Investigating Enzyme-Substrate Interactions with Vitamin B₆

nzymes are proteins that accelerate chemical reactions in cells that would proceed too slowly, or not at all, in their absence. Computational chemist Mary Jo Ondrechen and her colleagues at Northeastern University are investigating how four different enzymes interact with a Schiff base, a molecular compound made from an amino acid and vitamin B₆. From the study, her group hopes to learn more about how enzymes interact with substrates, a general term that describes anything on which an enzyme reacts. Their work is important to the development of drugs to treat such diseases as tuberculosis, which has had resurgence in recent years, particularly among populations of the homeless and individuals with active AIDS.

Under normal conditions, natural chemical barriers should prevent many enzyme-assisted reactions from occurring. To perform these same reactions in the laboratory requires extreme conditions, such as high temperature or low or high acidity, but enzymes make them happen within the body's normal temperature and neutral acidity.

THE VERSATILE COENZYME

BUILDING ENZYME MODELS

GAMESS ENZYMES PLAY

The Schiff base formed from vitamin B₆ and an amino acid inside the active site of the enzyme alanine racemase. Image by Mary Jo Ondrechen; structure courtesy of G.F. Stamper, A.A. Morollo and D. Ringe.

THE VERSATILE COENZYME

Some enzymes also require a coenzyme, a non-protein molecule such as a vitamin, to catalyze their reactions. The enzymes that Ondrechen's group is working with all require the coenzyme vitamin B₆, which is contained in a Schiff base formed of the vitamin and an amino acid. They hypothesize that such enzymes work by attaching the Schiff base to a specific site and in a specific orientation inside the enzyme. The group believes that the force then exerted by nearby atoms, as well as the electrostatic forces exerted by charged groups in the enzyme that are farther away, sharply reduces the energy needed to cleave and re-form the appropriate chemical bonds. The study of enzymes and Schiff base interaction reveals processes that likely also occur between enzymes and other substrates.

The goal of Ondrechen's work is to understand, using computational simulation on NPACI resources, just how the enzyme interacts with the substrate. She starts with an enzyme and a Schiff base in their initial states, then models what occurs as each enzyme reacts with the Schiff base to form new molecular structures.

The vitamin B₆ Schiff base is ideal for Ondrechen's study because it serves as a substrate in four different kinds of reactions. Also, depending on the enzyme with which it interacts, four distinctly different products are created. The group can therefore examine a single substrate and its reaction pathways as a function of the enzymatic environment.

In an earlier project, Ondrechen's group used NPACI resources to model the reaction that occurs when the Schiff base is formed with the simple amino acids alanine or serine. The results of that study gave them high-quality parameters for molecular modeling and mechanics of the Schiff base. "That served as a starting point for our current study," said Ondrechen. "We needed to have complete knowledge of the structure of the Schiff base so that we could then consider its interaction with the enzymes."

Since four classes of enzyme are known to act upon the Schiff base substrate, a representative enzyme with known structure data was required from each of those classes. The group focused their attention on proteins with available structure data, and an inhibitor with molecular structure similar to that of the Schiff base located in the active site and in reactive orientation to the substrate. Ondrechen selected D-amino acid aminotransferase and alanine racemase for the next stage of the study, and working with collaborator Dagmar Ringe of Brandeis University, crystallographic structure data for each was acquired (Figure 1). In addition, structural information for two other enzymes, dialkylglycine decarboxylase and tryptophan synthase--representing the remaining two classes of enzymes--were available from the Protein Data Bank.

BUILDING ENZYME MODELS

These four enzymes were then used to model the four types of reactions. Transamination moves the amino groups; racemization redistributed the enantiomers; decarboxylation removes the COOH groups to form carbon dioxide; and tryptophan synthase is involved with various side-chain reactions. These reactions are important to the researchers because each type represents the attachment of the substrate to the active center of a particular enzyme, and corresponding changes in the electronic properties of both. The differences from one to the next help them understand how substrates and enzymes interact.

Before investigating the models computationally, the group used the structure data of the enzymes as a visual starting point from which to proceed with the study. "From the crystallographic data, one can see that the shape of the active site cavity inflicts conformational constraints on the substrate," Ondrechen said. "In particular, the substrate has a number of low-frequency 'soft' degrees of freedom, namely internal rotations, along which the molecule can move to fit into a particular active site." After geometry optimization was achieved on the free substrate in the earlier study, the dihedral angles of the optimized structure were adjusted to match each enzymatic environment. Then, the electronic structure was repeated on the enzymatic conformation.

The group created quantum mechanical-molecular mechanical models to describe the enzyme-substrate interaction in the four different enzymatic environments. Each enzyme-bound substrate model had three parts: the substrate, the charged and hydrogen-bonding side chains in the immediate region of the substrate, and the rest of the protein (Figure 2).

"Including the side chains within bonding distance of the substrate in our electronic structure calculation allowed us to specifically address the close-neighbor interactions," Ondrechen said. These were treated quantum mechanically. The rest of the enzyme, which Ondrechen termed a modified protein matrix, was treated as a complex 3-D assembly of charges. The electric field gradients in the active sites of the modified protein matrix determine many of the subsequent interactions.

"I can't think of a more computationally intensive problem than ours," Ondrechen said. "It would be impossible to conduct our research without the computing power provided by the NPACI allocation."

GAMESS ENZYMES PLAY

Since the Schiff base has a total charge of –2 and an organic phosphate group, high-quality ab initio calculations were required to obtain accurate, reliable results. The GAMESS software was selected because it has all of the necessary features built in for the electronic structure phase of the project, and because an optimized, parallelized version of the code is available. Running the various models on the NPACI Cray T3E at the University of Texas at Austin took approximately 1,000 hours apiece.

"We were particularly interested in discovering what the relative importance of purely electrostatic effects is, and of bonding interactions with specific active-site residues," Ondrechen said. "We're also pursuing understanding of the interactions with other species, such as ions, in the active site, the presence or absence of solvent molecules, and the molecular conformation. The modeling system we undertook allowed us to segregate these effects and study them for the same substrate located in four different enzymes."

By looking at the transition states and the influence of the enzyme on them, the research also helped the group understand the mechanism of enzyme catalysis. "We learned that just as the enzyme acts upon the substrate, the substrate also has an influence on the enzyme," Ondrechen said. "This doesn't necessarily impart any major change on the enzyme itself, but it is interesting to note that there is strong interaction between the two."

The changes in enzymes and substrates revealed by this work ultimately impede or propel the resulting activity within cells. "An excellent example of a benefit derived from understanding this process is the development of pharmaceuticals," Ondrechen said. "Drugs are often intended to block enzyme-substrate interaction. A big problem facing society today is that many forms of disease-causing bacteria have become resistant to antibiotics. Two enzymes that I'm working with specifically, alanine racemase and D-amino acid aminotransferase, are involved in bacterial cell wall construction. If we can gain a better understanding of the enzymatic environment of these enzymes, it's quite possible that we can design a new system of antibiotics to impede their reactivity." --AF *