Reciprocity between O-GlcNAc and O-phosphate on the carboxyl terminal domain of RNA polymerase II.

The carboxyl terminal domain of RNA polymerase II has multiple essential roles in transcription initiation, promoter clearance, transcript elongation, and the recruitment of the RNA processing machinery. Specific phosphorylation events are associated with the spatial and temporal coordination of these different activities. The CTD is also modified by beta-O-linked GlcNAc on a subset of RNA Pol II molecules. Using synthetic CTD substrates, we show here that O-GlcNAc and phosphate modification of the CTD are mutually exclusive at the level of the enzymes responsible for their addition. In addition, we show that O-GlcNAc transferase and CTD kinase have different CTD repeat requirements for enzymatic activity. The Km values of the two enzymes for CTD substrates are in a similar range, indicating that neither enzyme has a distinct kinetic advantage. Thus, the in vivo regulation of O-GlcNAc and phosphate modification of the CTD may involve the differential association of these two enzymes with the CTD at specific stages during the transcription cycle. Furthermore, direct competition between OGT and CTD kinase in vivo could generate multiple functionally distinct isoforms of RNA Pol II. Taken together, these results suggest that O-GlcNAc may give rise to additional functional states of RNA Pol II and may coordinate with phosphorylation to regulate class II gene transcription.

Characterization of a mouse monoclonal antibody specific for O-linked N-acetylglucosamine.

beta-O-linked N-acetylglucosamine (O-GlcNAc) is an abundant posttranslational modification of resident nuclear and cytoplasmic proteins in eukaryotes. Increasing evidence suggests that O-GlcNAc plays a regulatory role in numerous cellular processes. Here we report on the production and characterization of a highly specific mouse monoclonal antibody, MAb CTD110.6, that specifically reacts with O-GlcNAc. The antibody recognizes O-GlcNAc in beta-O-glycosidic linkage to both serine and threonine. We could detect no cross-reactivity with alpha-linked Ser/Thr-O-GlcNAc, alpha-linked Ser-O-linked N-acetylgalactosamine (O-GalNAc), or N-linked oligosaccharides on ovalbumin and immunoglobulin G. The monosaccharide GlcNAc, but not GalNAc, abolishes immunoreactivity, further demonstrating specificity toward O-GlcNAc. Furthermore, galactose capping of O-GlcNAc sites also inhibits CTD110.6 immunoreactivity. Enrichment of GlcNAc-containing glycoproteins using the lectin wheat germ agglutinin dramatically enriches for CTD110.6-reactive proteins. The antibody reacts with a large number of proteins from cytoplasmic and nuclear extracts and readily detects in vivo changes in O-GlcNAc modification. These studies demonstrate that CTD110.6 is highly specific toward O-GlcNAc, with no cross-reactivity toward similar carbohydrate antigens or toward peptide determinants.

Dynamic O-glycosylation of nuclear and cytosolic proteins: cloning and characterization of a neutral, cytosolic beta-N-acetylglucosaminidase from human brain.

Dynamic modification of cytoplasmic and nuclear proteins by O-linked N-acetylglucosamine (O-GlcNAc) on Ser/Thr residues is ubiquitous in higher eukaryotes and is analogous to protein phosphorylation. The enzyme for the addition of this modification, O-GlcNAc transferase, has been cloned from several species. Here, we have cloned a human brain O-GlcNAcase that cleaves O-GlcNAc off proteins. The cloned cDNA encodes a polypeptide of 916 amino acids with a predicted molecular mass of 103 kDa and a pI value of 4.63, but the protein migrates as a 130-kDa band on SDS-polyacrylamide gel electrophoresis. The cloned O-GlcNAcase has a pH optimum of 5.5-7.0 and is inhibited by GlcNAc but not by GalNAc. p-Nitrophenyl (pNP)-beta-GlcNAc, but not pNP-beta-GalNAc or pNP-alpha-GlcNAc, is a substrate. The cloned enzyme cleaves GlcNAc, but not GalNAc, from glycopeptides. Cell fractionation suggests that the overexpressed protein is mostly localized in the cytoplasm. It therefore has all the expected characteristics of O-GlcNAcase and is distinct from lysosomal hexosaminidases. Northern blots show that the transcript is expressed in every human tissue examined but is the highest in the brain, placenta, and pancreas. An understanding of O-GlcNAc dynamics and O-GlcNAcase may be key to elucidating the relationships between O-phosphate and O-GlcNAc and to the understanding of the molecular mechanisms of diseases such as diabetes, cancer, and neurodegeneration.

O-GlcNAc and the control of gene expression.

Many eukaryotic proteins contain O-linked N-acetylglucosamine (O-GlcNAc) on their serine and threonine side chain hydroxyls. In contrast to classical cell surface glycosylation, O-GlcNAc occurs on resident nuclear and cytoplasmic proteins. O-GlcNAc exists as a single monosaccharide residue, showing no evidence of further elongation. Like phosphorylation, O-GlcNAc is highly dynamic, transiently modifying proteins. These post-translational modifications give rise to functionally distinct subsets of a given protein. Furthermore, all known O-GlcNAc proteins are also phosphoproteins that reversibly form multimeric complexes that are sensitive to the state of phosphorylation. This observation implies that O-GlcNAc may work in concert with phosphorylation to mediate regulated protein interactions. The proteins that bear the O-GlcNAc modification are very diverse, including RNA polymerase II and many of its transcription factors, numerous chromatin-associated proteins, nuclear pore proteins, proto-oncogenes, tumor suppressors and proteins involved in translation. Here, we discuss the functional implications of O-GlcNAc-modifications of proteins involved in various aspects of gene expression, beginning with proteins involved in transcription and ending with proteins involved in regulating protein translation.

