In ALS motor neurons selectively degenerate, resulting in loss of muscle innervation and eventual, fatal paralysis. To assess the severity of the ALS phenotype in the Ubqln2^P497S mouse model, 12-month-old animals with significantly denervated NMJs (Supplementary Figure 1) had their tissues collected, serially cryosectioned and immunolabeled. Motor neuron cell bodies in the ventral horns of the spinal cord in the lumbosacral enlargement were immunolabeled using Neuronal Nuclei (NeuN), a general neuron marker, and choline acetyltransferase (ChAT), a marker of cholinergic motor neurons. Slides were imaged on a confocal microscope and motor neurons were subsequently counted and averaged by the number of spinal cord sections evaluated. When compared by genotype, the Ubqln2^P497 cohort had a significant reduction in the average number ChAT+ motor neuron cell bodies compared to wild-type (Ubqln2^+/+) controls (Figure 1A). The reduction of cell body counts in the Ubqln2^P497 animals from approximately 17.2 to 8.2 ChAT+ motor neurons per spinal cord section indicates a greater than 50% decrease in motor neuron survival in these ALS animals compared to wild-type controls.

Given that the previous characterization of this Ubqln2^P497 line reported reduced nuclear TDP-43 staining coupled with strong cytoplasmic staining of granular aggregates in motor neurons at 8-months-of-age (22), we immunostained 12-month Ubqln2^P497 spinal motor neurons to assess TDP-43 expression and its nuclear clearance (Figure 1B). Nearly all motor neurons expressed basal levels of TDP-43 in their nuclei regardless of genotype. However, some motor neurons in Ubqln2^P497 mice did display TDP-43 in their cytoplasm with accompanying granularizations (white arrow), although this phenotype was not widespread like previously reported (5–10%).

The NMJs in EDL muscles of these mice were examined to determine if this loss of cell bodies in the spinal cord coincided with a reduction in innervation of skeletal muscle. Post-synaptic acetylcholine receptors were labeled using α-bungarotoxin (BTX) conjugated with a fluorophore and presynaptic terminals were labeled with antibodies to SYNAPSIN (Figure 1C). The labeling of BTX and SYNAPSIN in the merged panel of Ubqln2^+/+ animals showed co-localization of both channels in all NMJs, displayed in yellow. In the Ubqln2^P497S merged image there were a number of NMJ’s showing only BTX labeling (arrows), indicating a loss of innervation in some NMJs of these animals. To quantify innervation levels in these experiments, regions of interest (ROIs) were drawn around each NMJ throughout the EDL collected from each mouse and analyzed using the Leica LAS X software analysis package (Teaneck, NJ) as described in the methods. Regardless of genotype, the size of each NMJ was similar for both cohorts (Figure 1D). The number of SYNAPSIN pixels was quantified and averaged within their respective cohorts, showing a significant reduction in SYNAPSIN in the NMJs of Ubqln2^P497S mice as compared to Ubqln2^+/+ mice (Figure 1E). To control for synapse size, SYNAPSIN pixel sums were divided by the overall ROI area for each NMJ, and after this normalization there again was a significant decrease in innervation in Ubqln2^P497S NMJs from EDLs (Figure 1F). Taken together, these data indicate that, at 12 months of age, the Ubqln2^P497S mouse model recapitulates the motor neuron loss and reduction in skeletal muscle innervation observed in ALS.

To determine if the loss of motor neuron survival and skeletal muscle denervation observed in Ubqln2^P497S mice translated to the expected motor phenotypes associated with ALS pathology, a battery of strength and locomotor tests were conducted on these animals. All mice were weighed at 12 months of age just prior to testing. Ubqln2^P497S males showed a significant reduction in weight compared to Ubqln2^+/+ mice, however female Ubqln2^P497S mice maintained the same weight as wild-type littermates. When sexes were combined and compared by genotype alone, there was a trend toward a reduction in the weight of the Ubqln2^P497S cohort, although not statistically significant (Figure 2A).

