Research
Jump to Research NavigationMolecular and Solid State Modelling Group
The IPI Molecular and Solid State Modelling Group develops, validates and applies computational chemistry and molecular modelling approaches to predict the behaviour, structure and properties of pharmaceutical molecules and crystals.
Currently, the Group consists of five members, from left to right in the photograph:
Themes
Computational chemistry and molecular modelling can be valuable tools in pharmaceutical development. Potential applications include:
Overview
Most fine chemicals are crystallised at some stage during their manufacturing process, even if the final formulation does not contain solid particles, crystallisation is a large scale and cost effective means of purification and separation. Most drugs are administered in crystalline form, formulated in tablets, capsules or even inhalers. In order to optimise a crystallisation process or to obtain a product with optimal properties, it is imperative to gain detailed knowledge of solid state structures and properties, not only in the pharmaceutical industry, but also in the production of other fine chemicals such as pigments, explosives, agrochemicals, non-linear optical materials and foodstuffs.
Key physicochemical properties depend on how the molecules pack together in the solid state, including hardness, compressibility, stability, dissolution rate, solubility, colour, morphology and surface chemistry. Many solid state properties can be predicted from the crystal structure, thus allowing the rationalisation of crystallisation problems and design of optimal products. There are also compelling business reasons to determine the atomic structure of crystals as this knowledge allows for maximum protection in patenting and registration of new entities.
Polymorphism, which is the ability of a compound to crystallise in more than one distinct crystal structure, complicates matters further. Some organic compounds exhibit only one stable polymorph, but many have more than one stable form. Solid state properties depend on crystal structure and can therefore vary among polymorphs. There are many examples where polymorphism has a dramatic effect on the performance of a product. Chocolate, for instance, can be found in many polymorphic forms, only one of which has the optimal melting point for human consumption; other polymorphs would not melt in the mouth.
Computational chemistry and molecular modelling tools can be used to gain structural information on molecules and how they form crystals, in particular conformational analysis of isolated molecules followed by accurate quantum mechanical calculations and crystal structure prediction simulations.
Resources
The IPI Molecular and Solid State Modelling Group is based in a dedicated computational lab. Computational resources include a 48 node Beowulf Linux cluster, with each node comprising dual 2.8 GHz Intel® Xeon processors and 2 GB of memory. All nodes communicate via Gigabit Ethernet. In addition, 12 nodes are linked by a faster, Myrinet®, communication channel. Visualisation resources include several high-end graphic workstations and a large screen three-dimensional visualisation system. A variety of software packages and databases are used in the various research projects, including commercial, academic and in-house software.
Most bioactive molecules are flexible, and detailed knowledge of the most stable molecular conformations is essential for many of the simulations described below. A variety of techniques can be applied; including molecular mechanics, semi-empirical quantum mechanics, ab initio quantum mechanics, genetic algorithms, Monte Carlo techniques, grid scan approaches and molecular dynamics. Although the topic of conformational analysis might sound boring to some, it can be a devilishly complex problem if a molecule has more than four or five flexible torsions or a flexible ring system! The figure shows a contour map of energy as a function of two flexible torsions in the ephedrine molecule.
The application of Quantum Mechanics to aid the interpretation of analytical spectra has become a routine requirement in Infrared, Raman and NMR spectroscopies. Calculations of the intensities and frequencies of transitions provide a useful insight into the molecular origin of the transitions observed experimentally. Recently, with the advent of suitable radiation sources, the field of Terahertz spectroscopy has become extremely active and solid state density functional calculations of intensities and frequencies of the low frequency vibration have been a useful method for interpreting the nature of phonons in organic crystals. The IPI Molecular and Solid State Modelling Group is at the forefront of this exciting new development. The figure shows the simulated Terahertz spectrum for urea.
The manner in which molecules pack together in the solid state determines the external shape of the crystalline particle. Particle shape, or morphology, can be a crucial factor in the performance of a fine chemical, such as the colour of a pigment or the ease with which a pharmaceutical formulation can be pressed into a tablet. In chemical engineering, plate-like and needle-shaped crystals are undesirable because the former block filters and the latter can clog up pipes. Molecular modelling is a powerful tool to predict crystal morphology. Computational approaches range from the simple geometric Bravais-Friedel-Donnay-Harker approach to the more complex Hartman-Perdok method and attachment energy calculations. The figure shows the predicted morphology of form I acetaminophen (paracetamol).
