Cellular Inclusion Bodies: Structure and Mechanisms

Inclusion bodies (IBs) are insoluble aggregate structures that commonly form upon overexpression of proteins in heterologous hosts with relevance to biomedicine and biotechnology. These structures are of broad interest in many industries: a near-pure source of protein, they are commonly used in protein purification. Additionally, their high mechanical and thermal stability, low toxicity, and resistance to proteases have made IBs an attractive candidate for being adapted as a vaccine-type molecule. Despite their applicability, the molecular mechanisms governing IB formation and structures remain obscure. The inhomogeneity, insolubility, and large size of protein aggregates makes traditional methods of structural determination such as nuclear magnetic resonance (NMR) or X-ray crystallography impossible or limited to low resolution, obstructing research into aggregation mechanisms and structures. To address this issue, we undertook a multi-mutant study of the aggregation of Cu, Zn-Superoxide dismutase (SOD1) with the goal of determining if IB formation depends on protein primary sequence and stability, establish how IB structure can be modulated by varying cell growth conditions, and ultimately assemble a model of IB structure and formation. Initial studies described in Chapter 2 involved measuring the aggregation propensity of eighteen SOD1 mutants under six different growth conditions (3 different temperatures, with and without metal cofactors to simulate maturation of the protein) using SDS-PAGE experiments combined with gel densitometry. Compelling correlations arise from these data: aggregation propensity is strongly correlated with the melting temperature of immature SOD1 (R = -0.77), as well as with the Gibb’s free energy of dimer dissociation (R = -0.68) and monomer unfolding (R = -0.58), supporting previous findings that ALS mutations form aggregation-prone monomers, while also suggesting the contribution of dimer dissociation to aggregation. Correlations between these data and other biophysical parameters such as folding rate and disease duration were not observed, nor were correlations between aggregation propensity and normalized aggregation scores from several aggregation predictors. These results highlight how multiple variables can be involved in aggregation propensity. To analyze aggregate structure, we used quenched hydrogen-deuterium exchange NMR (qHDX) measurements, a methodology that has been used extensively on purified proteins but whose application to protein aggregation in a cellular context has been little explored. The optimization of this method is discussed in detail in Chapter 3. It is then applied to the aggregates of 9 SOD1 mutant proteins under a single growth condition with predominant population of the most immature form of v the protein. The results point to overarching similarities in aggregate structure for all mutants. Extensive protection against exchange is seen throughout SOD1 regardless of a region’s native hydrogen bonding or hydrophobicity. Closer examination of specific residues shows stretches of protection in regions previously shown to form protein-protein interfaces in both native and non-native SOD1 assemblies. Taken together, these data point to an ensemble of aggregation pathways in equilibrium and a heterogeneous aggregate. Lastly, we aimed to expand the method to other protein systems, both to see if the procedure developed could be broadly applied, as well as to learn more about IB structure (Chapter 5). The first application was to Adnectins, an engineered family of proteins based on the 10th fibronectin type III domain. Adnectins have a beta-sheet sandwich structure with variable loops. In applying the qHDX method, we found that all conditions established for SOD1 readily applied to Adnectins, including 3D assignment strategies. We further found that, like SOD1, Adnectins show protection of many amides throughout the entire protein, beyond the limited regions that are protected by native structure or predicted to aggregate/form amyloid in the sequence. Combined with previous work showing that Adnectins may form native-like assemblies, these preliminary results similarly point to an ensemble model of IB formation. In contrast to the Adnectins and SOD1, apomyoglobin is a helical protein that has been seen to readily form IBs in Escherichia coli. In this case, rather than studying the effects of point mutations, we are evaluating different C-terminal truncation mutations to examine the impact of this change on the resulting aggregate structure. Preliminary results show that in vivo IB structures mimic previously studied in vitro aggregates, and that the protein becomes more native with increasing chain length. Overall, this thesis shows the utility of the qHDX method and provides a framework for understanding protein aggregation mechanisms and the resulting aggregate structures