Cluster Computing:

BUTTERFLY BENDING AND OTHER DELICATE MATTERS
by Kathleen Ricker, NCSA Science Writer

The chemical reactions that make up metabolism are accelerated and controlled by catalysts known as enzymes. Composed mostly of protein, enzymes also often contain a cofactor. The cofactor can be made of metal ions, such as iron, or can be a vitamin derivative, such as a flavin (a derivative of riboflavin, also known as vitamin B2).

The flavin may accelerate a reaction by brokering electron transfers between reagents to convert them to product. For example, flavins are crucial for metabolism of fatty acids. Flavin enzymes allow a kind of "controlled burn," in which the fatty acids are oxidized, allowing the energy released to be captured for use by the body. "Flavin enzymes play many crucial roles in human metabolism," says Anne-Frances Miller, associate professor of chemistry and biochemistry at the University of Kentucky. "Deficiency with regard to one of the flavoenzymes essential for fatty acid metabolism has been implicated as a possible cause of sudden infant death syndrome."

It is their versatility that makes flavins especially interesting to biochemists like Miller. "Although many different enzymes contain flavins, and as a class, they do a wide variety of chemistry," says Miller, "a given enzyme catalyzes a unique reaction. So, somehow or other, the flavin cofactor is being used for different purposes in different contexts."

Two types of flavin cofactors are most common: flavin mononucleotides (FMN) and flavin adenine dinucleotides (FAD). "The flavin is rather a beautiful molecule," says Miller. "It's an intense, brilliant yellow. The optical properties are yet another reflection of the very versatile electronic properties of the flavin ring."

This versatility makes it both important and challenging to understand the flavin's exact mechanism and reactivity. How the mechanism and reactivity are determined by active sites--the protein environments in which a given chemical reaction occurs--is also critical. Different enzymes may apply the same flavin to different types of reactions, and, according to Miller, the interaction between the flavin and a given protein may be correlated with the resulting chemistry.

How flavins bond

Flavins are generally not covalently bound to enzymes. Rather, proteins generally hold and control them via constellations of weaker non-covalent interactions including charge-charge interactions and partial sharing of labile protons. Miller hypothesizes that the function a given flavin performs may be determined by the unique non-covalent chemical interactions occurring at a given active site and that recurring patterns or sets of interactions will be found in enzymes that use their flavins in similar ways. These interactions, she says, might shape the flavin's reactivity by changing the distribution and energy of the flavin's valence electrons.

To test this hypothesis, Miller is running a series of high-level calculations on the Alliance's HP Superdome cluster at the University of Kentucky. She is focusing on two very specific kinds of non-covalent interactions that take place between the flavin and the enzyme: hydrogen bonding (H-bonding) and butterfly bending. A hydrogen bond consists of a hydrogen ion (H+) bound between two electron-rich atoms, such as oxygen or nitrogen. Thus, both groups are drawn together by the partial sharing of the H+.

A butterfly bend is somewhat more unusual. "The flavin looks like three sections of a honeycomb or three hexagonal bathroom tiles," explains Miller. "If you were to fold the central segment right down the middle, the two outside rings would look a little like the wings of a butterfly. It's not a butterfly that can sit on a branch and fold its wings right up so that the tips actually touch--it will only bend by some 25 degrees. The wing tips do not come close to each other, but the ring system ceases to be flat, and that is crucial." This bending is important because it is associated with reshaping of the valence orbitals, and it generates tension of a sort that can translate into chemical reactivity.

Although the flavin butterfly doesn't have a large range of motion, significant electron density movement can be achieved by H-bonding taking place along the edge of the "wings" as well as bending. Miller likens this phenomenon to what happens when a standing wave is created in a small space. "If you sat in a bathtub and pushed yourself with your feet from one end, so that your body slid up to the other end, a big wave of water would follow you," she explains. "The wave would hit the back of the bathtub behind you and would then sweep you forward again toward the faucet. In this case, the water is the electron density, and it's pretty fluid--it does have the possibility of moving back and forth in the flavin system."

The flavin electron density doesn't really slosh back and forth in time, Miller points out, but it is spatially spread out over the whole flavin ring system just as the standing wave makes use of the whole tub even though the crest itself is only in one place at a time. But the extent to which the valence electron density is able to run the full length of the tub is related to how bent the ring is, and the extent to which the valence electron density will prefer to stay at one end of the tub depends on the H-bonding at that end.

For different cases of hydrogen bonding and butterfly bending, Miller calculates the ionization potentials, or energy required to remove an electron from reduced forms of the flavin. Her team also calculates what effects these interactions will have on the nuclear magnetic resonance shielding experienced at each carbon and nitrogen atom of the flavin ring. Finally, they compare the computational results with sets of experimental results obtained by nuclear magnetic resonance (NMR) spectroscopy, which provides information about the distribution of electrons in a molecule. Thus, her group validates calculations of electron density distributions in the flavin ring and uses them to predict and understand reactivity.

Chemistry in small spaces

Miller's research showcases the advantages of integrating computational and experimental biochemistry. However, both the subtlety of the interactions involved and the large variety of enzyme functions which the flavins assist are especially challenging. "In order to capture the particulars of an enzyme active site," she explains, "you need to include a few of the surrounding [protein] groups." It is crucial to identify features that are important to the flavin behavior without including inessential detail that will tend to blunt the focus of the conclusions and enormously increase the computational cost.

Another concern is the ability of existing experimental methods to support the extraordinarily fine structural detail needed to correctly compute electron density distribution. Miller says that typical current crystallographic structures available for proteins, at 1.8 angstrom resolution, do not have sufficient resolution for studies of flavin electron density. "That accuracy is as good as is available for proteins, but it's still very crude for quantum mechanical calculations. 0.1 angstrom makes a huge difference in the energy and the electronic structure calculated."

The reward of all this careful calibration and concentration on the infinitesimal world of flavin enzyme electron density, Miller hopes, will be a greater understanding of how the versatile flavins operate--and how they can be engineered to perform yet more useful tasks. "We would like to be able to design changes in an existing enzyme in order to make a new version that would catalyze desired chemistry," she explains. Engineered flavin enzymes are being used in the areas of drug metabolism and waste detoxification treatment and could even provide new gene therapies for congenital birth defects.

In the end, such essential biological processes come down to the valence electrons.

Funding statement: This research was supported in part by the provost's office at the University of Kentucky and now by the National Institutes of Health.