| dc.description.abstract | Eukaryotic cell organisation is classically described in terms of a cytoplasm and discrete, membrane-bound organelles enclosed within a plasma membrane. However, more recently it has been recognised that cells also organise their contents within membraneless organelles, including mRNP granules, which are dynamic, macromolecular assemblies of protein and RNA. These granules stochastically form and disassemble by a process of reversible phase separation, which is governed by the collective sum of weak but multivalent protein and RNA interactions amongst their components. Granules, being both distinctive in their composition and dynamic in their assembly/disassembly properties are well suited to their numerous roles in the regulation of activity, storage, and protection of molecules within cells.
Unexpectedly it has emerged that the biophysical properties that provide useful characteristics to granules also pose an inherent risk for cells. Pathological protein and RNA aggregates are a defining feature of a wide array of as yet untreatable neurodegenerative diseases. Considerable recent evidence suggests that disease-associated aggregates may arise as a result of toxic hijacking of canonical granule-forming mechanisms. With this in mind, elucidating the means by which physiological granule formation and disassembly is regulated is of considerable clinical interest.
Recent proteomic studies of RNP granule components indicate the existence of a network of protein and RNA chaperones with the ability to modulate granule-forming interactions. However, the full extent of granule associated chaperones, and the mechanisms by which they regulate RNP granule condensation is not yet clear. Additionally, it remains to be seen whether these granule regulators may be able to modulate toxicity in the context of disease. In light of this, the overarching aim of this thesis was to identify, in a Drosophila system, protein and RNA chaperones associated with stress granules (SGs), to investigate their capacity to regulate granule formation and to assess their effects in the context of neurodegenerative disease models. Further, I aimed to examine how principles learned from studies of mRNP granules could be extended to other classes of biomolecular condensate.
To this end, in the work outlined below, I have collated a list of candidate protein chaperones and RNA helicases of interest and assessed their physical colocalisation with SGs in Drosophila cell culture. While it appears that many of these candidates may interact with granules transiently or in a context-specific way, this work highlights the DEAD-box RNA helicases Me31B and Belle as consistent stress granule components.
Following on from this, I have assessed the capacity of Me31B and Belle to regulate stress granule formation. My results indicate that while neither helicase is required for granule formation, both facilitate a more efficient disassembly of granules. In addition, Belle is sufficient to induce granule formation in the absence of stress. Subsequently the exact role of Belle in granules was studied in more detail by assessing granule formation in the context of various Belle mutants. Many DEAD-box helicases, including Belle, have intrinsically disordered protein regions (IDRs) which can form transient, promiscuous interactions that are proposed to support granule formation. In line with this, my results indicate that the Belle IDRs contribute to granule formation and mediate localisation of Belle to granules. Conversely ATP hydrolysis by DEAD-box helicases has a proposed regulatory role and is likely involved in limiting condensation and maintaining appropriate granule dynamics. Here also, I found that Belle ATP hydrolysis activity is required to limit extensive SG formation. Overall, this work outlines a novel role for Belle as a regulator of SGs in Drosophila.
Following this, I next aimed to investigate whether Belle could modulate toxicity in a Drosophila ALS disease model characterised by overexpression of GR dipeptide repeats. To do so I made use of a well-established eye degeneration phenotype and also demonstrated for the first time that expression of this model in circadian clock neurons results in a circadian rhythm behavioural defect. I found that overexpression of wild-type Belle was protective in aged flies in the eye degeneration assay, but not in newly eclosed flies in either the eye degeneration or circadian rhythm assays. Conversely, loss of Belle in the context of the disease model resulted in pupal lethality when expressed in eyes and a stronger circadian rhythm phenotype during light:dark conditions when expressed in clock neurons. Together this suggests that Belle may be protective in the context of neurodegeneration. Interestingly, I found that expression of an ATP hydrolysis-deficient Belle mutant induced a circadian rhythm defect even in the absence of the ALS model and also enhanced degeneration in the eye when co-expressed with the model. Given that this mutant caused more extensive SG formation in cells, this may allude to a further connection between granule condensation and disease toxicity.
Having shown that a stress granule associated helicase can suppress toxicity in an ALS model, in the final part of this thesis I employed a similar strategy to assess neurodegeneration in a Drosophila tauopathy model. In this context I aimed to investigate the modulatory capacity of both the VCP cofactor, FAF2, which has previously been implicated in RNP granule disassembly, and a polyserine sequence which is associated with nuclear speckles, another type of membraneless organelle. I found that expression of FAF2 suppressed tau-induced toxicity and of considerable interest, this suppression was more apparent when FAF2 was linked to polyserine, which is proposed to localise proteins to both nuclear speckles in an endogenous context and tau aggregates in a disease context. This work reflects a novel role for FAF2 in suppressing tau toxicity and also demonstrates a proof of concept for a strategy with potential therapeutic value, whereby enhancing targeting of a protective protein to a pathological aggregate enhances the efficiency of its protective effect. This concept may be generally applicable to a number of diseases defined by inclusion pathology. It also highlights the importance of continuing to study RNP granules in order to identify both components which mediate targeting to condensates and components which can either limit condensation or enhance disassembly and/or degradation.
Overall, this work provides novel insights into the role of DEAD-box helicases in physiological granule regulation and their potential to suppress toxicity in the context of neurodegenerative diseases associated with pathological protein aggregation. It also alludes to a novel strategy which could be used to optimise the effective clearance of toxic inclusions. | en |
