Issue Date: February 7, 2011
Enabling Enzyme Studies
Three ambitious and dynamic research concepts that promise to accelerate the study and applications of enzyme mechanisms were presented at the 22nd Enzyme Mechanisms Conference last month in St. Pete Beach, Fla. Professor of chemical physiology Benjamin F. Cravatt III of Scripps Research Institute, in La Jolla, Calif., discussed a technique for finding inhibitors for hundreds of enzymes at a time that could speed drug discovery. Associate professor of chemistry Kate S. Carroll of Scripps Florida described a new way to assess enzyme damage from oxidative stress, which is associated with many human diseases. And professor of biopharmaceutical sciences Patricia C. Babbitt of the University of California, San Francisco, explained computational methods to determine the function of unknown enzymes found in genome projects—to understand how enzymes make the body work and learn how useful new enzymes can be engineered.
Cravatt’s strategy for ferreting out new inhibitors for multiple enzymes simultaneously is a good example “of how a strong understanding of enzyme mechanisms can serve as a starting point for studies in chemical biology,” said John P. Richard, chair of the meeting’s organizing group and a specialist on mechanisms of enzyme-catalyzed reactions at the State University of New York, Buffalo. Inhibitors are useful tools for probing enzyme functions, and they are potential drug candidates. But most inhibitors are found in a slow, one-by-one manner—for example, by screening a library of compounds for the ability to shut down a specific enzyme.
Called competitive activity-based protein profiling (competitive ABPP), the technique by Cravatt and coworkers could accelerate inhibitor discovery by making it possible to identify selective inhibitors for multiple enzymes that catalyze similar reactions.
Competitive ABPP is an extension of the ABPP technique that Cravatt and coworkers developed earlier. In ABPP, a fluorescence- or biotin-based tag is linked to an active-site functional group found in each member of an entire class of enzymes, such as serine hydrolases. These tags thus identify members of the enzyme class in proteomic analyses.
For example, the researchers recently used a cysteine-specific reagent to identify, in proteomic mixtures, enzymes that have highly reactive cysteine residues (Nature, DOI: 10.1038/nature09472; C&EN, Nov. 29, 2010, page 8). In another ABPP study, they used reagents specific for serine hydrolase active sites to label and identify serine hydrolases in proteomic mixtures.
“Our original idea was to use ABPP to compare normal and cancer cells, or normal and Alzheimer’s brain cells, to find aberrant enzyme activities relevant to the pathology,” Cravatt said. “The next thing you want to do is inhibit the [disease-related] enzymes” to probe their functional roles, “but few inhibitors are available. For example, there are about 250 serine hydrolases in humans, but until recently inhibitors had been identified for only a handful of them,” Cravatt said.
To systematically identify inhibitors for whole families of enzymes, such as serine hydrolases, Cravatt and coworkers devised competitive ABPP. “The goal would be to have 250 inhibitors that inhibit selectively 250 serine hydrolases,” Cravatt said. “We’re not there yet, but we’re making pretty good progress.”
In competitive ABPP of serine hydrolases, a library of potential inhibitors competes with a fluorophosphonate to react with the enzymes’ active sites (Proc. Natl. Acad. Sci. USA, DOI: 10.1073/pnas.1011663107). Fluorophosphonates are known to react selectively with hydroxymethyl side chains in serine hydrolase active sites. The most effective inhibitors, and the hydrolases they target, are then identified. Competitive ABPP can be extended to other enzyme families by developing other appropriate active-site-specific reagents, Cravatt noted.
Competitive ABPP thus identifies inhibitors for multiple enzymes in a single experiment. The technique also streamlines inhibitor discovery because it doesn’t require prior purification or recombinant expression of enzymes or prior identification of enzyme substrates and products—two time-consuming aspects of conventional techniques.
Cravatt and coworkers hope the new inhibitors discovered through competitive ABPP will, in fact, make it easier to identify enzymes’ substrates, products, and biological roles from observations of metabolic, physiological, and proteomic changes that occur when they use the inhibitors to squelch the activity of the enzymes.
The group’s newfound inhibitors could also be promising drug candidates. Serine hydrolase inhibitors discovered in the past include cholinesterase inhibitors to treat Alzheimer’s disease and dipeptidyl peptidase-4 inhibitors to treat diabetes.
At Scripps Florida, Carroll and coworkers have been probing the effects of reactive oxygen species on enzymes, an area of growing research interest. Oxidative stress—unchecked imbalances of reactive oxygen species in cells—has profound effects on biological systems and is associated with diseases such as cancer and diabetes. But the effects of oxidative stress on cells are not yet completely understood, and current techniques to study them are inadequate. A technique by Carroll and coworkers to identify and quantify protein oxidation could advance efforts to understand oxidative effects.
