RNA modulates the assembly pathway and structure of amyloids formed by the RNA binding protein hnRNPA1

Neurodegenerative diseases such as Amyotrophic Lateral Sclerosis (ALS) are associated with the formation of aberrant protein aggregates which co-localize with stress granules (SGs). These biomolecular condensates, rich in RNA, form and dissolve dynamically in response to cellular stressors. Despite the recognized connection between phase separation and pathological protein aggregation, it is still unclear how SGs promote amyloid fibril formation. In particular, although several aggregation-prone proteins are RNA binding proteins (RBPs), the role of RNA on the molecular mechanisms leading to the formation of fibrils and on their final structure remains elusive. In this thesis, we address this question focusing on heterogeneous nuclear ribonucleoprotein A (hnRNPA1A), a ubiquitously expressed human protein involved in several steps of RNA metabolism, including transcription, splicing and RNA shuttling in between the nucleus and cytoplasm. Mutations in hnRNPA1A are associated with the onset of diseases such as ALS and Multisystem Proteinopathy (MSP), which are characterized by cytoplasmic inclusions of hnRNPA1A and other RBPs. Here, we analyze how hnRNPA1A aggregation is modulated by bimolecular condensates and RNA in vitro. We adopt a biochemical approach, using minimal reconstituted systems composed of purified proteins and ribonucleic acids. Fibril assembly is evaluated using a combination of biophysical methods ranging from fluorescence-based techniques such as confocal microscopy and fluorescence correlation spectroscopy (FCS), to electron scattering methods including transmission electron microscopy (TEM) and cryo-transmission electron microscopy (cryo-EM). Our results show that in heterotypic condensates, similarly to recent observations in homotypic condensates, the interface of biomolecular condensates promotes fibril assembly. We find that RNA partitioning into condensates accelerates interface-mediated aggregation, possibly by modulating the material properties of the dense phase. In excess of RNA, condensate formation is suppressed, and amyloid formation occurs following a distinct molecular pathway. Our results highlight a potential pathological role of SGs, where the presence of non-specific RNA may lower the kinetic barrier for amyloid nucleation. In addition to the kinetics of fibril formation, we investigated the effect of RNA on the structure of amyloids. We show that RNA governs fibril morphology independently of the molecular pathway of aggregation. Leveraging cryo-EM, we evaluate with atomic resolution the structural changes induced by RNA on the amyloid architecture. Moreover, we report the fibril structure of the disease-relevant hnRNPA1A variant D262V in presence of RNA, that yet creates a distinct amyloid fold compared to the wild type. Overall, our data shows that the sequence responsible for hnRNPA1A aggregation is conserved, but its fold within the fibril core depends on the presence of RNA and sequence mutations. To gain molecular insights in RNA modulation of hnRNPA1A phase transitions, we extended our experiments from the micro to the nanoscale. Our analysis reveals that hnRNPA1A and RNA create a rich phase diagram composed of complexes characterized by largely different sizes and properties. These species are metastable and fibrils form over time in all conditions, although following distinct pathways. In conclusion, in this thesis we evaluated the effect of biomolecular condensates and RNA on the fibril assembly pathway of hnRNPA1A both from a kinetic and structural perspective. Our results consolidate that RNA is a relevant modulator of hnRNPA1A phase behaviour, and indicate how the behaviour of proteins and nucleic acids should be investigated together to understand their physiological and pathological roles.