Expression analysis of bait proteins for interactome analysis. A Schematic representation of GFP-tagged C21ORF2, KIF5A, NEK1, TBK1, TUBA4A and FUS. NTD—N-terminal domain; CTD—C-terminal domain; CCD—coil-coiled domain; KD—kinase domain; BD—basic domain; ULD—ubiquitin-like domain; SDD—scaffold and dimerization domain; QGSY—Glu-Gly-Ser-Tyr-rich region; G rich—Gly-rich region; E—nuclear export signal; RRM—RNA recognition motif; RGG—Arg-Gly-rich region; Z—zinc finger motif; L—nuclear localization signal. B–H Left panel: Expression of GFP and GFP-tagged proteins in Neuro2A (N2a) cells assessed by immunohistochemistry (DAPI in blue). Exogenous proteins show an endogenous distribution pattern. C–H Right panel: Western blots showing GFP-tagged (black arrowhead) and endogenous (white arrowhead) proteins. Anti-β-actin is used as loading control. Scale bar B-H: 5 µm

Analysis of the binding partners of six ALS-associated proteins. A–F Scatter plots showing the interactors of the indicated bait proteins. Each graph shows all proteins identified in the GFP-tagged protein sample (y-axis) versus proteins in a corresponding GFP control sample (x-axis). Each protein is plotted according to its intensity value (abundance) in the sample. The blue line represents the intensity value that is the same for both GFP-tagged protein and GFP control samples. The red line indicates the cutoff used (4 times enrichment in the GFP-tagged sample) to detect probable interactors. These interactors are also colored according to their predicted function (GO analysis using PANTHER software). For each graph the position of the GFP-tagged (bait) protein is indicated (arrow). G Heatmap summarizing the shared interactors of 11 ALS-associated bait proteins analyzed in the current study and our previous work (Blokhuis et al., 2016)

ALS-associated C21ORF2 variants cause neuronal apoptosis. A Coronal sections of adult mouse motor cortex and spinal cord were co-immunostained for C21ORF2 and CTIP2 (upper panel) or CHAT (lower panel) to mark upper and lower MNs, respectively. Squares in left panels are shown at higher magnification in the right three panels. B Co-immunostaining for C21ORF2 and tubulin of mouse primary cortical neurons at 5 days in vitro (DIV). C Immunostaining for C21ORF2 and cleaved-Caspase-3 (cCAS3) of DIV5 mouse primary cortical neurons exogenously expressing C21ORF2-V58L. D Quantification of the percentage of cCas3-positive primary cortical neurons in cultures expressing GFP, GFP-tagged wild type C21ORF2 or GFP-tagged C21ORF2 variants. C21ORF2 variants are related to ALS (V58L, R211*, A222, d263), ciliopathies (L224P, C61Y) or both (R73P, V111M). Results from three independent experiments are plotted as means with SEM. **** P < 0.0001 Kruskal–Wallis test, Dunn’s multiple comparison post-hoc test. Scale bars: A, 200 µm (low magnification); 20 µm (high magnification); B, 10 µm; C, 5 µm

