If you have an ACS member number, please enter it here so we can link this account to your membership. (optional)

ACS values your privacy. By submitting your information, you are gaining access to C&EN and subscribing to our weekly newsletter. We use the information you provide to make your reading experience better, and we will never sell your data to third party members.


Biological Chemistry

Enzymology Embraces Biology

Some mechanistic enzymologists go over to the protein function side

by Stu Borman
February 12, 2007 | A version of this story appeared in Volume 85, Issue 7

Credit: John Gerlt & Coworkers
Mandelate racemase (shown) and other members of the enolase superfamily have two domains: a capping domain, which includes residues (cyan) that control substrate specificity, and a barrel domain containing the enzyme's active site (primarily yellow), which includes a magnesium ion (magenta). An inhibitor (green) is bound to the active site.
Credit: John Gerlt & Coworkers
Mandelate racemase (shown) and other members of the enolase superfamily have two domains: a capping domain, which includes residues (cyan) that control substrate specificity, and a barrel domain containing the enzyme's active site (primarily yellow), which includes a magnesium ion (magenta). An inhibitor (green) is bound to the active site.

MECHANISTIC ENZYMOLOGY has traditionally focused on step-by-step explanations of the way enzymes do their jobs. To a large degree it still does, and there's nothing wrong with that. Such studies are enormously useful for designing new enzyme-inhibiting drugs, for example.

But some mechanistic enzymologists are adopting a broader perspective, investigating not just how enzymes work individually but also how they function in the context of the cell. An understanding of enzyme mechanisms informs the work but is no longer its main focus.

Not all mechanistic enzymologists subscribe to this shift in emphasis, but the effort expended on studies of enzyme function seems to be growing. The work is motivated in part by genomic sequencing projects, which have turned up many mysterious proteins of unknown function. Their biological roles are just waiting to be discovered, like ripe fruit waiting to be picked. The move toward enzyme-function studies was a theme at the 20th Enzyme Mechanisms Conference held last month in St. Pete Beach, Fla.

"One of the current trends in the field of mechanistic enzymology is a shift in experimental focus from mechanism to cell function," said Karen N. Allen, professor of physiology and biophysics at Boston University School of Medicine. She noted "a major adjustment in the field ... as enzymologists move beyond structure-function analysis to add new tools to their arsenal." With these tools, researchers are better able to nail down the locations of different enzymes in cells, determine the substrates and protein partners with which those enzymes interact, and probe the impact of enzymes on cell metabolites.

"Sequence databases contain more unknown proteins than known proteins, and that is a real problem," said John A. Gerlt, professor of biochemistry, chemistry, and biophysics at the University of Illinois, Urbana-Champaign. It's not known, for example, whether these unknown proteins catalyze reactions, and if so, what their substrates are.

Funding agencies now recognize that determining the functions of unknown enzymes in different organisms, a pursuit known as functional genomics, warrants more encouragement and better support.

"The National Institutes of Health has put a lot of money into research on determining structures of proteins," says biochemistry professor Richard N. Armstrong of Vanderbilt University. An example is NIH's Protein Structure Initiative (PSI), a collaborative federal, academic, and industry effort to make three-dimensional protein structures more readily available. In the past, functional genomics didn't get as much funding attention as structure determination. Part of the reason is that determining the functions of proteins is more exploratory and technically much more difficult than determining their structures, "but NIH is starting to take a greater interest in this area," Armstrong said.

Determining functions of multiple proteins in cells is also a key part of systems biology, the study of relationships and interactions among various parts of biological systems or pathways. So studying cellular functions of proteins as ensembles is not a new idea. But mechanistic enzymologists bring a different perspective and their own unique expertise to the game.

Tag, You're It
A biotin-derivatized ATP can be used to tag lysine residues in the ATP-binding sites of most active kinases. Kinases so labeled can then be isolated, identified, and quantified.
View Enlarged Image

At the enzyme mechanisms conference last month, Allen, Gerlt, Armstrong, and several other enzymologists described how their research programs are expanding into functional genomics. Gerlt, for example, tries to assign functions to enzymes in families whose members have similar structures and mechanisms. He discussed recent efforts by his group to ferret out the functions of enolase enzymes, which are related structurally and mechanistically and therefore constitute an enzyme superfamily.

