isogenic TBK1 G217R^REV/WT cells exhibited higher fluorescence intensity via flow cytometry, suggesting that they contained more lysosomes or aberrantly larger lysosomes. This difference between isogenic iPSCs was abrogated upon chloroquine treatment, which inhibits lysosomal acidification and autophagic degradation, confirming the specificity of this assay (Figures 6B and 6C). We next assessed whether this lysosomal accumulation was due to defective lysosomal activity, using a fluorescence readout for the activity of a lysosomal protease, cathepsin B (Figure 6D). TBK1 KO iPSCs had decreased cathepsin B activity compared to control cells (Figure 6E). Similar findings were observed in TBK1 G217R^SNV/SNV cells vs. isogenic TBK1 G217R^REV/WT cells (Figure 6F). As a control, chloroquine treatment negatively affected lysosomal activity, regardless of genotype. Similarly, complete loss of OPTN was associated with increased abundance of lysosomes and lysosomal proteases, cathepsins (Figures 2B; S7A and S7B).

Figure 6. TBK1 regulates lysosomal function.

(A) Experimental schematic: control, TBK1 KO, TBK1 G217R^REV/WT, and TBK1 G217R^SNV/SNV hiPSC were treated without or with chloroquine, a lysosomal inhibitor, and then incubated with reagents to detect lysosomes and analyzed by flow cytometry.

(B) Histograms representing fluorescence intensity of lysotracker (APC-A) per cell in TBK1 KO (blue) and control (gray) hiPSCs from one representative experiment (~2,000–2,500 cells assayed per condition). Three independent experiments gave similar results. Statistical significance was assessed between conditions via chi-squared statistical test in FlowJo.

(C) Histograms showing lysotracker (APC-A) fluorescence intensity in TBK1 G217R^SNV/SNV (purple) and TBK1 G217R^REV/WT (gray) hiPSCs, under no treatment and chloroquine treatment (100 μM, 24 h). Data presented from one representative experiment; three independent experiments gave similar results.

(D) Experimental schematic: cells were incubated with a reagent to assess lysosomal protease activity and then analyzed by flow cytometry.

(E) Histograms of fluorescence intensity of lysosomal activity sensor (PE-CF594-A) per cell in TBK1 KO (blue) and control (gray) hiPSCs (~1,200 cells assayed per condition). Data are shown for cells treated with or without chloroquine (100 μM, 24 h). One representative experiment is shown; three independent experiments gave similar results.

(F) Histograms of fluorescence intensity of lysosomal activity sensor (PE-CF594-A) in TBK1 G217R^SNV/SNV (purple) and TBK1 G217R^REV/WT (gray) hiPSCs, under no treatment and chloroquine treatment (100 μM, 24 h). Data represent one individual experiment; three independent experiments gave similar results.

(G) Immunofluorescence of SiR-lysosome staining and brightfield in control and TBK1 KO iNeurons (scale bar, 20 μm).

(H) Quantification of SiR-lysosome fluorescence intensity with 60 neurons per genotype analyzed. Pairwise comparisons to wild-type control were performed using a two-tailed t test (***p < 0.001 and as shown on plot).

Finally, we sought to confirm this phenotype in the disease-relevant cellular context of iPSC-derived excitatory neurons. Given that flow cytometry of neurons is often not feasible, we assessed lysosomal abundance using an alternative lysosomal probe and found that loss of TBK1 also resulted in increased lysosomal abundance in neurons (Figure 6G and 6H). Taken together with impaired phosphorylation of RAB7A with TBK1 and OPTN variants (Figure 5D), this strongly suggests that loss of TBK1 function causes dysregulation of lysosome function.

Of note, we identified a novel phosphorylation site on NCOA7 (nuclear receptor coactivator 7), a putative estrogen receptor-binding protein, that was regulated by TBK1 (Figure S8A). pS441 of NCOA7 corresponded to a TBK1 consensus phosphorylation motif with a phospho-site prediction model highlighting TBK1 as the top kinase for this residue (Table S11). Interestingly, recent studies have suggested that NCOA7 plays a role in the regulation of endo-lysosomal V-ATPase in the brain.⁵⁸ To initially test this relationship, cells were transfected with full-length NCOA7-GFP or GFP alone as a control. Overexpression of NCOA7 resulted in increased fluorescence intensity for LysoTracker, similar to the TBK1 LoF phenotypes (Figure S8B). Overall, our findings support a model where TBK1 regulates both autophagy and lysosomal activity through the regulation of an array of substrates linked to ALS/FTD-associated cellular pathways (Figure 7).

Figure 7. Model of the function of TBK1 on autophagy and endo-lysosomal pathways.

This scheme summarizes the multifaceted roles that TBK1 plays in selective autophagy and endo-lysosomal pathways in neurons. TBK1 interacts with its adapter proteins in neurons. Through phosphorylation of its targets, including selective autophagy cargo receptors and other autophagy factors, TBK1 likely contributes to autophagy initiation, cargo recruitment, and autophagosome-lysosome fusion. Importantly, we provide evidence that TBK1 may also directly regulate lysosomal activity via RAB7A in neurons and possibly even related retromer function via RAB11B (Note: while not observed here in neurons, the G217R variant modestly affected IRF3 phosphorylation in iPSCs, and TBK1 is well known to regulate the interferon response pathway in non-neuronal subtypes).

