A new home for membrane proteins
A new database that will help researchers understand the structure of proteins in a membrane is announced in the journal Structure.
Membrane proteins in bilayers. There are now over 2000 membrane protein structures embedded within lipid membranes and housed within the MemProtMD database (Click to enlarge)
Phill Stansfeld, with Mark Sansom, Jo Parker and Simon Newstead in the department, and collaborators in Oxford and Dublin, describe their pipeline 'MemProtMD' and its importance to the research community (1, 2).
The new resource is an automated simulation pipeline for predicting the location of a membrane protein structure in a lipid bilayer. It will give researchers greater insight into how a protein functions within its local membrane environment.
As Dr Stansfeld explains, the group was aiming to create a resource like the Protein Data Bank (PDB), specifically for membrane proteins. 'Whilst there have been methods which take a protein structure and insert it into a lipid bilayer, this has only been done manually up to now. With MemProtMD we have automated this process.'
Another novel feature of the tool is that it uses an 'explicit lipid membrane', doing more than simply inserting the protein into an artificial 'slab' membrane representation. MemProtMD determines the dynamic behaviour of protein in the membrane, taking into account how the protein may deform the lipid and where the protein-lipid interactions might be.
The tool is set up to automate the coarse-grained molecular dynamics (CGMD) methodology, Dr Stansfeld explains. Researchers continuously deposit protein structures into the PDB and every Wednesday new structures are released. 'MemProtMD identifies all membrane protein structures upon their release, puts them into the automated pipeline and sets up MD simulations to allow incorporation into the lipid bilayer.'
There are over 2000 membrane protein structures in PDB, with the numbers increasing on a steady basis.
In their paper, Dr Stansfeld and colleagues analysed a number of recently determined membrane protein structures to predict their location within a membrane. One of these was the plant nitrate transporter NRT1.1 that Dr Simon Newstead and colleagues solved in 2014. The group was able to improve the protein's structure in light of information derived from simulations using their new tool.
The MemProtMD Pipeline. Membrane protein structures are automatically downloaded from the PDB. The protein is then submerged in lipids and solvent. CGMD are used to build the membrane around the protein. The entire system is then converted to full atomic detail (Click to enlarge)
By analysing simulations of all known membrane protein structures, the group has also revealed details that will help efforts to model this important group of proteins. They identify new rules for membrane proteins, such as where the amino acid side chains like to sit within the membrane and the protein-induced membrane deformation.
Dr Stansfeld anticipates MemProtMD being a useful, collaborative tool for enhancing experimental studies - crystallography, NMR or cryo-EM - on membrane proteins. 'It will allow us, for example, to take lower resolution structures cryo-EM structures and identify whether residues in a helix are in the wrong place with respect to the membrane. The structural biologists would come to us, ideally with raw structure, and we would work closely to hone in on regions that may be suspect. Together, we would be able to resolve the structure.'
The tool has been made as simple as possible so that researchers can see for themselves how the protein resolves within the bilayer. 'We aim to place simulation coordinates and parameters in the database so that people can download and run their own molecular simulations, using MemProtMD as the starting point,' he comments.
Dr Stansfeld says that in the future they would like to move to more physiological compositions of membrane. 'We currently use a basic lipid bilayer but would ideally like to tailor the membrane to the protein. We could then look at interactions such as cholesterol binding.'
The group is also interested in adapting the methodology to look at another important group of proteins - peripheral proteins. These lie on the peripheral regions of the lipid bilayer and are often anchored via lipids.
With new membrane protein structures continuously feeding into MemProtMD, the tool will help an increasing number of researchers to see their protein structures in a more relevant environment.