Hydrogen bonding and drug designs - Term Paper Example

? Hydrogen Bonding & Drug Solubility Predictions – A Review November 15, Prediction of the solubility of a drug is vital for efficient drug design. It is essential for a drug molecule to be water-soluble, in order to be able to pass through biological membranes.1 While less soluble drugs are difficult to be absorbed by the body, highly soluble drugs are quickly absorbed distributed. Accurate prediction of the solubility of a compound during drug screening may immensely help the pharmacological industry by assisting in the selection of the right drug candidate that is sufficiently soluble and bioavailable.2 Thus, a number of computational and theoretical models have been proposed to predict the solubility of a compound. The extent of solubility of a solute depends on its hydrophilic nature. A molecule is more soluble if it can form numerous hydrogen bonds with its water solvent.1 Hydrophobic interactions between hydrophobic regions of the binding site of a ligand and its lipophilic surfaces are important predictors of its interactions in a biological system, which may affect water solubility to such an extent that may even result in unfavorable pharmacokinetic properties.2 It is thus evident that the number of hydrogen bonds made by a drug is an important determinant of its solubility, and ultimately, its bioavailability and effectiveness in a living system. This paper reviews three studies on the hydrogen bonding properties of drugs in an attempt to investigate the utility of hydrogen bonding studies in drug solubility prediction and drug design. Methodology The aim of this paper is to investigate the effects of hydrogen bonding on the molecular properties of drugs dissolved in water. Additionally, other aspects of the role of hydrogen bonding in drug design will also be dealt with. For the purpose of this review, studies relevant to the aims of this paper were searched via Google Scholar. Three most relevant studies were chosen based on the correlation of their subject of study and the aims of this paper. Results The first study assessed the impact of hydrogen bonding on molecular properties, namely Infrared (IR) and Raman spectra, of three drug molecules (caffeine, aspirin, and ibuprofen) dissolved in water.3 The second study investigated various computational and experimental models for the prediction of aqueous drug solubility.4 The third study presented a theoretical calculation of the strength of hydrogen bonding of drug molecules. The findings of these studies are systematically reviewed here. As the first study3 investigated changes in spectra upon dissolution of drugs in water, a frequency calculation was done for three drug molecules (caffeine, aspirin, and ibuprofen) in both gas phase and aqueous solution. IR and Raman spectra of the molecules were analyzed to investigate the influence of both long and short-range interactions on spectral intensity and vibrational frequency. The investigated frequencies were divided into three regions, namely – high, middle and low frequencies. Four different calculations were carried out. The first two calculations began with gas phase geometry optimization followed by calculation of IR and Raman frequencies in the gas phase (GS) and in PCM (Polarized continuum model) using water as solvent. The last two calculations were also done in two stages – geometry optimization with the water molecules located where they are assumed to form hydrogen bonds, followed by calculation of IR and Raman frequencies with and without polarized dielectric continuum. The structures of the three drug molecules with hydrogen-bonded water are shown in figures 1, 2 & 3. Figure 1 — Structure of aspirin hydrogen bonded to water molecules Figure 2 — Structure of caffeine hydrogen bonded to water molecules Figure 3 — Structure of Ibuprofen hydrogen bonded to water molecules In case of caffeine, there was a very small IR absorbance in the high frequency region except for peaks of higher absorbance indicating presence of OH bonds in aqueous solution. In the high frequency region, several peaks were observed for Raman spectra but these were non-water related. The middle region gave most of the observable peaks for IR spectra but the intensities for Raman spectra in the middle region were fairly small when compared to those observed in the high frequency region. The frequency was slightly decreased with and without PCM. The calculated frequencies were in agreement with experimental values given by Ohnsmann et al3. Figures 4, 5, 6 and 7 show a comparison of the calculated and experimental values of caffeine IR and Raman spectra at high and middle frequencies. Similar calculations and observations were made for aspirin and ibuprofen and the calculated values were found to be in good agreement with experimental values. Figure 4 — Caffeine IR spectra at high frequencies Figure 5 — Caffeine IR spectra at middle frequencies (Calculated values are shown in solid lines and experimental values are shown in dotted lines). There is a fair degree of agreement between calculated and experimental values. Figure 6 — Caffeine Raman spectra at high frequencies (Calculated values are shown in solid lines and experimental values are shown in dotted lines). There is a fair degree of agreement between calculated and experimental values. Figure 7 — Caffeine Raman spectra at middle frequencies (Calculated values are shown in solid lines and experimental values are shown in dotted lines). There is a fair degree of agreement between calculated and experimental values. The objective of the second study4 reviewed in this paper was to devise computational and experimental models for the prediction of drug solubility in aqueous solution. Solubility studies were performed using the shake flask method. For the development of a computational model, this study analyzed several molecular descriptors