DISCUSSION

Heterozygous LoF variants in TBK1 are causative for ALS, FTD, and combined ALS/FTD, while homozygous variants are associated with an early-onset systemic autoinflammatory syndrome.^10-14,59 However, the precise targets of TBK1 in cells and in neurons have remained elusive. Here, through unbiased proteome-wide profiling, we reveal that TBK1 primarily regulates the phosphorylation of autophagy and endo-lysosomal proteins in both proliferating stem cells and post-mitotic neurons (Figure 7). TBK1 targets a broad swath of serine residues falling within SL consensus motifs, especially in ubiquitin-binding SACRs, far beyond expectations from prior studies.^16,60 For instance, TBK1 regulates NCOA4 and TAX1BP1, which are mediators of ferritinophagy,^46,47 and GABARAPL2, though not related Atg8 family orthologs, such as MAP1LC3B, which is a common marker of canonical macroautophagy. We demonstrate that TBK1 regulates the late endosome GTPase RAB7A at serine 72 in excitatory neurons, which was previously identified as a target of both TBK1 and LRRK1 in cancer cell lines.^38,39,61 Interestingly, heterozygous missense RAB7A variants are consistently associated with a form of the peripheral neuropathy, Charcot-Marie-Tooth disease with sensory and motor deficits.⁶² Additionally, the cumulative effect of these proteome perturbations is the impairment of neurite outgrowth and lysosomal dysfunction. Overall, we delineate the cell-autonomous effects of loss of TBK1 activity on endo-lysosomal pathways that contribute to neurodegeneration.

Variants in C9ORF72, SQSTM1, and OPTN, encoding autophagy factors, have been consistently associated with familial ALS/FTD.^16,63 We identified novel TBK1-dependent phosphorylation sites in p62/SQSTM1 and in SMCR8, a constitutive C9ORF72-interacting partner that is involved in autophagy initiation.^55,56 Specifically in the cellular context of human neurons, we demonstrate that TBK1 acts to maintain the phosphorylation of OPTN (pS177). In contrast, loss of OPTN alone was insufficient to impair the phosphorylation status of either p62/SQSTM1 or the C9ORF72-SMCR8 heterodimer.

Additionally, our results suggest that TBK1 regulates NCOA7, a relatively uncharacterized interactor of the lysosomal v-ATPase. NCOA7 and OXR1 belong to the TLDc domain protein family and may promote resistance to oxidative stress, which has been implicated in neurodegeneration.⁶⁴ Mice lacking NCOA7 exhibit neurobehavioral abnormalities and impaired lysosomal function in neurons,⁵⁸ and OXR1 variants were recently associated with an autosomal-recessive neurological disease with lysosomal dysfunction.⁶⁵ Overall, we propose that TBK1 is at the nexus of a protein-protein interaction network of ALS/FTD-implicated autophagy proteins (Figure 7).

We found that TBK1 sustains the abundance of its adapter proteins, including AZI2/NAP1, which is also involved in interferon signaling and suppressing inflammation,⁶⁶ potentially analogous to the role of NEMO in IKK signaling.⁶⁷ In human neurons, we demonstrate that TBK1 interacts with and regulates the phosphorylation of AZI2/NAP1, TANK, and TBKBP1/SINTBAD. Loss of TBK1 was not associated with any detected alterations in RIPK1 signaling,⁷ although we did not examine the role of TBK1 in immune cells or in tissue samples for possible non-cell autonomous effects in disease. Further studies might delineate how these adapters contribute to the recruitment of TBK1 to distinct subcellular compartments and its activation toward specific substrates to govern autophagy and inflammation in various cellular contexts. Nonetheless, we demonstrate how the substrates of TBK1 diverge from the closely related kinase IKKE.

In neurons, partial or complete reduction of TBK1 alone did not result in TDP-43 pathological changes, such as the induction of the disease-associated STMN2 alternative “cryptic” exon,^24,32 suggesting that additional factors, such as aging, contribute to the onset of neurodegeneration in LoF variant carriers. Nonetheless, we did observe a trend toward increased TDP-43 phosphorylation in neurons with loss of TBK1 function (Figure S6). As previously shown, disruption of protein homeostasis induces a rapid, yet reversible shift in the localization of TDP-43 from the nucleus to cytoplasm in stem cell-derived neurons.^24,68 Therefore, we speculate that decades of sustained haploinsufficiency for TBK1 kinase activity likely have detrimental effects on excitatory neurons through the dysregulation of a wide array of phospho-proteins involved in autophagy and endo-lysosomal pathways, ultimately leading to a loss of protein homeostasis, which then triggers TDP-43 pathology and neural dysfunction. Hence, restoring TBK1 levels or promoting activation of TBK1 may be potential therapeutic strategies to reverse TDP-43 pathology in ALS/FTD, not only in TBK1 LoF variant carriers.¹⁶

Increased aberrant phosphorylation of several key neurodegeneration proteins, including TDP-43, alpha-synuclein, and tau are thought to promote their misfolding and aggregation.^31,42,44 Unexpectedly, we could detect phospho-TDP-43, phospho-tau, and phospho-alpha synuclein in nascent neurons, suggesting that certain phospho-residues are not necessarily age- or disease-dependent as their functional consequences in neurons continue to be investigated. Additionally, for the modeling of late-onset adult neurodegenerative diseases, our results demonstrate that more pronounced phenotypes may be more readily discernable in iPSCs with complete deletion of a given gene of interest vs. those harboring a single nucleotide variant, which more closely resembles at-risk carriers. Finally, we illustrate a functional genomics approach that would be immediately applicable for the identification of novel targets of additional protein kinases.

