)
We are using both bioinformatics and simulation based approaches to help us understand the relationship between a proteins structure and its dynamical behaviour in relation to function. Detailed, atomic level information on protein internal motions usually either comes from X-ray crystallography, Nuclear Magnetic Resonance experiments, or simulation. In the case of X-ray crystallography, for example, different conformations are observed when an open structure of an enzyme is found in the absence of a bound substrate, or substrate analogue, and the closed structure in the complexed state. In these cases one has detailed information about the internal rearrangements that occur upon substrate binding.
The DynDom Program
If one has two such conformations of a protein then these can be analysed for domain motions. Hayward and Berendsen[1] developed a unique and rigorous methodology based in rigid-body kinematics and implemented in the DynDom program for the specific purpose of analysing movements in proteins in terms of the relative motions of quasi-rigid parts. The figure shows the domain decomposition of citrate synthase in going from an unliganded state to a liganded state. This domain decomposition is based on movement not structure. Further to the domain decomposition the program determines the hinge axis and the hinge-bending regions.
DynDom is available from the DynDom website and the Collaborative Computational Project 4, CCP4 (X-ray crystallography) website
DynDom3D
We have recently extended the basic DynDom methodology to deal with large complex biomolecules. The new program is called "DynDom3D"[3]. DynDom3D can be applied to proteins comprising many individual chains, or even biomolecules that comprise both protein and DNA or RNA. It has been tested on a range of biomolecules including the 70S ribosome.
See the DynDom3D webpage for more information.
The DynDom Database
The DynDom database has two main parts. The first part is a user-created database that stores the results of successful DynDom runs initiated by visitors to the DynDom website [2]. At the website the visitor can select a pair of protein structures (identified by their PDB codes). The results are then inserted into the database which can be browsed and searched using the web interface. The second part is a comprehensive non-redundant database that is a result of the application of a multi-step method applied to the whole of the database of protein structures [4]. For more information please visit the webpage.
Database of ligand-induced domain movements in enzymes
Recently we have refined non-redundant database down to a set of enzymes where the domain movement in triggered by the binding of a functional ligand. This data will be used in order to help us understand the mechanism by which a ligand can cause such large scale movements in enzymes. For access to this database, please visit the enzyme webpage.
Simulation of Domain Motions
Experimental data can only tell us so much. In particular the X-ray data often provides us with only isolated structures. Even if it gives us the liganded and unliganded structures there are many unknowns, e.g. what path is taken between these two structures, or what is the free energy profile along that path? A question that has interested us is whether the closed liganded domain conformation is accessible to the open unliganded enzyme. With colleagues in Japan and Germany we are working to answers these types of questions using molecular dynamics simulation. One of our proteins of interest is liver alcohol dehydrogenase. Extensive simulations on this enzyme have been carried out in collaboration with Dr A. Kitao at the University of Tokyo[6]. A careful analysis of the simulations combined with an analysis of all known X-ray structures has revealed this enzyme's intriguing mechanism. In short it has a latch that locks the domains in an open conformation in the absence of NAD+ the ligand that causes domain closure. When NAD+ binds it the moves the latch to an unlocked position allowing the domains to closed. The enzyme was also found to have a cooperative mechanism operating between its two identical subunits. This finding actually explains the role of the latch as it can prevent the occurrence of an unproductive closed subunit when the other subunit has already closed with NAD+.
Simulations on citrate synthase suggest that for this enzyme also the domains are locked open to receive its substrate and only once the substrate binds does it close. The probable reason for this is to make enzymes more efficient: ones that continually open and close would not allow the ligand access to the binding site when in the closed state.
Amyloid and the alpha-sheet
Amyloid is an alternative form of protein that causes diseases such as Alzheimer's, Parkinson's, Huntington's, Type II diabetes and the prion diseases such as BSE. Mature amyloid forms fibrils but prior to that it exists and a so-called prefibrillar intermediate that is thought to be the main agent of cell damage. Recent simulations the group of Daggett have shown that proteins known to form amyloid under certain conditions can undergo a transition whereby some of their beta-sheet structure forms a little known secondary structure, termed "alpha-sheet". The alpha-sheet structure was predicted by Pauling and Corey in 1951 but is hardly found in native protein structures. The transition from beta-sheet to alpha-sheet was found to occur by one of the 32 different types of peptide-plane flips identified in our group [8]. This particular flip was searched for in the protein data bank structures and was found to occur in a significant proportion of cases, indicating that it can occur quite readily[9].
Although the structure of the beta-sheet is well known the structure of the alpha-sheet has not be described in any detail. For the beta-sheet the individual strands are tightly twisted but straight. We calculated the helical properties of the alpha-sheet strand and found it to be radically different to that of the beta-sheet strand. In contrast to beta-sheet the alpha-sheet strand forms a curved structure[10]. Therefore unlike beta-sheet the strands do not need to coil in order to maintain interstrand hydrogen bonding. This may mean the alpha-sheet can form quite flat sheets.
References
Hayward, S. Berendsen,H.J.C.Systematic Analysis of Domain Motions in Proteins from Conformational Change; New Results on Citrate Synthase and T4 Lysozyme Proteins, Structure, Function and Genetics, 30, 144-154, 1998.
Lee,R.A., Razaz,M., Hayward, S. The DynDom database of protein domain motions Bioinformatics, 19, 10, 2003.
Poornam, G.P., Matsumoto, A., Ishida, H., Hayward, S. A method for the analysis of domain motions in large biomolecular complexes submitted.
Qi, G. , Lee, R. Hayward, S. A comprehensive and non-redundant database of protein domain movements Bioinformatics, 21(12):2832-2838, 2005.
Qi, G., Hayward, S. Database of ligand-induced domain movements in enzymes: The role of interdomain bending regions submitted.
Hayward, S., Kitao A. Molecular Dynamics Simulations of NAD+-induced domain closure in horse liver alcohol dehydrogenase Biophys. J. , 91, 1823-1831, 2006.
Roccatano, D., Mark, A.E., Hayward, S. Investigation of the Mechanism of Domain Closure in Citrate Synthase by Molecular Dynamics Simulation J. Mol. Biol. , 310, 1039-1053, 2001.
Hayward, S. Peptide-plane flipping Prot.Sci., 10,2219-2227, 2001.
Milner-White, E.J., Watson,J.D., Qi, G., Hayward, S. Amyloid formation may involve alpha- to beta-sheet interconversion via peptide plane flipping Structure , 14, 1369-1376, 2006.
Hayward, S., Milner-White, E.J. The geometry of alpha-sheet: Implications for its possible function as amyloid precursor in proteins Protein, Structure, Function and Bioinformatics , 71, 415-425, 2008.
Research Team
Dr. Steven Hayward, Daniel Taylor