including lipophilicity, number of hydrogen bond donors and acceptors, total hydrogen bonds, molecular surface area, polar surface area, non-polar surface area, etc. While PLS (Partial least squares projection to latent structures) analysis showed that solubility is negatively correlated with lipophilicity, size and presence of non-polar atoms, it found only one hydrogen bond descriptor, i.e. surface area of nitrogen-bound hydrogen atoms as an important predictor of solubility. An important finding of the PLS analysis of this study is that hydrogen bond descriptors do not lead to improvements in solubility prediction models. This finding contradicts several other studies including those by Jorgensen & Duffy and Abraham & Le.4 The third and final study5 of this review deals with a computational model for the prediction of the strength of hydrogen bonding for different donors and acceptors. The study was based on DFT calculation of the interaction energy between donors and acceptors of hydrogen bonds. The proposed model accurately predicted the hydrogen bonding energies as there was a linear correlation between calculated and experimental values. Discussion Several aqueous solubility models based on hydrogen bonding as the main descriptor have been proposed until date.1 The role of hydrogen bonding as an important predictor of solubility has been stressed in many studies.2 The first study reviewed here examined the additive behavior of hydrogen bonding groups in drug molecules and reveals that additive model is predictable to a certain extent.1 The additive model is relevant in cases where the hydrogen bonding groups are isolated, while in the case of ?-conjugated hydrogen bonds, the additive model induces errors in solubility prediction. Most importantly, the study reveals that solvation effect influences IR and Raman spectra. Both IR and Raman spectra are important tools for identification and quantitative analysis of drug molecules.3 By characterizing the changes in IR and Raman spectra when the drug molecules are dissolved in water, this study sets the stage for the development of theoretical models for examining changes in the "bonding environments of drug molecules".3 This would be helpful in investigating the relationships between structures and properties of drug molecules.3 This would also assist in correlating the changes in spectral properties resulting from solvation or binding of the molecule to the type of intermolecular interactions that occur in the aqueous phase.3 While hydrogen bonding interactions have been accepted as important predictors of drug solubility, the results obtained by Bergstrom et al4 indicate otherwise. Their findings suggest that hydrogen-bonding descriptors are not significant in solubility prediction models based on PLS analysis. However, this finding contradicts with earlier findings and so, a more thorough investigation is warranted. It is also suggested that investigations on larger and more diverse drug molecules will have to be made in order to obtain more generalizable findings.4 While this study attempted to review and investigate the importance of hydrogen bonding as an important predictor of drug solubility, it suffers from several limitations. The number of studies reviewed is small and the review itself is narrow. A more expansive review will have to be attempted to encompass other factors that are influenced by hydrogen bonding upon dissolving of drug molecules in aqueous solutions. Conclusions The interaction between drug molecules and the surrounding media is a "key research problem" in drug design, and such interactions are difficult to predict based on theoretical models because very small free energies and complex kinetics are involved.3 The use of hydrogen bonding interactions, and computational and theoretical models for the calculation of hydrogen bond strengths can however be an important tool for the prediction of drug solubility, an important aspect of drug research. Moreover, computational models that predict the strength of hydrogen bonding for different donors and acceptors may be helpful for the evaluation of the impact of inhibitor binding geometry and steric hindrance on the strength of hydrogen bonding in drug design.5 Theoretical modeling based on changes in IR and Raman spectra may prove to be an important tool for understanding bonding environments of drug molecules in aqueous solutions. It is essential to accurately predict the solubility of a particular drug in water, as drugs with low solubility result in low drug concentrations in the body, leading to very low drug absorption that renders the drug useless.4 As the number of candidate drugs for screening is very high, experimental calculation of solubility places huge pressure on resources and time. Therefore, the development of theoretical and computational models for the accurate prediction of drug solubility is an immediate requirement in drug research. The reviewed studies take a step further in this direction. Bibliography 1. Laban Bondesson (2006). Microscopic Views of Drug Solubility. Theoretical Chemistry, Royal Institute of Technology. 2. Hugo Kubinyi (2007). Hydrogen Bonding: The Last Mystery in Drug Design? In Pharmacokinetic Optimization in Drug Research: Biological, Physicochemical, and Computational Strategies. Zurich: Verlag Helvetica Chimica Acta. 3. Laban Bondesson, Kurt V. Mikkelsen, Yi Luo, Per Garberg, & Hans Agren (2007). Hydrogen bonding effects on infrared and Raman spectra of drug molecules. Spectrochimica Acta Part A 66, 213–224. 4. Christel A. S. Bergstrom, Ulf Norinder, Kristina Luthman, & Per Artursson (2002). Experimental and Computational Screening Models for Prediction of Aqueous Drug Solubility. Pharmaceutical Research 19, 182-188. 5. Ming-Hong Hao (2006). Theoretical Calculation of Hydrogen-Bonding Strength for Drug Molecules. Journal of Chemical Theory and Computation 2, 863-872. Read More