Limitations of the study

To generate vast quantities of neurons for our multi-omics approach, we focused on NGN2-induced excitatory neurons, a widely used model to study neurodegeneration.³⁶ It will be interesting to explore TBK1-associated pathways in additional CNS cell types, such as iPSC-derived microglia.⁶⁹ We acknowledge the possibility that a subset of TBK1-dependent phosphoserine residues may be regulated by additional kinases downstream of TBK1. Though interactions between kinases and their substrates are often transient, further biochemical studies on TBK1 as a complex with its adapter proteins may help to understand how it targets selective autophagy proteins.^16,66,70 Validation of CRISPR-based gene editing in a well-characterized reference iPSC line facilitated our sensitive delineation of the impact that TBK1 and OPTN LoF has on the phospho-proteome and lysosomal function. Yet, it is plausible that TBK1 variants would have distinct phenotypes across various genetic backgrounds, which may harbor uncharacterized modifiers of endo-lysosomal pathways. It will be important for future studies to incorporate materials from TBK1-ALS/FTD or autoimmune disease patients, when publicly available in cell line and tissue repositories.

STAR★METHODS

EXPERIMENTAL MODEL AND SUBJECT PARTICIPANT DETAILS

Culture and validation of isogenic human induced pluripotent stem cell lines

The use of human iPSC cell lines was approved by the Human Gamete, Embryo and Stem Cell Research (GESCR) Committee at UCSF. Isogenic iPS cell lines and the parent control line KOLF2.1J were obtained from the Jackson Laboratory (JAX). The reference human stem cell line, a male line, was originally described in full detail by the Merkle lab and collaborators.⁷¹ ALS/FTD-associated variants (SNVs) in isogenic cell lines were confirmed by PCR followed by Sanger sequencing. Genetic deletions (knockout) of protein coding regions were confirmed by PCR, RNA-sequencing, and immunoblotting with antibodies against the corresponding proteins. The genetic background was not authenticated. A revertant derivative line (REV/WT), whereby the disease-associated SNV was previously corrected back to the wild-type allele to control for any potential off-target CRISPR gene editing, was used as a reference line for some experiments and analyses where indicated.

Stem cells were cultured in StemFlex Medium (Life Technologies, A3349401) on Geltrex-coated (LDEV-Free, hESC-Qualified Geltrex basement membrane matrix, Gibco, A1413301) tissue cell culture plates at 37°C in 5% CO₂. ROCK inhibitor (Y-27632, DNSK) was added to media at the time of passaging and removed the next day. All cell lines were routinely tested for mycoplasma with the MycoAlert PLUS Mycoplasma Detection Kit (Lonza, LT07-703), and were negative.

Differentiation of iPSCs to human induced neurons

Human iPSCs were differentiated into neurons by viral transduction using a lentiviral vector carrying doxycycline inducible NGN2 (Neurogenin 2). Lentivirus was prepared by combining NGN2 lentivirus and reverse tetracycline transactivator (rTA) lentivirus based on pIN differentiation with FUW-TetO-Ngn2-Puro and FUW-rtTA (gifts from Marius Werning). After dissociation with Accutase, cells in suspension were incubated with the virus in a 1.5 mL tube for 15 min at room temperature, along with 300 μL of mTESR medium containing ROCK Inhibitor (RI). The cells were then plated into a well of a 6-well plate. Media was changed 24 h after transduction.

Patterned induced neurons were differentiated using doxycycline to induce NGN2 expression rapidly, in combination with small molecule patterning factors, as previously described (Nehme et al., and modified in Pintacuda et al.), except for omitting normocin or geneticin selection.^37,72 On days 5–7 post-induction, neurons were passaged with Accutase treatment onto plates coated with poly-L-ornithine hydrobromide (PLO) and laminin (PLO: Sigma, Cat. No. P3655; laminin: Life Technologies, Cat. No. 23017015). Neurons were cultured in neuron maintenance medium, consisting of Neurobasal medium (Gibco), 0.2% N-2 supplement (Life Technologies), 2% B-27 supplement (Life Technologies), 1% Glutamax, and 1% MEM non-essential amino acids (NEAA) solution (Gibco). This medium was supplemented with 1:10,000 doxycycline, neurotrophic factors (1:10,000 brain-derived neurotrophic factor (BDNF), 1:10,000 ciliary neurotrophic factor (CNTF), 1:10,000 glial cell-derived neurotrophic factor (GDNF), Bio-Techne), and 2 μM 5-Fluoro-2′-deoxyuridine (FUdR) to remove residual proliferating cells. From day 7 onwards, neurons were fed by replacing half of the media with fresh neuron maintenance medium supplemented with neurotrophic factors every 2–3 days.

Cell lines

Validated U2OS cells were obtained from the university cell repository and were cultured in DMEM/F12 with 10% FBS (Fisher Cytiva HyClone, SH3008803). For the transfection of U2OS cells plasmids expressing NCOA7 fused with GFP or GFP alone, were diluted in Opti-MEM media and transfected with Lipofectamine 3000 reagent (Invitrogen) according to the manufacturer’s standard protocol. Cells were harvested 2 days post transfection for FACS analysis.

