Final Report Summary - IMPSCORE (Introducing stacking and halogen bonding effects into ligand-target interaction energy calculations)
I. Summary description of the project objectives.
Our primary intention is to improve the accuracy of enthalpy calculation in docking studies including special weak interactions like halogen bond or stacking. High level quantum chemical investigations can provide accurate enthalpy energy which can be compared to experimental data. The experimental data can be found in the literature and determined by isothermal titration calorimetric measurements. So, the enthalpy term in a scoring function should be improved with the help of quantum level calculation results and experimental data. Moreover, the high level quantum chemical analysis enables us to investigate the character of the halogen bond in selected systems.
It is worth to mention that data-mining in large and freely available structure repositories or investigation of the effects of the solvation on the docking results are also part of the project.
II Description of the work performed since the beginning of the project
We have investigated selected systems based on weak interactions like hydrogen or halogen bonds with different level quantum chemical methods. High level DFT methods were applied for energy decomposition analyses (EDA) as well as charge analyses technique. Lower level semi-empirical DFT method (DFTB) was also tested to calculate the interaction energy for selected complexes. Implementation of new analyses technics into the applied program code has been also started as well as new docking target generating method was developed in collaboration with other research groups.
III Description of the main results achieved so far
The project can be divided into independent sub-tasks, all ultimately directed towards an improved docking calculation. These independent directions can be summarized in the following points:
Target preparation for docking studies
Docking calculations regularly use crystallographic structure taken from the protein database. In collaboration with Hungarian research groups we have examined how a crystallographic target can be improved. A molecular dynamics based preparation process was suggested which generates an ensemble of targets. This method was applied and examined for selected receptor model.
Crystal structures from the pdb database (www.pdb.org) were collected which contains fragment size ligand with halogen atom(s). Whenever possible, the complexes in our database were augmented with experimental enthalpy value taken from isothermal titration calorimetric measurements.
High level interaction energy calculations including energy decomposition analyses (EDA) for special model systems.
To clarify the role and character of the halogen bond or dispersion interaction in ligand binding we applied EDA technics and charge analyses method. In the first step, case studies were performed on selected model systems (infinite guanine and xanthine chains, quadruplex structures with combined halogen and hydrogen bonds) and a special quantum effect (cooperativity) was examined. The cooperative effect had been pointed out earlier in hydrogen bonded quadruplex systems. The important question that emerges from these earlier findings is whether such effects may also arise in the corresponding but much less understood halogen-bonded complexes. Moreover, our results also show that the weight of halogen bond in the total interaction is much more significant compared to the hydrogen bond. This latter statement is in contrast to the general belief that hydrogen bond rules the formation of a complex in the presence of both interactions which is an interesting novelty. In paralell with the previous investigations, new model systems were also introduced. The selection of the new systems based on the important protein–ligand interactions, that is, all the possible side chain–ligand connections were considered. The first results of the EDA calculations are in line with the work of Wolters and Bickelhaupt, which means considerable covalent character and charge transfer between halogenated ligand and the selected side chain fragment. Moreover the strength of the interaction energy was also comparable to the interaction energy of the hydrogen bond. In conclusion, we would like to underline that halogenation of drug candidates should not be handled as a simple increase of the hydrophobic character of a ligand compound. The halogen bond should be considered as a real auxiliary option to the hydrogen bond in the field of drug design. This must be taken into account in future forcefield development, as well.
Interaction energy calculations and solvent effects
Several aspects of the solvent effects were investigated in the frame of the project. As the methodological development of the combination of EDA with implicit solvent calculations (COSMO) still in progress, the effect of implicit solvent was examined by cation affinity order investigations in model systems (bilayer quadruplex structures). Our simulation could reproduce the consolidated experimental cation affinity sequence with a combined gas phase and implicit solvent calculation. The results of these simulations were presented at three conferences.
The cooperation with a research group in Hungary related to explicit water investigations also continued. The Hungarian group initiated a new crystal water predicting method (www.mobywat.com) with which the correlation could be improved between calculated interaction energies and experimental enthalpy values according to preliminary results. The verification of the robustness of the method has not finished yet, and the test systems are without halogen bond. However, the inclusion of explicit crystal waters in docking is essential, as the presence or absence of waters in the binding region can drastically affect the biding of the ligand concerning either the pose or the interaction energy.
IV. The expected final results and their potential impact and use (including the socio-economic impact and the wider societal implications of the project so far).
Halogen bond has emerging importance in drug design and the expected outcome of the work is an improved understands of this secondary interaction. Our quantum level investigations proof the similar character and behave of halogen bond to the well-known hydrogen bond. A potential impact of these results is that halogen bonds should no longer be considered as purely electrostatic interactions with some hydrophobic effect but more as a weak covalent chemical bond that augments hydrogen bonds as tuning parameter in the field of drug design. These findings are of direct relevance of the development of better scoring functions which are key to accurate docking studies.
