Simulation of Protein Properties; Structure of N-glycans and Conformations of Disordered Peptides; Simulation of Enzyme Mechanisms, and Protein Dynamics, Structures and Properties

Although attachment of an oligosaccharide sidechain (glycosylation) to proteins is the normal state for most extracellular proteins, how such "decoration" affects the structure, dynamics, and functional properties of proteins is largely unknown. Indeed it is largely ignored, as nearly all high-resolution structural and other experimental molecular studies are performed on protein, i.e. without attached glycans. For these reasons, computer simulation offers a particularly rich opportunity to complement the large amount of protein structural data, by studying the properties of proteins reclothed with their sugar chains. Another protein problem intractable to experimental study is definition of the conformational manifold of disordered peptides or peptide segments within folded proteins, and variation of this manifold with environmental conditions such as temperature. We are applying computational simulation to study both these effects in prion protein (PrP). PrP is a very unusual protein involved in diseases such as mad-cow disease (BSE) and CJD in humans. Although the correlation of PrP glycosylation state with disease characteristics is well-known, NMR structures and other experimental molecular studies have all been done on unglycosylated protein. We are characterizing the changes in the conformational flexibility and surface properties of the C-terminal domain of PrP after addition of the normally attached sugar chains using MD simulations and linear-scaling all-QM (MOZYME) calculations. We are also investigating the conformational distributions of the N-terminal repeat region of PrP, which high-resolution NMR structural solution shows is disordered, using MD simulations.

Principal Investigator

Jill Gready
Computational Proteomics and Therapy Design, JCSMR
Australian National University

Project

x04, d55

Co-Investigators

Vladislav Vassiliev
ANU Supercomputer Facility
Australian National University

Peter Cummins
Johannes Zuegg
Computational Proteomics and Therapy Design, JCSMR
Australian National University

RFCD Codes

250106, 250699, 270105, 270199, 250606, 270108, 270304, 270402, 280210, 320305, 320302, 320403

Significant Achievements, Anticipated Outcomes and Future Work

Calculations have been undertaken or are in progress within the two subprojects on PrP:
(a) Electrostatic Potentials of C-terminal Domain. As earlier MD simulations on PrP models with attached sugars and GPI anchor indicated electrostatics appear critical for PrP structure and function, we extended these studies by using a combination of both MD and the linear-scaling semi-empirical QM method, MOZYME. MOZYME offers the unique opportunity to calculate molecular electrostatic potentials (MEPs) of the whole protein at a QM level. An extensive simulation series for models of wild-type and 13 mutants of human PrP was undertaken. The results did not show differences in MEPs which would seem likely to contribute to the disease association of the mutants. However, the main value of the study, so far, is that it showed significant differences between MEPs calculated from Amber (i.e. MM) and AM1 (i.e. semiempirical QM in MOZYME) charges. The full implications of these results are still unclear but, in general, they sound a warning against drawing conclusions about protein-ligand or protein-protein interactions from force-field calculations where electrostatic (i.e. long-range) components dominate.

(b) Simulations and FRET. In a novel combination of simulation and experiment we have undertaken MD simulations to interpret FRET (fluorescence resonance energy transfer) results for, initially, an 11-residue peptide containing a decarepeat of a marsupial PrP. FRET can be used as a spectroscopic ruler to measure distances between two points either within or between molecules; for the PrP peptide the donor chromophore was the natural trypophan (Trp) residue and the acceptor chromophore a dansyl group attached to the N-terminus. To sample effectively the conformational space of the peptide in solution we performed REMD simulations within MOPS, with 24 parallel trajectories at temperatures distributed exponentially from 280 to 630 K for total 6 ns. Analysis of sampled coordinates (1,440,000 sets) showed excellent agreement with both the experimental FRET distance (i.e. macroscopic average) and temperature dependence (5-85 oC). However, the main value of the simulations is in accessing the complete conformational distributions and their dependence on temperature; PCA analysis is being used to analyse in detail the nature of the semi-folded structures apparent from the distance versus abundance plots.

 

Computational Techniques Used

Replica-exchange molecular dynamics (REMD) simulations: This molecular simulation method involves exchanging small amounts of data (simulation temperatures) between randomly chosen pairs of processors, i.e. we do a Monte Carlo simulation in the space spanned by the specified number of processors. The method is covered by Mitsutake A, Sugita Y, Okamoto Y. (2001) "Generalised-ensemble algorithms for molecular simulations of biopolymers." Biopolymers 60, 96-123. It has been implemented into our MOPS program for use with both MM and QM/MM potentials.

Conventional molecular dynamics simulation: standard MD, simulated annealing and analysis programs. Uses Amber suite of programs.

Linear-Scaling Quantum Chemical calculations: local molecular orbital (LMO) implementation within MOZYME module of MOPAC2000, with local modifications.

 

Publications, Awards and External Funding

External Funding and Awards

None.

Publications

J. Zuegg, A. A. Bliznyuk and J. E. Gready, Comparison of electrostatic potential around proteins calculated from Amber and AM1 charges: Application to mutants of prion protein, Mol. Phys. 101, 2003, 2437-2450.
M. Gustiananda, J. R. Liggins, P. L. Cummins and J. E. Gready, Conformation of prion protein repeat peptides probed by FRET measurements and MD simulations, Biophys J., in press.