METHOD DETAILS

Genotyping

Genomic DNA was extracted from 2 million human iPSCs using the Qiagen Tissue and Blood DNA Extraction Kit (Cat. No. 69504). DNA amplification by PCR was carried out using the Promega GoTaq G2 Green Master Mix (Cat. No. M7822) with the following primers for the targeted mutations. See primer list in Table S9. Sanger sequencing was performed by Quintara Biosciences to confirm the presence of the target mutations.

Quantitative PCR

Reverse transcription of RNA to complementary DNA was performed using the iScript Advanced cDNA Synthesis Kit (Bio-Rad, Cat. No. 1725038). Quantitative RT-PCR (qRT-PCR) was conducted using diluted cDNA, primers synthesized by Integrated DNA Technologies, and the Maxima SYBR green qPCR master mix 2X (Thermo Fisher, Cat. No. K0253). Reactions were run on a Real-Time PCR System (CFX96, Bio-Rad). All reactions were run in technical duplicates or triplicates on the thermocycler and averaged for analysis. Gene expression levels were normalized to GAPDH expression unless otherwise specified. Normalized gene expression values are presented relative to the control sample. Primer pair sequences were obtained from prior publications or the MGH PrimerBank, and sequences are provided in Table S9 (Document S1).

Immunoblotting

For Western blotting, stem cells and neurons were lysed in Pierce RIPA lysis buffer (PI89900 Thermo Scientific) supplemented with a phosphatase and protease inhibitor cocktail (Halt Protease and Phosphatase Inhibitor Cocktail 78440 ThermoFisher Scientific). After protein concentration determination by Pierce BCA Protein assay kit (Fisher PI23227) to adjust the quantity of protein loaded per well to 30μg. Cleared lysates were combined with sample buffer containing DTT (Cell Signaling Technology) and run on gradient proteins gels (Mini-PROTEAN TGX Precast Gels, Biorad) and transferred to a PVDF membrane using the iBlot2 system. Following blocking, primary antibodies were applied-, anti-GAPDH (clone 6C5 mouse, Millipore Sigma, MILL-CB1001), TBK1/NAK (clone E8I3G Rabbit, Cell Signaling, 38066), OPTN (clone E4P8C Rabbit, Cell Signaling, 70928) - then anti-mouse or anti-rabbit LICOR IRdye secondary antibodies. Signal detection and quantification was performed on the LICOR ODYSSEY system.

Microscopy

Cells were fixed with ice-cold 4% paraformaldehyde (PFA) for 15 min at 4°C, permeabilized with 0.2% Triton Xin 1x PBS for 1 h, and blocked with 10% donkey serum in 1x PBS-T (0.1% Triton X-) for 1 h. The following primary antibodies: OCT4 (mouse, Santa Cruz, sc-5279), TDP-43 (proteintech, 10782-2-AP, 1/500), and Tuj1 (R&D systems, MAB1195) were diluted in the blocking solution and incubated overnight at 4°C. After at least four washes (5 min each) with 1x PBS-T, cells were incubated with secondary antibodies for 1 h at room temperature (diluted in blocking solution). Nuclei were stained with DAPI. OCT4 images were acquired using a Leica DMi8 Inverted Microscope Widefield with an HC PL APO 20x/0.80 objective and TDP-43 images with the CSU-W1 SoRA Spinning Disk confocal microscope confocal with 40×air objective.

To label lysosomes, iNeurons were plated at a density of 100K cells per well in a glass bottom 24 well plate. At D21, the SIRLYSOSOME (CY-SC012) dye was diluted at 1 μM in the neuron media and incubated 30min at 37°C, 5% CO2. Images were acquired on lived cell with a Leica DMi8 Inverted Microscope Widefield equipped with an HC PL APO 40x. Images were processed and quantified with the Fiji (ImageJ) software. After applying a threshold to define lysosomes, the intensity of fluorescence was measured.

Neurite growth analysis

For longitudinal tracking, neurons from each genotype were plated at the same cell density, 7,000 neurons per well, in a 96-well plate and monitored for 5 days in the incubator (37°C). The plate was scanned every 30 min using a 10X objective with an Incucyte SX1 Live-Cell Analysis System (Sartorius, Göttingen, Germany). Four phase-contrast images per well were captured, and the cell body cluster, neurite length, and neurite branches were quantified using the Incucyte Neurotrack Analysis Software Module (9600-0010). Each condition was tested in quadruplicate, and three independent experiments were performed. In parallel, a neuron survival measurement was performed using images acquired with the Incucyte on the same experiment. Cell counts were manually performed on images selected at 12-h intervals, across 11 time points with 4 wells and 5 images per well analyzed per genotype. The experiment was independently replicated to ensure reproducibility.

