Figure 1: TDP-43 is subject to lysine acetylation within the RNA-binding domain a) QBI-293 cells were transfected with TDP-43 in the presence or absence of the acetyltransferase CBP and acetylated TDP-43 was evaluated by immunoprecipitation (anti-tdp-43, clone 171) followed by immunoblotting using anti-acetylated lysine antibody. b) Cells transfected with wild-type TDP-43 were pulse-labeled with [ 3 H]- acetate for 2 hr followed by immunoprecipitation and analysis by SDS-PAGE and autoradiography for 2 weeks. Total cellular lysates were immunoblotted using anti-tdp- 43 and GAPDH antibodies. c) TDP-43 acetylation was determined by immunoblotting using either wild-type TDP-43 or the cytoplasmic mutant TDP-43- NLS in the presence of the acetyltransferase CBP. Samples were analyzed by either immunoprecipitation or direct immunoblotting using acetylated lysine and total TDP-43 (1038) antibodies.
Supplementary Figure 2: TDP-43 acetylation of a panel of ALS-associated TARDBP mutations a) A schematic of the TDP-43 protein depicting the nuclear localization sequence (NLS), RNArecognition motifs (RRMs), and C-terminal glycine-rich domain harboring the majority of the ALS-associated genetic mutations. b) Cells were co-transfected with Creb-binding protein (CBP) and the following cytoplasmic targeted expression plasmids: unmodified TDP-43-ΔNLS (positive control for acetylation), TDP-43-ΔNLS containing ALS-associated mutations (A90V, G294A, G295S, G298S, R361S), or TDP-43-ΔNLS containing 4FL RNA-binding deficient mutations in RRM1 and RRM2 (F147/149/229/231L). Cell lysates were evaluated by immunoprecipitation using anti-tdp-43 (171) and western analysis using an anti-acetyl-lysine antibody. Overall, there were no appreciable differences in TDP-43 acetylation status among the various cytoplasmic targeted mutants compared to TDP-43-ΔNLS, as determined by acetyllysine antibody detection.
Supplementary Figure 3: Mass spectrometry of immunoprecipitated TDP-43 upon co-expression with Creb-binding protein (CBP) acetyltransferase QBI-293 cells were transfected with vector alone or TDP-43- NLS in the absence or presence of CBP and total TDP-43 was immunoprecipitated using anti-tdp-43 clone 171. Samples were separated by SDS-PAGE followed by gel excision and analyzed by mass spectrometry, which identified the Lys-145 and Lys-192 acetylated peptides shown in Supplementary Figure 4.
Supplementary Figure 4: Ion scores from acetylated Lys-145 and Lys-192 peptides identified from mass spectrometry analysis Acetylated peptides were identified by mass spectrometry analysis from immunoprecipitated TDP-43-ΔNLS protein. Listed in the table are the ion scores that correspond to the m/z spectrums obtained. The numbers listed in red bold illustrate statistical significance and the presence of acetylation at Lys-145 (top) and Lys-192 (bottom). Mass spectrometry data were acquired with Xcaliber software (Thermo Fisher) and analyzed using PEAKS 6.0 (Bioinformatics Solutions Inc.) and Scaffold 3 software systems.
Supplementary Figure 5: TDP-43 acetylation occurs on Lys-145 and Lys-192 within the RRM domains Cytoplasmic TDP-43-ΔNLS or a comparable construct containing K R mutations at Lys-145 and Lys-192 (TDP-43-ΔNLS-2KR) were analyzed by immunoblotting in the presence of CBP and HDAC/Sir2 family inhibitors (overnight incubation with 5 mm nicotinamide/ 1 m Trichostatin A) to promote full TDP-43 acetylation in cultured cells. The acetyl-lysine positive bands >40 kd reflects total cellular acetylated proteins that are readily detected with this antibody.
