Opendata, web and dolomites

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - RAMP (RAtionalising Membrane Protein crystallisation)

Teaser

Over the past two decades, advances in automation for crystallization of soluble proteins, diffraction data collection and data analysis have made structure solution of soluble proteins almost routine. Crystallization – however – remains a major bottleneck for membrane...

Summary

Over the past two decades, advances in automation for crystallization of soluble proteins, diffraction data collection and data analysis have made structure solution of soluble proteins almost routine. Crystallization – however – remains a major bottleneck for membrane proteins. The additional challenge with membrane proteins is two-fold: first, the protein starting material itself - is hard to produce in large quantities and tends to be unstable when not in a lipid environment. Second, the phase space is even more complex than for soluble proteins by several parameters - the identity of detergent(s), their concentration(s), use of lipid(s) and their type(s) etc. The crystallization process is thus even more complex and poorly understood. Membrane proteins are currently crystallized by using brute-force screening to search this high-dimensional parameter space to find initial conditions, followed by trial-and-error optimisation to grow crystals suitable for diffraction studies. Though over 85% of known drug targets are membrane proteins, fewer than 650 unique membrane protein structures have been determined. We therefore urgently require better methods to crystallize membrane proteins reliably. RAMP has created a unique training network that brings together three strands: (1) the development of a microfluidics-based technology with an unprecedented ability to control membrane protein crystallisation, and the ability to sample the large parameter space of crystallisation conditions rapidly; (2) the introduction to membrane protein crystallisation optimisation via study and modelling of the phase diagram, a technique used with great success for soluble protein crystallisation; and (3) the application of these developments to medically and biologically important membrane protein targets. RAMP also trains the students on two new and emerging techniques that pose challenges for the crystallisation of membrane proteins. While serial crystallographic methods increasingly used at synchrotron sources as well as at rapidly developing ultra-bright free-electron laser sources, require crystals in the 1-20 μm size to solve structure of previously intractable proteins. Neutron protein crystallography requires much larger crystals (> 0.01 mm3) since neutrons interact very weakly with matter. This is, however, the only way to visualise all of the protons in a structure, important information for drug design.

Work performed

In the first strand of the RAMP network, the UGA has developed an easy and inexpensive way to fabricate microchips that cover the whole pipeline from crystal growth to beam eliminating the need for crystal handling prior to the diffraction experiment. The chips have been fully characterized for their mechanical and flow properties as well as for their transparency to X-rays. The optically transparent resin used for the microchip fabrication allows for the monitoring of the crystal growth via optical microscopy. On the other hand, the new flowing reservoir dialysis set-up for the crystallisation bench (Figure 1) has been designed to overcome the problems of large crystal growth of membrane proteins for neutron protein crystallography. Soluble model proteins have been successfully crystallized in both the microchips and the flowing reservoir dialysis set-up. The rational protocols for crystallisation optimisation have been established. In the future, RAMP students will use these crystallisation tools to model the experimental data like nucleation kinetics and solubility in strand 2 to control the size of grown crystals for serial synchrotron and neutron crystallography as well as the target membrane proteins in strand 3. At Surrey University, as a part of strand 2 of the RAMP network, the model for mixing the precipitants in the aforementioned UGA crystallisation bench has been established. The mass transport that forms supersaturation gradients and therefore affects both crystal size and diffraction quality of generated crystals has been investigated. This work feeds back to strand 1, allowing better control of the kinetic pathways induced in the new fluidic tools. Lipids are critical in stabilising membrane proteins during crystallisation. However, neither the in meso nor the HiLiDe method, both using lipid environment in membrane protein crystallisation, have been fully investigated and rationalized, and it therefore remains unclear why they are extremely successful for some targets and totally unsuccessful for others. The work for rationalizing membrane protein crystallisation by the HiLiDe method at Aarhus University has progressed significantly. The HiLiDe method (Figure 2) is based on using high concentrations of lipids and detergents to relipidate the protein prior to crystallisation and a simple and systematic screening of lipid-protein-detergent ratios combined with regular crystallisation screens. A phase diagram has been successfully used to guide for protein crystal optimisation of MhsT with the HiLiDe method. In strand 3 three universities study new structures of membrane transporters. The University of Leeds has progressed with integral membrane pyrophosphatases (mPPases), looking to establish a structural understanding of potassium activation. The structure of the non-potassium dependent P. aerophilum mPPase (PaPPase), when compared with potassium-dependent pyrophosphatases, such as T. maritima (TmPPase) and V. radiata mPPases (VrPPase), highlights potential mechanistic differences. Imperial College London has successfully expressed and purified A. thaliana borate transporter as well as single-point AtBOR1 mutants in order to lock the transporter conformation and improve crystal packing and thus their diffraction quality/resolution. Finally, University Aarhus solved the structure of P-type ATPase SERCA1 at room temperature (RT). The overall differences between the RT structure and the previously published structure collected at cryotemperature (100 K) are small. The promising result is the demonstration of the feasibility to obtain large well diffracting crystals of SERCA represented in Figure 2 that may be suitable for future neutron diffraction experiments.

Final results

RAMP provides interdisciplinary, intersectoral research training to the next generation of membrane biologists, producing leaders who will be able to advance the field academically and be able to make leading contributions for drug discovery research.
Our rational approach for optimising on-chip protein crystallization via chemical composition and temperature control allows to tailor crystal size, number and quality. Combining transparent microfluidics and dialysis provides precise control over the experiment and reversible exploration of the crystallization conditions. We have for the first time shown that in addition to diffusion in solution volume, solute-driven convection was driving mixing, greatly reducing both the mixing time and the concentration gradients during equilibration. Our results in using a phase diagram as a general guide for membrane protein crystallisation are extremely promising, particularly for the membrane protein targets studied by RAMP students in strand 3. Successful mechanistic studies using new structures of membrane proteins have been carried out, mainly on integral membrane pyrophosphatases and ATPase SERCA. For SERCA we have shown it feasible to obtain large well diffracting crystals suitable for neutron crystallography in the future.

Website & more info

More info: https://ramp-itn.eu/.