Protein Dynamics by Studied by Quasi-Elastic Neutron Scattering (QENS)

Introduction

Proteins in the intracellular environment occur in highly concentrated "crowded" aqueous solutions of different macromolecules and salts. Both the molecular crowding and the presence of different salts affect the mobility and in particular the diffusion of the proteins. Salts also affect protein aggregation and induce complex phenomena such as the reentrant condensation, i.e. a solubility which is not monotonously related to the ionic strength. We investigate these issues on simplified model systems by applying neutron spectroscopy and complementary techniques to study globular proteins and salts with different valency in aqueous solution.

Reentrant condensation

Proteins in solution form monodisperse colloidal suspensions. In addition to their biological role, proteins in solution are therefore of fundamental interest in the context of soft matter science. Proteins, however, differ in one important aspect from many simple colloidal systems: The distribution of charges on the surface of a protein is in general inhomogeneous. This inhomogeneous surface charge distribution can in turn be assumed to have a fundamental biological relevance in controlling for instance aggregation phenomena and biological activity such as docking processes. Characteristic of proteins in their native environment is the macromolecular crowding - i.e. relatively large volume fractions being occupied by proteins - and the aqueous solvent containing salt ions. These salt ions are crucial for the understanding of the effective interactions of proteins and the resulting structures as well as indeed the dynamics.
While mean-field based concepts for the description of charge effects in colloidal suspensions have been known for some time, there are surprisingly few quantitative studies of the effective interactions of proteins depending on their charge. Recently, new effects have been found, most notably the reentrant condensation induced by the addition of higher-valent salts [1]. The reentrant condensation is characterised by different regimes defined by the protein volume fraction φ and salt concentration c: For low salt concentrations c<c* with a characteristic concentration C*(φ), the proteins are in solution, whilst for c*<C<c** with a second characteristic concentration c**(φ), a phase separation regime occurs, which re-enters into a solution regime for c>c** (Fig. 1). The reentrant condensation has previously been observed for DNA or polyelectrolytes and a charge inversion theory has been proposed to explain the observations [2]. However, this phenomenon has never been directly observed in a protein system before. Monitoring the effect of the salt type and concentration on protein dynamics therefore promises a deeper understanding of both the salt-protein and protein-protein interactions, including phenomena such as the binding of salts, protein aggregation and re-dissolution. Protein aggregation by itself is a phenomenon which in particular also has to be explored in its dynamic aspects. Controlled aggregation, namely crystallization, is still one of the "holy grails" of protein science.

Experimental results

For this project we apply quasi-elastic neutron scattering using for instance the cold neutron backscattering spectrometer IN16B and the neutron spin-echo spectrometers IN15 and WASP within the Spectroscopy Group at the ILL. In addition, we employ small angle X-ray and neutron scattering, dynamic light scattering and other complementary techniques.

Here, we study static and dynamic aspects of proteins in aqueous solutions containing salts. These address the issues of crowding and of the effects of monovalent salts for the model protein Bovine Serum Albumin (BSA) [3], using a combination of small-angle X-ray scattering (SAXS), and the quasi-elastic neutron scattering (QENS) techniques backscattering (BS), and spin-echo (NSE). SAXS thereby accesses the static structure on protein-protein nearest neighbour distances, whereas BS and NSE provide information on different regimes of the protein diffusion on nanosecond time scales.

From SAXS data we find a qualitative change with rising protein concentration from an uncorrelated to a strongly correlated solution. For weaker charge screening (i.e. less salt) this correlation is found already for lower protein concentrations. We conclude that below a volume fraction of approximately 10% crowding is induced by unscreened charges. In this case the SAXS correlation peak disappears by the addition of NaCl due to the salt screening effect (Fig. 2), i.e. with increasing ionic strength the surface potential decays faster with distance and reduces the long-ranged repulsion between protein molecules. By contrast, above that volume fraction crowding is dominated by the excluded-volume contribution, and the SAXS correlation peak is conserved upon salt addition (data not shown). The conservation of the correlation peak also for higher salt concentrations implies the absence of aggregation in this highly concentrated protein solution.

Neutron backscattering and spin-echo probe different regimes of diffusion due to the different scattering vector ranges accessed by these techniques and different sensitivity to coherent and incoherent scattering. The scattering vectors accessed by spin-echo are approximately commensurate with those accessed by SAXS, whilst backscattering measures at larger vectors corresponding to intramolecular length scales. From the backscattering and spin-echo data we find a continuously changing behaviour of the self-diffusion of the proteins due to the excluded-volume effect. The addition of salt has little or no effect on the apparent diffusion coefficients observed in backscattering (Fig. 3), although charge screening is assumed to change both interaction time and coupling strength. In contrast to backscattering data, we see an increase of diffusion upon addition of salt in neutron spin echo data (Fig. 4), whereas the dependence on protein concentration remains qualitatively the same, i.e., a decrease of apparent diffusion upon increasing protein concentration. In the protein concentration range thus far covered by our experiments, i.e. from approximately 4% to 27% volume fraction, our data are in agreement with a continuous decrease of the apparent diffusion constants with the protein concentration. In contrast to the static data, our dynamic data show no distinct value where crowding due to the excluded-volume contribution sets in.


Macromolecular crowding, caused by a variety of hydrodynamic and direct interactions, is an important feature of cellular environments and has to be studied in further detail. With our study we prepare the ground for future experiments using multivalent salts and the extension to other protein systems. In order to clarify the physical picture of protein solutions, the combination of different experimental techniques recording both static and dynamical information is required. The X-ray and neutron data is complemented by techniques available in the ILL/ESRF soft matter laboratories such as dynamic light scattering.

References

[1] F. Zhang, M. Skoda, R. Jacobs, et al., Phys. Rev. Lett. 101 (2008) 14 .
[2] A. Y.Grosberg, T. T. Nguyen, B. I. Shklovskii, Rev. Modern Phys. 74 (2002) 329.
[3] F. Roosen-Runge, M. Hennig, T. Seydel, F. Zhang, M. W.A. Skoda, S. Zorn, R. Jacobs, M. Maccarini, P. Fouquet, F. Schreiber; BBA - Proteins and Proteomics 1804 (2010) 68.