Supplementary Figure 6: TDP-43 acetylation causes accumulation of insoluble aggregates Soluble (S) or insoluble (I) fractions from cells expressing acetylation mimics (KQ mutants) and non-mimics (KR mutants) (a-b) or CBP-mediated acetylated TDP-43 (c) were analyzed by immunoblotting using the indicated myc-9e10, TDP-43 (1038), phospho-tdp-43 (P409/410), or GAPDH antibodies.
Supplementary Figure 7: TDP-43 acetylation-mimics promote TDP-43 aggregation in differentiated Neuro2A cells a) Neuro2A cells were transfected with cytoplasmic TDP-43 (TDP-43-ΔNLS), or plasmids containing acetylation mimic mutations (TDP-43-ΔNLS-K145Q or TDP-43-ΔNLS- K145/192Q) for 48 hrs followed by differentiation by serum deprivation for an additional 24 hrs to extend neuritic processes. Soluble (S) and insoluble (I) cell lysates were analyzed by immunoblotting using P-409/410, myc (9E10), and GAPDH antibodies. b) Differentiated Neuro2A cells expressing non-mimic TDP-43 (left panel) showed limited P-409/410 immunoreactivity within neurites (dashed box region is shown at higher exposure within inset to avoid image over-exposure). In contrast, the acetylation-mimic mutant (right panel), displayed strong accumulation of P-409/410-positive TDP-43 aggregates within the cell soma and distally within extended neuritic processes. See white arrows highlighting transfected cells containing myc-positive TDP-43 aggregates in both cellular compartments. Scale bar represents 50 m.
Supplementary Figure 8: TDP-43 acetylation impairs RNA binding a) Recombinant full length wild-type TDP-43, non-mimic TDP-2KR and acetylation mimic TDP-2KQ mutants were purified by a sequential extraction detailed in the methods and analyzed by Coomassie staining. b) Cells expressing vector control, TDP-43-shRNA, wild-type TDP-43, TDP-K145Q, or RNA-binding deficient 4FL plasmids were cotransfected for subsequent analysis of CFTR minigene splicing. TDP-43 protein levels were determined by immunoblotting. c) Cross-linking immunoprecipitation (CLIP) assay was performed on cells expressing wild-type TDP-43, K145Q, 2KQ (K145/192Q), or TDP-4FL plasmids. Levels of the TDP-43 proteins were assessed by immunoblotting using anti-tdp-43 (1038) and GAPDH as a loading control.
Supplementary Figure 9: Oxidative stress induces TDP-43 acetylation Immunoblotting of cell lysates from soluble and insoluble fractions indicates accumulation of acetylated TDP-43-ΔNLS upon co-transfection of CBP and exposure to 0.2 mm sodium arsenite for 1 hr.
Supplementary Figure 10: Characterization of an acetylated TDP-43 antibody (Ac- K145) a-b) ELISA assays were performed with the indicated antibodies against chemically acetylated (a) or unmodified (b) Lys-145-containing peptides consisting of the following amino acid sequence: TGHSKGFGFVR. Shown is a representative ELISA assay from N=3 independent biological replicates, and error bars represent standard deviation (SD) among triplicate experimental samples. As shown, anti-acetyl-lysine and acetylated TDP- 43 antibody (Ac-K145) specifically detected the Lys-K145 acetylated peptide, but not the unmodified peptide, indicating a site-specific acetylated TDP-43 antibody. c) Recombinant WT TDP-43 or TDP-2KR (K145/192R) proteins were incubated in an acetylation reaction lacking cofactor (mock), or containing 0.4 mm coenzyme A (coa) or acetyl coenzyme A (Ac-coA) in the presence of 0.5 μg recombinant CBP (Enzo Life Sciences) for 1 hr at 37 followed by immunoblotting with Ac-K145. Acetylated WT TDP-43 protein was specifically detected by Ac-K145, while the TDP-2KR mutant protein (K145/192R) showed diminished CBP-mediated acetylation. Lys-145 and Lys- 192 residues possess similar flanking amino acid sequences and thus the TDP-2KR mutant was employed to abolish a common acetylated epitope.
