Mitochondrial defects in the respiratory complex I contribute to impaired translational initiation via ROS and energy homeostasis in SMA motor neurons

Spinal muscular atrophy (SMA) is a neuromuscular disease characterized by loss of lower motor neurons, which leads to proximal muscle weakness and atrophy. SMA is caused by reduced survival motor neuron (SMN) protein levels due to biallelic deletions or mutations in the SMN1 gene. When SMN levels fall under a certain threshold, a plethora of cellular pathways are disturbed, including RNA processing, protein synthesis, metabolic defects, and mitochondrial function. Dysfunctional mitochondria can harm cells by decreased ATP production and increased oxidative stress due to elevated cellular levels of reactive oxygen species (ROS). Since neurons mainly produce energy via mitochondrial oxidative phosphorylation, restoring metabolic/oxidative homeostasis might rescue SMA pathology. Here, we report, based on proteome analysis, that SMA motor neurons show disturbed energy homeostasis due to dysfunction of mitochondrial complex I. This results in a lower basal ATP concentration and higher ROS production that causes an increase of protein carbonylation and impaired protein synthesis in SMA motor neurons. Counteracting these cellular impairments with pyruvate reduces elevated ROS levels, increases ATP and SMN protein levels in SMA motor neurons. Furthermore, we found that pyruvate-mediated SMN protein synthesis is mTOR-dependent. Most importantly, we showed that ROS regulates protein synthesis at the translational initiation step, which is impaired in SMA. As many neuropathies share pathological phenotypes such as dysfunctional mitochondria, excessive ROS, and impaired protein synthesis, our findings suggest new molecular interactions among these pathways. Additionally, counteracting these impairments by reducing ROS and increasing ATP might be beneficial for motor neuron survival in SMA patients.

Fig.S1 Complex I de ciency leads to dysfunctional and fragmented mitochondria.
a Volcano plot of proteins identi ed in whole proteome analysis and reported in the MitoCarta2.0 database; plotted p-values (-log 10 ) against fold changes (log 2 , SMA/ WT). Four independent samples of WT MNs and three independent samples of SMA MNs were used for analysis. P-values were determined using unpaired two-sided t-test. Proteins with p < 0.05 are highlighted in blue (32 down-regulated) or red (29 up-regulated).
b Representative images and quanti cation of mitochondrial size and circularity of 10DIV WT and SMA MNs labelled with anti-TOM20 antibody or MitoTracker ® . Enlarged areas are highlighted with yellow boxes in the corresponding image of the whole neuron. Scale bar in enlarged images: 10µm; Scale bar in whole neuron images: 20µm. Each dot represents the quanti cation of individual neurons (n=30). To compare WT (blue circles) and SMA (red squares), two-way ANOVA with Tukey HSD post hoc analysis was used on independent biological replicates (N=3) to determine statistical signi cance. Bar graphs depict the mean ± s.d. *p <0.05, **p <0.01. b Optimization of Complex I activity using different amounts of mitochondria extract isolated from 10DIV WT MNs. Linear regression shows that complex I activity and reaction time is proportional over 1 h.
c Complex I activity using 20 µg mitochondria extract isolated from 10DIV WT and SMA MNs. Linear regression shows that complex I activitiy from WT and SMA MNs is proportional over 1 h, with a reduced slope for SMA MNs.
d Complex I activity rate using 300 µg protein extracts of heart from P7 WT and SMA mice (N=4). Quantification represents the increase of mean OD450 nm/ h. Two-way ANOVA with Tukey HSD post hoc analysis was used on independent biological replicates (N=4) to determine statistical significance. n.s. p >0.05.
e Glucose uptake in 10DIV WT and SMA MNs. Each dot represents data from biological replicates (N=7). Two-tailed unpaired t-test was used to determine statistical significance. **p <0.01.
f Pyruvate uptake is increased in WT MNs and SMA after 1 h supplementation (N=3). Oneway ANOVA with Dunnett post hoc analysis was used to compare each timepoint with the control. **p <0.01, ***p<0.001.
g Supplementation of WT MNs and SMA MNs with 10 mM/ 50 mM lactate or 10mM/ 50 mM pyruvate for 1 h. 50 mM pyruvate treatment shows a significant increase of ATP levels in SMA MNs (N=6). Two-tailed unpaired t-test was used on independent biological replicates to determine statistical significance. Means of two groups were compared, *p <0.05, **p <0.01.