Oxidative stress is toxic to cells because highly reactive oxygen species can damage proteins, lipids, DNA, and other biomolecules. For example, high cellular levels of hydrogen peroxide (H2O2) interfere with cell function by modifying the catalytic activity, DNA-binding activity, and stability of enzymes. But peroxide also plays an important role in activating signaling pathways. So cells have to regulate peroxide levels carefully.
In enzymes and other proteins, cysteine residues react with peroxide to form S-hydroxylated cysteines, also known as sulfenic acids. These can be detected by treating protein samples with dimedone, which reacts specifically with sulfenic acid to form an adduct that can be detected mass spectrometrically. However, determining the level of stress, which is the fraction of cysteines oxidized to sulfenic acids, requires assays not only for sulfenic acid but also for cysteine.
Carroll and coworkers have now made such determinations possible by developing iododimedone. This new reagent reacts with cysteine thiols to give the same adduct as that formed from dimedone and sulfenic acids. Labeling dimedone’s methyl groups with deuterium provides a 6-dalton difference that allows adducts from the two reactions to be distinguished by mass spectrometry (Angew. Chem. Int. Ed., DOI: 10.1002/anie.201007175). In principle, the method can determine the relative level of thiol and sulfenic acid side chains in any enzyme sample.
“This capability has never been reported before,” Carroll said. “It’s an extremely facile and enabling technology.”
Carroll’s “group has made a very simple change in the structure of dimedone—iodination—that changes the reagent from a nucleophile that reacts with electrophilic sulfenic acids to an electrophile that reacts with nucleophilic thiols,” Richard commented. “This work provides an incredibly simple and elegant solution to a difficult problem.”
The technique should help scientists prioritize proteins for further characterization and functional analysis related to oxidative stress. It could also be used to assess protein modifications caused by redox-modulating drugs and to look for protein markers of oxidation in various disease states for diagnostic purposes.
So far, Carroll and colleagues have used the new approach only in vitro, “but the door is wide open for cellular studies,” she said. “We’re excited about what lies ahead.”
Meanwhile, UCSF’s Babbitt and coworkers have been using sequence- and structure-based computational methods to discover the physiological functions of unknown enzymes discovered in genome projects and thus understand the role of those enzymes in health and disease.
Assigning functions to poorly characterized enzymes is a major problem. For example, the sequences of more than 6,000 enzymes in the enolase superfamily are known. But the functions of about half those enzymes remain unknown, despite extensive efforts by several research groups.
To organize what is known about superfamilies, researchers divide them into subgroups of related families that catalyze different reactions but share common functional properties. They then use sticklike phylogenetic tree diagrams to display probable or known evolutionary relationships of different enzymes in each family.
Babbitt and coworkers have adopted a different form of representation that they believe has advantages over traditional tree models. They use a program called Cytoscape to display sequence or structure data on enzyme family members in “similarity networks.”
Data in similarity networks, such as evolutionary distances between family members, are consistent with similar data in phylogenetic tree diagrams. But similarity networks are more interactive and compact, provide information not available in trees, and can easily handle many more sequences or structures. Similarity networks provide a better intuitive sense of differences in rates of evolution for different families at a glance and indicate the relative size of each reaction class more effectively, Babbitt said.
Researchers are also using similarity networks—together with other methods such as in silico docking, structural characterization, and experimental screening—to develop a general strategy for assigning functions to enzymes discovered in genome projects. For example, Babbitt; John A. Gerlt of the University of Illinois, Urbana-Champaign; Matthew P. Jacobson of UCSF; Steven C. Almo of Albert Einstein College of Medicine; and coworkers recently identified a cluster of enzymes of unknown function in a sequence similarity network as new variants of muconate lactonizing enzymes and N-succinylamino acid racemases (Biochemistry, DOI: 10.1021/bi802277h; Nat. Chem. Biol., DOI: 10.1038/nchembio.2007.11).
Other applications of sequence and structure similarity networks include discovery of enzyme substrates and products and of homologous relationships among different enzymes. They could also help “guide the choice of starting scaffolds for enzyme engineering,” Babbitt said. By taking enzymes that nature has retooled over and over to evolve new chemistry and using them as structural templates for enzyme engineering and design, Babbitt said, “we may be able to more successfully engineer new enzymatic reactions in the lab.”
- Chemical & Engineering News
- ISSN 0009-2347
- Copyright © American Chemical Society