Cells contain a multitude of different chemical reaction paths running simultaneously and quite independently next to each other. This amazing feat is enabled by molecular recognition, the ability of biomolecules to form stable and specific complexes with each other and with their substrates. A better understanding of this process, i.e. of the kinetics, structures and thermodynamic properties of biomolecule binding, would be invaluable in the study of biological systems. In addition, as the mode of action of many pharmaceuticals is based upon their inhibition or activation of biomolecule targets, predictive models of small molecule receptor binding are very helpful tools in rational drug design. Since the goal here is normally to design a new compound with a high inhibition strength, one of the most important thermodynamic properties is the binding free energy DeltaG(0). The prediction of binding constants has always been one of the major goals in the field of computational chemistry, because the ability to reliably assess a hypothetical compound's binding properties without having to synthesize it first would save a tremendous amount of work. The different approaches to this question range from fast and simple empirical descriptor methods to elaborate simulation protocols aimed at putting the computation of free energies onto a solid foundation of statistical thermodynamics. While the later methods are still not suited for the screenings of thousands of compounds that are routinely performed in computational drug design studies, they are increasingly put to use for the detailed study of protein ligand interactions. This review will focus on molecular mechanics force field based free energy calculations and their application to the study of protein ligand interactions. After a brief overview of other popular methods for the calculation of free energies, we will describe recent advances in methodology and a variety of exemplary studies of molecular dynamics simulation based free energy calculations.

The primary electron donor P700 of photosystem I is a dimer comprised of chlorophyll a (P(B)) and chlorophyll a' (P(A)). P(A) is involved in a hydrogen bond network with several surrounding amino acid residues and a nearby water molecule. To investigate the influence of hydrogen bond interactions on the properties of P700, the threonine at position A739, which donates a putative hydrogen bond to the 13(1)-keto group of P(A), was replaced with valine, histidine, and tyrosine in Chlamydomonas reinhardtii using site-directed mutagenesis. Growth of the mutants was not impaired. (i) The (P700(+)* - P700) FTIR difference spectra of the mutants lack a negative band at 1634 cm(-1) observed in the wild-type spectrum and instead exhibit a new negative band between 1658 and 1672 cm(-1) depending on the mutation. This band can therefore be assigned to the 13(1)-keto group of P(A) which is upshifted to higher frequencies upon removal of the hydrogen bond. (ii) The main bleaching band in the Q(y)() region of the (P700(+)* - P700) and ((3)P700 - P700) absorption difference spectra is blue shifted for the mutants by approximately 6 nm compared to that of the wild type. A blue shift is also observed for the main bleaching in the Soret region. (iii) The (P700(+)* - P700) CD difference spectrum of the wild type reveals two bands at 694 nm (positive CD) and 680 nm (negative CD) of approximately equal area. For each mutant, these two components are blue-shifted to the same extent. The results strongly suggest that a blue shift of the Q(y) absorption band of P(A) is responsible for a blue shift of the exciton bands. (iv) Redox titrations yielded a decrease in the midpoint potential for the oxidation of P700 by 32 mV for the exchange of Thr against Val. (v) ENDOR spectroscopy shows that the hfc of the methyl protons at position 12 of the spin-carrying Chl P(B) is decreased due to the removal of the hydrogen bond to P(A). This indicates a redistribution of spin density in P700(+)* compared to that in the wild type. This gives evidence for an electronic coupling between the two halves of the dimer in the wild type and mutants.

Using a confocal fluorescence microscope with an avalanche photodiode as detector, we studied the fluorescence of the tetramethylrhodamine labeled F I part of the H + -ATPase from Escherichia coli, EF I , carrying the Q QT106-C mutation [Aggeler, J.A. and Capaldi, R.A. (1992) J. Biol. Chem. 267, 21355^21359] in aqueous solution upon excitation with a modelocked argon ion laser at 528 nm. The diffusion of the labeled EF I through the confocal volume gives rise to photon bursts, which were analyzed with fluorescence correlation spectroscopy, resulting in a diffusion coefficient of 3.3U10^U cm P s^I. In the presence of nucleotides the diffusion coefficient increases by about 15%. This effect indicates a change of the shape and/or the volume of the enzyme upon binding of nucleotides, i.e. fluorescence correlation spectroscopy with single EF I molecules allows the detection of conformational changes.z 1998 Federation of European Biochemical Societies.

In this study we show that a combination of different theoretical methods is a viable approach to calculate the binding affinities of new ligands for the human neutrophile elastase. This protease degrades elastin and likely aids neutrophils in fulfilling their immunological functions. Abnormally high human neutrophil elastase (HNE) levels are involved in several diseases; therefore, inhibitors of HNE are of interest as targets for drug design. A recent study has revealed that cinnamic acid and bornyl ester derivatives bind to HNE, but DeltaG0 values from ligand docking results exhibited no correlation with those calculated from the IC50 values. To accurately compute binding affinities, we generated possible protein ligand complex structures by ligand docking calculations. For each of the ligands, the 30 most likely placements were used as starting points of nanosecond length molecular dynamics simulations. The binding free energies for these complex structures were estimated using a continuum solvent (MM-PBSA) approach. These results, along with structural data from the molecular dynamics runs, allowed the identification of a group of similar placements that serve as a model for the natural protein ligand complex structure. This structural model was used to perform thermodynamic integration (TI) calculations to obtain the relative binding free energies of similar ligands to HNE. The TI results were in quantitative agreement with the measured binding affinities. Thus, the presented approach can be used to generate a probable complex structure for known ligands to HNE and to use such a structure to calculate the effects of small ligand modifications on ligand binding, possibly leading to new inhibitors with improved binding affinities.

No abstract

No abstract

In native reaction centers (RCs) from the photosynthetic bacterium Ahodobacter sphaeroides, the charge recombination of D+QAQB-proceeds indirectly via the intermediate state D+QA-QB. We show from a detailed kinetic analysis that, in mutant RCs in which Asp (L213) was replaced by Asn, direct recombination of QBpredominates below pH 8. Between pH 9 and 10, the direct and indirect pathways are about equally effective. We find the charge recombination from QB-(~B D ) to increase from 0.04 s-1 at pH 6.0 up to 0.4 s-1 at pH 10. This change in k g~ arises from a change in energy of the D+QAQB-state due to protonation of amino acid residues. The charge recombination from QB-, kBD, is about 2 orders of magnitude smaller than from QA-AD).We attribute the large difference between k g~ and kAD to a difference in the reorganization energy of the QA and QB sites; this is reasonable, as the environment of the QB site is more polar than that of the QA site. By using the classical Marcus theory of electron transfer, we fit the experimental data with a reorganization energy XBD = 1.23 eV, which is considerably larger than that found for the QA-site in native RCs (XAD = 0.64 eV).The pH profile of kBD in native RCs is also deduced from the Marcus theory.

No abstract