C21ORF2-V58L causes apoptosis, reduced mitochondrial activity and impaired DNA damage responses. A–L Expression and functional analysis of iPSC-derived motor neuron (MN) cultures at days in vitro 25 (DIV25, 7 days after plating). C1 and C2, healthy controls; P1, ALS patient with C21ORF2-V58L mutation; P2, ALS patient with C21ORF2-V58L/C61Y mutations. A–E C21ORF2 protein and mRNA expression is unchanged in mutant MNs A, B Co-immunostaining for CHAT and C21ORF2 in control and mutant. B Quantification of immunostainings as in A. C RT-qPCR analysis of C21ORF2 gene expression in control and mutant MNs. For relative mRNA expression the ΔCt values (ΔCt = Ct(C21ORF2) – Ct (ACTINB)) for each line are shown. D, E Western blot analysis using anti-C21ORF2 and anti-β-ACTIN antibodies. β-ACTIN was used as loading control. D Representative Western blot. E Quantification of C21ORF2/actin intensity ratio in Western blot experiments as in D. F, G The number of cCas3-positive MNs is increased in mutant cultures. F Co-immunostaining for cleaved-caspase-3 (cCas3) and CHAT in control and mutant cultures. G Quantification experiments as in F. H Representative kinetics graphs of mitochondrial oxygen consumption in cultures. I, J Quantification of maximum respiration and spare respiratory capacity, both of which show a significant reduction in mutant cultures. K, L Co-immunostaining for CHAT and yH2AX, a marker of the DNA damage response, after 1 h of neocarzinostatin (NCS) treatment. L Quantification of yH2AX staining intensity in MN nuclei after NCS treatment, showing a significant decrease in mutant MNs. B–E, G, I, J, L Results of at least three independent experiments are plotted as means with SEM. *p < 0.05; ** p < 0.01; ****p < 0.0001, G, I, J, L Kruskal-Wallis test, Dunn’s multiple comparison post-hoc test. Scale bars: A, 10 µm; F, 100 µm; K, 5 µm

C21ORF2-V58L mutant phenotypes are rescued in isogenic control cultures. A–G Functional analysis of iPSC-derived motor neuron (MN) cultures at days in vitro 25 (DIV25, 7 days after plating). C1, healthy control; P1, ALS patient with C21ORF2-V58L mutation; iso, isogenic control 1 generated from P1. A Co-immunostaining for cleaved-Caspase-3 (cCAS3) and CHAT in control, mutant and isogenic cultures. B Quantification experiments as in A. Data from 3 independent experiments per line were quantified. C Representative kinetics graphs of mitochondrial oxygen consumption rate. Quantification of maximum respiration (D) and spare respiratory capacity (E). Raw data from 3 independent experiments and at least 3 replicates per line were quantified. F, G Co-immunostaining for CHAT, DAPI and yH2AX, a marker of the DNA damage response, after 1 h of neocarzinostatin (NCS) treatment. G Quantification of yH2AX staining intensity in MN nuclei after NCS treatment. Data from 3 independent experiments per line were quantified. Results of three independent experiments are plotted as means with SEM *p < 0.05; ****p < 0.0001. B, D, E, G Kruskal–Wallis test, Dunn’s multiple comparison post-hoc test. Scale bars: A, 100 µm; F, 5 µm

Higher spontaneous activity but lower capacity for further stimulation-driven excitability in C21ORF2-V58L motor neurons. A Representative voltage traces in response to 20 pA or 80 pA depolarizing current injections across the three conditions: Healthy control 1 (C1, black), C21ORF2-V58L patient 1 (P1, red), and isogenic control 1 of P1 (iso1, blue). B Average current-action potential relationships in the three conditions: C1 (black; ncells = 48), P1 cells (red; ncells = 46), iso1 (blue; ncells = 31). Two-Way Repeated Measures ANOVA. Current × Genotype interaction. F(38,2318) = 2.082, p = 0.0001. C Left: Bar graph with mean and SEM for action potential number in response to 20pA across the three conditions. One-Way ANOVA: F(2122) = 5.148, p = 0.0071. Tukey post-hoc tests. C1 versus P1, p = 0.0192. C1 versus Iso1, p = 0.9517; P1 versus Iso1, p = 0.0197. Right: Same but in response to 80 pA. One-Way ANOVA: F(2,122) = 4.926, p = 0.0088. Tukey post-hoc tests. C1 versus P1, p = 0.0068. C1 versus Iso1, p = 0.1712; P1 versus Iso1, p = 0.6045. D Bar graph plotted as means with SEM for rheobase across the three conditions. One-Way ANOVA: F(2,122) = 2.317, p = 0.1029. E Bar graphs with the percentage of spontaneously active cells at 0 pA injection steps across the three conditions. C1: 19/48 cells = 40%; P1: 32/46 cells = 70%; Iso1: 18/31 cells = 58%. Kruskall Wallis test = 8.606, p = 0.0135. Dunn’s post-hoc tests. C1 versus P1, p = 0.0108; C1 versus Iso1, p = 0.3245; P1 versus Iso1, p = 0.9647. *p < 0.05; **p < 0.01

