Yan, Tingxiang

AbstractBackgroundUFMylation is an understudied ubiquitin‐like post‐translational modification (PTM). Like ubiquitin, UFM1 is conjugated to substrates via a catalytic cascade involving a UFM1‐specific E1 (UBA5), E2 (UFC1), and an E3 ligase complex (UFL1, DDRGK1 and CDK5RAP3). UFMylation is reversible, and this is mediated by UFSP2. UFMylation plays essential roles in several biological processes including the DNA damage response, endoplasmic reticulum homeostasis, unfolded protein response, autophagic functions, and immune response. Importantly, UFMylation is also crucial for healthy brain development. Alzheimer’s disease (AD), a complex neurodegenerative disorder, is characterized by the accumulation of pathologic tau and beta‐amyloid proteins and has been associated with abnormalities in the aforementioned processes. Intriguingly, the potential link between AD and UFMylation has not been investigated yet. Considering that pathological tau plays a critical role in AD, and tau also plays an important role for the DNA damage response in neurons under physiological conditions, we speculated that tau itself could be modified with UFM1.MethodWe performed Western blot and Meso Scale Discovery based ELISA to measure the protein level of UFM1, UFSP2 and other UFMylation pathway components in RIPA‐soluble and insoluble fraction in human post‐mortem frontal cortex. To test whether tau can be modified by UFM1, we performed co‐immunoprecipitation (Co‐IP), Proximity Ligation Assays (PLA) and liquid chromatography‐tandem mass spectrometry (LC‐MS/MS).ResultWe found a significant reduction of soluble UFSP2, but an increase of soluble total UFM1 in AD compared to controls. We found that tau interacts with UFL1 and DDRGK1. In addition, using immunoprecipitation of denaturing lysates we found that tau can indeed be covalently modified by UFM1 when overexpressed. We detected strong signal of HA‐UFM1‐tau in neurons but not in negative controls using PLA. Furthermore, we immunoprecipitated denatured UFM1 from human autopsy brain, and detected two specific UFM1‐tau bands with tau antibodies. Together, this data strongly indicates that tau is UFMylated under endogenous conditions. Using LC‐MS/MS, we identified 17 UFMylation sites in tau proteins.ConclusionUFMylation is dysregulated in AD; tau protein can be modified by UFM1 both in vitro and in vivo, which enables a more comprehensive understanding of tau PTMs.

AbstractBackgroundNeuropathologically, Alzheimer’s disease (AD) is characterized by the presence of senile plaques composed of beta‐amyloid (Aβ) and neurofibrillary tangles made of hyperphosphorylated tau. Aβ and p‐tau have been shown to impact endoplasmic reticulum homeostasis, DNA damage response and autophagic functions. Intriguingly, alterations in these systems have also been connected to dysfunctions in UFMylation, a ubiquitin‐like posttranslational modification. Analogous to ubiquitin, UFM1, is conjugated to substrates via a catalytic cascade involving UFM1‐specific set of E1 (UBA5), E2 (UFC1), and a complex that consists of E3 ligase (UFL1), DDRGK1 and CDK5RAP3. UFMylation is reversible, and this is mediated by UFSP2. The UFMylation pathway is essential for brain development as complete loss of function of any of its components result in severe neurodevelopmental disorders. However, alterations in UFMylation and their role in and significance for age‐related AD remains unknown.MethodWe performed Western blot and Meso Scale Discovery based ELISA to measure the protein level of UFM1, UFSP2, tau, p‐tau and other UFMylation pathway components in human post‐mortem brain. Analyses were performed with RIPA‐ and 2% SDS‐soluble fraction from frontal cortex. To validate our finding in human tissues, we knocked out UFM1 and UFSP2 in ReN cells, a neural progenitor cell model that has already been used to study AD and that can be terminally differentiated in neurons.ResultWe found a significant reduction of RIPA UFSP2, but an increase of RIPA and SDS UFM1 in AD compared to controls. Strikingly, RIPA UFSP2 had a strong negative correlation with both RIPA and SDS UFM1. Furthermore. UFM1 was significantly increased in UFSP2‐KO neurons compared to WT, mirroring the negative correlation between UFM1 and UFSP2 observed in human brain. In addition, RIPA and SDS UFM1 had a strong positive correlation with p‐tau, indicating that dysfunctional UFMylation might play an important role for disease pathogenesis.ConclusionReduction of RIPA soluble UFSP2 in human AD frontal cortex might result in the abnormal accumulation of UFM1 and this might further facilitate the AD pathogenesis. Because of its role for several central AD‐relevant pathways, the UFMylation pathway might be a potential therapeutic target for AD.