Local Water Structures Govern the Mixing Thermodynamics of Glycerol-Water Solutions

Abstract

Glycerol is a major cryoprotective agent and is widely used to promote protein stabilization. Through a combined experimental and theoretical study, we show that global thermodynamic mixing properties of glycerol and water are dictated by local solvation motifs. We identify three water populations, i.e., bulk water, bound water H-bonded to hydrophilic groups of glycerol and wrap water hydrating hydrophobic moieties. Each population provides distinct spectroscopic fingerprints in the THz/FIR spectral range, which allow to quantify their respective abundance and their partial contributions to the mixing enthalpy. We uncover a 1:1 connection between the number of bound waters and the mixing enthalpy, as deduced from experiments as well as from simulations. The balance between local hydrophobic wrap and hydrophilic bound contributions at the molecular level dictates macroscopic thermodynamics of mixing. This offers opportunities to rationally design polyol water mixtures to optimize technological applications by tuning mixing enthalpy and entropy based on spectroscopic screening.

Liquid–Liquid Phase Separation? Ask the Water!

“…where ΔH wrap,bound and ΔS wrap,bound are constant scaling factors (independent of the solute) that define the linear correlation between Δα wrap,bound and hydration entropy and enthalpy. 20,27,31,41 Such linear correlations could be first deduced from measurements on hydrated alcohol chains. 20,27,31 The correlation factors were fitted to reproduce the standard calorimetry data.…”

“…The total hydration free energy of a protein (or any solute) is given by the sum of the free energy terms associated with these two steps: normalΔ G solv = normalΔ G cavity + normalΔ G bound As proposed before, THz-calorimetry is a new method to quantify these partial contributions to solvation entropy and enthalpy directly based on the changes in the measured amplitudes of cavity-wrap and bound spectroscopic populations. ,, By assuming a specific heat capacity for each population that differs from the bulk, it was possible to predict solvation entropy and enthalpy directly from experimental observables, i.e., changes in Δα wrap,bound : , .25ex2ex normalΔ G cavity = normalΔ α wrap normalΔ H̅ wrap − T normalΔ α wrap normalΔ S̅ wrap normalΔ G bound = normalΔ α bound normalΔ H̅ bound − T normalΔ α bound normalΔ S̅ bound where Δ wrap,bound and Δ wrap,bound are constant scaling factors (independent of the solute) that define the linear correlation between Δα wrap,bound and hydration entropy and enthalpy. ,,, Such linear correlations could be first deduced from measurements on hydrated alcohol chains. ,, The correlation factors were fitted to reproduce the standard calorimetry data. This THz-calorimetry approach was shown to provide total solvation entropy and enthalpy in excellent agreement with standard calorimetry ...…”

“…However, based on our previous studies ,, we are able to quantify the bound water contribution ( Δα bound ) from the slope of Δα in the frequency range >400 cm –1 . We previously demonstrated that the slope analysis is an accurate and simple method to quantify the bound water spectroscopic population, which does not require any assumption any particular lineshape (see Figure S5 and refs and for details). We carefully checked , that the choice of the frequency interval over which the slope is evaluated (in the range of water librations where the intensity rise) does not alter the trends observed for the bound population.…”

“…We previously demonstrated that the slope analysis is an accurate and simple method to quantify the bound water spectroscopic population, which does not require any assumption any particular lineshape (see Figure S5 and refs and for details). We carefully checked , that the choice of the frequency interval over which the slope is evaluated (in the range of water librations where the intensity rise) does not alter the trends observed for the bound population. In the diagram, both x -/ y -axes assume values very close to zero in cases where LLPS does not take place, as no significant changes in wrap and bound populations must be expected in the absence of LLPS.…”

“…For the bound ( y -axis), we cover only a part of the librations frequency range; thus, we cannot deduce a lineshape. However, based on our previous studies ,, we are able to quantify the bound water contribution ( Δα bound ) from the slope of Δα in the frequency range >400 cm –1 . We previously demonstrated that the slope analysis is an accurate and simple method to quantify the bound water spectroscopic population, which does not require any assumption any particular lineshape (see Figure S5 and refs and for details).…”