Home
》》》》》》》1
Key steps toward computational chemistry get three the 2013 Nobel Sat Nov 30, 2013 9:05 pm
Admin
Admin
[You must be registered and logged in to see this image.]
Enlarge / A mixture of physical detail and classical approximations, showing one of the proteins that won people the Medicine Nobel on Monday.
Brookhaven National Laboratory
Given billions of years of evolution, nature has come up with proteins that can perform staggering feats of chemistry—in some cases, a few grams of protein would be enough to outperform a large industrial facility. Understanding how these proteins work can have applications in areas from energy to industry to medicine.
But that understanding has been hard to come by. It's possible to take static pictures of what the enzymes look like when they are stuck in a crystal, but that only provides a limited picture. The enzymes actually work in a dynamic environment, undergo physical changes, and catalyze reactions that involve the rapid shuffling of chemical bonds and transfer of electrons.
This year's Nobel Prize in Chemistry recognizes three researchers who helped bring dynamism to our study of large molecules like proteins. Martin Karplus, Michael Levitt, and Arieh Warshel started studying how to simulate the activity that goes on inside a protein back in the 1970s, when computing power was extremely hard to come by. The descendants of the methods they developed are still in use today, even as our computational models have grown ever larger and more sophisticated.
To provide some sense of the scale of the problem, you can admire carbonic anhydrase, the fastest enzyme in the world. Although it only catalyzes a simple chemical reaction, some versions of the enzyme can turn over 1,000,000 reactions every second. And each one of those involves a bit of flexing in the enzyme itself, chemicals moving in and out of its active site, and a rearrangement of chemical bonds.
Yet for most of these catalysts, all we have is a static picture. Techniques like X-ray crystallography can tell us where all the atoms in a protein typically reside, but only if those atoms sit still for their portrait. The other major technique for determining structures, NMR imaging, works when the proteins are in solution, but it only provides an indication of what the atoms are doing on average.
So how do we get a glimpse of the dynamic world where enzymes really operate if all we have are single snapshots? Given the position of the atoms, it should be possible to use fundamental physics to tell us what they would do as they interact with one another while bumping into water and other molecules in their environment. But that requires some pretty hefty computations, and computers aren't always up to the task. This situation was an even more significant problem back when the prize winners started their work in the 1970s.
Early computers had the ability to model molecules as classical systems—the ball-and-stick view of atoms and their bonds, which did allow them to flex and bounce off each other. But each bounce involves the interactions of the outermost shells of electrons held by these atoms, and their behavior is governed by quantum mechanics. This is even more true for the reactions catalyzed by proteins, which generally involve the transfer of electrons and/or protons. The problem is that describing the quantum behavior of these systems adds a tremendous amount of computational complexity, enough to overwhelm the computers of the time.
The researchers are being honored for their development of a hybrid quantum-classical model of molecular behavior. A key event in this process came when Arieh Warshel (now at USC) spent time in the lab of Martin Karplus, then and still at Harvard. Warshel had experience with classical modeling, while Karplus brought expertise in the quantum behavior of molecules. Together, they built a model of a large organic molecule: retinal, the molecule that absorbs photons within the photoreceptors that provide us with vision. Retinal has a complicated structure that includes rings and alternating double and single bonds, which allows some of the outermost electrons to become delocalized and orbit the molecule as a whole.
Warshel and Karplus' model treated most of the molecule using easy-to-calculate classical mechanics. But the delocalized electrons were given a quantum treatment. The combination allowed the calculations to be simple enough that they would run on the computers available at that time but still capture the most important features of a dynamic molecule.
By the mid-'70s, Warshel was collaborating with Michael Levitt (now at Stanford). By improving and generalizing the techniques used on retinal and taking advantage of the growing power of computers, the two were able to create a hybrid classical-quantum model that was able to describe the behavior inside a protein's active site. For proteins, quantum calculations are used for the area near where the chemical reactions take place. The parts of the protein distant from that have atoms, or even clusters of atoms, that are treated as single classical objects.
