Protein Folding And Aggregation In The Presence Of The Hsp70 Chaperone

Most life on earth depends on ribosome-assisted biosynthesis and on the generation and preservation of correct protein structure. Molecular chaperones and their cochaperones act co- and post-translationally to promote de novo protein folding, overcome protein damage upon stress and even disaggregate protein aggregates. Hsp70, a ubiquitous and highly conserved 70 kDa heat shock protein, is a particularly important and well-studied chaperone. It is often referred to as a central "hub" due to its myriad of functions and its profound effect on cell viability. While the Hsp70 chaperone cycle has been well-documented in the literature, there is still much to be understood about the interplay between Hsp70 and its client-proteins, including the kinetic and thermodynamic client-protein characteristics required for interaction with Hsp70. The Hsp70 chaperone is nucleotide-dependent and derives part of its driving force for assisting protein folding from ATP hydrolysis. The Hsp70-related studies carried out to date bear an apparent inconsistency. Namely, some proteins were reported to attain their native state more slowly in the presence of the Hsp70 chaperone than under chaperone-free conditions. On the other hand, aggregation-prone proteins routinely acquire a bioactive native state faster, in the presence of Hsp70. Part of the work carried out in this thesis attempts to explain this apparent inconsistency. In addition, we explore the kinetic and thermodynamic client-protein characteristics necessary for interaction with the Hsp70 chaperone. Finally, we address the relation between protein aging and Hsp70-chaperone activity.The thesis is divided into six chapters. Chapter 1 delves into the current literature and summarizes what is known about protein folding and how the folding process is influenced by the Hsp70 chaperone cycle. This chapter further discusses the structure and function of Hsp70 and how these characteristics affect the conformation and dynamics of chaperone-bound client proteins. The chapter also provides a brief overview of the current computational approaches to predict the timecourse of Hsp70-assisted protein folding. Chapter 2 focuses on the development of CHAMPION70, a computational model able to perform Chaperone-Mediated Protein folding kinetic Simulations involving Hsp70. We then apply CHAMPION70 to four classes of client proteins with different kinetic (fast- or slow-folding) and thermodynamic (stable or unstable) stabilities in the presence of either no aggregation, weak aggregation or strong aggregation propensities. We find that, in the absence of aggregation, unstable client proteins capture (i.e., stay bound to) the Hsp70 chaperone indefinitely. This is a clear disadvantage unless Hsp70 serves as a transport machine, for these proteins. Conversely, in the presence of weak or strong aggregation propensities, it is very beneficial for client-proteins to interact with the Hsp70 chaperone system. Specifically, slow-folding and thermodynamically stable client proteins experience the greatest aggregation-prevention advantages in the presence of Hsp70, especially if the class of client proteins is strongly aggregation-prone. However, Hsp70 is unable to assist the folding of strongly aggregation-prone and thermodynamically unstable proteins. Importantly, we also predict that the E. coli Hsp70 chaperone system is unable to prevent protein aggregation over long time spans long-term (i.e., greater than ca 60 years). This result suggests that one of the consequence of protein aging is the intrinsic failure of the bacterial Hsp70 chaperone machinery. Of course E. coli bacteria double in only a few minutes and "old proteins" likely persist in the progeny (i.e., daughter cells). Yet these old proteins progressively become more and more dilute, hence less-aggregation-prone. This phenomenon may rescue bacteria from disaster. Yet one wonders whether this effect may have a more severe impact on eukaryotic Hsp70s. In summary, the CHAMPION70 simulator is a powerful tool to enable the prediction of client-protein behavior in the presence of one of the most amazing cellular machines, the Hsp70 chaperone system. Chapter 3 provides simple computational tools to discriminate folded from intrinsically disordered proteins (IDPs) under physiologically relevant conditions, solely based on protein amino-acid composition. This tool only requires knowledge on protein hydrophobicity-per-residue and net-charge-per-residue. The net-charge-non-polar (NECNOP) algorithm results in 95% accuracy, and this value increases for proteins of more than 140 residues. Chapter 4 delves into influence of the E. coli ribosome on both co- and post-translational protein folding in the absence typical molecular chaperone systems (DnaK, trigger factor) and in the presence of aggregation. In this experimental investigation, translation through the ribosome is found to promote nascent-protein solubility even in the absence of cotranslationally active molecular chaperones. This work also shows that the E. coli trigger factor and DnaK molecular chaperones increase the solubility of nascent chains emerging from the ribosomal exit tunnel and minimize co- and post-translational aggregation. Most importantly, this work shows the importance of immediately post-translational kinetic partitioning of nascent proteins between native-state and aggregates, upon release form the ribosome. This partitioning is dramatically sensitive to subtle variations in amino-acid sequence, including single-point mutations. Chapter 5 demonstrates the increased sensitivity of the NMR hyperpolarization technique known as low-concentration photochemically induced dynamic nuclear polarization (LC-photo-CIDNP). This technique is used for detection of aromatic amino acids in the presence of both a photosensitizer dye (fluorescein) and a cryogenic probe. Experiments rapidly detect the amino acids tryptophan (Trp) and tyrosine (Tyr) at unprecedented concentrations (200 nM). Detection of the model protein drkN SH3 (which bears Trp, Tyr and His) at 500 nM on a 600 MHz spectrometer via LC-Photo-CIDNP leads to a 30-fold better S/N relative to conventional 2D experiments performed at higher magnetic field (900MHz spectrometer). Spectral editing of the model protein allowed for secondary and tertiary structure analysis. In contrast to regular photo-CIDNP, LC-photo-CIDNP does not heavily depend on laser intensity, thus allowing for safer and more cost-effective experiments. Chapter 6 further develops the investigations of Chapter 5 on LC-photo-CIDNP. A major limitation of LC-photo-CIDNP is that a limited number of scans (up to ca 200) can typically be collected before sample degradation takes over. The signal-to-noise (SN) ratio becomes progressively weaker as the number of scans increases. This disadvantage strongly limits the ability to perform long-term experiments. Two reductive radical quenchers - ascorbic acid (vitamin C) and 2-mercaptoethylamine (MEA) - were employed in this study, to minimize the extent of photodamage in NMR samples. This technique both enhanced the S/N by over 100% and allowed for more transients to be acquired for amino-acid and protein samples in solution.