Segmental Conformational Disorder and Dynamics in the Intrinsically Disordered Protein α-Synuclein and Its Chain Length Dependence

Abstract

Conformational ensembles of fully disordered natural polypeptides represent the starting point of protein refolding initiated by transfer to folding conditions. Thus, understanding the transient properties and dimensions of such peptides under folding conditions is a necessary step in the understanding of their subsequent folding behavior. Such ensembles can also undergo alternative folding and form amyloid structures, which are involved in many neurological degenerative diseases. Here, we performed a structural study of this initial state using time-resolved fluorescence resonance energy transfer analysis of a series of eight partially overlapping double-labeled chain segments of the N-terminal and NAC domains of the α-synuclein molecule. The distributions of end-to-end distance and segmental intramolecular diffusion coefficients were simultaneously determined for eight labeled chain segments. We used the coefficient of variation, C_v, as a measure of the conformational heterogeneity (i.e., structural disorder). With the exception of two segments, the C_vs were characteristic of a fully disordered state of the chain. Subtle deviations from this behavior at the segment labeled in the NAC domain and the segment at the N termini reflected subtle conformational bias that might be related to the initiation of transition to amyloid aggregates. The chain length dependence of the mean segmental end-to-end distance followed a power law as predicted by Flory, but the dependence was steeper than previously predicted, probably due to the contribution of the excluded volume effect, which is more dominant for shorter-chain segments. The observed intramolecular diffusion coefficients (< 10 to ∼ 25 Ǻ²/ns) are only an order of magnitude lower than the common diffusion coefficients of low molecular weight probes. This diffusion coefficient increased with chain length, probably due to the cumulative contributions of minor bond rotations along the chain. These results gave us a reference both for characteristics of a natural unfolded polypeptide at the moment of initiation of folding and for detection of possible initiation sites of the amyloid transition.

Introduction

The challenge of studying protein folding is to describe the transition from a fully disordered polypeptide to its native active state. While much is known about various forms of folded and even misfolded states of protein molecules, this is not the case for the starting point of the folding transition. Description of the characteristics of ensembles of unfolded or disordered polypeptides and protein molecules is of interest for three major reasons. First, any description of the folding transition of proteins should include the ensemble of disordered conformers, which is the starting point for the folding transition. In addition, understanding the structural transitions of intrinsically disordered proteins (IDPs) and peptides depends on the characterization of their ensembles in the disordered states.¹ Finally, transient structures in otherwise disordered segments of globular proteins are essential in the transport of proteins across membranes and for their turnover in the cell.² Our understanding of the unfolded state is based primarily on a statistical model, the random-coil model, which was developed largely by Flory³ in the 1950s and 1960s. According to this model, a polypeptide chain is considered to be in the fully disordered conformation when the energy differences between accessible backbone conformers are of the order of kT (where k is Boltzmann's constant and T is the absolute temperature) and there are no energetically preferred conformations.

The “random coil” model for self-avoiding (real) chain molecules, introduced by Flory in 1964, predicts a power law dependence of the RMS chain end-to-end distance, 〈r²〉^1/2, on the number of the chain's statistical elements; 〈r²〉^1/2 is proportional to n^ν, where n is the number of monomers and ν is a scaling factor. The scaling factor in this power law accounts for interactions between the chain and the solvent and between monomers, and ranges from 0.3 for a collapsed state to 0.6 for a fully disordered chain. Early measurements by Tanford provided the experimental support of the random-coil model,⁴ and since then, this model has been considered as the best available approximation of the conformational properties of unfolded polypeptides.

Rose and coworkers challenged the commonly accepted random-coil model of ensembles of denatured and apparently disordered hetero-polypeptide chains. Rose emphasized the fact that chains in which the persistence length exceeds one physical link can be treated as random coil of Kuhn segments, while Flory's isolated pair model is not applicable.⁵ Furthermore, using simulations, Fitzke et al. showed that a polypeptide chain can appear to behave as a random coil even if most of the sequence (92%) is composed of secondary-structure elements.⁶ Consequently, Gaussian statistics of the dimensions of a hetero-polypeptide backbone cannot be necessarily interpreted as an indication for a random mode of local conformations.

