Introduction
Intracellular post-translational modifications, particularly protein phosphorylation, are critical modulators of mammalian signal transduction pathways. To be regulatory, a post-translational modification needs to be both inducible and dynamic. Regulatory post-translational modifications are typically widespread, occurring on multiple components in various signaling pathways, which allows attenuation of the signal and cross-talk between different signaling cascades. Achieving this complexity requires a cellular system to modify particular sites on given proteins in response to a specific stimulus without altering the same modification on other proteins. For phosphorylation, this is accomplished in mammalian cells by producing literally hundreds of different protein kinases—the largest enzyme family in the human genome. In this issue of Nature Chemical Biology, Khidekel et al. reveal that the O-linked 
-N-acetylglucosamine (O-GlcNAc) modification of proteins, which like phosphorylation modifies serines and threonines of intracellular proteins, is not only dynamic but also differentially regulated: the stoichiometry of certain O-GlcNAc sites is extensively altered by a given stimulus, whereas other sites are not significantly affected1.
Khidekel et al. recently developed a GlcNAc-specific labeling strategy2 that borrows on the concept of analog-sensitive variants from the kinase community3. This chemoenzymatic approach relies on specifically modifying proteins containing a terminal GlcNAc moiety by using a -1,4-galactosyltransferase that has been engineered to transfer a ketone-containing galactose to the C4 hydroxyl of a GlcNAc acceptor. The ketone functionality can then be tagged with an aminooxy biotin derivative for purposes of enrichment and identification (Fig. 1). In the paper presented here, the authors add an additional step to their workflow that allows for relative quantitative mass spectrometry–based analysis. Using a method developed by Hsu et al.4, peptides are labeled using isotope tagging of primary amines via dimethyl (light/heavy) groups. The authors have termed their combined approach quantitative isotopic and chemoenzymatic tagging (QUIC-Tag) (Fig. 1).
Figure 1: The QUIC-Tag strategy for quantitative analysis of O-GlcNAc.
O-GlcNAc–modified proteins (blue squares) from two samples are specifically labeled with a ketone-containing galactose (yellow circles), which is then further reacted with a biotin moiety (green hexagon). After tryptic digestion, light (purple) or heavy (red) methyl groups are added to the amino groups of the peptides, and the O-GlcNAc–modified peptides are enriched via avidin. Both identification (most robustly performed by ETD fragmentation) and relative quantification can then be achieved.
Full size image (53 KB)Combining this labeling approach with the power of electron-transfer dissociation (ETD)—a relatively new mass spectrometry fragmentation approach for identifying labile post-translational modifications that was pioneered by coauthor J. Coons5—permits site mapping for multiple O-GlcNAc modifications. The addition of the ETD approach is an important component, as it overcomes one of the main limitations of the original chemoenzymatic approach: relying on collision-assisted dissociation (CAD) to map labile modifications. In CAD, the weakest bond (that is, the bond to the serine/threonine post-translational modification) cleaves first. Alternatively, ETD has proven robust in fragmenting post-translationally modified proteins while retaining the modification on the hydroxyl amino acid. Though other quantitative site-mapping strategies exist for O-GlcNAc, such as BEMAD (-elimination followed by Michael addition with DTT)6, the QUIC-Tag approach combined with ETD seems to now offer the best strategy in terms of enrichment, specificity, quantification and site mapping.
Importantly, Khidekel et al. apply their newly developed method to biological systems: cultured cortical neurons and in vivo–stimulated rodent cerebral cortex. In these systems, the true power of developing chemical biology approaches is revealed—the uncovering of new biological phenomena. Although multiple investigators have demonstrated that O-GlcNAc levels can be globally elevated or decreased in response to various stimuli7, this new method reveals for the first time that whereas certain sites of modification may undergo tremendous changes in occupancy in response to a particular perturbation or stimulus, other sites remain virtually unchanged. This differential regulation illustrates an important way in which O-GlcNAc behaves, in a fashion analogous to that of phosphorylation, like a regulatory post-translational modification.
Like all good scientific work, this new discovery leaves those of us in the O-GlcNAc field with many more questions to answer and avenues to explore. It seems that evolution overcame the problem of how to differentially phosphorylate proteins by generating hundreds of different protein kinases with different donor specificity, expression, localization and regulatory-associated proteins. However, for O-GlcNAc, there is only one apparent animal O-GlcNAc transferase. Thus, how is O-GlcNAc modification being differentially regulated? Perhaps instead of taking the route of gene duplication and diversification (as for phosphorylation), the O-GlcNAc modification is controlled much as mRNA expression is regulated. Like O-GlcNAc transferase, only one RNA polymerase II exists and is responsible for mRNA transcription, but expression is exquisitely regulated by protein-protein associations, localization, substrate availability and other post-translational modifications. More experimental data will need to be acquired before the regulation of O-GlcNAc modification can be addressed. However, aided by advancements such as those presented here by Khidekel et al., we are well on our way to a better understanding of the biology of this small carbohydrate modification of nuclear and cytosolic proteins. One final comment: just as phosphorylation was not fully appreciated until the work of Krebs and Fischer defined the first of many clear functions for phosphate modification on a protein8, O-GlcNAc cannot be classified as a key regulatory modification until we discover the 'smoking gun'—evidence that O-GlcNAc modification at a specific site on a given protein alters its biological properties. The QUIC-Tag method gives the community a valuable tool for conducting this hunt.
