Skip to main content
European Commission logo print header

Computational Studies of Proton Dynamics in Hydrogen Bonded Systems and Enzymes

Final Report Summary - COSPRODYN (Computational Studies of Proton Dynamics in Hydrogen Bonded Systems and Enzymes)

The aim of this computational project was to improve the understanding of biocatalysis and receptor triggering by studying the quantum nature of nuclear motions in model hydrogen bonded systems and proteins. The research was centred on neurotransmitters, which are small organic molecules, such as histamine, dopamine and serotonin, that are synthesised in the brain and used for the signal transmission from neurons to target cells.

In the first phase of the project, we were investigating how histamine behaves in the aqueous solution and applied several computational approaches to model its interaction with the water solvent molecules. This is significant, since under physiological conditions most of biologically relevant systems are surrounded by water molecules, and the way the former compounds interact with solvent through a series of hydrogen bonds is analogous to how they are recognised by larger biological systems such as transporters and receptors. Microsolvation analysis at the quantum-mechanical level, employing supramolecule / cluster approach, revealed that at least five explicit water molecules are required for the trans N-H conformer to overcome the larger intrinsic stability of the gauche counterpart, the former being experimentally determined as the most dominant structure of histamine in water. We proceeded with the analysis of the vibrational spectra, since vibrational spectroscopy is among the most powerful techniques for characterising medium-sise molecules in condensed phases. In such medium strength and in strongly hydrogen-bonded systems in particular, the X-H infrared (IR) absorption bands are relatively broad, intense and exhibit complex structure, and their proper assignment is still far from being straightforward. Therefore, in order to improve the overall agreement with the experiment in terms of the position and the shape of the N-H stretching envelopes and to develop an efficient and accurate computational method to calculate IR spectra in aqueous solution, we performed ab initio Car-Parrinello molecular dynamics simulations of histamine placed within a box of explicit solvent molecules and introduced an a posterior quantisation of specific N-H stretching motions by solving the vibrational Schroedinger equation beyond harmonic approximation. In that way, specific solvent-solute interactions as well as long range electrostatic effects were explicitly included in the simulation, together with the nuclear quantum effects due to the low mass of the proton. We obtained excellent agreement with the experimental spectra, measured in collaboration with colleagues at the National Institute of Chemistry in Ljubljana, Slovenia, for both water and D2O solutions. Our calculations clearly demonstrated that the ring amino group absorbs at higher frequencies than the remaining three amino N-H protons of the protonated aminoethyl group implying that the latter group forms stronger hydrogen bonding with the surrounding solvent molecules. In this way, the results of the simulation complemented experiments that cannot distinguish between the two sets of protons. These findings provide important information for the rationalisation of the biological role and function of histamine under physiological conditions. Our new methodology is of general applicability to strongly correlated systems and represents a powerful tool for the computational support to vibrational spectroscopy.

In the second phase of the project, we investigated how neurotransmitter molecules or other biogenic amines are metabolised by monoamine oxidase (MAO) group of flavoenzymes. MAOs catalyse the conversion of amines into the corresponding imines, which are then nonenzymatically hydrolysed to the final carbonyl compounds and the ammonia. Insufficient stimulation of certain parts of the brain by serotonin and dopamine was determined to cause psychiatric conditions such as depression and Parkinson disease, respectively. As such, selective inhibition of MAO became the primary pharmacological strategy to relieve the symptoms of both illnesses. Still, despite tremendous research efforts devoted to MAOs over several decades, neither the catalytic nor the inhibition mechanisms of MAO have been unambiguously established, although nowadays compounds such as rasagiline and selegiline are in practical use as irreversible MAO inhibitors, meaning that they form covalent bond with the enzyme's flavin cofactor. Both of mentioned acetylenic compounds exhibit several adverse effects, thus the design and the synthesis of novel and efficient compounds is highly desirable. Using quantum-chemical methods, we investigated mechanisms of flavin-inhibitor bond formation using seven different forms of the inhibitor, which suggested the anionic mechanism, involving nucleophilic attack of the terminally deprotonated anionic inhibitor onto the flavin N5 atom, as the most feasible, the resulting complex matching the available X-ray structures. In addition, the calculated activation free energies for both rasagiline and selegiline were in an excellent agreement with experimentally determined values. These results were selected to feature a graphic on the cover page of the European Journal of Organic Chemistry. We also performed a careful and consistent study of the catalytic activity of MAO and proposed a new two-step mechanism associated with the activation free energy of 24.4 kcal/mol. Our QM/MM calculations on the empirical valence bond (EVB) level suggested that the enzyme catalysis reduces the activation free energy around 8 - 10 kcal/mol relative to the corresponding gas-phase reaction. We demonstrated that our hydride mechanism is in the agreement with available experiments and provided evidence against both traditional polar nucleophilic and single-electron radical pathways. The calculated pKa values of the potential catalytic residues and the dopamine molecule within the MAO active site, obtained using the semimacroscopic protein dipole / Langevin dipole approach in its linear response approximation version (PDLD / S-LRA) employing the full dimensionality of the protein, are found line with the proposed mechanism. All of these results supply researchers with valuable insight for the mechanistic studies on other flavoenzymes and pave the way for the design of novel antidepressants and antiparkinsonian drugs as transition state analogues.