Our primary intention is to improve the accuracy of enthalpy calculation in docking studies including special weak interactions like halogen bond or stacking. High level quantum chemical investigations can provide accurate enthalpy energy which can be compared to experimental data. The experimental data can be found in the literature and determined by isothermal titration calorimetric measurements. So, the enthalpy term in a scoring function should be improved with the help of quantum level calculation results and experimental data. Moreover, the high level quantum chemical analysis enables us to investigate the character of the halogen bond in selected systems.
It is worth to mention that data-mining in large and freely available structure repositories or investigation of the effects of the solvation on the docking results are also part of the project.
II Description of the work performed since the beginning of the project
We have investigated selected systems based on weak interactions like hydrogen or halogen bonds with different level quantum chemical methods. High level DFT methods were applied for energy decomposition analyses (EDA) as well as charge analyses technique. Lower level semi-empirical DFT method (DFTB) was also tested to calculate the interaction energy for selected complexes. Implementation of new analyses technics into the applied program code has been also started as well as new docking target generating method was developed in collaboration with other research groups.
III Description of the main results achieved so far
The project can be divided into independent sub-tasks, all ultimately directed towards an improved docking calculation. These independent directions can be summarized in the following points:
Target preparation for docking studies
Docking calculations regularly use crystallographic structure taken from the protein database. In collaboration with Hungarian research groups we have examined how a crystallographic target can be improved. A molecular dynamics based preparation process was suggested which generates an ensemble of targets. This method was applied and examined for selected receptor model.
Crystal structures from the pdb database (www.pdb.org) were collected which contains fragment size ligand with halogen atom(s). Whenever possible, the complexes in our database were augmented with experimental enthalpy value taken from isothermal titration calorimetric measurements.
High level interaction energy calculations including energy decomposition analyses (EDA) for special model systems.
To clarify the role and character of the halogen bond or dispersion interaction in ligand binding we applied EDA technics and charge analyses method. In the first step, case studies were performed on selected model systems (infinite guanine and xanthine chains, quadruplex structures with combined halogen and hydrogen bonds) and a special quantum effect (cooperativity) was examined. The cooperative effect had been pointed out earlier in hydrogen bonded quadruplex systems. The important question that emerges from these earlier findings is whether such effects may also arise in the corresponding but much less understood halogen-bonded complexes. Moreover, our results also show that the weight of halogen bond in the total interaction is much more significant compared to the hydrogen bond. This latter statement is in contrast to the general belief that hydrogen bond rules the formation of a complex in the presence of both interactions which is an interesting novelty. In paralell with the previous investigations, new model systems were also introduced. The selection of the new systems based on the important protein–ligand interactions, that is, all the possible side chain–ligand connections were considered. The first results of the EDA calculations are in line with the work of Wolters and Bickelhaupt, which means considerable covalent character and charge transfer between halogenated ligand and the selected side chain fragment. Moreover the strength of the interaction energy was also comparable to the interaction energy of the hydrogen bond. In conclusion, we would like to underline that halogenation of drug candidates should not be handled as a simple increase of the hydrophobic character of a ligand compound. The halogen bond should be considered as a real auxiliary option to the hydrogen bond in the field of drug design. This must be taken into account in future forcefield development, as well.
Interaction energy calculations and solvent effects
Several aspects of the solvent effects were investigated in the frame of the project. As the methodological development of the combination of EDA with implicit solvent calculations (COSMO) still in progress, the effect of implicit solvent was examined by cation affinity order investigations in model systems (bilayer quadruplex structures). Our simulation could reproduce the consolidated experimental cation affinity sequence with a combined gas phase and implicit solvent calculation. The results of these simulations were presented at three conferences.
The cooperation with a research group in Hungary related to explicit water investigations also continued. The Hungarian group initiated a new crystal water predicting method (www.mobywat.com) with which the correlation could be improved between calculated interaction energies and experimental enthalpy values according to preliminary results. The verification of the robustness of the method has not finished yet, and the test systems are without halogen bond. However, the inclusion of explicit crystal waters in docking is essential, as the presence or absence of waters in the binding region can drastically affect the biding of the ligand concerning either the pose or the interaction energy.
IV. The expected final results and their potential impact and use (including the socio-economic impact and the wider societal implications of the project so far).
Halogen bond has emerging importance in drug design and the expected outcome of the work is an improved understands of this secondary interaction. Our quantum level investigations proof the similar character and behave of halogen bond to the well-known hydrogen bond. A potential impact of these results is that halogen bonds should no longer be considered as purely electrostatic interactions with some hydrophobic effect but more as a weak covalent chemical bond that augments hydrogen bonds as tuning parameter in the field of drug design. These findings are of direct relevance of the development of better scoring functions which are key to accurate docking studies.