Tryptophan residues in wtTDP-43 and TDP-43^ΔNLSmodulate the aggregation of mutant SOD1^G85R-GFP in cultured cells.A, schematic representation of the TDP-43 structural domains and location of its six Trp residues. B, TDP-43 variants (red) cross-seed aggregation of SOD^G85R-GFP reporter protein (green), viewed 48 h after co-transfection into HEK293FT cells. TDP-43^ΔNLS with all Trp residues mutated to serine (Trpless-TDP-43^ΔNLS) is abundantly expressed but does not induce SOD1^G85R-GFP aggregation. Scale bar represents 20 μm. C, time-course analysis of induced SOD1^G85R-GFP aggregate abundance in HEK293FT cells from 24 to 72 h post co-transfection with TDP-43^ΔNLS, wherein combinations of Trp residues are mutated to serine (Ser). Trp68 and Trp172 in TDP-43 modulate the induction of SOD1^G85R-GFP aggregation. E, as per transfections of panel C, except using WT TDP-43 plasmid variants. D and F, percent of cells with detectable SOD1^G85R-GFP aggregates quantified by flow cytometry shows more cross seeding of SOD1^G85R-GFP by TDP-43^ΔNLS (D) or WT TDP-43 (F) constructs when compared to their variants bearing Trp to Ser mutation(s). Statistical significance was determined using one-way ANOVA and Dunnett’s test for multiple comparisons (∗∗p < 0.01; ∗∗∗∗p < 0.0001). Error bars represent SEM; between, 8 and 34 biological replicates were performed for each construct. LCD, low-complexity domain; NES, nuclear export signal; NLS, nuclear localization signal; NTD, N-terminal domain; RRM, RNA recognition motif; RRM, RNA recognition motif; TDP, TAR DNA-binding protein.

Tryptophan residues in human SOD1 and TDP-43 proteins mediate their synergistic impact on motor neuron pathology in vivo.A, mnx1:GFP zebrafish (36 hpf, hours post-fertilization) expressing GFP in the motor neurons. Abnormal primary axons (arrowheads) with branches above the ventral notochord (dashed line). Muscle actin counterstained in magenta. Scale bars represent 0.5 mm (top left and top right); 200 μm (bottom four panels). B, human WT SOD1 and WT TDP-43 synergistically induce primary motor neuron axonopathy. Ser68 and Ser172 TDP-43 mutations reduced SOD1-TDP-43–induced axonopathy (p = 3.57 × 10⁻⁵). The SOD1 Trp32Ser mutation also abrogated axonopathy (p = 4.54 × 10⁻⁵) and in combination with Trpless-TDP-43 (p = 8.94 × 10⁻⁵). Error bars indicate SEM. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 (Kruskall–Wallis test with Mann–Whitney pairwise comparisons): sample sizes are noted at the base of each bar. Mnx1, motor neuron and pancreas homeobox protein 1; SOD1, superoxide dismutase; TDP, TAR DNA-binding protein.

Trp68 is exposed in cytoplasmic TDP-43 aggregates in cultured cells and in sporadic ALS motor neurons.A, representative centroid structure of the RRM1 domain of TDP-43, obtained from equilibrium molecular dynamics (MD) simulation. Trps 113 and 172 are significantly exposed compared to most native structures in the protein databank (11). B, representative centroid structure of the N-terminal domain of TDP-43, obtained from equilibrium molecular dynamics simulation. Trp68 is nearly fully buried: only about 9% of its tripeptide Gly-Trp-Gly value of 247 Angstrom² is exposed in the centroid structure. C, the affinity-purified rabbit polyclonal anti-Trp68 antibody (red) was tested for reactivity and specificity in cells transfected with different TDP-43 constructs, then fixed and stained 48 h later. A mouse pan-TDP-43 antibody against the C-terminal domain and a chicken anti-HA-tag antibody were used to test co-localization with TDP-43 (yellow) and HA-TDP-43 (green), respectively. The anti-Trp68 antibody specifically recognizes mislocalized cytoplasmic TDP-43 aggregates in TDP-43^ΔNLS-transfected cells but not those lacking Trp68 (Ser68 TDP-43^ΔNLS) nor does it recognize nuclear TDP-43. The composite images are a merge between nuclear, anti-Trp68, and anti-HA-tag staining. Scale bar represents 20 μm. D, representative images of TDP-43 pathology in ALS cervical spinal cord sections from two subjects immunostained with anti-Trp68 in paraffin-embedded tissue. The left panel shows the anti-Trp68 antibody and right the commercial pTDP-43 antibody. Motor neurons show thread-like inclusions using the anti-Trp68 antibody and the commercial pTDP-43 antibody, but no immunoreactivity is observed of natively folded TDP-43 in neuronal or glial nuclei Scale bar represents 50 μm. ALS, amyotrophic lateral sclerosis; RRM, RNA recognition motif; TDP, TAR DNA-binding protein.

