Molecular mechanisms underlying LSM12-dependent post-transcriptional regulation in neural physiology

Abstract

Post-transcriptional regulation is emerging as an essential mechanism to diversify eukaryotic gene expression. Neurons take great advantage of several regulatory mechanisms, including RNA stability, poly-A tailing, gene silencing, and translational activation, to diversify their functional properties. RNA-binding proteins (RBPs) may thus have evolved neuron-specific regulatory mechanisms to support complex and prompt changes in neuronal gene expression and the relevant physiologies such as circadian clocks and neurodegeneration. The translational regulator TWENTY-FOUR (TYF) associates with the Drosophila homolog of ATAXIN-2 (ATX2) to activate the translation of the rate-limiting circadian clock gene period (per), thereby defining a post-transcriptional co-activator function of ATX2 in Drosophila circadian clocks. ATX2 has also been increasingly implicated in neurodegenerative diseases. However, it is poorly understood how ATX2-associating factors assemble into a specific protein complex and how they contribute to the physiological function of the ATX2 complex in vivo. To address these questions, we first performed an in vivo genetic screen for the clock function of ATX2-associating proteins. This approach identified LSM12 (like-Sm protein 12) and ME31B, as two key components of the ATX2 complex sustaining circadian periodicity and rhythm amplitude, respectively. We established a Drosophila model of Lsm12 deletion by imprecise excision of a P element insertion. Our Lsm12-deletion flies exhibited abnormal circadian rhythms with long-period, as similarly seen in tyf mutants and Atx2-deficient flies that were caused by dampened PER oscillation in circadian pacemaker neurons. In contrast, ME31B-depleted flies exhibited poor rhythmic behaviors via a PER-independent ATX2 pathway. These results suggest that LSM12 and ME31B have distinct roles in post-transcriptional regulation by the ATX2 protein complex. A series of biochemical analyses revealed that LSM12 acts as a molecular adaptor of the ATX2-associating protein complex to recruit TYF. The ATX2-LSM12-TYF complex is associated with 5’ cap-binding proteins in an ATX2-dependent manner, thereby supporting TYF-dependent translational activation. On the other hand, ME31B/DDX6 facilitated the selective association of ATX2 with a scaffold protein NOT1 of the CCR4-NOT deadenylase complex to support NOT1-mediated gene silencing activity. We thus propose that the ATX2 protein complex switches between activator and repressor modes of post-transcriptional regulation via its associating factors LSM12 and ME31B. It defines PER-dependent and -independent clock functions of ATX2 that contribute to 24-hour periodicity and high-amplitude rhythms, respectively, in Drosophila circadian behaviors. Since the biochemical association of LSM12 with ATX2 is well conserved between Drosophila and humans, we next sought to explore whether LSM12 would play similar roles in other ATX2-relevant physiology such as neurodegeneration. We performed an in vivo genetic screen with transgenic Drosophila models of neurodegenerative diseases, including C9ORF72-associated amyotrophic lateral sclerosis and frontotemporal dementia (C9-ALS/FTD), spinocerebellar ataxia 3 (SCA3), and Alzheimer’s disease. Interestingly, we observed C9-ALS/FTD-induced neurodegeneration was exacerbated in Drosophila mutants of Lsm12, indicating its neuroprotective function. Nucleocytoplasmic transport (NCT) defects are increasingly implicated in the pathogenesis of C9-ALS/FTD. We thus asked whether the disruption of NCT would explain LSM12-dependent molecular pathogenesis of C9-ALS/FTD. We identified a neuroprotective pathway of LSM12 and EPAC1 (exchange protein directly activated by cyclic AMP 1) that sustained the nucleocytoplasmic RAN GTPase (RAN) gradient, a key NCT regulator, and thereby suppressed NCT dysfunction by the C9ORF72-derived poly(glycine-arginine) protein. LSM12 depletion in human neuroblastoma SH-SY5Y cells aggravated poly(GR)-induced impairment of NCT and nuclear integrity while promoting the nuclear accumulation of poly(GR) granules. Overexpression of ALS-associated LSM12V135I mutant comparably increased poly(GR) toxicity, indicating dominant-negative effects. Transcriptome and reporter analyses revealed that LSM12 post-transcriptionally up-regulated EPAC1 expression, whereas EPAC1 overexpression rescued the RAN gradient and NCT defects in LSM12-deleted cells. Poly(GR)-induced neurodegeneration was consistently exacerbated in Drosophila Epac mutants. Finally, we employed induced pluripotent stem cell (iPSC)-derived neurons from C9-ALS patients (C9-ALS iPSNs) to validate the physiological relevance of our findings. Lentiviral overexpression of LSM12 or EPAC1 indeed rescued NCT-relevant pathologies in C9-ALS iPSNs. These findings support our conclusion that LSM12 and EPAC1 constitute a neuroprotective pathway for sustaining the RAN gradient and NCT in the pathophysiology of C9ORF72-associated ALS/FTD. A series of biochemical experiments revealed that EPAC1 was essential for associating the RAN-importin β1 complex with the cytoplasmic nuclear pore complex, thereby dissipating the nucleocytoplasmic RAN gradient critical for NCT. These findings define a conserved role of the LSM12-EPAC pathway as an important suppressor of the NCT-related pathologies in C9-ALS/FTD. Taken together, my thesis research has elucidated the molecular principles underlying LSM12-dependent post-transcriptional gene regulation and provided insights into the molecular pathogenesis of circadian disruption and ALS.