Projects in Scientific Computing

Download the PDF version of this article as it appeared in Projects in Scientific Computing, 2003.

THE FIRST STEP-BY-STEP PICTURE OF HOW DNA OPENS UP TO INTERACT WITH AN ENZYME.

Feb. 28, 1953: Two young men walk into a dingy pub in Cambridge, England called the Eagle. To a lunchtime crowd they announce that they’ve discovered the secret of life. They have.

The two young men were Watson and Crick. Fifty years later, their success at deciphering the structure of DNA stands as the founding event of molecular biology. The elegant spirals of this structure and the phrase that denotes it, “the double helix,” have become ingrained in our culture. School lessons teach us that the rungs of the DNA spiral staircase are bonded pairs of chemical bases — A and T, C and G — letters that shape the destiny of all known forms of life.

As this new science has progressed, we’ve learned much about how the sequence of bases exerts its mighty influence, and we know that DNA doesn’t act alone. Enzymes are the deus ex machina of DNA drama, coming into the scene and inciting change. Enzymes interact with the bases to facilitate cell division and protein-making, and as a first step in these processes the base pairs must fold out from their sheltered space inside the double helix, a structural shift called base flipping.


						Alex MacKerell, Niu Huang, University of Maryland

Alex MacKerell, Niu Huang, University of Maryland

“The information in DNA is hidden,” says University of Maryland biophysicist Alex MacKerell, “and for DNA to perform its biological function, DNA has to open up so the information can be accessed. Base flipping is a simple structural change that may be the first step in replication and transcription of DNA and is essential for other processes in which enzymes interact with the bases.”

Laboratory studies have shown the structure of flipped-out DNA, but laboratory work tells virtually nothing of exactly what happens to initiate the shift and what intermediate states occur along the way. With the availability of LeMieux, Pittsburgh Supercomputing Center’s terascale system, MacKerell and his research team tackled these questions with an extensive series of simulations. Their results — reported in the Proceedings of the National Academy of Sciences (Jan. 7, 2003) — provide the first atom-by-atom, step-by-step picture of enzyme-facilitated DNA base flipping.

Which Groove?

Top-Down View Represents the Unflipped Dna Helix Compared to the Flipped State

Click Image to Enlarge

This top-down view represents the unflipped DNA helix (right) compared to the flipped state (left), with the target base for methylation, cytosine (atoms in red), fully turned out. The "orphaned" guanine (green atoms) remains within the helix. Arrows indicate the major groove versus minor groove pathway.

Although DNA base flipping happens in all organisms from plants to people, researchers first confirmed it in bacteria. Laboratory studies have shown that an enzyme called methyltransferase attaches to cytosine, the C of A,T,C and G, and chemically changes it, by adding a methyl group (CH3-). This relatively simple chemical change, called methylation, is thought to be widespread in DNA interactions. “We’re starting to understand,” says MacKerell, “that chemical modifications of certain bases are involved in the regulation of the expression and transcription of DNA.”

The base has to flip out for methylation to occur, and the flipped-out DNA structure has been identified in laboratory work. Still, what was known before MacKerell’s work was a bare outline of the process, like seeing the opening scene of a romantic movie and falling asleep until the noisy wedding at the end.

A central unanswered question had to do with how much the enzyme is involved in the base flipping. Does the enzyme help the base to flip out, or does it bind after it’s already flipped? Experiments gave no clear answer. “How do we understand,” asks MacKerell, “going from the normal duplex DNA shape to the flipped out shape?”

Another question had to do with DNA’s grooves. Because of the way paired bases stack up, an intact double helix of DNA has a groove on each side, one smaller than the other, aptly named the minor groove, and a larger one called the major groove. Through which of these grooves do the bases turn as they flip outward? Structural evidence suggested the minor groove, but some experimental evidence suggested the major groove.

“THE COMPUTER IS INVALUABLE, BECAUSE IT ALLOWS US TO LOOK AT EVENTS, WHICH IN EXPERIMENTAL TIME FRAMES YOU CAN’T SEE. IN THE COMPUTER WE CAN SEE WHAT HAPPENS.”

“This is where the computer is invaluable,” says MacKerell, “because it allows us to systematically change the structure and look at events, which in experimental time frames happen so fast that you can’t see them. In the computer, using our mathematical models, we can see what happens.”