From a structural standpoint, each member of the enolase superfamily is composed of two domains: a capping domain, which determines the enzyme's specificity for different substrates, and a barrel domain, which contains the enzyme's active site. Enolases catalyze enolization, a reaction in which a proton next to a carboxylate in the substrate is extracted by a base in the active site to generate an enolate anion. The enolase superfamily consists of some 1,700 different members, and the cellular functions of about 500 of these are unknown.

To determine the cellular function of an enolase, it helps to know what its substrate is. Gerlt and coworkers try to determine this by screening enolases with libraries of potential substrates. However, "we've screened a number of unknown proteins and have failed to determine their function," he said. "We believe this is because the libraries are simply not large enough, not comprehensive enough, to be generally useful for defining function."

A better way would be to predict substrate specificity simply on the basis of the sequence and structure of specific proteins, Gerlt said. He and his coworkers, in collaboration with assistant professor of pharmaceutical chemistry Matthew Jacobson's group at the University of California, San Francisco, have tried to pursue this better path by using homology modeling (comparison with structures of known proteins) and computational docking to predict likely active sites and substrates.

In one case, they were able to confirm experimentally their theoretical prediction of an unknown enzyme's structure and substrate. Identifying this enzyme's substrate also enabled them to identify its function. "This is a very encouraging example of the ability to use homology modeling and docking, rather than libraries, to determine unknown function," Gerlt said.

Whereas Gerlt and coworkers have been determining the functions of enzymes that are similar structurally and mechanistically, Armstrong and coworkers have been doing so for members of the glutathione transferase (GST) superfamily, which are related structurally but not mechanistically.

GSTs use a common architecture to accomplish widely varying functional ends. "Similar GST structures can do many different things that are completely unrelated mechanistically," Armstrong said. And the functions of many GSTs have never been assigned. "My background in mechanistic enzymology led me to ask the questions, What are these proteins doing, and how are they doing it?" Armstrong said.

He and his coworkers use several experimental approaches to predict answers to these questions. They use gene knockouts, disabling or deleting GST genes one by one in the organism, to assess the resulting functional effects when GST enzymes encoded by those genes are missing.

They also examine the degree of expression of different GST enzymes under a variety of conditions. For example, a gene that is important for function in a specific set of circumstances, such as starvation or oxidative stress, often will be upregulated when the condition exists, and that response can be an indicator of function, Armstrong said. And they sometimes obtain crystal structures of new enzymes to look for structural clues of enzyme function.

"It turns out that GSTs are very divergent in what they do," Armstrong said. Not only do they have diverse catalytic functions, but one even turned out to regulate RNA polymerase. This function suggests that it's essentially a regulatory protein and may not actually catalyze a reaction.

Family Ties
Credit: Richard Armstrong & Coworkers
Members of the glutathione transferase superfamily share a common structure as well as a set of consensus residues (shown as spheres, where green is carbon, red is oxygen, and blue is nitrogen).
Credit: Richard Armstrong & Coworkers
Members of the glutathione transferase superfamily share a common structure as well as a set of consensus residues (shown as spheres, where green is carbon, red is oxygen, and blue is nitrogen).

Allen is also interested in finding the functions of unknown enzymes. Her group and that of her collaborator, chemistry professor Debra Dunaway-Mariano at the University of New Mexico, recently focused their efforts on the haloalkanoic acid dehalogenase (HAD) superfamily. This superfamily has about 3,800 members in prokaryotes and eukaryotes, and each organism has multiple members. Humans have about 56, for example.

Many HADs are phosphatases, enzymes that cleave phosphate groups. They all tend to have structurally related active sites that catalyze the same mechanism of phosphoryl transfer: an aspartate on the enzyme attacks and expels a substrate phosphoryl group, a phosphoenzyme intermediate forms, and a molecule of water hydrolyzes the intermediate and releases the enzyme and phosphate.

Although most HAD active sites are similar, the enzymes do differ significantly in their cap domains, where specificity-determining groups that cause the enzymes to interact with different substrates are located. "The cap domain drives the evolution of biochemical function within the HAD superfamily," Allen said.