An open field test was utilized to assess distance traveled (Figure 2B), movement time (Figure 2C) and vertical movement time (hind limb rearing, Figure 2D). Interestingly, the Ubqln2^P497S cohort outperformed the wild-type controls for all three measurements when analyzing by genotype. If sex of the animals was also considered, Ubqln2^P497S and Ubqln2^+/+ females had a significantly higher distance traveled compared to males of both genotypes (Figure 2B), and both male and female Ubqln2^P497S mice traveled greater distances compared to their wild-type littermates. For total movement time and vertical movement time, females were not significantly different by genotype, and males were only different by genotype for total movement (Figures 2C,D). When grip strength was assessed, there were no significant differences observed between cohorts for forelimb grip strength (Figure 2E), regardless of sex or genotype. However, all limb grip strength trials indicated that Ubqln2^P497S mice had reduced grip strength compared to Ubqln2^+/+ mice, and this difference was driven by Ubqln2^P497S males that had a significant reduction in grip strength compared to Ubqln2^+/+ males (Figure 2F). Lastly, the rotarod test (Figure 2G) showed no significant differences by sex or genotype. Overall, only the all-limb grip strength test showed a reduction in performance by Ubqln2^P497S animals, indicating that this mouse model does not develop the significant motor phenotypes observed in Sod1^G93A mice. It should be noted that male Ubqln2^P497S mice had reduced weight and all-limb grip strength compared to Ubqln2^+/+ males, and these deficits were not observed in females, which could be due to the Ubqln2 gene being X-linked.

In order to purify and sequence mRNAs that are translated in motor neurons during the progression of ALS, the Ubqln2^P497S mutation was crossed into a mouse line containing both RiboTag (Ribo) (24) and Chat-Cre to drive the expression of HA-tagged RPL22, and thus HA-tagged ribosomes, in a motor neuron-specific manner, as we have done previously (23). Western blotting of the protein precipitates following immunoprecipitation showed that only Chat-Cre + mice expressed RPL22-HA, demonstrating genotype specificity for the ribosomal tagging of spinal motor neurons (Supplementary Figure 2A). Additionally, there was a RiboTag gene dosage-dependent increase in RPL22-HA expression in homozygous mice compared to heterozygous Ribo; Chat-Cre + mice. Extracted RNA was analyzed by bioanalyzer and all samples had a high purity with RIN values of greater than 8, and concentrations greater than 20 ng/μl, indicating the RNA samples were well suited for downstream sequencing applications.

Samples from 12-month old mice were collected and the mRNA isolations were subjected to Illumina-based sequencing and subsequent bioinformatic analysis. Two groups of samples, both of which consisted of ribosome IPs from Ubqln2^P497S and Ubqln2^+/+ mice, were collected on different dates and sequenced separately, which required batch correction prior to bioinformatic analysis (Supplementary Figure 2B). As we have observed previously (23), sequencing of the translatomes showed an enrichment of motor neuron-specific transcripts and a depletion of transcripts associated with glial cell types, demonstrating the specificity of our RiboTag IP protocol (Supplementary Figure 2C). To further evaluate the enrichment of motor neuron genes as compared to the input, differential gene analysis was performed on RiboTag IPs from Ubqln2^+/+ mice as compared to their input, which was total spinal cord RNA. Volcano and smear plots indicate many genes that are differentially expressed with an FDR < 0.05 (Supplementary Figure 3A). Many of these genes were neural and/or motor neuron-specific in the IPs, whereas glial genes were prevalent in the input. Collectively these data confirm the specificity of the ribosome IPs for motor neuron genes, as we reported previously (23). Differential gene analysis was performed on RiboTag IPs from Ubqln2^P497S mice compared to their input as well, again showing enriched motor neuron genes in the IPs (Supplementary Figure 3B).

