SPECTRINROUGHNESS · Direct Investigation of a Folding Energy Landscape and Its Implications for the Unexplained Folding Behaviour of Spectrin Domains

Direct Investigation of a Folding Energy Landscape and Its Implications for the Unexplained Folding Behaviour of Spectrin Domains

Spectrin domains R15, R16 and R17 are found in tandem arrays in proteins of the cytoskeleton, and spectrin itself is believed to be responsible for maintaining the shape of red blood cells by imparting elasticity to the cell membrane. In spite of very similar thermodynamic stability and transition state structure (as inferred through phi-value analysis) (1-2), R16 and R17 reach their native state (i.e. they “fold”) over a thousand times slower than R15 (3). The intriguing piece of evidence that triggered this project was the finding that such behaviour is likely to reflect differences in the pre(-exponential) factor of the equation describing the folding reaction (4). In a Kramers’-like formulation of folding kinetics this pre-factor is essentially dependent on the intrachain diffusion coefficient (D) of the denatured polypeptide chain, a number which is now possible to measure with a combination of single-molecule Förster resonance energy transfer (smFRET) and fluorescence correlation spectroscopy (FCS), and which reflects the fine structure of the energy landscape underpinning the folding reaction (5). In this framework, a “smooth” energy landscape would be associated with fast folding, while a “rough” landscape would probably originate slower folding kinetics, due to the presence of the so called “internal friction” slowing down the folding process. The term “internal friction” indicates a general ensemble of intra-chain interactions, non-productive for the folding reaction, whose detectable effect is indeed a deceleration of folding (6). It is plausible that slow-folding spectrin domains R16 and R17 have rough energy landscapes and a non-negligible internal friction, whereas the fast folding domain R15 has a smooth energy landscape and no internal friction. I planned to study these spectrin domains using a combination of smFRET and FCS, in conjunction with newly developed data analysis methods and a diffusive description of the folding reaction from Kramers theory (7), in order to investigate the energy landscape that underpins the folding reaction of these proteins. The aim was to highlight and quantify the roughness in the landscape of these proteins, trying to clarify the intriguing features of their folding mechanism and, in turn, deepening our knowledge of the elementary properties of the protein folding reaction. In order to perform the proposed investigation it is necessary to prepare the proteins by labelling them with fluorescent molecules, namely to covalently attach suitable fluorophores to the polypeptide chain: such fluorophores will be the reporters of protein dimensions and movements (the dynamics). To do so, appropriate amino acid residues (“X”) of the polypeptide chain have to be identified and mutated into Cysteines. Single and double X→Cysteine mutants of R15, R16 and R17 domains were designed, produced, lebelled with the appropriate fluorophore(s) Alexa 488 and Alexa 594 (Invitrogen) and purified. Thermodynamic stability, as well as folding and unfolding kinetics of the double mutants 6-99 and 39-99 of R15, R16 and R17, has been assessed through guanidinium chloride (GdmCl)-induced equilibrium and stopped-flow denaturation experiments, monitoring the fluorophores’ fluorescence signal or the intrinsic triptofan fluorescence, when possible. These same variants and the singly-labelled variants of the same domains have been tested for possible anisotropy of the attached dyes. All of the variants were found to be suitable for the proposed spectroscopic investigation, with exclusion of the doubly-labelled R16 39-99, which was therefore was not included in the following study.

Data: CORDIS, © European Union