Supplementary Figure 11: TDP-43 acetylation is associated with insoluble and phosphorylated aggregates a) Immunoblotting with anti-tdp-43 (1038) depicts the robust accumulation of insoluble TDP-43 upon arsenite stress. Immunoprecipitating TDP-43 from the urea solubilized fraction with Ac-K145 followed by immunoblotting with a TDP-43 monoclonal antibody (ProteinTech) allowed increased sensitivity for detection of acetylated TDP-43 in the absence of CBP co-transfection. b) Immunoblotting analysis of TDP-43- NLS and the double lysine mutant, TDP-43- NLS-2KR, was evaluated in an arsenite-induced aggregation time-course (0-3hr, 0.1mM sodium arsenite), which resulted in slightly delayed accumulation of phosphorylated (P-409/410) TDP-43- NLS-2KR aggregates compared to the cytoplasmic control, TDP-43- NLS.
Supplementary Figure 12: The deacetylase HDAC6 is sufficient to mediate TDP-43 deacetylation under basal conditions in QBI-293 cells Cells were co-transfected with TDP-43-ΔNLS, CBP, and either wild-type HDAC6 or the partially catalytically inactive mutant, HDAC6-H611A, and TDP-43 acetylation was determined by immunoblotting with acetyl-lysine antibodies.
Supplementary Figure 13: A stress-inducible interaction between TDP-43 and HDAC6 a) Cells transfected with myc-tagged TDP-43- NLS and FLAG-tagged HDAC6 were analyzed by co-immunoprecipitation reactions under basal or stimulated conditions or in the presence of 0.1 mm arsenite or 1 mm hydrogen peroxide stress for 2 hr. Input, representing 5% of total lysate volume, was analyzed by immunoblotting using the indicated antibodies. b) Co-immunoprecipitation analysis using either TDP-43- NLS or acetylation mimic mutants (nuclear TDP-43-K145Q or cytoplasmic TDP-43- NLS- K145Q) and wild-type HDAC6 were performed in the absence or presence of arsenite, and similarly analyzed by immunoblotting.
Supplementary Figure 14: Induction and activation of CHIP and Hsp70 correlate with cytoplasmic TDP-43 aggregation in response to arsenite stress Protein levels of the HDAC6/CHIP/Hsp70 complex were determined by western analysis of soluble and insoluble fractions using HDAC6, CHIP, Hsp70, and Hsp40 specific antibodies. Note, CHIP dimers suggest activation of CHIP E3 ligase function.
Supplementary Figure 15: Acetylated TDP-43 aggregates recruit and physically associate with the CHIP/Hsp70 complex a) Double-labeling immunofluorescence microscopy illustrates that TDP-43 aggregates induced by either exposure to 0.2 mm arsenite for 1 hr (top rows) or acetylation-mimic mutations in the absence of stress (TDP-43-ΔNLS-K145Q, bottom row) are sufficient to recruit both CHIP and Hsp70 to cytoplasmic TDP-43 aggregated foci (see merged panels). Scale bar represents 25 μm. b) HDAC6-FLAG was immunoprecipitated followed by detection of myc-tagged cytoplasmic TDP-43 (TDP-43-ΔNLS) or endogenous CHIP. The data indicate formation of an arsenite-inducible protein complex (HDAC6/CHIP/TDP-43) that occurs independently of HDAC6 catalytic activity (compare lanes 2 and 3) (CD=catalytically-dead).
Supplementary Figure 16: HDAC6 depletion leads to increased TDP-43 phosphorylation and aggregation Depletion of HDAC6 was performed using a specific HDAC6 sirna followed by arsenite treatment, and solubility of TDP-43-ΔNLS and TDP-43-ΔNLS-2KR proteins were analyzed by immunoblotting with total and phosphorylated TDP-43 antibodies. HDAC6 sirna knockdown efficiency was validated by assessing total levels of HDAC6 (anti-hdac6 H-300, Santa Cruz) and acetylated tubulin (611B1, Sigma).