NMR relaxation studies on the hydrate layer of intrinsically unstructured proteins

Abstract

Intrinsically unstructured/disordered proteins (IUPs) exist in a disordered and largely solvent-exposed, still functional, structural state under physiological conditions. As their function is often directly linked with structural disorder, understanding their structure-function relationship in detail is a great challenge to structural biology. In particular, their hydration and residual structure, both closely linked with their mechanism of action, require close attention. Here we demonstrate that the hydration of IUPs can be adequately approached by a technique so far unexplored with respect to IUPs, solid-state NMR relaxation measurements. This technique provides quantitative information on various features of hydrate water bound to these proteins. By freezing nonhydrate (bulk) water out, we have been able to measure free induction decays pertaining to protons of bound water from which the amount of hydrate water, its activation energy, and correlation times could be calculated. Thus, for three IUPs, the first inhibitory domain of calpastatin, microtubule-associated protein 2c, and plant dehydrin early responsive to dehydration 10, we demonstrate that they bind a significantly larger amount of water than globular proteins, whereas their suboptimal hydration and relaxation parameters are correlated with their differing modes of function. The theoretical treatment and experimental approach presented in this article may have general utility in characterizing proteins that belong to this novel structural class.

FIGURE 1

Illustration of the method applied to measure the fraction of the unfrozen water component x_unfrozen (MAP2c solution, T = −10.4°C). The slowly decaying part of the time-domain FID signal was extrapolated to t = 0 by applying Lorentzian approximation (dashed line). The extrapolated signal intensity was then compared to the signal intensity measured above 0°C when the whole sample is in liquid state, to get the x_unfrozen value. The inset shows the typical spread in time of FID signals produced by ice protons, protein protons, and unfrozen water protons.

FIGURE 2

Typical magnetization versus pulse spacing curve of π-t-π/2 experiments (inversion-recovery method). Spin-lattice relaxation rate R₁ was obtained by fitting the equation M₀ − M(t) = 2 M₀ × exp(−t × R₁) (line) to the experimental data (circles).

FIGURE 3

¹H spin-lattice relaxation rate (circles) and unfrozen water fraction (squares) in MAP2c solution (50 mg/ml) at ω₀/2π = 44.14 MHz. (Solid line) Redfield-Slichter model fitted to R₁ data; dotted lines are guides to the eye. SEs are represented by the size of the symbols.

FIGURE 4

¹H spin-lattice relaxation rate (circles) and unfrozen water fraction (squares) in CSD1 solution (50 mg/ml) at ω₀/2π = 44.14 MHz. (Solid line) Redfield-Slichter model fitted to R₁ data; dotted lines are guides to the eye. SEs are represented by the size of the symbols.

FIGURE 5

¹H spin-lattice relaxation rate (circles) and unfrozen water fraction (squares) in ERD10 solution (25 mg/ml) at ω₀/2π = 82.57 MHz. (Solid line) Redfield-Slichter model fitted to R₁ data; dotted lines are guides to the eye. SEs are represented by the size of the symbols.

FIGURE 6

¹H spin-lattice relaxation rate (circles) and unfrozen water fraction (squares) in BSA solution (50 mg/ml) at ω₀/2π = 44.14 MHz. (Solid line) Redfield-Slichter model fitted to R₁ data; dotted lines are guides to the eye. SEs are represented by the size of the symbols.