Differential gene expression analysis determined that there were 467 enriched transcripts and 558 reduced transcripts when comparing Ubqln2^P497S ribosome IPs to Ubqln2^+/+ IPs, at a p-value of <0.05 (Figure 3A; Supplementary Table 1). When a more stringent statistical significance cutoff was used, FDR <0.1, there were 3 enriched and 22 depleted transcripts for Ubqln2^P497S versus Ubqln2^+/+ mice (Figure 3B; Supplementary Table 1). Interestingly, when the most highly enriched and depleted mRNAs were analyzed using a heat map (Supplementary Figure 4) muscle related transcripts were among the most downregulated when comparing Ubqln2^P497S to wild-type animals, including the genes Actin Alpha-1 (Acta1), Desmin (Des), and Troponin T3 (Tnnt3). Acta1 is a skeletal muscle-specific member of the actin family of proteins responsible for cell structure and motility, acting as a major component of muscle contraction (25). Des is another gene that encodes a key microfilament involved in skeletal, cardiac and smooth muscle contraction (26). Tnnt3 encodes one of the three regulatory proteins that form the trimeric troponin complex in fast-twitch skeletal muscle fibers, and this complex initiates muscle contraction when bound with Ca²⁺ (27). We went on to perform GSEA analysis followed by Gene Ontology and found that Wnt receptor activity and Wnt protein binding were two of the most depleted terms when considering molecular function (Figure 3C). When looking at biological processes, respiration and energy metabolism terms were the most enriched, which was also seen in the analysis of molecular function terms with enriched terms relating to ATP synthesis and metabolism (Figures 3C,D). Consistent with the results from Heat Mapping, muscle associated terms like myofibril assembly, striated muscle cell development and actinomycin structure organization were among the most depleted Gene Ontology terms (Figure 3D).

To evaluate what genes are most differentially expressed in total spinal cord mRNA, differential gene analysis was performed on the input mRNA from Ubqln2^P497S versus Ubqln2^+/+ mice (Supplementary Figure 3C). There were 214 genes that were upregulated, and 192 genes that were downregulated, in total spinal cord mRNA from Ubqln2^P497S mice versus Ubqln2^+/+ mice. The number of categories that were identified as changed in disease spinal cords as compared to controls were limited because of the limited number of differentially expressed genes, but 5 of the 6 identified categories were immune-related, such as tumor necrosis factor biological processes, in the GSEA analysis (Supplementary Figure 3D). Interestingly, the most significantly upregulated and downregulated genes (Heat Map analysis, Supplementary Figure 4) were not identified as differential in the IP samples, except for Thy-1, which was the promotor used to drive transgene expression. This observation demonstrates the utility of ribosomal profiling of motor neurons in that genes that are affected in diseased motor neurons are not likely to be detected in total spinal cord RNAseq because of their low abundance compared to all of the other contributing spinal cord cell types.

Following sequencing, a subset of differentially expressed genes were evaluated using fluorescent in situ hybridization (FISH). If the amount of a mRNA associated with ribosomes changes between Ubqln2^P497S and Ubqln2^+/+ mice, this may be due to changes in the expression levels of this transcript. Because differences in mRNA expression levels should be detectible by FISH, we evaluated mRNA expression changes for some of the differentially-enriched genes within the cell bodies of motor neurons. G protein-coupled receptor 50 (Grp50) and Tenascin N (Tnn) expression increased in Ubqln2^P497S animals compared to controls (Figure 4A), supporting the mRNA sequencing results showing their upregulation in ALS mice. Interestingly, Tnn functionally is involved in the downregulation of canonical Wnt signaling pathways, supporting the depleted Wnt-related categories observed in our GSEA and GO analyses (Figure 3D) (28). Because Des, Tnnt3, and Acta1 are thought to be muscle-specific genes not expressed in motor neurons, we evaluated their expression using FISH. In support of the specificity of the ribosomal profiling of motor neurons, expression of Des, Tnnt3 (Figure 4B), and Acta1 (Figure 4C) were clearly expressed in motor neuron cell bodies of WT mice in the ventral horn of the spinal cord. Interestingly, all three of these genes were dramatically reduced in Ubqln2^P497S mice compared to Ubqln2^+/+ mice, confirming the reduction in the levels of muscle transcripts observed in Ubqln2^P497S motor neurons in the ribosomal profiling screen.

