Towards time-resolved structural studies of membrane transporters

Abstract

This work describes progress towards carrying out time-resolved crystallographic studies of membrane transporters, using the bacterial sodium symporters LeuT and Mhp1 as model systems. Specifically, a new approach to reaction initiation was explored, where a disulfide crosslink was used to lock the protein into an inactive state. The disulfide could then be released by UV-laser excitation or by rapid mixing with a reductant solution. The degree of crosslinking was quantified, and a protocol developed to deliver a near 100% crosslinked sample for crystallization. Several spectroscopic methods were then tested to identify a protocol which could be used to monitor the on-set of ligand binding and transporter function upon crosslink cleavage. Finally, first steps towards the optimization of microcrystal growth were taken.

In parallel a number of other studies were made, some of which are also presented in this thesis.

The optimization of large-scale Mhp1 production, in order to allow SAXS studies of Mhp1 conformational changes and oligomerization under different buffer conditions are presented. In addition, Mhp1 datasets from a range of ligand complexes were reanalyzed and the utility of a novel torsion angle based model and ensemble refinement were compared to standard atom-based modelling and refinement.

Preliminary work is presented on enabling time-resolved studies of a second class of membrane protein target, diacylglycerol kinase (DgKa), where reaction initiation will be via decaging of the substrate ATP. This includes first SSX diffraction data collections as well as an exploration of the compatibility of LCP-crystallized samples with laser excitation. Here the effects of light refraction in solid targets as well as laser induced phase changes in the LCP were explored.

As part of a wider programme to develop improved methods for membrane protein crystallization, a study of detergent diffusion rates across dialysis membranes, commonly used for diffusion-based crystallization, was made. This showed that not only were the diffusion rates of different detergents quite distinct, but also that the dialysis membranes strongly absorbed detergents, resulting in a lower actual detergent concentration than expected. This potentially explains some of the challenges in using diffusion-based approaches for membrane protein crystallization.

Finally, the large-scale production of HRV-3C protease and concentrative nucleoside transporter (CNT) proteins are described.