Investigating Structure and Dynamics of Proteins in Amorphous Phases Using Neutron Scattering

Abstract

In order to increase shelf life and minimize aggregation during storage, many biotherapeutic drugs are formulated and stored as either frozen solutions or lyophilized powders. However, characterizing amorphous solids can be challenging with the commonly available set of biophysical measurements used for proteins in liquid solutions. Therefore, some questions remain regarding the structure of the active pharmaceutical ingredient during freezing and drying of the drug product and the molecular role of excipients. Neutron scattering is a powerful technique to study structure and dynamics of a variety of systems in both solid and liquid phases. Moreover, neutron scattering experiments can generally be correlated with theory and molecular simulations to analyze experimental data. In this article, we focus on the use of neutron techniques to address problems of biotechnological interest. We describe the use of small-angle neutron scattering to study the solution structure of biological molecules and the packing arrangement in amorphous phases, that is, frozen glasses and freeze-dried protein powders. In addition, we discuss the use of neutron spectroscopy to measure the dynamics of glassy systems at different time and length scales. Overall, we expect that the present article will guide and prompt the use of neutron scattering to provide unique insights on many of the outstanding questions in biotechnology.

Keywords: Freeze-dried proteins; Frozen protein solutions; Glasses; Molecular dynamics; Neutron scattering; Protein dynamics; Protein structure.

Fig. 1

X-ray and neutron scattering cross sections and coherent scattering lengths (scattering factors) for different elements. Circles and bars are drawn to scale.

Fig. 2

Schematic of a small-angle neutron scattering experiment.

Fig. 3

A. Form factor P(q) for various shapes. B. Structure factor S(q) for interacting spheres.

Fig. 4

A. Different views of an atomistic monoclonal antibody structure and the corresponding ensembles represented by density plots. B. Scattering profiles are calculated for the ensemble of structures and compared with experimental data. Error bars represent ± 1 standard deviation.

Fig. 5

A. Schematic of a contrast variation experiment in SANS for a protein–DNA complex using different ratios of H₂O/D₂O. Protein–DNA structure was generated from PDB 1LMB  . B. Scattering length densities for various biomolecules were adapted from reference [28] with permission from the author.

Fig. 6

SANS profiles of lysozyme in amorphous phases. A. SANS profiles of lysozyme solutions during the freezing process. B. SANS profiles of frozen lysozyme solutions as a function of temperature. C. Schematic representation of the morphology of lysozyme in frozen solutions. D. SANS profiles of lyophilized lysozyme powders with controlled water content. Error bars represent ± 1 standard deviation.

Fig. 7

SANS profile of sorbitol–lysozyme frozen solutions. A. No sorbitol. B. Sorbitol & high water content. C. Sorbitol & low water content. Water refers to D₂O. Protein to sorbitol ratio is equivalent in samples containing sorbitol. Error bars represent ± 1 standard deviation.

Fig. 8

A. Theoretical intermediate scattering function $I(\widehat{\mathbf{q}},t)$ of non-exchangeable protons calculated directly from a molecular dynamics trajectory of the protein myoglobin in solution and after multiplying the MD result by various resolution functions corresponding to different instrumental resolutions. B. Incoherent dynamic structure factors $S(\widehat{\mathbf{q}},\mathrm{E})$ of the protein α-lactalbumin at selected magnitudes of momentum transfer ($\left|\widehat{\mathbf{q}}\right|=$0.5, 1.0, and 1.5 Å ⁻¹ from bottom to top) measured by quasielastic neutron scattering (boxes) and predicted from MD simulations of the protein in solution (lines).

Fig. 9

A.  <u² > from neutron scattering of binary trehalose glasses as a function of glycerol mass fraction. Inset shows enzyme deactivation times in trehalose glasses with varying glycerol content. Higher deactivation times are observed at the same glycerol fraction at which  <x² > is smallest at most temperatures, as pointed by the black arrows. Error bars represent ± 1 standard deviation. B. Correlation between  <u² > and aggregation and chemical degradation rates of freeze-dried proteins in sugar glasses is from Ref.  . Each label represents a different protein and/or sugar matrix for either aggregation or chemical destabilization.

Fig. 10

A. Snapshot of hydrated ribonuclease A in glycerol and trehalose glass. Protein density is ∼7 mM. B.  <x² > from MD simulations for ribonuclease A and trehalose in a binary glass. A minimum  <x² > is found at a particular mass fraction of glycerol, in agreement with experimental data [115] .

Fig. 11

Origins of protein dynamics in glassy systems. Residence time-correlation functions for A glycerol and B water for a 4 Å water shell from the surface of the protein as a function of glycerol content. C. Average hydrogen bonds between protein and: protein (P-P), water (P-W), trehalose (P-T), and glycerol (P-G). D. Interaction energies for protein–solvent and trehalose–glycerol.