Control of crystal morphology and a number of other solid state properties may be achieved by tailor made additives and / or solvent selection. For instance, needle shaped crystals can be made more isometric by selecting additive or solvent molecules that interact strongly with the fast growing faces of the crystalline particle and less strongly with the slow growing faces. Once the crystal structure of a compound is known, and its morphology determined or predicted, molecular modelling tools can be used to analyse surface chemistry and to study the interaction of solvents, additives and impurities. The same approach can be used to identify cleavage planes in solids, which play an important role in pharmaceutical formulation. The figure shows an ethanol molecule on the {001} surface of form I acetaminophen.
Controlling the release of an active pharmaceutical ingredient from a crystalline particle embedded in a polymer matrix requires a fundamental understanding of the diffusion of small molecules in polymers. The application of molecular dynamics simulations to the problem reveals the issues which control the glass transition temperature of the polymer matrix and aid in understanding how diffusion is affected by such factors as chemical composition of the matrix, humidity, temperature and concentration.
Accurate, reliable crystal structure prediction has been described as the Holy Grail in crystallography. Dr. Marcus Neumann of Avant-garde Materials Simulation (AMS) is at the forefront of developments in this area, as demonstrated by his recent successful participation in the fourth blind test of crystal structure prediction, in collaboration with Drs. Leusen and Kendrick of the IPI Molecular and Solid State Modelling Group. The Blind Test is organised by the University of Cambridge and hosted by the Cambridge Crystallographic Data Centre. For the first time in the history of this international event, Drs Neumann, Leusen and Kendrick correctly predicted the crystal structures of all four test compounds, as reported in Nature (Nature 2007, Vol 450, Page 771). They applied AMS proprietary technology which encompasses a novel hybrid method integrating a tailor-made force field with a solid state density functional theory approach. The IPI Molecular and Solid State Modelling Group also collaborates with Accelrys Inc. to validate the application of their Polymorph Predictor technology to complex molecules, and to suggest future enhancements.
Most pharmaceutical compounds are chiral, and many cannot be synthesised as optically pure compounds. If only a racemic mixture of a chiral drug can be produced, it is imperative that the two enantiomers are separated because of their distinct biological effects. Crystallisation is by far the most cost effective separation technique for racemate resolution. Ideally, if the enantiomorph is more stable than the racemic solid, the racemic mixture can be resolved spontaneously, i.e., without the addition of any resolving agents or other additives. In most cases, however, the resolution is achieved by the selective crystallisation of diastereomeric salts, the so-called classical resolution first described by Pasteur in the middle of the 19th century. This resolution process is used in the production of about one third of all drugs on the market today. The researchers of the IPI Molecular and Solid State Modelling Group have pioneered the application of computational chemistry and molecular modelling tools to the prediction of both spontaneous resolution as well as classical racemate resolution. The figure shows the crystal structure of a diastereomeric salt.
Although detailed structural knowledge of molecules and crystals is a tremendous asset in pharmaceutical development, it does not explain the full dynamics of the crystallisation process. Crystallisation is not an equilibrium process and is driven by kinetic and external factors such as temperature, entropy, solvent, impurities, pressure, etc. As a consequence, the thermodynamically most stable polymorphs may not always be observed in reality because there may be no kinetic pathways to crystallise these structures. To capture these kinetic and external factors it is essential to study the crystallisation process itself in addition to the structures of the molecules and polymorphs. The crystallisation process consists of two stages: crystal nucleation and crystal growth. At the nucleation stage, molecules dissolved in a (super)saturated solution interact with each other to form an ‘embryonic’ particle. During the growth stage, this nucleus grows into a crystal. To gain full understanding of crystallisation kinetics, the IPI Molecular and Solid State Modelling Group aims to extend its research to study crystal nucleation and growth. These ambitious projects require additional computational resources as well as man power, for which funding is being sought.
Selected peer reviewed publications
2007
2006
2005
2003
2002
2001
2000
