Computational Biology – STFC Computational Science and Engineering Department

Computational Biology Group

Computational Biology is a new theme of the Computational Science and Engineering Department. It builds on considerable biological work done elsewhere in the department, in particular in the Computational Chemistry and CCP4 groups. It also exploits links with experimental groups in STFC.

The transition
mode of the enzyme PHBH modelled with QM/MM

Dr Martyn Winn

Current projects

Atomistic MD simulations. MD simulations help us understand the dynamical properties of biological systems at the molecular, fully atomistic level. Projects we are currently involved in are described in the following.

We investigate membrane receptors, e.g. the epidermal growth factor receptor (EGFR) and ErbB family. ErbB receptors transduct signals from the cell surface through the membrane to the cytosolic tyrosine kinase where the signal is propagated in a cascade further down to launch certain cell processes. The ErbB receptor family is involved in growth, differentiation, and apoptosis of the cell. This signalling network is implicated in the development of most human cancers and hence is a prime target of anti–cancer molecular therapeutic agents (see Multi-million funded experiments could lead to revolution in cancer treatment).
We use modelling and MD simulations to build and investigate models to explain FRET measurements taken by Marisa Martin-Fernandez' collaborating group.

CHO cells with cytosolic GFP and extracellular EGF-Alexa546

The glutamine binding protein (GlnBP) is involved in the first stage of transporting glutamine (nutrient uptake) across the inner membrane of E. coli. GlnBP is the periplasmic – though not covalently bound – component of the corresponding ABC (ATP binding cassette) transport system. Upon ligand binding the 2–domain protein connected by a common hinge region undergoes an extreme conformational change. We are interested in the collective motions of GlnBP that are linked to its biological function. We also want to elucidate the closing mechanism of this protein. This project has been carried out in collaboration with Prof. Akio Kitao, Laboratory of Molecular Design, the University of Tokyo.

Coarse–Grained MD simulations. Whereas atomistic simulations can today reasonably be done for several hundred thousands to a few million atoms on the time scale of tens to hundreds of nano seconds, larger systems on a longer time scale are obviously still out of reach. To address this problem we apply CG simulations complementary to atomistic simulations. At the moment we are looking into the recent MARTINI force field for lipids and proteins and assessing its value for membrane–receptor systems.

Using ab initio models in experimental structure determination. Molecular Replacement (MR) is one of the key methods available for determining protein structure from X–ray diffraction data. The method uses a trial structure to provide initial estimates of the phases. Historically, ab initio models have been too inaccurate to give suitable phase estimates, and the method uses instead experimentally determined structures if these are available. The past few years have seen the progressive maturation of ab initio modelling, and we have been investigating the use of such models in MR to see if they are now reaching the required accuracy. While their use is clearly limited to certain sizes and classes of proteins, we have achieved some successes. (Read more ...)

Software

We have extended the functionality of the ptraj utility (AMBER9/AmberTools 1.x) to include a few handy routines like computation of mean residence times and our own diffusion code. (more info)

We have a few Tcl jiffies to help analyse the geometries of proteins and lipids, for use with VMD. (more info)

We have prepared a CHARMM script to do MM-PBSA calculations and also provide results from a sample simulation to play with. (more info)

Recent publication highlights

J. Kaestner and P. Sherwood, Mol. Phys., 108 293–306 (2010)
– "The ribosome catalyzes peptide bond formation by providing high ionic strength"

Hannes H. Loeffler and Martyn D. Winn, DL Technical Reports, DL-TR-2009-002 (2009)
– "Large Biomolecular Simulation on HPC Platforms I. Experiences with AMBER, Gromacs and NAMD"

Hannes H. Loeffler and Akio Kitao, Biophys. J., 97 2541–2549 (2009)
– "Collective Dynamics of Periplasmic Glutamine Binding Protein upon Domain Closure" (Featured Article)
http://dx.doi.org/10.1016/j.bpj.2009.08.019

Johannes Kaestner, Hannes H. Loeffler, Selene K. Roberts, Marisa L. Martin-Fernandez and Martyn D. Winn J. Struct. Biol., 167 117–128 (2009)
– "Ectodomain orientation, conformational plasticity and oligomerization of ErbB1 receptors investigated by molecular dynamics"
http://dx.doi.org/10.1016/j.jsb.2009.04.007

D.J Rigden, R.M Keegan and M.D Winn, Acta Cryst. D64 1288–1291 (2008)
– "Molecular Replacement using ab initio polyalanine models generated with ROSETTA"
http://journals.iucr.org/d/issues/2008/12/00/fw5187/fw5187bdy.html

Johannes Kaestner and Paul Sherwood, J. Chem. Phys. 128, 014106 (2008)
– "Superlinearly converging dimer method for transition state search"
http://dx.doi.org/10.1063/1.2815812

Stephen E.D. Webb, Selene K. Roberts, Sarah R. Needham, Christopher J. Tynan, Daniel J. Rolfe, Martyn D. Winn, David T. Clarke, Roger Barraclough, and Marisa L. Martin-Fernandez, Biophys. J. 94, 803–819 (2008)
– "Single molecule imaging and FLIM show different structures for high and low-affinity EGFRs in A431 cells"
http://dx.doi.org/10.1529/biophysj.107.112623

R.M.Keegan and M.D.Winn, Acta Cryst. D63, 447–457 (2007)
– "Automated search-model discovery and preparation for structure solution by molecular replacement"
http://dx.doi.org/10.1107/S0907444907002661

J Kaestner, S Thiel, HM Senn, P Sherwood, W Thiel, J. Chem. Theory Comput. 3, 1064–1072 (2007)
– "Exploiting QM/MM Capabilities in Geometry Optimization: A Microiterative Approach Using Electrostatic Embedding"
http://dx.doi.org/10.1021/ct600346p