To evaluate the expression of “muscle-specific” genes in motor neurons at the protein level, immunofluorescence labeling was performed using antibodies to Desmin. Within the spinal cord, some Ubqln2^+/+ and not Ubqln2^P497S cell bodies expressed Desmin, although this protein expression was not widespread (Figure 5A). Interestingly, Desmin expression was heavily concentrated in the NMJ endplates of EDLs from both genotypes, and overlapped most substantially with Bungarotoxin labeling, indicating it was postsynaptic expression (Figures 5B,C).

UBQLN2 (B6.Cg-Tg Thy1-UBQLN2*P497S, Stock # 029968), RiboTag (B6N.129-Rpl122tm1.1Psam/J, Stock # 011029) and Wild-type C57BL/6J (Stock # 000664) mice were obtained from The Jackson Laboratory (Bar Harbor, ME). All procedures performed were approved by the Institutional Animal Care and Use Committee (IACUC) of Indiana University School of Medicine, and all animal facilities are accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC). Mice were maintained on a 12-h light/dark cycle and provided food and water ad libitum in a specific pathogen-free facility with environmental enrichment. RiboTag mice (24) were crossed with ChAT-Cre and Ubqln2 mice to generate Rpl22^HA/HA; ChAT^Cre/+; Ubqln2^P497S/+ (Ribo^HOM;ChAT^Cre;Ubqln2^HET) and Rpl22^HA/HA; ChAT^Cre/+; Ubqln2^+/+ (Ribo^HOM;ChAT^Cre;Ubqln2^WT) experimental mice. Animals were genotyped using publicly available PCR protocols with the primer sequences provided by the Jackson Laboratory.

Subjects were 26 mice (6 male Ubqln2^+/+, 9 male Ubqln2 ^P497S/+, 5 female Ubqln2^+/+, and 6 female Ubqln2 ^P497S/+). Four male Ubqln2^+/+ and two male Ubqln2 ^P497S/+ were tested in October 2021, and the others were evaluated in December 2022; in total 11 Ubqln2^+/+ mice and 15 Ubqln2 ^P497S/+ mice were evaluated with the three tests described below. Prior to testing in each study, mice were moved from the main colony to a room adjacent to the testing facility and placed under low light conditions with a white noise machine for a 30-min acclimation period.

Mice were tested using an Omnitech Electronics open field apparatus (Columbus, OH). Animals were evaluated up to 6 at a time with one in each testing area. Mice were placed inside the darkened shelter within the arena and were allowed to explore freely once the chamber door closed. Photobeams tracked animal x, y, and z axes to allow the quantification of movement, distance, and position. Between the testing of each animal, all components of the apparatus were cleaned with Clidox (VWR, #MSPP-96118F). Testing was performed with overhead room lights off and an active white noise machine.

Animals were measured using the BioSeb BIO-GS3 Grip Strength Meter (Pinellas Park, FL). Forelimb grip strength was tested by lowering the mice onto the metal testing grid so their forepaws would grasp the mesh near the cross bar. Once the grip was established, the mouse was pulled back in a single, rapid motion by its tail with its body perpendicular to the ground until the grid was released. Each animal was tested twice with the strength scores averaged together. For all-limb measurements, the animals were placed on the metal grid with all four paws below the crossbar with the head of the animal facing the sensor apparatus. The animal was then pulled back in a single, rapid motion by its tail to make sure that all four paws released simultaneously. The force required for release was recorded, with each animal being tested twice and the scores averaged. All grip strength testing occurred with overhead room lights off without a white noise machine.

Rotarod testing was performed using the Omnitech Electronics AccuRotor EzRod. Mice underwent training prior to recorded trials. They were placed on the rotarod at 4 rpm for a total of 60 s per training trial. Mice were replaced on the rotarod each time they fell until they achieved 60 total seconds of successful locomotion (timer stops while not on rotarod). Each animal received three training trials with at least a 5-min resting time in their home cages between each trial. These animals were then tested on the following day for the research trials, being placed on a rotarod that started at 4 rpm and accelerated at a constant rate to 40 rpm over 300 s. The latency to fall was recorded for each animal in two separate trials and averaged, with at least a 10-min rest period in their home cage between trials. The rotarod was cleaned between each animal with Clidox and each animal was tested individually. Testing occurred under low light conditions (~5 lux) without a white noise machine.

