Endocytosis regulates TDP-43 toxicity and turnover

Abstract

Amyotrophic lateral sclerosis (ALS) is a fatal motor neuron degenerative disease. ALS-affected motor neurons exhibit aberrant localization of a nuclear RNA binding protein, TDP-43, into cytoplasmic aggregates, which contributes to pathology via unclear mechanisms. Here, we demonstrate that TDP-43 turnover and toxicity depend in part upon the endocytosis pathway. TDP-43 inhibits endocytosis, and co-localizes strongly with endocytic proteins, including in ALS patient tissue. Impairing endocytosis increases TDP-43 toxicity, aggregation, and protein levels, whereas enhancing endocytosis reverses these phenotypes. Locomotor dysfunction in a TDP-43 ALS fly model is also exacerbated and suppressed by impairment and enhancement of endocytic function, respectively. Thus, endocytosis dysfunction may be an underlying cause of ALS pathology.

Fig. 1

TDP-43 toxicity and turnover in yeast are only modestly affected by autophagy. a Turnover of TDP-43-YFP protein in WT and null strains in mid-log phase. Time indicates period following GAL1-transcriptional shut off. TDP-43-YFP protein level was quantified relative to a Pgk1 loading control. b Vacuolar degradation of TDP-43-YFP in WT cells in mid-log phase, +/−4 h nitrogen starvation (N-S), assessed by accumulation of free YFP fragment. Protein levels quantified as above. c Turnover of TDP-43-YFP in WT cells in mid-log phase, +/−N-S. Time indicates period following CHX addition (0.2 mg/ml). Protein levels quantified as above. d WT strain was transformed with TDP-43-mRuby2 and GFP-Atg8. % value indicates co-localization of TDP-43 foci with Atg8. Scale bar = 2 µm. e Indicated strains transformed with TDP-43-YFP or Pab1-GFP were harvested and vacuolar free YFP/GFP was detected. Protein levels quantified as above. f Turnover of TDP-43-YFP indicated strains (vacuolar protease mutants) examined as in c. g Serial dilution growth assay of WT and indicated isogenic null strains expressing vector or TDP-43-YFP

Fig. 2

Yeast PI3K complex II affects TDP-43 toxicity and turnover. a Serial dilution growth assay of indicated isogenic null strains expressing vector and TDP-43. b, c Indicated strains expressing TDP-43-YFP were examined/quantified for overall TDP-43-YFP levels as in Fig. 1e (b) or foci intensity (c). Significance was assessed by one-way ANOVA. Repeated data are shown as mean ± s.e.m. d TDP-43-YFP turnover in indicated strains, as in Fig. 1a. e WT strain co-transformed with TDP-43-mRuby2 and Vps34-GFP, Vps38-GFP, or Atg14-GFP. % value indicates co-localization of TDP-43 foci with Vps34, Vps38, or Atg14. Scale bar = 2 µm

Fig. 3

TDP-43-induced toxicity and endocytosis defects are rescued by Vps38 overexpression in yeast. a WT cells transformed with vector or TDP-43-YFP, and vps9Δ cells were stained with FM4-64 dye (8 µM) for indicated time and examined. Vacuolar membrane staining intensity was analyzed. Significance was assessed by one-way ANOVA. b WT co-transformed with TDP-43-YFP and either multi-copy empty vector, Vps38, or Atg14-expressing plasmids. c, d Same strains were quantified for TDP-43-YFP foci intensity (Supplementary Fig. 3) (c) and protein levels (d), which were assessed relative to Pgk1 loading control. Significance in c assessed via one-way ANOVA. n.s. = no significance. Repeated data are shown as mean ± s.e.m. e, f WT cells transformed with indicated plasmids were examined as in a. Significance was assessed by two tailed Student’s t test. Data are shown as mean ± s.e.m

Fig. 4

Endocytosis affects TDP-43 toxicity and turnover in yeast. a Schematic of endocytosis and trans-Golgi vacuolar trafficking, highlighting all genes/complexes tested in yeast; green text (Sec7, Scd5) indicates GFP co-localization experiments only conducted. b WT and vps9Δ (Rab5 GEF) strains expressing vector or TDP-43-YFP. c WT, vps18Δ, and vps16Δ strains expressing vector or TDP-43-YFP. d, e Strains expressing indicated GFP-tagged proteins were transformed with TDP-43-mRuby2. % value indicates co-localization of TDP-43-mRuby2 foci with GFP-tagged proteins. Scale bar = 2 µm. f Turnover of TDP-43-YFP in WT and indicated null strains; conducted as in Fig. 1c

