Fig 2: Kapß2 and H8-E disperse heterotypic FUS:RNA condensates.A–E, to assess the ability of Kapß2 and truncation constructs to disperse FUS:RNA droplets, we mixed 1 µM Cy3-labeled U40 RNA with 1 µM FUS, which form liquid-like droplets after 4 h. Taking images using fluorescence microscopy (A) or brightfield microscopy (B) 2 h after the addition of buffer or the indicated Kapß2 construct demonstrates that FUS:RNA droplets are unaffected by the addition of buffer (scale bar represents 20 µm). By contrast, FL Kapß2 (0.75 µM) effectively disperses FUS:RNA droplets, reducing them in number (C), overall area of the coverslip occupied by condensates (D), and the average size of each droplet (E). H1-H13 can also reduce the number and total area of FUS:RNA droplets, but is not as effective as FL Kapß2, and does not decrease individual droplet size. However, an equal concentration of H8-E can reduce the number and area of FUS:RNA droplets to a similar extent as FL Kapß2. Data points in (C–E) represent individual images, and bars represent means ± SD, n = 5. An ordinary one-way ANOVA with Dunnett’s multiple comparisons test was used to compare the value for each condition to that of buffer alone. ****p < 0.0001. Additionally, an unpaired t test was used to compare FL and H8-E. **p < 0.004. FL, full-length; Kapß2, Karyopherin-ß2.

Fig 3: Kapß2 and H8-E robustly suppress FUS fibrillization.A, 5 µg of purified His-SUMO-FL Kapß2, cleaved FL Kapß2, H1-H13, H8-E, His-SUMO-H8-H16, and H9-E were processed for SDS-PAGE and stained with Coomassie. B, to measure the ability of His-SUMO-tagged Kapß2 to prevent FUS fibrillization, we incubated 3 µM FUS with 1 µg TEV protease and the indicated concentration of His-SUMO-tagged Kapß2 and monitored turbidity at 395 nm. Increased turbidity indicates FUS fibrillization and each curve represents the mean of 2 trials ± SD. For each experiment, the maximal value of FUS turbidity in the absence of Kapß2 was set to 100 and all data for that trial were normalized to that value. C, to compare uncleaved and cleaved FL Kapß2 chaperone activity, we calculated the area under the curve (AUC) for each concentration and set the AUC for FUS alone as 100 to normalize all other conditions to produce percent aggregation. Overlaying the plot of concentration versus aggregation (%) for cleaved and uncleaved FL Kapß2 shows that the uncleaved protein works as well as cleaved Kapß2 in preventing the fibrillization of FUS, suggesting that the His-SUMO tag itself neither hinders nor enhances Kapß2 chaperone activity in this assay, although it is likely stabilizing H8-H16. D, we assessed the ability of Kapß2W460A:W730A (WWAA) to chaperone FUS. We incubated 3 µM FUS with 1 µg TEV protease and the indicated concentration of Kapß2W460A:W730A and monitored turbidity at 395 nm. Each curve represents the mean of 2 to 3 trials ± SD. The maximal value of FUS turbidity in the absence of Kapß2 was set to 100 and all data for that trial were normalized to that value. E, in Figure 4, we assessed the ability of FL and truncated Kapß2 constructs to chaperone FUS. We incubated 3 µM FUS with 1 µg TEV protease and the indicated concentration of each Kapß2 construct and monitored turbidity at 395 nm for 90 min. Here, only the final normalized absorbance reading (where the maximum value for FUS alone is set as 100) is plotted for each Kapß2 construct at each of the indicated concentrations. Data points represent individual trials and bars represent means ± SD. An ordinary one-way ANOVA with Dunnett’s multiple comparisons test was used to compare the final time point of each condition to FUS alone for that construct. *p < 0.03; **p < 0.008; ***p < 0.0003; ****p < 0.0001.

Fig 4: Phospho-mimetic substitution at Ser505 variably enhances the effects of ALS-linked mutations. (A) Intracellular localization of TLS and its mutants with or without S505D substitution in N2a (top) or COS-7 (bottom) cells. Bars represent average NCI values (mean ± SD, n = 3). For each mutant, open and red bars indicate non-substituted S505 and S505D, respectively. *P < 0.05 in comparison with S505 in two-tailed t-test. (B) Intracellular localization of S505D/H509Q (left panels) and S505D/H509P (right panels) in N2a cells. Scale bar indicates 10 µm. (C) S505D substitution modulates splicing regulation of E1A/mut12s in N2a cells. Bars represent the ratio of the 9s spliced product as in Figure 5 (mean ± SD, n = 3). TLS, S505D and S505D/H509Q showed significantly higher ratio of 9s compared to EGFP (P < 0.05). On the other hand, S505D and S505D/H509Q was not significantly different from TLS. In addition, S505D/H509P and S505D/R510K showed significantly decreased activity compared to TLS, with a greater extent for S505D/H509P (**P < 0.01 and *P < 0.05, ANOVA and Tukey's test).

Fig 5: FUS shows a postsynaptic accumulation in immature (smaller) synapses of rat MNWTFUS at DIV 14. (A) The overview SMLM images shows several small synapses (marked with white arrowheads) labeled for Bassoon(N), FUS, and Homer1. (B–D) Exemplary images of individual synapses and histograms for the analyzed distances between the COM of the two protein populations are shown for (B) Bassoon(N) and Homer1, (C) Bassoon(N) and FUS, and (D) Homer1 and FUS. The lower 60% of the distances are shown in gray and the upper 40% in black. Scale bar is 500 nm for the overview and 100 nm for single synapses.