In parallel to my IT work, I continue to develop and apply new simluation techniques to describe chemical and biochemical systems. The following topics are of current interest:
Early work on this topic included investigation of the structural origins of the observed 'A-states' in IR-spectra of bound MbCO. Combinations of CHARMM molecular dynamic simulations and QM/MM ab initio calculations were used to eliminate a number of proposed mechanisms, but a definitive origin could not be identified. Work on this subject is continuing.
As a later study, we then addressed development of a force field that is able to accurately describe vibrational coupling and hence vibrational relaxation of the CO ligand in photoexcited MbCO. It was demonstrated that the introduction of anharmonic bonded terms into CHARMM to describe the adjacent Fe-C and C-O stretching modes is sufficient to encapsulate anharmonic coupling between the modes, despite the wide separation in their vibrational frequencies.
Finally, I have been contributing to a collaborative project examining the thermal conductivity of the protein myoglobin (Mb). We performed simulations in collaboration with the experimental group of Prof. P. Hamm in order to describe the transfer of energy from excited solvent molecules at the protein surface to a CO probe molecule located in the Xe4 pocket. The model was used to determine both the energy pathway and the mechanism for local energy transfer from the protein to the free CO molecule, allowing validation of the model originally proposed from measured experimental data.
This work focused on determining what terms are necessary in a classical force field to adequately describe the interaction of a pair of bonded atoms with an external electric field. It has been shown that a detailed description of the change in interaction energy between the field and the bonded pair with bond-length during molecular vibration is crucial in describing the effect on vibrational frequency. In addition to a suitably calibrated, anharmonic bonded term, atomic multipoles that fluctuate as a function of bond-length and in some systems a bond-length-dependent polarizability term are necessary to encapsulate the Vibrational Stark Effect found in complex electric fields, such as those experienced by a probe in a protein binding site.
I continue to work alongside Dr. N. Gresh and Prof. J.P. Piquemal on the construction of a general set of parameters for the polarizable 'SIBFA' force field, capable of modeling the unusual chemistry of Pb-complexes. The sterically active lone pair of the central Pb cation in these complexes leads to the formation of so-called 'hemi-directed' geometries, with uneven distribution around the central metal. We have helped to demonstrate that the lone pair is forced into an off-central position by the net electric field generated by asymmetrically distributed surrounding ligands. This polarization of the cation charge density provides the necessary stabilization of the hemi-directed structure. Encapsulation of such electronic effects is difficult within the framework of a classical force field, but the detailed polarization term and short-range corrections included in SIBFA provide the necessary functionality. Following parametrization of Pb-H2O complexes we are now in the process of generalizing the force field for a range of biologically relevant ligands, with the aim of producing a set of parameters that can be used to aid in the development of a therapeutic agent to treat lead-poisoning, and to provide models that help to explain the mechanism of action of Pb2+ in the human body.
In collaboration with the group of Prof. P. Popelier at the University of Manchester, we are continuing to develop the 'Quantum Isostere Database'. This web-based tool, based on the Quatum Chemical Topology (QCT) partitioning method, is designed to extract intuitive chemical descriptors from an ab initio wave function, which are subsequently used to find pairs of fragments that will exhibit similar chemistries in a given lead compound scaffold. Such bioisosteric fragment pairs can be used to tailor different properties of an active compound while leaving the underlying biological activity in tact.