Julia M. Goodfellow

The pore dimensions of gramicidin A.

The ion channel forming peptide gramicidin A adopts a number of distinct conformations in different environments. We have developed a new method to analyze and display the pore dimensions of ion channels. The procedure is applied to two x-ray crystal structures of gramicidin that adopt distinct antiparallel double helical dimer conformations and a nuclear magnetic resonance (NMR) structure for the beta6.3 NH2-terminal to NH2-terminal dimer. The results are discussed with reference to ion conductance properties and dependence of pore dimensions on the environment.

Hydrophobic solvation in aqueous trifluoroethanol solution

The titration of an aqueous solution of a de novo designed peptide with trifluoroethanol (TFE) shows complete helix formation with the addition of only 30% TFE. A molecular simulation of the peptide, in which a single shell of TFE molecules initially surrounds the peptide, reveals preferred sites of solvent interaction. The TFE molecules show greater preference for the hydrophobic compared with hydrophilic side chains. The helix-enhancing ability of TFE in aqueous solution may be rationalized in terms of stabilizing the hydrophobic collapse of apolar side chains of the formed helix.

A knowledge-based model of DNA hydration

The aqueous hydration of DNA is an important aspect of its structure, which is of direct relevance to mechanisms of radiation damage. We have made a quantitative analysis of solvent interactions within hydrogen bonding distance of polar atoms of oligonucleotides using 12 B-DNA oligonucleotide crystal structures. The distribution of water molecules around the four bases, the sugar residues and the phosphate groups were generated and analysed both qualitatively and quantitatively. These data have then been used in a knowledge-based method to generate the likely hydration sites around a canonical B-DNA conformation in order to generate models of use in track studies of radiation damage.

AQUARIUS2: knowledge-based modeling of solvent sites around proteins

The program AQUARIUS2 calculates probable positions for water molecules within the first hydration shell of any protein for which atomic coordinates are known. Like its predecessor, AQUARIUS, it uses a knowledge of water molecules sites from crystallographically determined protein structures. Energy calculations are not employed. It differs substantially from the original program in that a 3‐D probability map (for solvent sites) is generated around the surface of the protein instead of the previously used discrete points. The accuracy of the program has been gauged by comparison with experimentally derived water molecule positions for proteins not used in the knowledge base of the program. It has also been tested by combining the probability density maps with crystallographically determined electron density maps for the protein porphobilinogen deaminase. This procedure filters the most likely solvent electron density peaks from the background noise and has been used in the determination of the solvent structure around the protein nerve growth factor.

Unfolding crystallins: The destabilizing role of a β‐hairpin cysteine in βB2‐crystallin by simulation and experiment

The thermodynamic and kinetic stabilities of the eye lens family of βγ‐crystallins are important factors in the etiology of senile cataract. They control the chance of proteins unfolding, which can lead to aggregation and loss of transparency. βB2‐Crystallin orthologs are of low stability and comprise two typical βγ‐crystallin domains, although, uniquely, the N‐terminal domain has a cysteine in one of the conserved folded β‐hairpins. Using high‐temperature (500 K) molecular dynamics simulations with explicit solvent on the N‐terminal domain of rodent βB2‐crystallin, we have identified in silico local flexibility in this folded β‐hairpin. We have shown in vitro using two‐domain human βB2‐crystallin that replacement of this cysteine with a more usual aromatic residue (phenylalanine) results in a gain in conformational stability and a reduction in the rate of unfolding. We have used principal components analysis to visualize and cluster the coordinates from eight separate simulated unfolding trajectories of both the wild‐type and the C50F mutant N‐terminal domains. These data, representing fluctuations around the native well, show that although the mutant and wild‐type appear to behave similarly over the early time period, the wild type appears to explore a different region of conformational space. It is proposed that the advantage of having this low‐stability cysteine may be correlated with a subunit‐exchange mechanism that allows βB2‐crystallin to interact with a range of other β‐crystallin subunits.