In the last phase of the project realisation, we investigated the effect of deuteration on binding affinities of histamine H2 receptors towards agonists such as histamine, and antagonists such as tiotidine. Significant changes in the binding and the receptor activation were observed both computationally and in experiments performed at the Faculty of Medicine, University of Ljubljana, Slovenia, which is, to the best of our knowledge, the first such study. Computations involved the construction of H2 receptor binding site and the quantum chemical modelling of the binding free energies with included simplified quantisation of the proton / deuteron motion. Both sets of data convincingly reveal that upon deuteration the affinity towards agonists increases, whereas it decreases for antagonists, but only to a small extent, the relevant binding free energies being changed between 0.5 - 1.0 kcal/mol. We concluded that changes in the differentiation between agonist and antagonist could be rationalised by altered strength of the hydrogen bonding induced by deuteration that is known as the Ubbelohde effect. These results should help understand the nature of the receptor activation, which is one of the central points of molecular pharmacology. Any method that would be able to predict the nature of binding and/or activation is of tremendous importance since this would give rise to the decreased number of (animal) experiments and would, at the same time, direct the design of novel ligands with desired properties.

During project duration, we also investigated features of an enzyme soybean lipoxygenase 1 (SLO 1), which catalyse the dioxygenation of polyunsaturated fatty acids containing cis,cis-1,4-pentadiene fragment. In agreement with previous results found in the literature, our analysis showed that the rate limiting step involves hydrogen atom transfer from the CH bond of an acid to the OH radical facilitated with the Fe2+ ion. This reaction is associated with one of the largest kinetic isotope effects (KIE), kH/kD, being 81 at the room temperature. We performed path integration molecular dynamics simulation (PIMD) on this system and observed no major temperature dependence of the H/D KIE, thus tying in with the situation found in most similar enzymes exhibiting pronounced values of the H/D KIE, the latter giving strong evidence of tunnelling in the rate limiting step.

In summary, we can safely conclude that all of the goals specified in the project proposal were fully realised. This work already appeared in five scientific papers, while two additional manuscripts are submitted for consideration. Both the researcher, Dr Vianello, and the scientist in charge, Dr Mavri, presented the obtained results at several conferences, meetings, workshops, popular events and national Marie Curie InfoDays, and the awareness of the general public towards these achievements was brought about through interviews appearing in daily newspapers of both countries, as specified in the next section. During the project, Dr Vianello acquired the necessary knowledge of the advanced methods of computational (bio)chemistry, which enhanced his competence and skill diversification. The added knowledge gives Dr Vianello the ability to quickly assess a diverse array of novel scientific problems with the promising outlook for industrial and pharmaceutical applications, thus substantially contributing to the European excellence and competitiveness in the field of computational enzymology.

Contact information:

Dr Janez Mavri
Tel: +38-614-760309
Fax: +38-614-760300

Dr Robert Vianello
Tel: +38-514-561117
Fax: +38-514-680084