Motor neuron differentiation

Differentiation of hiPSC into lower induced motor neurons (liMoNes/liMNs) was performed through a combination of doxycycline to rapidly induce NGN2 expression and small molecule patterning factors as previously described (Limone et al.).⁷³

Global proteomics and phospho-proteomics

Protein extraction and in solution digestion

Cells (20–40 million cells per sample) were harvested and pelleted at 200g, washed 3 times in PBS, and after removing the supernatant, pellets were flash-frozen and stored at −80°C until processing. For protein extraction and tryptic digestion, samples were thawed and homogenized in 1000 μL 8 M Guanidine Hydrochloride (GndHCl), using a probe sonicator (ThermoFisher Scientific). Protein concentration in the extracts was measured using the Pierce BCA Protein Assay Kit (Thermo Scientific) in triplicates, taking 2 μL aliquots of this solution. Aliquots of the samples containing 300 μg of protein were added 9X the sample volume of ethanol and allowed to precipitate at −80C overnight. After that, pellets were centrifuged at 15000xg 10min, the supernatant was removed, and the samples were allowed to dry. The pellets were resuspended in 12.5 μL 6.25 M guanidine hydrochloride containing 6.25% V/V of Sigma-Aldrich Phosphatase inhibitors cocktails 2 and 3, plus 78 mM triethylammonium bicarbonate (TEAB) pH 8.0, and 7.8 mM Tris(2-carboxyethyl)phosphine hydrochloride. Samples were sonicated in a bath sonicator for 5 min, then incubated at 56°C for 15 min, followed by a 30-min incubation at room temperature in the dark with 20 mM iodoacetamide. The samples were added 62.5 μL TEAB containing 2 μg Lysyl endopeptidase (LysC) (FUJIFILM-Wako), and incubated at 37C o/n. After that, the samples were added 5% (W/W) modified trypsin (TPCK-Trypsin, Thermo Scientific), and incubated overnight at 37°C.

Digested Samples were labeled according to TMTProTM-18 label plex kit instructions (ThermoFisher Scientific), with some modifications. Briefly, TMT reagents were dissolved in acetonitrile at 25 μg/μL, and 20 μL of these stocks added to the samples (500 μg reagent, 1.7-fold the peptide mass amount). After incubation for 1 h at room temperature samples were quenched with 2 μL 5% hydroxylamine, then all samples were combined adding them over 20 mL 0.1% formic acid, and desalted using a C18 SepPak cartridge (Waters). The Sep Pak eluate was dried in preparation for phosphopeptide enrichment.

Enrichment of phosphorylated peptides using titanium dioxide

Phosphopeptide enrichment was performed in an AKTA Purifier (GE Healthcare, Piscataway, NJ) using 5 μm titanium dioxide (TiO2) beads (GL Sciences, Tokyo, Japan) in-house packed into a 4.0 mm × 3 cm analytical column (Upchurch Scientific, Oak Harbor, WA). Combined TMT labeled tryptic digests were resuspended in 240 μL buffer containing 35% MeCN, 200 mM NaCl, 0.4% TFA and loaded onto the TiO₂ column at a flow rate of 2 mL/min. The non adsorbed peptides (non-phosphorylated material, flow though fraction) were recovered. The column was then washed for 2 min with 35% MeCN, 200 mM NaCl, 0.4% TFA to remove non phosphorylated peptides. Phosphopeptides were eluted from the column using 1 M potassium phosphate monobasic (KH₂PO₄) at a flow rate of 0.5 mL/min for 30 min, directly onto an on-line coupled C18 column (Luna 10 μm C18(2) 100 Å , LC Column 30 × 4.6 mm, Phenomenex) This column was washed with 0.1% trifluoroacetic acid (TFA) for 14 min, and the adsorbed material was eluted in 400 μL of 50% MeCN, 0.1% TFA at a flow rate of 0.25 mL/min. The eluate was solvent evaporated and then resuspended in 240 μL 20 mM ammonium formiate pH 10.4 for fractionation of the peptide mixture by high pH RP chromatography. The flow through fraction was solvent evaporated, desalted in a C18 SepPak cartridge (Waters), then resuspended in 240 μL 20 mM ammonium formiate pH 10.4 for fractionation of the peptide mixture by high pH RP chromatography.

High pH reverse phase chromatography

The phosphopeptide enriched sample was fractionated on an AKTA purifier system utilizing a Phenomenex Gemini 5u C18 110A 150 × 4.60 mm column, operating at a flow rate of 0.550 mL/min. Buffer A consisted of 20 mM ammonium formate (pH 10), and buffer B consisted of 20 mM ammonium formate in 90% acetonitrile (pH 10). Gradient details were as follows: 1% to 30% B in 49 min, 30% B to 70% B in 4 min, 70% B down to 1% B in 4 min. Peptide-containing fractions were collected, evaporated and resuspended in 0.1% formic combining early and late chromatographic fractions to produce between 48 and 52 samples for LCMSMS analysis.

Mass spectrometry analysis

Aliquots (containing around 3 μg of digested material) coming from RP fractionation were run onto a 2 μm, 75mm ID x 50 cm PepMap RSLC C18 EasySpray column (Thermo Scientific). 3-h MeCN gradients (2–25% in 0.1% formic acid) were used to separate peptides, at a flow rate of 200 nL/min, for analysis in a Orbitrap Exploris 480 or a Orbitrap Lumos Fusion (Thermo Scientific) in positive ion mode with the following settings. MS spectra were acquired between 375 and 1500 m/z with a resolution of 120000. For each MS spectrum, multiply charged ions over the selected threshold (2E4) were selected for MSMS in cycles of 3 s with an isolation window of 0.7 m/z. Precursor ions were fragmented by HCD using stepped relative collision energies of 30, 35 and 45 in order to ensure efficient generation of sequence ions as well as TMT reporter ions. MSMS spectra were acquired in centroid mode with resolution 60000 from m/z = 120 in the Exploris, 50000 from m/z = 110 in the Lumos. A dynamic exclusion window was applied which prevented the same m/z from being selected for 30s after its acquisition.