Allen, Dunaway-Mariano, and coworkers have been investigating how variations in enzyme structure affect substrate selectivity and lead to diverse cellular functions. By using molecular modeling to dock different potential substrates into the "cage" generated by rolling a solvent molecule around the active site of different HAD enzymes, they have been able to propose, and test experimentally, functions for the enzymes.

For example, the structure of the bacterial HAD enzyme YidA had been determined previously by the New York Structural Genomics Research Consortium, part of NIH's PSI, but YidA had no known function. "We used the solvent cage derived from our analysis of the structure, in conjunction with the electrostatic and steric features of the active site, to predict potential substrates," Allen said. When they tested and validated the substrates through enzyme kinetics, the substrate with greatest activity was 5-phospho-d-arabinonate. The result suggests that YidA's cellular function involves sugar metabolism.

"It is a good example of how the marriage of structural and functional analysis yields rich results," Allen said. "I believe that in the future, enzymologists will contribute to the understanding of cell metabolism by assigning functions to those proteins whose structures are solved" as part of structural genomics efforts.

Overall, the team's functional studies of HAD superfamily members have shown how minor changes in a common structural scaffold—such as the HAD cap domain—can lead to a variety of functions. The studies also have demonstrated that genetic regulation, enzyme localization, metabolite availability, and other factors influence cell functions of different HAD enzymes, Allen noted.

John Kozarich, chairman and president of ActivX Biosciences, La Jolla, Calif., is also engaged in functional analysis of enzymes. Kozarich and coworkers have developed a technique to fish out and identify most protein kinase enzymes that are active in specific cell types.

ActivX focuses on kinases because they have widely varied cellular roles and are extraordinarily important in drug discovery. For example, approved drugs that are kinase inhibitors include the anticancer agents Gleevec and Iressa.

Mammalian cells harbor more than 500 kinases. But they're not all readily available for drug-screening purposes, their activity in different cell types is largely unknown, and knowledge of the selectivity of their interactions with different inhibitors is fragmentary.

Protein kinases transfer a phosphate group from a donor compound like adenosine triphosphate (ATP) to a protein substrate. Kozarich and coworkers synthesized a biotin-containing ATP derivative that reacts with lysine residues in the ATP-binding sites of most active kinases and labels them with biotin (Biochemistry 2007, 46, 350). The labeled kinases can then be isolated, identified, and quantified.

Division of Labor
Credit: Karen Allen & Coworkers
HAD enzymes have two domains: a cap domain (top half), where a loop structure (blue) controls substrate specificity, and a core domain (bottom half), which contains the active site (red) and a key magnesium ion (magenta).
Credit: Karen Allen & Coworkers
HAD enzymes have two domains: a cap domain (top half), where a loop structure (blue) controls substrate specificity, and a core domain (bottom half), which contains the active site (red) and a key magnesium ion (magenta).

Once identified, the selectivity of inhibitors for the various kinases can be assessed and the cellular functions of individual kinases can be studied. More than 400 different protein kinases—over 80% of the predicted human kinome—"have been identified and in most cases functionally assayed in various mammalian tissues and cell lines using this method," the researchers note in their paper.

Despite the efforts of researchers like Armstrong, Gerlt, Allen, and Kozarich, the contributions of mechanistic enzymology to functional genomics are not universally appreciated. "Biologists often do not understand the value of enzymology," said biochemistry professor Christian R. H. Raetz of Duke University Medical Center. "Deans of medical schools are likewise often unaware of the relevance of enzymology. In a sense, we have failed to communicate the value of enzymology to the broader scientific community."

Consequently, Raetz said, many scientific administrators "are spending huge amounts of money on genomics and systems biology without including people who can think at the level of chemical structure and mechanism." That may change, because chemists are now demonstrating increasingly that expertise in enzyme mechanisms can indeed provide an essential platform from which to assess the cellular functions of proteins revealed in the genomics studies of biologists.

To researchers like chemistry reader Adrian Mulholland of the University of Bristol, in England, the value of mechanistic enzymology to functional genomics and systems biology is evident. "In the U.K., we have something of a systems biology mania setting in," Mulholland said. "We are being told now that we should all be systems biologists. So it's great to come to a meeting where people are passionately interested in the fundamental chemistry of biology. This is really what we need."


This article has been sent to the following recipient:

Chemistry matters. Join us to get the news you need.