The Nobel Prize in Chemistry 2013

‘…for the development of multiscale models for complex chemical systems’

Figure 1: Illustration of the QM/MM model.

Figure 1: Illustration of the QM/MM model.

This is how the Nobel Committee explains its decision to award the 2013 Nobel Prize in Chemistry to Martin Karplus, Université de Strasbourg and Harvard University, Michael Levitt, Stanford University, and Arieh Warshel, University of Southern California. But what do these words mean? Multiscale models generally come into play when using computers to study chemical systems spanning large temporal and spatial scales. Protein folding is an example of a process with multiple temporal scales. The electrons in the protein move on an attosecond scale. The atoms move on a femtosecond scale. Smaller parts of the protein move on a picosecond scale, while the entire protein structure folds on a much longer time scale, maybe milliseconds. The Laureates were awarded for having developed methods enabling scientists to include processes that occur over multiple temporal and spatial scales in one and the same computer calculation.

The complex systems that the Laureates have studied involve mainly macromolecules of relevance in biochemistry, or the chemistry in living organisms. In this work, different parts of the biomolecule receive different levels of attention. An assessment is made of which area is of most importance for the process in question, and then this area is given the most attention. The parts of the molecule that fall outside this area are treated with a higher degree of approximation.

The main methods that led to the Nobel Prize are called QM/MM, or Quantum Mechanics/Molecular Mechanics (Figure 1). Quantum mechanics implies that we solve equations describing the distribution of electrons in the system, usually the Schrödinger equation:

This gives us the potential energy (the energy of position) for the chosen positioning of atoms. This is often done repeatedly in an attempt to find the geometrical structure of atoms that implies the least energy. This process requires significant computational capacity. The development of methods and computer programmes for these types of electronic structure calculations yielded the Nobel Prize in Chemistry in 1998.

In the part that is treated with MM, the potential energy is calculated from simpler analytical functions. When the potential energy is described with simple functions, it is in these contexts referred to as a classic potential. The parameters in the analytical functions are often adapted to experimentally available information. It also happens that the parameters in the analytical functions are adapted to data from quantum mechanical calculations. Since simpler functions are used, the potential energy calculated in the MM region ends up more approximate than that calculated in the QM part. However, the calculations are much faster, implying that larger and more complex systems can be studied.

Proteins do not exist in a vacuum, but in a solution. Studies have shown that often it is important to take the surrounding solution into account.  This is usually accomplished in a more approximate manner where the solution is treated as a continuum characterised by a dielectricity constant.

Dividing the calculations into regions that are analysed with different levels of scrutiny is one of two important contributions from the Nobel Prize Laureates. The other contribution is so-called coarse-graining. This is illustrated in Figure 2 and implies that atoms are not considered individually but rather together as a larger unit in select parts of a molecule. This may for example occur for the side chains of amino acids in a protein. In effect, you end up with considerably fewer degrees of freedom when you want to minimise the energy to determine the equilibrium structure of the protein, the calculation takes less time and it enables studies of much bigger systems.

THE laureates’ methods are used frequently in physical chemistry at the University of Gothenburg. In issue 1, 2013 of the Science Faculty Magazine, we wrote about Leif Eriksson’s research. Eriksson uses these methods in studies of enzymatic mechanisms to model how liposomes filled with medicine interact with cell membranes. They are also used in the development of new medicines for cancer and other diseases.

Gunnar Nyman’s research team in physical chemistry uses the QM/MM model to study how molecules are formed in interstellar clouds, or the areas between stars where new starts are born. Coincidentally, these types of research projects started earlier this year when a postdoc, Chamil Sameera, was recruited to study how small radicals diffuse on icy surfaces (Figure 3), which are common in the very cold interstellar clouds. The radical and parts of the icy surface are then included in the QM region, while a considerably larger part of the ice is treated with the MM method. In so doing, Chamil has tried to decide which quantum methods and which force fields for the molecular mechanics work well for the interstellar applications.

Within QM/MM, it is common not to follow the movement of the molecule over time but instead calculate only the most probable structures and a reaction path. Sometimes, however, the protein is allowed to move dynamically, by using classical mechanics. A true union of quantum physics-based atomic motion and classical physics does not occur in the QM/MM model, though. Still, this combination is a very exciting research field that a number of universities around the world, including the University of Gothenburg, have become involved in. In August 2014, an international conference, MOLEC, centred around this research field will be held in Gothenburg.