Peptide and protein identification and TMT quantitation

Peak lists were generated using PAVA software.⁷⁴ All generated peak lists were searched against the human subset of the SwissProt database (SwissProt.2019.07.31), using Protein Prospector.⁷⁵ Analytical chemistry 71, 2871–2882) with the following parameters: Enzyme specificity was set as Trypsin, and up to 2 missed cleavages per peptide were allowed. Carbamidomethylation of cysteine residues, and TMTPro16plex labeling of lysine residues and N terminus of the protein were allowed as fixed modifications. N-acetylation of the N terminus of the protein, loss of protein N-terminal methionine, pyroglutamate formation from of peptide N-terminal glutamines, oxidation of methionine and phosphorylation on serine, threonine and tyrosine were allowed as variable modifications Mass tolerance was 5 ppm in MS and 30 ppm in MS/MS. The false discovery rate was estimated by searching the data using a concatenated database which contains the original SwissProt database, as well as a version of each original entry where the sequence has been randomized. A 1% FDR was permitted at the protein and peptide level, for the sets of unphosphorylated, monophosphorylated and doubly phosphorylated peptides. For quantitation only unique peptides were considered; peptides common to several proteins were not used for quantitative analysis. Relative quantization of peptide abundance was performed via calculation of the intensity of reporter ions corresponding to the different TMT labels, present in MS/MS spectra. Intensities were determined by Protein Prospector. Summed intensity per sample on each TMT channel for all identified spectra were used to normalize individual intensity values. Relative abundances were calculated as ratios vs. the average intensity levels in the channels corresponding to the relevant control samples. Spectra representing replicate measurements of the same peptide were kept. Individual spectral ratios were aggregated to the peptide level using median values of the log2 ratios. Non-phosphorylated peptides were used to estimate relative protein levels. For total protein relative levels, peptide ratios were aggregated to the protein levels using median values of the log2 ratios. Phosphopeptide ratios were corrected by the available protein ratios. Statistical significance was calculated at the phosphopeptide and protein levels by comparing the values for the 3 biological replicates of the relevant groups in the TMT experiment with a 2-tailed t test.

Gene set enrichment analysis (gene ontology (GO) and KEGG) of ranked proteins (or corresponding proteins from ranked peptides based on significance) for proteomic results were performed using g:Profiler.⁷⁶

For a given phospho-peptide containing a phosphoserine residue, the scoring and rankings of predicted protein kinases were performed with the Kinase Library developed by the Cantley and Yaffe labs.¹ Mapping of the TBK1 dimer and its phosphorylation sites on predicted structures of selective autophagy cargo receptors was performed using RSB protein databank (PDB) and alphafold.⁷⁷

AP-MS

Control iNeurons were harvested at D17 for the TBK1 experiment with 3 million cells collected per sample. Proteins were extracted from the cells using Pierce IP Lysis Buffer (Thermo Scientific), which was supplemented with protease and phosphatase inhibitors. The cell pellet was resuspended in 1 mL of lysis buffer, and the mixture was incubated on ice for 30 min to allow for protein extraction. After incubation, the lysate was clarified by centrifugation at maximum speed (15,000 × g) for 15 min at 4°C. The supernatant was collected as the cleared lysate for subsequent immunoprecipitation.

TBK1

To enrich for the protein of interest, 0.5 μg of primary antibody was added to 1 mL of cleared lysate. The antibodies used included the IgG control (Biovision, catalog #1268-100), the TBK1/NAK antibody (Cell Signaling, catalog #38066; E8I3G), or the TBK1 antibody (Abcam, catalog #ab109735). The lysates were then incubated with washed Protein A/G magnetic beads (Thermo Scientific, Cat# 88802) to capture the antibody-protein complex. The incubation was carried out overnight at 4°C with gentle rotation. Following incubation, the beads were washed four times with 300 μL of lysis buffer. Each wash step involved incubating the beads on ice for 5 min, followed by placing the sample on a magnetic rack for 1 min to separate the beads from the solution. The supernatant was discarded after each wash.

Sample preparation for LC-MS analysis

Bound antigens were eluted by adding 1 μg of trypsin in 100 μL of 50 mM TEAB (pH 8.5) and incubating at 37°C with shaking at 800 rpm for 1 h. Beads were separated using a magnetic stand, and the eluate was transferred to a fresh microcentrifuge tube. Proteins were reduced by adding 2 μL of 500 mM DTT and incubated at 56°C for 30 min. After cooling to room temperature, 3 μL of 375 mM IAA was added, followed by incubation for 30 min in the dark. Trypsin digestion was performed by adding 1 μg of trypsin per sample in 50 mM TEAB and incubating at 37°C for 6 h with shaking at 500 rpm. Digestion was stopped by adding 10% trifluoroacetic acid (TFA), adjusting the pH to ~3. Samples were centrifuged at 15,000 × g for 2 min to remove insoluble material, and the supernatant was transferred to a clean tube. Peptides were desalted using C18 tips, dried using a SpeedVac concentrator, and resuspended in 0.1% formic acid in water before LC-MS analysis.