As computing power grows, more and more of the molecules are able to be given a sophisticated treatment, and further details like the presence of water molecules can be added to the model. But the basic approach pioneered by the researchers being honored today is still widely in use.]
Enlarge / A mixture of physical detail and classical approximations, showing one of the proteins that won people the Medicine Nobel on Monday.
Brookhaven National Laboratory
Given billions of years of evolution, nature has come up with proteins that can perform staggering feats of chemistry—in some cases, a few grams of protein would be enough to outperform a large industrial facility. Understanding how these proteins work can have applications in areas from energy to industry to medicine.
But that understanding has been hard to come by. It's possible to take static pictures of what the enzymes look like when they are stuck in a crystal, but that only provides a limited picture. The enzymes actually work in a dynamic environment, undergo physical changes, and catalyze reactions that involve the rapid shuffling of chemical bonds and transfer of electrons.
This year's Nobel Prize in Chemistry recognizes three researchers who helped bring dynamism to our study of large molecules like proteins. Martin Karplus, Michael Levitt, and Arieh Warshel started studying how to simulate the activity that goes on inside a protein back in the 1970s, when computing power was extremely hard to come by. The descendants of the methods they developed are still in use today, even as our computational models have grown ever larger and more sophisticated.
To provide some sense of the scale of the problem, you can admire carbonic anhydrase, the fastest enzyme in the world. Although it only catalyzes a simple chemical reaction, some versions of the enzyme can turn over 1,000,000 reactions every second. And each one of those involves a bit of flexing in the enzyme itself, chemicals moving in and out of its active site, and a rearrangement of chemical bonds.
Yet for most of these catalysts, all we have is a static picture. Techniques like X-ray crystallography can tell us where all the atoms in a protein typically reside, but only if those atoms sit still for their portrait. The other major technique for determining structures, NMR imaging, works when the proteins are in solution, but it only provides an indication of what the atoms are doing on average.
So how do we get a glimpse of the dynamic world where enzymes really operate if all we have are single snapshots? Given the position of the atoms, it should be possible to use fundamental physics to tell us what they would do as they interact with one another while bumping into water and other molecules in their environment. But that requires some pretty hefty computations, and computers aren't always up to the task. This situation was an even more significant problem back when the prize winners started their work in the 1970s.
Early computers had the ability to model molecules as classical systems—the ball-and-stick view of atoms and their bonds, which did allow them to flex and bounce off each other. But each bounce involves the interactions of the outermost shells of electrons held by these atoms, and their behavior is governed by quantum mechanics. This is even more true for the reactions catalyzed by proteins, which generally involve the transfer of electrons and/or protons. The problem is that describing the quantum behavior of these systems adds a tremendous amount of computational complexity, enough to overwhelm the computers of the time.
The researchers are being honored for their development of a hybrid quantum-classical model of molecular behavior. A key event in this process came when Arieh Warshel (now at USC) spent time in the lab of Martin Karplus, then and still at Harvard. Warshel had experience with classical modeling, while Karplus brought expertise in the quantum behavior of molecules. Together, they built a model of a large organic molecule: retinal, the molecule that absorbs photons within the photoreceptors that provide us with vision. Retinal has a complicated structure that includes rings and alternating double and single bonds, which allows some of the outermost electrons to become delocalized and orbit the molecule as a whole.
Warshel and Karplus' model treated most of the molecule using easy-to-calculate classical mechanics. But the delocalized electrons were given a quantum treatment. The combination allowed the calculations to be simple enough that they would run on the computers available at that time but still capture the most important features of a dynamic molecule.
By the mid-'70s, Warshel was collaborating with Michael Levitt (now at Stanford). By improving and generalizing the techniques used on retinal and taking advantage of the growing power of computers, the two were able to create a hybrid classical-quantum model that was able to describe the behavior inside a protein's active site. For proteins, quantum calculations are used for the area near where the chemical reactions take place. The parts of the protein distant from that have atoms, or even clusters of atoms, that are treated as single classical objects.
As computing power grows, more and more of the molecules are able to be given a sophisticated treatment, and further details like the presence of water molecules can be added to the model. But the basic approach pioneered by the researchers being honored today is still widely in use.]
