Protein folding and misfolding is often described as diffusion of a molecular ensemble over a funnel-shaped energy landscape. Yet, detailed experimentally derived energy landscapes are rare, especially in the case of multi-domain proteins which represent the majority of the eukaryotic proteome. This knowledge-gap is partly due to our inability to unveil the details of folding mechanisms that can be buried in the ensemble-averaged output of traditional bulk methods. Single-molecule techniques have provided a perspective beyond the ensemble average and enable studying the folding trajectories of protein molecules in unprecedented detail. These methods can, in principle, detect rare folding or misfolding events, and ultimately lead to a reconstruction of the free energy landscape. In this thesis, the folding mechanism of both single- and double-domain proteins is unraveled using single-molecule optical tweezers.
We first focused on the mechanical properties and unfolding pathway of the four-helix acyl-CoA binding protein (ACBP). Contrary to previous studies which have shown protein native states to be brittle, we observed extraordinary compliance for ACBP along two orthogonal pulling axis, with transition states located almost halfway between the native and unfolded states. When pulled from the N- and C-termini, both experiments and simulations suggested that the molecule populates a transition state that resembles that observed during chemical denaturation, with respect to structure and position along the reaction coordinate. The results call for a modified view of the mechanical response of native states which may be relevant in an evolutionary sense for functionality and for biomolecular design.
A vastly more complex folding behavior was observed for the two-domain neuronal calcium sensor 1 (NCS1). The NMR solution structure of NCS1, in combination with fluorescence spectroscopy and mutational analysis, suggested a novel role for the C-terminal tail in regulating conformational stability. On the single-molecule level, the C-domain folded through a partially folded intermediate state followed by slow rearrangement of the sensory site, crucial for the subsequent folding of the N-domain. If the sensory site rearrangement failed, the molecule populated misfolded states at lower forces in a process that could be modulated with relaxation speed. Remarkably for a calcium sensor, high calcium levels increased the lifetime of the misfolded states and slowed folding to the native state. We propose a multi-dimensional and rugged energy landscape for NCS1, and speculate on a direct link between protein misfolding, calcium dysregulation and neurodegeneration.