LC-MS

An Orbitrap Fusion Lumos Tribrid mass spectrometer (Thermo Fisher Scientific) connected to an Ultimate 3500 RSLC nano system (Thermo Fisher Scientific) was used for analysis of peptides samples. Mobile phases A and B consisted of 0.1% (v/v) formic acid and acetonitrile, 0.1% (v/v) formic acid, respectively. Samples were loaded in 2% acetonitrile, 0.1% formic acid on an Easy-Spray column (C18,15 cm, 50 μm ID, 2 μm particle size, 100 Å pore size) running at 0.3 μl/min flow and kept at 45°C. Analytes were eluted with the following gradient: 2%–40% buffer B in a linear 60-min gradient, followed by an increase to 95% buffer B over 2 min. The column was then maintained at 95% buffer B for 10 min. Mass spectra (m/z = 375–1500) were acquired with mass resolving power = 120,000 for MS1 and 30,000 for MS/MS in positive mode. The MS/MS analyses were performed by m/z 1.6 isolation with the quadrupole, a normalized higher-energy collision dissociation (HCD) collision energy of 30%. Dynamic exclusion was set to 60 s. Monoisotopic precursor selection (MIPS) was set to Peptide. The maximum injection time was set to 100 ms.

Data analysis

Database searching was performed on Proteome discoverer 3.0 (Thermo Fisher Scientific) with SEQUEST HT search engine. Database search parameters were set as follows: precursor ion mass tolerance was set 10 ppm; product ion mass tolerance was set 0.02 Da; UniProt human proteome database (UP000005640, reviewed) was used as database. Trypsin was set for enzyme digestion with maximum 2 missed cleavage. Oxidation on methionine, acetylation on protein N-terminal and deamination on asparagine or glutamine were set as variable modifications. Carbamidomethylation on cysteine was set as fixed modification. The peptide FDR was set as 1%. For the assessment of candidate TBK1 interacting proteins, the list of proteins detected by mass spec. was narrowed based on abundance ratio of greater than 3 (anti-TBK1 relative to IgG control), greater than 2 peptides detected, and mapping to the human proteome. Additionally, proteins that were predicted to be common interactors in mass. spec. experiments from a database of 716 experiments in the CRAPome 2.0 were filtered.

TBK1 co-IP

To confirm the AP-MS results, TBK1 co-immunoprecipitation was performed. After incubation of the beads, antibodies, and proteins, the samples were washed with lysis buffer. Bound protein complexes were then resuspended in sample buffer containing DTT and denatured by incubating at 95°C for 5 min. Denatured samples were subjected to immunoblotting. For the detection of TBK1 interactors, PVDF membranes were incubated independently with the AZI2, TANK, CYLD, TBK1 primary antibodies.

Flow cytometry analysis for lysosome assays

For the lysosomal assay, human iPSCs were pretreated with 100 μM Chloroquine for 24 h, or left untreated. After 24 h, cells were harvested by incubating with Accutase for 10 min at 37°C, followed by a PBS wash. Cells were then incubated with Invitrogen LysoTracker Deep Red (Cat. No. L12492) at a final concentration of 0.5 μM (1/50,000 dilution in PBS) for 30 min at RT in the dark. After incubation, cells were washed with PBS and resuspended in 100 μL of 0.5% BSA in PBS for flow cytometry analysis. For control and autofluorescence measurements, non-stained cells were used to determine gating for fluorescence. The APC signal was measured for fluorescence intensity. To assess Lysotracker fluorescence in U2OS cells overexpressing NCOA7, the same protocol was followed, with two modifications: chloroquine treatment was limited to 4 h, and cells were harvested using trypsin instead of Accutase.

To assess Cathepsin B activity, human iPSCs were pretreated with 100 μM Chloroquine for 24 h, or left untreated. After 24 h, cells were harvested with Accutase, then incubated with the Cathepsin B Assay Kit (Magic Red, Abcam, Cat. No. ab270772) for 1 h at 37°C, protected from light. Immediately prior to staining, the Magic Red dye stock solution was diluted 1:10 with deionized water to form the staining solution. Cells were stained with a final 1/5 dilution (20 μL of the diluted solution in 100 μL of cell suspension). After incubation, cells were gently pelleted by centrifugation at 300 x g for 5 min at room temperature, and the supernatant was discarded. The cells were then resuspended in 100 μL PBS containing 0.5% BSA for flow cytometry analysis. The PE-CF594-A signal was measured for fluorescence intensity.

Flow cytometry analysis was performed using a BD Weill FACS Aria Fusion, with appropriate settings for fluorescence detection. Data were analyzed using FlowJo v10.10.0 software.

RNA sequencing

Total RNA was extracted from control, TBK1 KO, OPTN KO, TBK1 E696K^SNV/WT, and TBK1 E696K^SNV/SNV hiPSCs using the RNeasy Plus Mini Kit (Qiagen), and from differentiated iNeurons D21 using the RNeasy Micro Kit (Qiagen). Each sample consisted of 4 million cells, with RNA extracted from four independent passages. The RNA concentration for each hiPSC sample was standardized to 500 ng/μL, while RNA from the iNeurons met a minimum concentration threshold of 150 ng/μL. All purified RNA samples underwent quality control (QC), with RNA integrity numbers (RINs) exceeding 9, ensuring that all samples met the required quality for downstream analyses. Samples were sent to Novogene for library preparation and bulk RNA-sequencing analyses.

