Probing the early stages of polyglutamine aggregation with computational methods

Exonic CAG repeat diseases are a class of neurodegenerative age-of-onset diseases caused by an unstable trinucleotide expansion in a coding region of a gene. The most prominent example is Huntington's disease (HD) whose symptoms are characterized by loss of motor control and cognitive deficits. For all nine of the known CAG repeat diseases, pathology is ascribed to the mutant proteins which carry expanded stretches of glutamine residues (polyglutamine). The length of the polyglutamine segment is inversely correlated with the disease age-of-onset. Protein aggregates are routinely found in postmortem tissue samples of brains of HD patients. These findings suggest a prominent role for polyglutamine-mediated protein aggregation in disease pathogenesis. Subsequent studies characterized the intracellular aggregates as amyloid-like. In amyloids, the polypeptide backbone predominantly adopts conformations in the β-basin of the Ramachandran map, i.e., the aggregates have high net β-content. This has led to the hypothesis that β-rich conformers play a prominent role in mediating the aggregation process; specifically, it has been postulated that a β-rich form of polyglutamine acts as the monomeric nucleus from which fibrillar aggregates grow via a downhill elongation mechanism. This thesis investigates the intrinsic properties of polyglutamine during early stages of aggregation. We employ computer simulations to obtain a qualitative picture of the process at an atomistic level. Our results suggest the following: soluble polyglutamine is intrinsically disordered and forms collapsed globules in aqueous solution. These globules associate readily and randomly to form disordered dimers. We identified no structural requirements for association to occur. The conversion of monomeric polyglutamine to a conformation high in β-content, i.e., to a putative aggregation nucleus, is associated with a high free energy penalty. We detect no coupling between structure and associativity, but find a profound modulation of polyglutamine's intrinsic properties in the presence of wild-type flanking sequences. From our results, we postulate a model where polyglutamine forms large soluble and disordered oligomers which undergo a rate-limiting conformational conversion to a fibrillar precipitate. We conclude that structure-based drug designs may not prove a viable strategy for interfering with the early stages of polyglutamine aggregation and hence with disease pathology