By combining activity-based proteomics and metabolomics, researchers have developed a new systems biology strategy for characterizing enzymes in the context of metabolic networks.
By harnessing the power of mass spectrometry, researchers have been generating massive amounts of peptide and protein sequence data from just about every kind of biological sample. Although the databases containing these lists are certainly valuable resources, the effort to follow up with detailed studies to assign all of these proteins in biological contexts has lagged. Even in the very best-annotated protein classes, perhaps only 50% or so of the known members have been characterized.
But some researchers, such as Benjamin Cravatt at the Scripps Research Institute, find this part of the process the most exciting. Cravatt's group has been at the forefront of efforts to develop activity-based protein profiling (ABPP), or functional proteomics, which they use to identify biologically interesting enzymes, especially those relevant to human diseases such as cancer.
Though the ABPP method allows them to discover enzymes in various functional classes, characterization requires steps for further analysis. Cravatt explains, "Our next goal was to come up with an unbiased small-molecule profiling platform that would, in response to inactivating an enzyme, allow us to see what the metabolic consequences were in the broadest possible sense and hopefully link [the enzyme] into the metabolic network that it regulates." For their first demonstration of this strategy, they chose the target KIAA1363, an enzyme that is found in particularly high levels in aggressive cancer cells, but whose biological context was uncharacterized.
Using competitive ABPP, Cravatt and his colleagues identified a small-molecule inhibitor of KIAA1363, named AS115, that strongly and selectively inhibited the enzyme's activity. They treated several cancer cell lines with AS115 and compared the liquid chromatography–mass spectrometry metabolic profiles of the inhibitor-treated cells to those of untreated cells. The inhibitor treatment resulted in a reduction in the level of an unusual class of lipids, the monoalkylglycerol ethers, which were shown to be direct products of a KIAA1363-catalyzed reaction. Amounts of a secondary metabolite class, the lysophospholipids, also were reduced. In parallel, by performing short hairpin RNA–mediated knockdown of KIAA1363 (shKIAA1363), the researchers observed similar reductions in the same metabolite classes, but the downstream effects were even more dramatic than with small-molecule inhibition. "The advantage of the small-molecule inhibitor is that you can get a quick picture of what the acute effects are of inactivating an enzyme," explains Cravatt, whereas "the shRNA approach gives you a steady-state picture of what an inactivated enzyme looks like over many days," which will certainly be important for evaluating potential drug targets. Additionally, mice injected with shKIAA1363 cancer cells showed reduced tumor growth compared to those injected with control cancer cells.
This powerful approach allowed the researchers to place KIAA1363 at a central node in an ether lipid signaling network. "This paper really shows, I think for the first time, that one can go into a complex biological model like a human cancer cell and annotate a disease-relevant protein using purely systems biology methods. We didn't have to recombinantly express or purify the enzyme at all to figure any of this stuff out," says Cravatt. As opposed to the more classical 'test-tube' enzyme characterization, Cravatt explains, "being able to do these experiments in living systems is really valuable because it allows you to circumvent potential in vitro artifacts related to what an enzyme can do versus what it really does do."
