The science behind Michael Levitt's Nobel Prize
Modeling can predict a protein's molecular structure based on its amino-acid sequence.
In this video, Levitt’s Stanford colleague, Vijay Pande, PhD, shows a simulation of protein folding, and explains why computational biology is important to the future of medicine.
Michael Levitt, PhD, has dramatically advanced the field of structural biology by developing sophisticated computer algorithms to build models of complex biological molecules.
Applying known three-dimensional structures and basic principles of physical chemistry as complementary guidelines, this modeling can, for example, predict a protein's molecular structure on the basis of that protein's amino-acid sequence.
Proteins are long concatenations of chemical subunits called amino acids, and their job description is manifold. First, they do the vast bulk of every cell's physical work by catalyzing chemical reactions and shipping smaller molecules from place to place. They also are the key building blocks of the complex skeletons and scaffolds that maintain each cell's geometry. Further, a protein can serve as a messenger both within and between cells by contacting another protein with a complementary shape.
Biologists say that when it comes to proteins, "structure determines function." Like any precision machine, a protein can perform its job only when it is in just the right shape and has exactly the right electrochemical properties. When a protein is even slightly misshapen, its efficiency drops immensely or, worse, it actually becomes dangerous.
But no protein is born in its final shape, like the Greek goddess Athena springing full-grown from Zeus' head. Far from it — each protein begins its life as a linear assembly of amino acids, as many as hundreds or even tens of thousands of units long. Yet, in some mysterious way, a protein folds into its correct structure within a fraction of a second after its creation.
In the frenzied broth that is a cell's innards, however, proteins often get misshapen. Or they may be malformed at the outset — for example when, at any given position along the string, the correct amino acid (there are 20 different candidate amino acids, each with its own electrical-charge distribution and water-seeking or water-avoiding propensity) is replaced by one of the other 19 amino acids most typically as a result of a mutation in the gene whose instructions specify that protein's exact sequence.
In practice, biologists frequently don't know which proteins do which jobs, or how they do them. One way to find out is to learn how a protein is shaped and what makes its shape change and in what way, because a protein's structure determines its function. Proteins are so tiny that this is much easier said than done. Using radiological approaches such as X-ray crystallography allows scientists to "see" much better at this nanoscale, to reveal the structure — and a hint as to the probable workings — of important, but structurally complicated proteins.
But it is often the case that a protein can't be purified sufficiently, or lacks the biochemical characteristics, to be amenable to the finicky procedure that is X-ray crystallography. On the other hand, it is usually fairly easy to determine any protein's linear amino-acid sequence, either by direct analysis or by studying the gene dictating that sequence. The computer-simulation and molecular modeling techniques pioneered by Levitt reproduce the structural, thermodynamic and dynamic properties of a macromolecule in as accurate a way as possible, profoundly expanding the range of protein structures that can be discerned and unlocking the door to studying these proteins' function. Levitt's methods also permit prediction of the steps via which a protein molecule folds from its initial linear condition to assume its final, working form.
The techniques involved in this modeling start with simple but realistic expressions for the interactions between atoms and classical laws of motion. They are applicable not only to proteins but to other complex biomolecules such as DNA and RNA. Levitt's studies of DNA double-helix segments in solution preserve its classical double helix while still showing a wide range of possible motions on the part of various stretches of that helix. This is significance, as the "reading" of gene's instructions by the massive cellular machines that are crucial to protein production, as well as DNA replication, requires that the helix be flexible enough to accommodate their intrusions.
Using both molecular dynamics simulation and molecular modeling, Levitt and his associates have simulated the measurable static and dynamic properties of several different proteins surrounded by thousands of water molecules, at different temperatures. They have also used sophisticated computer programming and molecular-modeling methods to study, among other things, the immensely variable "tips" of antibody molecules. They are working on the question of how a single amino-acid change can destabilize a protein.
In early 2013, Levitt's group at Stanford employed novel methods to figure out the structure of an important class of molecules in a way that explains their function in greater detail that known before. The molecules under study were the most complex of a larger group of proteins called chaperonins, key "helper" proteins within all cells that act as midwives and monkey-wrenches to tease nascent or damaged proteins into their proper active shapes.