Bar graphs scatterplots depict the mean ± s.d.  b Representative western blot images and quanti cation of SUnSET assay in NSC-34 cells. Neither 50 mM pyruvate nor 10 µM NAC increased protein synthesis signi cantly. Each dot represents the quanti cation of individual biological replicates (N=5). Bar graphs depict the mean ± s.d. Two-tailed unpaired t-test with Holm-Bonferroni correction for multiple comparisons was used to determine statistical signi cance. n.s. p >0.05.  a Representative images of WT and SMA MNs after SUnSET assay. Anti-puromycin (rainbow color) and anti-Tau (white) antibodies are used. Tau positive neurites (axons) were selected with a segmented line, straightened and divided into 20 µm bins using the concentric circles plugin. MNs were treated with 10 µM NAC or 50 mM pyruvate or 100 µM menadione or 50 µM anisomycin for 1 h. Images con rm that protein synthesis is blocked by anisomycin or menadione and increased in SMA MNs by 10 µM NAC. Scale bar: 20 µm.
b Quanti cation of mean puromycin intensity pro les, corresponding to protein levels, against distance in discrete categories. Each dot represents the average quanti cation of 10 neurons. Data are obtained from 5 individual biological replicates (N=5). Bar graph depict the mean ± s.d. Two-tailed unpaired t-test with Holm-Bonferroni correction for multiple comparisons was used to determine statistical signi cance. n.s. p >0.05, *p <0.05.  a Graph representing 22 proteins changed in SMA MNs compared to WT MNs, and pyruvate treated WT MNs compared to non-treated WT MNs; plotted fold change (log2) comparing with and without pyruvate treatment.
b Volcano plot of whole proteome analysis after 50 mM pyruvate treatment to WT MNs for 1 h; plotted p-value (-log10) against fold change (log2). Four independent samples were used for analysis. P-values were determined using an unpaired two-sided t-test. Proteins with p < 0.05 are highlighted in blue (down-regulated) and red (up-regulated).
c Venn diagram showing quantity of proteins in WT MNs compared to SMA MNs treated with 50 mM pyruvate for 1 h.
d Graph representing 40 proteins changed in SMA MNs compared to WT MNs, and NAC treated WT MNs compared to non-treated WT MNs; plotted fold change (log2) comparing with and without NAC treatment.
e Volcano plot of whole proteome analysis after 10 µM NAC treatment to WT MNs for 1 h; plotted p-value (-log10) against fold change (log2). Four independent samples were used for analysis. P-values were determined using an unpaired two-sided t-test. Proteins with p < 0.05 are highlighted in blue (down-regulated) and red (up-regulated).
f Venn diagram showing quantity of proteins in WT MNs compared to SMA MNs treated with 10 µM NAC for 1 h.
g Volcano plot of whole proteome analysis after 100 µM menadione to WT MNs for 1 h; plotted p-value (-log10) against fold change (log2). Four independent samples were used for analysis. P-values were determined using an unpaired two-sided t-test. Proteins with p < 0.05 are highlighted in blue (down-regulated) and red (up-regulated).  a Volcano plot of translation related proteins comparing WT and SMA MNs; plotted p-value (-log 10 ) against fold change (log 2 , SMA/ WT). Four independent samples of WT MNs and three independent samples for SMA MNs were used for analysis. P-values were determined using an unpaired twosided t-test. Proteins with p < 0.05 are highlighted in blue (down-regulated) and red (up-regulated).
b Elongation speed is not altered in SMA MN compared to WT MNs. Each dot and each line represent the average of four independent biological replicates (N=4). Regression analysis comparing the least square means does not show any signi cant difference between WT MNs and SMA MNs.     a Representative western blot and quanti cation shows that 50 mM pyruvate increases SMN protein levels in NSC-34 cells after 1 h treatment (N=5). One-way ANOVA with Dunnett post hoc analysis was used to compare each timepoint with the control. n.s. p>0.05, *p <0.05, **p <0.01.
b Quantitative real-time PCR using gene speci c Smn primers con rms that pyruvate supplementation did not change Smn transcript levels in NSC-34 cells and WT MNs (N=4). Actb was used as a loading control.