Animals were transcardially perfused with phosphate-buffered saline (PBS) and spinal cords were extracted via hydraulic extrusion (49) followed by the collection of EDLs from the mouse hind-limbs. Extruded spinal cords were further dissected down to only include the lumbosacral enlargement. For histology samples, tissues were post-fixed using 4% paraformaldehyde (PFA, Electron Microscopy Sciences, # 50980494) overnight at 4°C. Next, tissues were transferred into a 30% sucrose solution for at least 24 h and then frozen at −80°C in optimal cutting temperature (OCT, Fisher Scientific, #23-730-571). Tissues used for RNA and Protein extraction were immediately processed for immunoprecipitation.

Spinal cords were dounce homogenized in IP buffer (50 mM Tris, pH 7.4, 100 mM KCl, 12 mM MgCl₂, 1% Nonidet P-40 [NP-40]) supplemented with 1 μl/ml RNasin (Promega, # N2611), 1 mg/ml heparin, 100 μg/ml cycloheximide, and protease inhibitor mixture (Sigma-Aldrich, # 11697498001). Following homogenization, samples were rotated at 4°C for at least 20 min before centrifugation at 10,000 × g at 4°C for 10 min. A 50 μl sample was taken from the supernatant of this centrifugation and the remaining supernatant was transferred to a new tube. Samples were pre-cleared by adding Protein G Magnetic Beads (1,8 vol,vol, ThermoFisher Scientific, # S1430S) and rotated for 1 h at 4°C. Beads were then separated from each sample using a magnetic rack, and the remaining supernatant was again transferred to a new tube and incubated with an HA antibody (1,240, BioLegend, #MMS-101P) with rotation at 4°C for 4 h. Protein G magnetic beads were then added (1,24) and the mixture was rotated overnight at 4°C. The next day, samples were separated by placing them on a magnetic rack. The supernatant was decanted, and the IP product bound to the magnetic beads was washed 3 times using a high salt buffer which consisted of the IP buffer with the addition of 300 mM KCL. RNA was then extracted from the beads following manufacturer instructions using the RNeasy Plus Micro Kit (Qiagen, # 74034). The flow through from the first RNA column separation step was added to chilled acetone (4X sample volume) and kept on ice for 30 min to precipitate the sample proteins. The protein-acetone mixtures were then centrifuged at 10,000 × g at 4°C for 10 min. The acetone was decanted, and the protein pellet was washed once with 100 μl of ice-cold 100% ethanol and then allowed to dry. Protein pellets were then frozen at −80°C for downstream applications.

Precipitated IP protein pellets were re-suspended in 100 μl of molecular grade water (Promega, # PR-P1193). 100 μl of 2X sodium dodecyl sulfate (SDS) sample buffer (4% SDS, 20% glycerol, 1% β-mercaptoethanol and bromophenol blue in TBS, pH 6.8) was added to the re-suspended samples. 400 μl of 2X SDS sample buffer was also added to the previously collected 50 μl input samples, and these protein samples were then denatured at 100°C for 10 min prior to electrophoresis.

