(page 2 of 2)

The Structure of Viruses

Brooks and Case also collaborate on a National Institutes of Health (NIH) center to develop multi-scale modeling tools for structural biology and to make these tools available to the scientific community for free.

"Our mandate is to build tools to help structural biologists better interrogate, explore, and understand structural models," says Brooks, who is director of the NIH center.

One of the center's objectives is to develop tools that might be used for genome-scale modeling and structure prediction. As more and more genomes are solved and annotated with gene prediction algorithms and proteomic techniques, the number of proteins of unknown structure and function is growing, which is creating a great demand for computational tools that can predict the structures—or partial structures—of these proteins.

Brooks and Case work on tools for problems including protein folding prediction and homology modeling—where the structure of a protein is predicted based on the similarity of its amino acid sequence to another, known protein—and test these tools for their ability to predict the fold of unknown proteins. The computations that they use are sometimes very intensive.

"We recently ran a calculation in which we used 2,500 nodes (processors) at the Pittsburgh Computing Center together with about 500 nodes of San Diego's machine and hundreds of nodes as a site in Virginia," says Brooks. "They were all working at once for one computation for a period of 24 to 48 hours."

That particular calculation, says Brooks, involved exploring a protein-folding landscape for a protein that is known to fold very quickly in order to understand how the sequence of amino acids in the chain determines the three dimensional structure of the folding protein.

In another area of research at the NIH center, Brooks and Case aim to make connections between atomic-level descriptions of molecules obtained from crystallography and NMR and lower-resolution pictures obtained with other techniques, such as electron microscopy (EM).

This is important because in solving structures, crystallographers and NMR spectroscopists often can only solve a small piece of a large structure, while EM can handle large structures, but not at high-resolution. Reconciling the two allows them to fit high-resolution pieces together, jigsaw-like, in order to obtain a complete picture with more atomic detail than would be possible using only one technique or the other.

They then take this still-static picture and make a sophisticated model out of it, modeling the dynamics of the molecule at a larger scale.

"Generally, the reason for doing this," says Case, "is to get biologically interesting states that are not observable at high resolution.

They have also worked in collaboration with TSRI Professor Jack Johnson to study the assembly of viruses—a subject that Johnson has been studying with x ray crystallography for a number of years.

"We're looking at understanding, at a molecular level, the process that swells and shrinks the particles," says Brooks.

Virus particles are dynamic in solution and undergo large structural changes throughout their "lifetime," and some of these changes are interesting biologically because they may be related to such issues as how the virus particle gets its genetic information into a cell that it infects. However not all the changes may be accessible experimentally, since the transition states may be unstable and therefore impossible to crystallize or study with NMR.

Brooks and Case are also trying to understand how nucleic acids are packaged in viruses. "Usually that cannot be seen at high resolution," says Case, "because the DNA or RNA is too disordered."

Still, he adds, the problem is not so simple computationally either. The insides of a virus are incredibly crowded, which makes computing difficult. Some of the work they do involves figuring out how to remove atomic detail from the structures so that they do not have to take it into account. Case, for instance is working on how to model a protein or DNA in continuum solvent to simplify the calculations.

"You keep an atomic-level protein or DNA model, but you remove all the atomic level descriptions of the water or ions," he explains.

The Fluctuating Ribosome

Brooks is particularly interested in the workings of the large molecular machines that carry out much of the work of the cell, such as the ribosome or actin/myosin.

"Nature effectively exploits the shape of these objects to provide robustness in the motions that they have to undergo," says Brooks.

In the case of the ribosome, the starting point for the study of these motions are the near atomic-resolution and atomic-resolution molecular structures that have been solved in the last couple of years using EM and x ray crystallography.

Brooks is working to create atomic-level models using these structures that give dynamic movement to them. He, with his postdoctoral collaborator Florence Tama, is building an elastomechanical model that captures the shape of molecular "objects" and allows dynamic behavior to emerge from normal vibrations and rotations associated with the atoms in the molecule.

These dynamic movements may be the key to some of the molecules' most complex behavior. In the model of the ribosome, the collective fluctuations may give rise to a ratchet-like motion that is involved with the phenomenon known as translocation.

Translocation is an important part of the ribosome function because it involves moving tRNA molecules loaded with amino acids from the site on the ribosome where the anticodon of the tRNA is paired with the codon of the mRNA to the site on the ribosome that catalyzes the formation of the peptide bond between the amino acid loaded on the tRNA and the growing protein chain. A major motion of the ribosome is associated with this movement from one location on the ribosome to another after an "effector" molecule binds to the ribosome.

In Brooks' dynamic model of the ribosome, the movement occurs quite naturally, as one of the "normal modes" of vibration. Physicists describe normal modes as fluctuations about a local energy minimum (preferred vibration) for simple harmonic oscillations—between two adjacent atoms, for instance. More collective motions in large molecular assemblies like the ribosome may describe functionally important dynamics.

"[The translocation] is happening only because of the shape of the molecule," says Brooks. "It seems as though nature somehow engineers these shapes so that it is a single mode that is functionally relevant."

These models are promising because they may be the only way to access atomic-level structural information about transition states of large structures like the ribosome. Such transition states may be inherently unstable and completely inaccessible to experimental techniques but nevertheless important to the operational cycle of these cellular machines—in other words, they may hold some of the secrets of life.

And perhaps of physics as well.


1 | 2 |



From machine to man: the physics of protein synthesis. Click to enlarge.