Edited by John Denu This article contains supporting information.

The past decade has seen a rapid acceleration in the discovery of new genetic causes of ALS, with more than 20 putative ALS-causing genes now cited. These genes encode proteins that cover a diverse range of molecular functions, including free radical scavenging (e.g., SOD1), regulation of RNA homeostasis (e.g., TDP-43 and FUS), and protein degradation through the ubiquitin-proteasome system (e.g., ubiquilin-2 and cyclin F) and autophagy (TBK1 and sequestosome-1/p62). It is likely that the various initial triggers of disease (either genetic, environmental and/or gene-environment interaction) must converge upon a common set of molecular pathways that underlie ALS pathogenesis. Given the complexity, it is not surprising that a catalog of molecular pathways and proteostasis dysfunctions have been linked to ALS. One of the challenges in ALS research is determining, at the early stage of discovery, whether a new gene mutation is indeed disease-specific, and if it is linked to signaling pathways that trigger neuronal cell death. We have established a proof-of-concept proteogenomic workflow to assess new gene mutations, using CCNF (cyclin F) as an example, in cell culture models to screen whether potential gene candidates fit the criteria of activating apoptosis. This can provide an informative and time-efficient output that can be extended further for validation in a variety of in vitro and in vivo models and/or for mechanistic studies. As a proof-of-concept, we expressed cyclin F mutations (K97R, S195R, S509P, R574Q, S621G) in HEK293 cells for label-free quantitative proteomics that bioinformatically predicted activation of the neuronal cell death pathways, which was validated by immunoblot analysis. Proteomic analysis of induced pluripotent stem cells (iPSCs) derived from patient fibroblasts bearing the S621G mutation showed the same activation of these pathways providing compelling evidence for these candidate gene mutations to be strong candidates for further validation and mechanistic studies (such as E3 enzymatic activity assays, protein–protein and protein–substrate studies, and neuronal apoptosis and aberrant branching measurements in zebrafish). Our proteogenomics approach has great utility and provides a relatively high-throughput screening platform to explore candidate gene mutations for their propensity to cause neuronal cell death, which will guide a researcher for further experimental studies.

TDP-43 is a nuclear RNA-binding protein that can undergo liquid-liquid phase separation (LLPS) and forms pathological insoluble aggregates in frontotemporal dementia and amyotrophic lateral sclerosis (ALS). Perturbations of TDP-43 function are linked to mislocalization and neurodegeneration. By studying TDP-43 in vivo, we confirmed for the first time that TDP-43 undergoes LLPS and forms biomolecular condensates in spinal motor neurons (MNs). Importantly, we discovered that interfering with the K136 SUMOylation site of TDP-43 altered its phase separation behavior, reducing cytoplasmic mislocalization and aggregation. Introduction of the ALS-linked mutation G294V did not alter these LLPS characteristics, indicating that posttranslational modifications such as lysine-specific alterations can modulate TDP-43 pathogenesis through regulating phase separation. Altogether, our in vivo characterization of TDP-43 confirms the formation of dynamic nuclear TDP-43 condensates in zebrafish spinal neurons and establishes a critical platform to validate the molecular grammar of phase separation that underpins TDP-43 aggregation in ALS and other proteinopathies.

No abstract

No abstract