Fig. 5

TDP-43 foci co-localize with human cell endosomes. a HEK293A cells transfected with vector, TDP-43-, TDP-43 Q331K-, and TDP-35-GFP plasmids were stained with transferrin-633 and endocytosis rate was determined by cellular incorporation (Supplementary Fig. 6a). Significance was assessed by one-way ANOVA. Repeated data are shown as mean ± s.e.m. b HEK293A cells were transfected with Rab5-mRFP and either TDP-43-, TDP-43 Q331K-, TDP-35-, or TDP-25-GFP plasmids and examined via immunofluorescence. Scale bar = 5 µm. c Immunofluorescence was performed in HEK293A cells transfected with TDP-43-GFP and TDP-35-GFP. The percentage of TDP-43 foci that co-localized with endogenous Rab5 or LC3 foci was calculated. Arrowhead indicates TDP-43 co-localization with Rab5 and LC3. Scale bar = 5 µm. d Comparison of TDP-43 foci size +/− Rab5 overexpression. Significance was assessed by two tailed Student’s t test. e Organelle fractionation was performed in TDP-43-transfected HEK293A cells. Detection of endogenous Rab5, GM130, and Calnexin indicate EE, Golgi, and ER components. The asterisk indicates full-length TDP-43-GFP

Fig. 6

Endocytosis and autophagy affect TDP-43 degradation. a, b HEK293A cells were transfected with TDP-43 WT-, TDP-43 Q331K-, TDP-35-, or TDP-25-GFP and treated with 40 µM dynasore for 24 h. Immunofluorescence was performed; TDP-43-GFP shown in a. Arrowhead indicates TDP-43 cytoplasmic foci. Scale bar = 5 µm. Percentage of cells with cytoplasmic foci quantified in b. Significance was assessed by two tailed Student’s t test. Repeated data are shown as mean ± s.e.m. c Stable HEK293A cell line including vector, WT, Q331K, and TDP-35 were constructed, treated with or without 40 µM dynasore or Rab5 overexpression and cell viability was compared after 48 h. Significance was assessed by one-way ANOVA; n.s. = no significance. Data are shown as mean ± s.e.m. d HEK293A cells were transfected with TDP-43 WT and Q331K-GFP plasmids, treated with or without 40 µM dynasore for 48 h, then harvested and subject to western blotting and quantification of TDP-43 levels relative to a Vinculin loading control. e Endogenous TDP-43 expression level was detected and quantified in control (CTL) and Rab5 knock down cells as above. f Rab5 or LC3 was co-transfected with TDP-43 WT or Q331K plasmids. Cells were harvested after 48 h culture and subject to western blotting and quantified as above. Note, higher-molecular weight bands in mRFP-Rab5 row, TDP-35 set (1st and 3rd column) are FLAG-TDP-35. g HEK293A cells were treated with ATG5 siRNA knockdown (efficiency indicated in second row) and/or 40 µM dynasore for 24 h. Endogenous TDP-43 and LC3 protein levels were examined and quantified as above. h HEK293A cells were treated with 0.3 µM thapsigargin for 24 h or combined with Rab5 transfection. Endogenous TDP-43 and LC3 protein levels were examined and quantified as above

Fig. 7

TDP-43-induced locomotor dysfunction and neurodegeneration is modulated by key endocytic proteins. a–c TDP-43 was specifically expressed in motor neurons +/− distinct Rab mutant backgrounds, and larval turning times (locomotor dysfunction) assessed. a Rab5 loss of function (LOF), b Rab7 dominant negative (DN), c two independent Rab5 overexpression (OE) lines. Significance was assessed by two tailed Student’s t test. Data are shown as mean ± s.e.m. d Fly retina neurodegeneration assessed by depigmentation; TDP-43 was expressed using GMR GAL4. e Fluorescence immunohistochemistry of control and ALS frontal cortex tissue, staining for endogenous TDP-43, Rab5, and DNA (DAPI). Arrowhead indicates TDP-43 co-localization with Rab5 foci. Scale bar = 5 µm. f Quantification of TDP-43 cytoplasmic foci exhibiting Rab5 co-localization in e five ALS patients and three control patients were examined. Significance was assessed by two tailed Student’s t test. Repeated data are shown as mean ± s.e.m