Trp68 is exposed in TDP-43 aggregates in gray and white matter in TDP43 subtypes of FTD. TDP-43 pathology in sections from the anterior cingulate cortex of FTD-TDP subtypes A, B, C, and E immunostained with anti-Trp68 (left column), commercial phospho-TDP-43 (central column), and commercial pan-TDP-43 (right column) in paraffin-embedded tissue. FTD-TDP-A shows thread-like inclusions in the white matter using the anti-Trp68 antibody (A) and the commercial pTDP-43 antibody (C). In the gray matter, cytoplasmic inclusions are predominantly stained using the anti-Trp68 antibody (B) and the pTDP-43 antibody (D) that shows abundant staining of cytoplasmic and extracellular aggregates. In FTD-TDP-B, mainly cytoplasmic inclusions are observed in the white and gray matter using the anti-Trp68 antibody (E and F) and the commercial pTDP-43 antibody (G and H). FTD-TDP-C is characterized by long threads in the white and gray matter. These long threads are seen using the anti-Trp68 antibody (I and J) and the pTDP-43 antibody (K and L). FTD-TDP-E is characterized by cytoplasmic and abundant granular TDP-43 aggregation. Using the anti-Trp68 antibody, cytoplasmic and thread-like inclusions are seen in white and gray matter (M and N), but no granular staining is observed. The pTDP-43 antibody shows cytoplasmic and granular staining in the white and gray matter (O and P). The right column (Q–X) shows pan-TDP-43 distribution, which does not well correspond to pathological antibodies. Scale bar represents 50 μm. FTD, frontotemporal domain; TDP, TAR DNA-binding protein.

TDP-43–mediated SOD1 misfolding is attenuated by anti-RRM1 single chain antibodies.A, representative image of endogenous SOD1 misfolding (3H1 antibody, red) in HEK293 cells co-transfected with WT TDP-43 (gray) and anti-RRM1 scFvs (green) (VH1Vk9 or VH7Vk9), control scFv (D1.3, anti-chicken lysozyme), or empty vector for scFv. Scale bar represents 10 μm. B, quantification of misfolded SOD1. Data are expressed as mean ± SEM (five fields of view were quantified per condition; n = 3 experiments per condition, represented as dots in the bar graph). For each field of view, the integrated density of misfolded SOD1 (3H1) was measured and normalized to the total number of cells. Normalized misfolded SOD1 in scFv-transfected cells was then expressed as fold change compared to cells transfected with the empty vector (no scFv). Statistical significance was established using one-way ANOVA followed by Tukey’s multiple comparison test, ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 versus empty vector (∗) or control scFv antibody (#). RRM, RNA recognition motif; SOD1, superoxide dismutase; TDP, TAR DNA-binding protein.

5-FUrd alleviates cross-seeding of SOD1 by TDP-43.A, immunocytochemistry of HEK293FT cells 48 h post co-transfection with TDP-43^ΔNLS (red) and SOD^G85R-GFP reporter protein (green) in the presence of vehicle (top) or 1 μM 5-FUrd (bottom), demonstrating a reduction in SOD1 aggregates with 5-FUrd. Scale bar represents 50 μm. B, time-lapse live-cell microscopy of SOD^G85R-GFP cross-seeding by TDP-43^ΔNLS in the presence of 1 or 5 μM 5-FUrd shows a reduction in aggregate accumulation proportionate to drug concentration. Images were acquired every 30 min for 30 h. C and D, flow cytometry of saponin-treated cells co-transfected with the reporter protein with either TDP-43^ΔNLS (C) or WT TDP-43 (D) showing a reduced number of SOD1^G85R-GFP aggregate containing cells when treated with 5-FUrd. For each experiment, error bars represent SEM of 8 to 13 biological replicates. Statistical significance was determined using one-way ANOVA and Dunnett’s test for multiple comparisons (∗∗p < 0.01; ∗∗∗∗p < 0.0001). SOD1, superoxide dismutase; TDP, TAR DNA-binding protein.

5-FUrd rescues axonopathy induced by co-expressing human SOD1 and WT TDP-43 in zebrafish. When human SOD1 and TDP-43 are co-expressed in zebrafish (A, exemplar cross-sections of spinal cord), the motor neurons exhibit axonopathy (B and C), and this axonopathy can be partially but significantly rescued with drugs that block human SOD1 Trp32. Scale bar represents 200 μm. B, 5-fluorouridine (5-FUrd, 1.5 μM) rescued axonopathy in embryos injected with SOD1 and WT TDP-43 (p = 0.011). C, a significant reduction was also observed with 5-FUrd in SOD1 and TDP-43^ΔNLS-injected embryos (p = 0.040). Both vehicle and 5-FUrd solutions contained 5 μM uridine and 0.2% DMSO. Error bars indicate SEM; ∗p < 0.05 (Mann-Whitney pairwise comparisons); sample sizes are noted at the base of each bar. SOD1, superoxide dismutase; TDP, TAR DNA-binding protein.