Proton transfer in the mechanism of triosephosphate isomerase.

Triosephosphate isomerase (TIM) catalyzes the reversible interconversion of dihydroxyacetone phosphate (DHAP) and glyceraldehyde 3-phosphate (GAP), with Glu-165 removing the pro-R proton from C1 of DHAP and neutral His-95 polarizing the carbonyl group of the substrate. During the TIM reaction, approximately 2% of the pro-R tritium from C1 of DHAP is conserved and appears at C2 of GAP [Nickbarg, E. B., and Knowles, J. R. (1988) Biochemistry 27, 5939]. In the "classical" mechanism, 98% of the pro-R tritium exchanges with solvent from Glu-165 at the intermediate state and the remaining 2% is transferred by Glu-165 to C2 of the same substrate molecule. This intramolecular transfer of tritium is therefore predicted to be independent of DHAP concentration. On the basis of NMR detection of a strong hydrogen bond between Glu-165 and the 1-OH of an analogue of the enediol intermediate [Harris, T. K., Abeygunawardana, C., and Mildvan, A. S. (1997) Biochemistry 36, 14661], we have suggested a "criss-cross" mechanism for TIM in which Glu-165 transfers a proton from C1 of DHAP to O2 of the enediol, and subsequently from O1 of the enediol to C2 of the product GAP. Since the pro-R proton is transferred to O2 instead of C2 in the criss-cross mechanism, no intramolecular transfer of label from substrate to product would be expected to occur. However, intermolecular transfer of label could occur if the label exchanges from O2 into a group on the protein and is transferred to GAP in subsequent turnovers. The extent of intermolecular tritium transfer in the criss-cross mechanism would be predicted to be dependent on DHAP concentration. The extent of tritium transfer was studied as a function of initial DHAP concentration using DHAP highly tritiated at the pro-R position. At 50% conversion to GAP, triphasic tritium transfer behavior was found. For phase 1, between 0.03 and 0.3 mM DHAP, a constant extent of tritium transfer of 1.19 +/- 0.03% occurred. For phase 2, between 0.3 and 1.0 mM DHAP, the extent of transfer progressively increased as a function of DHAP concentration to 2.17 +/- 0.15%. For phase 3, between 1.0 and 7.0 mM DHAP, the extent of transfer slightly decreased to 1.68 +/- 0.17%. In a direct test for intermolecular isotope transfer, doubly labeled [1(R)-D, 13C3]DHAP and 13C-depleted [1(R)-H,12C3]DHAP were synthesized, mixed in equal amounts, and incubated at 1 mM total DHAP with TIM, GAP dehydrogenase, NAD+, and arsenate until 50% conversion to 3-phosphoglycerate occurred. Electrospray ionization mass spectral analysis of the stable 3-phosphoglycerate product detected an extent of 1.4 +/- 0.4% of intramolecular D transfer from [13C3]DHAP to the 13C3 product, but no intermolecular transfer (

Selective detection and site-analysis of O-GlcNAc-modified glycopeptides by beta-elimination and tandem electrospray mass spectrometry.

Over the past decade, a number of nuclear and cytoplasmic proteins have been identified that are modified by single N-acetylglucosamine residues attached to the hydroxyl side chain of serines or threonines (O-GlcNAc). O-GlcNAc is a dynamic modification and therefore may act in a regulatory capacity analogous to phosphorylation. To undertake site-directed mutagenesis studies of O-GlcNAc's function, it is necessary to identify the sites of glycosylation on various proteins. The current method of site mapping, which involves galactosyltransferase labeling, generation of glycopeptides by proteolysis, purification by several rounds of HPLC, and gas-phase and manual Edman sequencing, is very tedious and requires about 10 pmol of pure, labeled glycopeptide. In this report, synthetic glycopeptides were generated and used to demonstrate that O-GlcNAc-modified peptides can be rapidly identified in complex mixtures by HPLC-coupled electrospray mass spectrometry due to the partial loss of the O-linked glycan (204 amu) at a modest orifice potential. Furthermore, the exact site of glycosylation was directly identified in the low picomole range by collision-induced dissociation (CID) of the glycopeptide after removal of the O-GlcNAc by alkaline beta-elimination. The conversion of glycosylserine to 2-aminopropenoic acid (2-ap) by beta-elimination both decreased the mass of the glycopeptide by 222 amu and resulted in a CID fragment ion representing the loss of 69 amu (2-ap) instead of 87 amu (Ser) at the position of the glycosylserine. Finally, we tested this method on an identical synthetic, alpha-linked O-GalNAc-modified peptide. Like O-GlcNAc, the O-GalNAc moiety was selectively removed at a modest orifice potential; however, the beta-elimination conditions that efficiently removed the O-GlcNAc only liberated about 20% of the O-GalNAc. We conclude that the selectivity and the sensitivity of this method will make it a powerful tool for determining the sites of O-GlcNAc modification on proteins of low abundance such as transcription factors and oncogenes.