C21ORF2-V58L impairs motor behavior in zebrafish embryos. A Representative images of 5 days post-fertilization (5 dpf) zebrafish embryos non-injected and injected with C21ORF2-WT or C21ORF2-V58L constructs. B Quantification of the number of deformed 5 dpf zebrafish embryos as in A. Pooled percentages from five independent experiments (total numbers of embryos analyzed—281 non injected embryos, 77 GFP injected embryos, 113 C21ORF2-WT injected embryos, 82 C21ORF2-V58L embryos). Data are shown as means with SEM *p < 0.05, **p < 0.01 Kruskal–Wallis test, Dunn’s multiple comparison post-hoc. C, D Zebrabox analysis of zebrafish embryo motor behavior at 5 dpf. Time spent swimming (seconds) and trajectory covered (millimeters) during 1 h were quantified. Pooled data from 4 independent experiments are plotted as means with SEM *p < 0.05, **p < 0.01, ****p < 0.0001 Kruskal–Wallis test with Dunn’s multiple comparison post-hoc test. Scale bar: A, 1 mm

C21ORF2-V58L affects NEK1 expression at the post-transcriptional level. A–L Analysis of NEK1 distribution and expression in iPSC-derived motor neuron (MN) cultures at days in vitro 25 (DIV25, 7 days after plating). C1 and C2, healthy controls; P1, ALS patient with C21ORF2-V58L mutation; P2, ALS patient with C21ORF2-V58L/R60W mutations; iso, isogenic control 1 generated from P1. A Co-immunostaining for NEK1 and CHAT/ISL1 to label MNs. B, C Quantification of NEK1 expression in soma and nucleus compartments based on immunostaining as in A. D Ratio of NEK1 immunostaining in nucleus compared to soma. E RT/q-PCR analysis of NEK1 gene expression. For relative mRNA expression the ΔCt values (ΔCt = Ct(NEK1) – Ct (ACTINB)) for each line are graphed. F, G Western blot analysis using anti-NEK1 and anti- β -actin antibodies. β-actin was used as loading control. F Representative Western blot. G Quantification of NEK1/actin intensity ratio in Western blot experiments as in F. H–L Immunostaining (H, I), RT-q-PCR (J), Western blot analysis (K, L) and corresponding quantifications using MNs derived from the C1, P1 and iso1 lines and performed as described for panels A–G. Results of three independent experiments are plotted as means with SEM **p < 0.01; ***p < 0.001; ****p < 0.0001, (G, L) One-Way ANOVA, Tukey multiple comparison post-hoc test. (B, C, I) Kruskal–Wallis test with Dunn’s multiple comparison post-hoc test. Scale bars: A, 5 µm; I, 20 µm

Proteasomal inhibition reverses NEK1 downregulation in C21ORF2-V58L motor neurons. A Immunocytochemistry for NEK1 in MNs from two control (C1, C2) lines and two ALS patient lines carrying C21ORF2 mutations (P1, P2) treated with the proteasome inhibitor MG132 (100 nM, 3 h) or non-treated at days in vitro 7 (DIV7). B Quantification of immunostaining of NEK1 in control and mutant neurons without or with treatment with MG132 as in A. C-D Western blot analysis of lysates of control and mutant MNs without or with MG132 treatment. Western blot analysis using anti-ubiquitin, anti-NEK1 and anti-β-actin antibodies. Beta-actin was used as loading control. C Representative Western blot. D Quantification of Western blot experiments as in C. Results of 3 (A, B) and 5 (C, D) independent experiments are plotted as means with SEM *p < 0.05; **p < 0.01; ****p < 0.0001. (B) Kruskal–Wallis test with Dunn’s multiple comparison post-hoc test. D One-Way ANOVA, Tukey multiple comparison post-hoc test. Scale bar: A, 10 µm