The determination of the structural characteristics of the fully disordered state of hetero-polypeptides is a major long-standing technical and conceptual challenge.5, 7, 8, 9, 10 The theoretical models most commonly used to interpret experimental data are also limited. Due to the multiplicity of conformations, only probability distribution functions can provide adequate descriptions of ensembles of conformers in the unfolded state.¹¹ Two closely related measures are commonly used for characterizing the dimensions of disordered polypeptides: the radius of gyration, R_g, and the end-to-end distance. Determination of the radius of gyration, R_g, by analysis of small-angle X-ray scattering (SAXS) experiments is currently the main source of data concerning the dimensions of disordered polypeptide chain.12, 13 While SAXS experiments can yield pair distributions and interdomain distances in folded or partially folded populations of protein molecules, in the case of fully disordered molecules, the results are limited to the radius of gyration. For more detailed characterization of ensembles of disordered proteins, the distributions of intramolecular distances are much more informative. Subdomain structures can be very effective in determining the available conformational space for partially folded protein molecules and may serve as nucleation sites in intermolecular interactions. In order to characterize the conformation of selected chain segments in polypeptide chains, we have developed a time-resolved fluorescence resonance energy transfer (trFRET)-based method for determination of distance distributions of short- or medium-size polypeptide segments within long natural or synthetic hetero-polypeptide molecules.10, 14, 15, 16 The parameters of the segmental end-to-end distance distribution, its mean and width, are sensitive indicators of the degree of conformational disorder of chain segments in ensembles of chain molecules.

The characterization of a novel class of proteins that are disordered under folding conditions17, 18, 19 and are involved in the pathology of several neurodegenerative disorders²⁰ enhanced the interest of researchers in the properties of disordered polypeptide backbone conformations in solution. Synthetic polypeptides were used as models of disordered polypeptide chains under folding conditions, but these did not provide a true simulation of the characteristics of segments of natural hetero-polypeptide at the moment of the initiation of its folding transition upon transfer from denaturing to folding conditions. This could be achieved by studying the chain dimensions of IDPs under physiological (i.e., folding) conditions.21, 22, 23 Here, we used fluorescently labeled mutants of human α-synuclein (αS) as a model. We wished to determine to what extent the backbone of the αS molecule is disordered in its native state. To address this question, we applied trFRET experiments and determined the distributions of intramolecular segmental end-to-end distance probabilities.16, 24

αS is a major component in the proteinaceous aggregates in motor neurons of the substantia nigra and locus coeruleus, the so-called Lewy bodies, which are characteristic of idiopathic Parkinson's disease (PD).²⁵ However, the cause of PD is as yet unclear, due in part to a complex etiology involving a combination of genetic susceptibility and numerous environmental factors.26, 27 Three mutations in αS (A30P, E46K, and A53T) have been identified in the autosomal dominantly inherited early-onset form of PD.²⁸ The 140 amino acid residues that form the backbone of the αS monomer are distributed in three domains²⁹: (i) the N-terminal section (residues 1–60, Fig. 1), which includes five imperfect repeats with the conserved hexamer motif KTKEGV (the three familial mutations that accelerate αS oligomerization have been assigned to this domain³⁰); (ii) the hydrophobic NAC domain (residues 61–95; Fig. 1); and (iii) the highly acidic C-terminal region (residues 96–140). The protein undergoes spontaneous conformational transitions and forms amyloid fibrils at an accelerated rate under conditions such as increased temperature, low pH,³¹ high salt concentration,³² and in the presence of osmolytes³³ and naturally occurring amines.³⁴ Although the biological function of αS is not clear, several specific interactions of this protein were established.35, 36, 37 Several recent experiments support a cytotoxic role in PD for protofibrils, early oligomers of αS.²⁷ Thus, we wished to characterize the structure of the soluble unstructured αS monomer, and to define in structural terms the mechanism of transition to the toxic oligomeric form. A study of the potentiation of aggregation by point mutations, ligand binding, or changes in solution conditions might suggest approaches to slow down the initiation of aggregation and its toxic outcome.

Earlier studies of αS by trFRET methods yielded a population distribution but did not account for the high backbone dynamics of αS,³⁸ an effect that is known to distort the results.39, 40 In the present study, we simultaneously determined the parameters describing the distributions of intramolecular distances and intramolecular diffusion coefficients of segment ends by application of trFRET measurements of double-labeled αS mutants.

These experiments give us an opportunity to study the characteristics of an unfolded protein under native conditions, to test models proposed for the dependence of the dimensions of polypeptides in the disordered state on chain length, and to study elementary rates of conformational dynamics that set the limit for the rate of the initial steps of protein folding.

An additional goal of this study was to distinguish between disordered segments and segments that are biased towards a distinct conformation; such segments may lead to the initiation of aggregation that is assumed to be involved in toxic effects in the brain.⁴¹

We characterized the monomeric state of the molecule and evaluated the degree of conformational heterogeneity by calculation of the coefficient of variance, C_v, which is defined by the ratio of the standard deviation to the mean of the distributions. In addition, we determined the chain length dependence of mean end-to-end distances. We found that in most of the N-terminal domain of the αS molecule, the chain segments are disordered and are characterized by wide distributions and fast fluctuations of the end-to-end distances of the segments. An exception to this observation is found in the N terminus (residues 4–18), and the central NAC domain (residues 66–90) which have a narrower distribution, with reduced segment end intramolecular diffusion coefficient relative to the inner N-terminal segments. This subtle bias of the ensemble towards a nonrandom segmental structure might be crucial for understanding the initiation of the oligomerization transition of the αS molecule, which is of great interest for Parkinson's disease research.