Samples were all resolved using 16% SDS—polyacrylamide electrophoresis gels (Invitrogen, #XP00165BOX) and dry transferred using the Invitrogen iBlot 2 Gel Transfer Device (Waltham, MA) to polyvinylidene fluoride (PVDF) membranes. Membranes were blocked with 5% milk in TBS (pH 7.4) containing 0.1% Tween-20 (TBST) at room temperature for 1 h with agitation. Primary antibodies were HA (1:1,000, Cell Signaling, #3724) and Actin (1:5,000, Santa Cruz, #SC-1616-G), and were added to 3% bovine serum albumin (BSA) in TBST to make the primary antibody solutions. After blocking with milk, membranes were washed with 1% BSA and then incubated overnight at 4°C in their respective primary antibody solutions. Membranes were again washed with 1% BSA and incubated with their respective horseradish peroxidase (HRP)-linked secondary antibodies (1:10,000, Jackson ImmunoResearch) in 3% BSA. Membranes were visualized using PICO plus chemiluminescent substrate (ThermoFisher Scientific, #34580) and imaged using the Amersham Imager 680 (Buckinghamshire, UK). Quantification was performed by densitometric analysis with ImageJ2 FIJI (NIH, V-2.14.0/1.54f). The intensity of signals was determined using arbitrary units for the amount of HA protein in the IP samples and ACTIN in the input samples, and the quantification of HA was normalized to ACTIN measurements for each sample. Eight animals were examined in these experiments; 2 Ribo^HOM;ChAT^Cre-, 3 Ribo^HET;ChAT^Cre+, and 3 Ribo^HOM;ChAT^Cre+.

20 μm transverse cryosections (spinal cords) and 40 μm longitudinal sections (EDLs) were cut using a Leica CM1950 cryostat (Teaneck, NJ). Antigen retrieval was performed on spinal cord sections prior to immunostaining by boiling slides in 0.1 M sodium citrate buffer, pH 6.0 for 1 min and 20 s and then washed with 1X PBS. Sections of spinal cords and EDLs were blocked in PBS containing 0.3% Triton, 1% BSA, 10% donkey serum (ThermoFisher, #nc9624464) and M.O.M blocking reagent (Vector Laboratories, #MKB2213) at room temperature for 1 h. Slides were then incubated at 4°C overnight in a primary antibody solution containing PBS with 0.3% Triton, 1% BSA and the following primary antibodies: NeuN (1:200, Millipore, #MAB377), ChAT (1:200, Millipore, #AB144P), Synapsin (1:200, Cell Signaling, #5297S), TDP-43 (1:500, Proteintech, #10782-2-AP) and Desmin (1:200, Abcam, # ab15200). Note that these antibodies have been thoroughly validated with relevant control tissues and only label the structure they are directed towards. The following day the sections were washed with PBS and labeled with the appropriate secondary antibodies (Biotium, 1:200) in PBS with 0.3% Triton at room temperature for 1 h. EDLs were also labeled with fluorescently conjugated α-bungarotoxin to analyze NMJs. Slides were again washed in PBS and then mounted using DAPI mounting medium (Southern Biotech, #0100-20). Imaging was performed using the Leica SP8 Lightning confocal microscope (Teaneck, NJ). 12 animals were utilized for immunostaining experiments (6 Ubqln2^P497S/+ and 6 Ubqln2^+/+). One Ubqln2^+/+ spinal cord was damaged during collection, so spinal cord counts compare 5 Ubqln2^+/+ (2 male and 3 female) and 6 Ubqln2^P497S/+ (3 male and 3 female) animals.

20 μm transverse cryosections (spinal cords) were cut using a Leica CM1950 cryostat. FISH was performed using RNAscope Multiplex Fluorescent Detection Reagents v2 (ACDbio, # 323110) according to the manufacturer’s protocol for fixed-frozen tissues. The RNAscope mouse probes used were Gpr50 (#318221), Des (#407921-C2), Acta1 (#808831-C3), Tnnt3 (#485511-C3), Tnn (#522321-C3), and Chat (#408731-C2). Imaging was performed using the Leica SP8 Lightning confocal microscope. For FISH analysis, 3 Ubqln2^+/+ animals were compared to 3 Ubqln2^P497S/+ animals.