May the Force Field Be With You

To produce a comprehensive base-flipping picture, MacKerell turned to a computational approach called molecular dynamics. In essence, MD treats a molecule as a dynamic structure of atoms interacting with each other and with nearby atoms. The computer tracks how each atom in the molecule moves by calculating the forces between it and every other atom at successive slices of time.

Over a period of years, MacKerell has helped to extend MD, first developed for proteins, to become a powerful tool for DNA. Much of his work has focused on “empirical force fields” — a way to express the quantum-mechanical energies between atoms as empirical constants. Deriving these empirical force fields, which approximate the probabilities of quantum theory, is the art and science of MD. It has the important benefit of making it possible to do MD simulations, which have proven ability to reveal the atomic-level details of biomolecular processes.

“To get these parameters to treat the chemical system accurately,” says MacKerell, “is a continual process in which we optimize the empirical force field to reproduce experimental data. We also use quantum mechanical data as part of the target data. The empirical force fields have become more sophisticated and more accurate with time.”

What Happens When the Enzyme Arrives

What Happens When the Enzyme Arrives
Click Image to Enlarge

What Happens When the Enzyme Arrives
MacKerell’s simulations show that the methyltransferase binding site — the part of the enzyme that interacts with the DNA — has a dramatic impact on the free energy of base flipping. When the enzyme (blue rods and coils) pairs with the DNA helix, amino acids within the enzyme’s binding-site loop (dark blue) interact directly with the cytosine (red) and guanine (green). The first frame (top) shows a serine (purple) competing for the hydrogen bonds (dotted lines) that bind the two bases. The second frame (above) shows the early flipped state, stabilized by the serine and a glycine (gold).

To arrive at his full-story picture of base flipping, MacKerell broke the process down into chunks he called “simulation windows.” Each window is a scene from the full scenario. All possible configurations of the DNA, from closed-to-flipped-to-closed, are represented as a circle, like a clock face, which is sub-divided into 72 five-degree arcs. Within each of these windows, MacKerell calculated relative free-energies, key information that tells what shape the molecule prefers, since it tends to assume the shape that requires the least expenditure of energy.

Using eight LeMieux processors for each free-energy window, MacKerell simulated four different configurations of a 12-base DNA sequence: An unflipped helix in a water solution (17,700 atoms), a flipped helix with methyltransferase in two different positions, and a flipped helix with methyltransferase and a third molecule, called a cofactor. For each five-degree window, he simulated 160 picoseconds (a trillionth of a second) of movement — with a snapshot of the action every two femtoseconds, 80,000 time slices per window.

With fourteen months of computing time, 80,000 single-processor hours, and much careful analysis, MacKerell and his colleagues had answers where before there was only mystery. The enzyme initiates flipping, and the base flips through the major groove pathway. “The presence of the enzyme destabilizes the DNA,” says MacKerell, “and then the base interacts further with the enzyme, until the enzyme-cofactor complex stabilizes the fully flipped state.”

These findings, MacKerell believes, suggest a process by which DNA and the enzyme are in cross talk with each other, like a molecular pas de deux. The enzyme, arms spread, approaches to begin binding, and the DNA in turn starts to open, which draws the enzyme closer, until it stabilizes the DNA in the flipped state.

Overall, it’s a result that highlights the power of computational methods to uncover the details of DNA-enzyme interactions, a field of study that’s still new. “Everyone has known for a long time,” says MacKerell, “that DNA has to change its shape to perform its function. We’ve been able to show for the first time how an enzyme actually facilitates the conformational change. And we’ve been able to see the atomic details of how it does that.”



Researchers:
Alex MacKerell, Niu Huang, University of Maryland.

Hardware:
Terascale Computing System.

Software:
User-developed code.

Related Material on the Web:
MacKerell Laboratory at the University of Maryland.

References:
Niu Huang, Nilesh K. Banavali & Alexander D. MacKerell, Jr., "Protein-facilitated base flipping in DNA by cytosine-5-methyltransferase," Proceedings of the National Academy of Science 100, 68-73 (2003).

Author:
Leah Kauffman, ed. M. Schneider.

Web Design:
Sean Fulton, Pittsburgh Supercomputing Center

Revised: October 27, 2003
URL: http://www.psc.edu/science/2003/mackerell/when_dna_flips_out.html