Proteomes take the electrophilic bait.

Proteomics researchers are constantly on the lookout for new and better ways to fractionate their samples since the mind-numbing complexity of virtually any proteome easily defeats the ability of even the most advanced mass spectrometers or other instruments to “see” all of the molecules in the sample. In this issue of Nature Chemical Biology, Weerapana et al. report new chemical tools for this purpose1. They demonstrate that particular electrophilic molecules that one might imagine would have similar reactivities in fact display largely nonoverlapping labeling of proteins in complex mixtures. These reagents thus constitute a set of chemical ‘baits’ to catch different kinds of protein ‘fish’. Most large-scale proteomics projects, for example those focused on biomarker discovery in serum or tissue homogenates or other interesting biological samples, use chromatographic fractional upfront to simplify the samples2. An alternative idea is to separate proteins based on their differential chemical reactivity instead of how tightly they bind to a chromatography column. This approach has been called activitybased protein profiling (ABPP)3. Initially, ABPP studies used molecules that were essentially suicide inhibitors of particular classes of enzymes, for example the reaction of fluorophosphonates with serine peptidases, which have an unusually reactive serine in their active site. Unfortunately, there are only a limited number of such mechanism-based enzyme-probe pairs, and so most of the proteome has been beyond the reach of ABPP. To begin to move beyond this limitation, Weerapana et al.1 turned to modestly reactive electrophiles (see Fig. 1). There is a long history in protein chemistry of using strongly electrophilic reagents to modify specific amino acid residues. For example, α-iodoacetamide will react with most surface-exposed cysteines in a protein. The goal here, however, was to ‘tone down’ the reactivity of the reagent such that one might observe differential reactivity based on the microenvironment of the nucleophilic amino acid in a particular protein, thereby making the molecule far more useful as an ABPP reagent. Preliminary experiments revealed that epoxide-containing molecules are too unreactive to be of utility, whereas an unsaturated ketone (UK) was quite reactive and an α-chloroacetamide (CA) and sulfonate ester (SE) displayed more moderate protein reactivity. Experiments using simple amino acid derivatives showed that SE and especially CA were both quite selective for coupling to cysteine thiols, whereas UK was more promiscuous, forming adducts with cysteine, histidine and lysine. Interestingly however, when these three reagents were incubated with a mouse liver extract, quite different results were obtained. Using a powerful analytical methodology that allowed specific capture of the modified peptide and the determination of the site of reaction of the electrophile on a protein, it was shown that CA and UK are quite specific for coupling to cysteines, whereas SE coupled to a variety of residues, especially glutamates. Moreover, an analysis of the identity of proteins modified by CA and UK in the liver extract showed that they are largely nonoverlapping. Taken together, these data argue strongly that the reactivities of electrophilic agents in the context of intact proteins can be very different than those exhibited for simple amino acids. Moreover, it demonstrates that the UK and CA probes, when applied to the proteome, are highly cysteineselective modifiers that are highly sensitive to the microenvironment of the cysteine. Thomas Kodadek is in the Division of Translational Research, Department of Internal Medicine, The University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75390-9185, USA. e-mail: [email protected] COO COO SE S O