Spinal cord and EDL sections were imaged with high resolution (2048 × 2048) at 20X magnification and maximally projected using the Leica LAS X software (V3.5.5.19976). Motor neuron cell bodies in the spinal cord were serially imaged about 60 μm apart and ChAT+ motor neurons were manually counted and divided by the total number of imaged/counted sections to yield the average number of ChAT+ neurons per section per animal. For NMJ analysis, regions of interest (ROIs) were drawn around each α-bungarotoxin-labeled endplate throughout each EDL and quantifications were done using the Leica LAS X software. ROI area was calculated as the number of pixels within each ROI. The presence of presynaptic Synapsin was utilized to quantify NMJ innervation. The Synapsin Pixel Sum represents the number of Synapsin pixels present within the NMJ (ROI) and the Synapsin mean value is a calculation of the number of Synapsin pixels divided by the total area of the ROI. For the EDL quantifications, the number of NMJs was normalized across animals to give each animal equal statistical power by averaging them into 6 data points (for each of 6 muscle sections imaged) per animal. For these experiments n = 10 animals were analyzed [5 Ubqln2^+/+ (2 male and 3 female) and 5 Ubqln2^P497S/+ (3 male and 2 female)]. Imaging and analysis settings were kept the same for all sections within each experiment and cell body counts were performed by individuals blinded to genotypes at the time of counting. NMJ innervation quantifications were automated by the LASX software, removing the potential for human bias and/or variability.

RNA samples were obtained and purified from 18 animals (5 Ubqln2^P497S/+ males, 5 Ubqln2^P497S/+ females, 4 Ubqln2^+/+ males and 4 Ubqln2^+/+ females). The total RNA samples were first evaluated for their quantity and quality using an Agilent Bioanalyzer 2100. All samples had good quality with RIN (RNA Integrity Number) values greater than 8. One nanogram of total RNA was used for library preparation with the SMART-Seq v4 Ultra Low Input RNA Kit (Clontech, Mountain View, CA) and the Nextera XT DNA Lib Kit (Illumina, San Diego, CA) following the manufacturer’s protocols. Each final uniquely dual-indexed library was quantified and quality accessed by Qubit and Agilent TapeStation, and multiple libraries were pooled in equal molarity. The pooled libraries were sequenced with 2 × 100 bp paired-end configuration on an Illumina NovaSeq 6000 sequencer.

The sequencing reads were first quality checked using FastQC (v.0.11.5, Babraham Bioinformatics, Cambridge, UK) for quality control. The sequence data were then mapped to the mouse reference genome mm10 using the RNA-seq aligner STAR (v.2.5) (50) with the following parameter: “--outSAMmapqUnique 60.” To evaluate quality of the RNA-seq data, the number of reads that fell into different annotated regions (exonic, intronic, splicing junction, intergenic, promoter, UTR, etc.) of the reference genome was assessed using bamutils (from ngsutils v.0.5.9) (51). Uniquely mapped reads were used to quantify the gene level expression employing featureCounts (subread v.1.5.1) (52) with the following parameters: “-s 0 -Q 10.” The data was normalized using TMM (trimmed mean of M values) method. Differential expression analysis was performed using edgeR (v.3.36.0) (53, 54). False discovery rate (FDR) was computed from p-values using the Benjamini-Hochberg procedure. RNAseq was performed on two separate sequencing runs, creating batch variation effects. Batch correction was performed but was unable to completely remove the differences in these samples, which is best observed in the MDS plots and heat maps (Supplementary Figures 2, 4). This likely reduced our ability to detect some significantly changed genes and necessitated using statistical cutoffs of FDR < 0.1 or p-value<0.05. It should be noted that FISH validation experiments of several identified genes demonstrated differences in expression between Ubqln2^P497S/+ and Ubqln2^+/+ mice using these cutoffs. In the sequencing analysis of input mRNA, this was total spinal cord mRNA and was not motor neuron specific. We were technically not able to evaluate the transcriptome of spinal motor neurons and, therefore, are not able to directly evaluate changes in translation to transcription. All sequencing data from this project will be available from the Gene Expression Omnibus (GEO).

Analysis of data was performed using Graphpad Prism 10 Software (v10.1.1). The Shapiro–Wilk Test was used to determine the normality of each dataset followed by t-tests or one-way ANOVAs depending on the number of groups compared. Tukey’s post hoc tests were used to identify significant differences in pairwise comparisons unless otherwise indicated in figure legends. All bar graph error bars represent the standard error of the mean (SEM). The sample size (n) and significance level (∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001, and ∗∗∗∗p ≤ 0.0001) are also included in the figure legend of each dataset.