Abstract
Toxic polyglutamine (polyQ) expansions in ataxin-2 (ATXN2) trigger neurodegenerative processes, causing spinocerebellar ataxia type 2, and enhancing TAR DNA-binding protein 43-dependent pathology in amyotrophic lateral sclerosis/frontotemporal dementia. Primary disease events can be compensated transiently, delaying disease manifestation. To define potential therapy targets, here we studied how cells modify phosphoprotein signals, using preferentially affected nervous tissue from end-stage Atxn2-CAG100-Knockin mice. The spinal cord phosphoproteome revealed massive hyperphosphorylations flanking the polyQ expansion in ATXN2 and for SQSTM1 and moderate hyperphosphorylations also for amyotrophic lateral sclerosis proteins, OPTN (optineurin), UBQLN2 (ubiquilin-2), TNIP1 (TNFAIP3 interacting protein 1), and TBK1-targeted TAX1BP1. Conversely, strong hypophosphorylations of WNK1 (protein kinase with no lysine 1), SPARCL1 (secreted protein acidic and cysteine rich-like 1), and PSMD9 (proteasome 19S regulator non-ATPase assembly chaperone P27) were found. Significant enrichments of SRC-homology domain type 3-containing proteins, autophagy/endocytosis factors, and actin modulators could be explained by N-terminal, polyQ-adjacent, proline-rich motifs in ATXN2, suggesting that spinocerebellar ataxia type 2 pathogenesis is highly similar to Huntington's disease, where neurotoxicity is mediated by abnormal polyQ-proline-rich motif-SRC-homology domain type 3 interactions. Validation of protein and mRNA levels was done in mouse spinal cord and embryonic fibroblasts or patient fibroblasts after bafilomycin or arsenite treatment, observing polyQ-dependent OPTN deficiency and SQSTM1 induction impairment. Overall, this phosphoproteome profile identified and quantified the main cellular efforts in adapting autophagy pathways to the aggregation propensity of the ATXN2-N-term.