QUANTIFICATION AND STATISTICAL ANALYSIS

General statistics

Statistical analyses were performed using GraphPad Prism 10.4.1 (GraphPad Software, La Jolla, CA, USA). For comparisons between two groups, a 2-tailed unpaired Student’s t-test was used, with a p-value of p < 0.05 considered statistically significant unless stated otherwise. For multiple group comparisons, a one-way ANOVA with appropriate multiple comparison corrections was applied. Data are presented as mean ± standard deviation (SD), unless otherwise noted. Significance levels are indicated as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Individual p values for selected comparisons are provided in Table S10 (in Document S1). Each dot represents an independent experiment or technical replicate as indicated in the figure legend.

Supplementary Material

Supplemental information can be found online at 10.1016/j.celrep.2025.116494.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
TBK1/NAK (E8I3G), Rabbit mAb	Cell Signaling Technology	Cat# 38066, RRID:AB_2827657
Anti-NAK/TBK1 antibody, Rabbit mAB	Abcam	Cat# ab109735, RRID:AB_10863562
Optineurin (E4P8C) Rabbit mAb	Cell Signaling Technology	Cat# 70928, RRID:AB_3073769
TANK antibody	Cell Signaling Technology	Cat# 2141, RRID:AB_2302999
CYLD (D1A10) Rabbit mAb	Cell Signaling Technology	Cat# 8462, RRID:AB_10949157
Oct3/4 Antibody (C-10)	Santa Cruz	Cat# sc-5279, RRID:AB_628051
AZI2/NAP1 antibody	Proteintech	Cat# 15042-1-AP, RRID:AB_2878103
Anti-GAPDH Mouse mAb (6C5)	Millipore Sigma	Cat# CB1001, RRID:AB_2107426
TDP-43 antibody, rabbit	Proteintech	Cat# 10782-2-AP, RRID:AB_615042
Alexa Fluor 488 Secondary anti-Mouse	Invitrogen (Thermo)	A-21202
Alexa Fluor 555 Secondary anti-Rabbit	Invitrogen (Thermo)	A-31572
Rabbit IgG (control)	Biovision	Cat# 1268-100, RRID:AB_478072
IRDye 680RD Goat anti-Mouse IgG	LI-COR Biosciences	Cat# 926–68070, RRID:AB_10956588
IRDye 800CW Goat anti-Rabbit IgG	LI-COR Biosciences	Cat# 926–32211, RRID:AB_621843
Chemicals, peptides, and recombinant proteins
Chloroquine (diphosphate salt)	Sigma-Aldrich	C6628
Accutase	Fisher	NC1793126
Geltrex	Thermo Fisher Scientific	A1413301
StemFlex medium	Thermo Fisher Scientific	A3349401
Neurobasal	Life Technologies	21103049
N2 supplement (Gibco)	Thermo Fisher Scientific	17502048
Non-essential amino acids (Gibco)	Thermo Fisher Scientific	11140035
Glutamax (Gibco)	Life Technologies	35050061
Doxycycline	Sigma-Aldrich	D9891-1G
B27 supplement (Gibco)	Thermo Fisher Scientific	17504044
Neutrotrophics: BDNF, GDNF, CNTF	Bio-Techne	11166-BD, 212-GD, 257-NT
DMEM/F-12	Fisher	11320082
Laminin	Life-Technologies	23017015
poly-L-ornithine (PLO)	Sigma-Aldrich	P3655
FBS	Fisher Cytiva HyClone	SH3008803
Critical commercial assays
TMTPRO 18-PLEX kit for TMT peptide labeling	Life Technologies	A52045
Deposited data
Raw proteomic reads (see Tables S1-S8 for data)	This paper	ProteomeXchange: PXD068150
Raw proteomic AP-MS reads (see Table S2 for data)	This paper	ProteomeXchange: PXD068535
RNA-sequencing reads	This paper	GEO GSE308008, GSE308151, SRA PRJNA1328198
Experimental models: Cell lines
KOLF2.1J iPSC and isogenic derivatives	Jackson Laboratory (INDI)	hPSCreg: WTSli018-B-12; https://www.jax.org/jax-mice-and-services/ipsc/
U2OS human bone osteosarcoma epithelial cells		RRID:CVCL_0042
Oligonucleotides
Primers for qPCR, see Table S9 in Document S1		N/A
Recombinant DNA
Custom NCOA7 plasmid: NCOA7 ORF (NM_181782; RC221673) insert in pLenti-EF1a-C-mGFP-P2A-Puro	Origene	CW311001
pLenti-EF1a-C-mGFP-P2A-Puro	Origene	PS100121
Software and algorithms
FlowJo™ v10.10.0	BD Life Sciences	https://www.flowjo.com/
GraphPad Prism v10.4.1	GraphPad Software, La Jolla, CA, USA	https://www.graphpad.com/
Fiji	ImageJ	https://imagej.net/software/fiji/
Incucyte® Neurotrack Analysis Software Module	Sartorius	https://www.sartorius.com/
Proteome Discoverer 3.0	Thermo Fisher Scientific	https://www.thermofisher.com/us/en/home.html
Xcalibur	Thermo Fisher Scientific	https://www.thermofisher.com/us/en/home.html
Protein Prospector	In-house/UCSF Mass Spectrometry Facility	https://prospector.ucsf.edu/prospector/mshome.htm

Highlights.

• Differential proteomics in isogenic iPSC lines identifies targets of the kinase TBK1

• TBK1 regulates ALS/FTD-associated selective autophagy proteins in iPSC-derived neurons

• Loss of TBK1 function has cell-autonomous effects on endo-lysosomal pathways

• Deep proteomics provides a reference phospho-protein atlas for human neurons