Simulated dynamics and biological macromolecules

We present a review of the use of molecular dynamics techniques to study the behaviour of proteins. The application of such methods to biological macromolecules has evolved directly from its use to study simpler physical and chemical systems. We describe the methods typically used in producing multiple nanosecond atomic trajectories. This technique is now so common that it is impossible to review the whole area. Therefore, we have focused on three areas, namely the application to proteins of biomedical importance, to folding of proteins from a random conformation to a stable well-defined tertiary structure and to the reverse process, that of unfolding. Finally, we describe some methods which have been developed to analyse complex trajectories with the aim of defining the most important features of protein dynamics and changes in conformation.

Monte Carlo Studies on Water in the dCpG/Proflavin Crystal Hydrate

The extensive water network identified in the crystallographic studies of the dCpG/Proflavin hydrate by Neidle, Berman and Shieh (Nature 288, 129, 1980) forms an ideal test case for a) assessing the accuracy of theoretical calculations on nucleic acid--water systems based on statistical thermodynamic computer simulation, and b) the possible use of computer simulation in predicting the water positions in crystal hydrates for use in the further refinement and interpretation of diffraction data. Monte Carlo studies have been carried out on water molecules in the unit cell of dCpG/proflavin, with the nucleic acid complex fixed and the condensed phase environment of the system treated by means of periodic boundary conditions. Intermolecular interactions are described by potential functions representative of quantum mechanical calculations developed by Clementi and coworkers, and widely used in recent studies of the aqueous hydration of various forms of DNA fragments. The results are analyzed in terms of hydrogen bond topology, hydrogen bond distances and energies, mean water positions, and water crystal probability density maps. Detailed comparison of calculated and experimentally observed results are given, and the sensitivity of results to choice of potential is determined by comparison with simulation results based on a set of empirical potentials.

Monte Carlo Computer Simulation of Water--Amino Acid Interactions

The sensitivity of computer simulated solvent structures to changes in both non-bonded (Lennard-Jones) coefficients and partial atomic charges has been investigated with use of amino acid hydrate crystals in which the water structure is well defined experimentally. The polarizable electropole (p. e.) model of water has been extended to describe water–protein interactions; thus, the cooperative nature of the hydrogen bond (i. e. non-pair additive effects) is allowed for through a polarizable dipole. By means of Monte Carlo calculations, the predicted water positions were found to be very sensitive to the input parameters used to define both the non-bonded and electrostatic interactions. Root mean square deviations between simulated and X-ray structures were not always adequate to describe these differences and so more detailed comparisons were made. Non-pair additive effects were shown to lead to large changes in water dipoles, the values of which depended specifically on the system under consideration.

What base pairings can occur in DNA? A distributed multipole study of the electrostatic interactions between normal and alkylated nucleic acid bases

Ab initio distributed multipole electrostatic calculations are used to predict likely nucleic acid base pair structures for both the gas phase and within a double helical backbone, as represented by simple constraints. The resulting structures are interpreted by comparison with an analysis of the experimental variation of base pair geometries found in oligonucleotide crystals. Our calculations on all pairs of the normal bases (G, A, T, C) correctly predict all the multiply hydrogen-bonded structures, in agreement with supermolecule SCF calculations, and also predict some new low-energy structures. Consideration of the helical constraints confirms that the Watson–Crick G · C and A · T pairings are most favourable for inclusion in DNA, but certain mismatch base pairs, G · T and G. · A, are also energetically favourable and their geometries correspond to the experimentally observed wobble conformations. This approach is also used to study the effect of the O6 methylation of guanine which can form a doubly hydrogen-bonded Watson–Crick-like structure with thymine. However, there are also a range of O6-methylguanine · cytosine structures which fit into the helical backbone and are energetically competitive. Thus the mutation-inducing effects of this base modification are likely to be very sensitive to the exact sequence and local conformation of the DNA.

Parallelization strategies for molecular simulation using the Monte Carlo algorithm

We describe the development of Metropolis Monte Carlo algorithms for a general network of multiple instruction multiple data (MIMD) parallel processors. The implementation of farm, event, and systolic parallel algorithms on transputer‐based computers is detailed and their relative performance discussed. Although the emphasis is on methodology, the application of such parallel algorithms will be important for addressing computational problems such as the determination of free energy differences in complex biologically important molecular systems.