Clarke AJ et al. (2008), Structural insights into mechanism and specific...

XB-ART-38857

EMBO J 2008 Oct 22;2720:2780-8. doi: 10.1038/emboj.2008.186.

Show Gene links Show Anatomy links

Structural insights into mechanism and specificity of O-GlcNAc transferase.

Clarke AJ , Hurtado-Guerrero R , Pathak S , Schüttelkopf AW , Borodkin V , Shepherd SM , Ibrahim AF , van Aalten DM .

Abstract
Post-translational modification of protein serines/threonines with N-acetylglucosamine (O-GlcNAc) is dynamic, inducible and abundant, regulating many cellular processes by interfering with protein phosphorylation. O-GlcNAcylation is regulated by O-GlcNAc transferase (OGT) and O-GlcNAcase, both encoded by single, essential, genes in metazoan genomes. It is not understood how OGT recognises its sugar nucleotide donor and performs O-GlcNAc transfer onto proteins/peptides, and how the enzyme recognises specific cellular protein substrates. Here, we show, by X-ray crystallography and mutagenesis, that OGT adopts the (metal-independent) GT-B fold and binds a UDP-GlcNAc analogue at the bottom of a highly conserved putative peptide-binding groove, covered by a mobile loop. Strikingly, the tetratricopeptide repeats (TPRs) tightly interact with the active site to form a continuous 120 A putative interaction surface, whereas the previously predicted phosphatidylinositide-binding site locates to the opposite end of the catalytic domain. On the basis of the structure, we identify truncation/point mutants of the TPRs that have differential effects on activity towards proteins/peptides, giving first insights into how OGT may recognise its substrates.

PubMed ID: 18818698
PMC ID: PMC2556091
Article link: EMBO J
Grant support: [+]

Species referenced: Xenopus
Genes referenced: akt1 aopep ogt tab1 tbx2 tpr

Article Images: [+] show captions

	Figure 1. OGT sequence conservation. Alignment of OGT sequences from Xanthomonas campestris pv. campestris (XcOGT), Caenorhabditis elegans (CeOGT), Xenopus laevis (XlOGT) and human (hOGT, isoform 1). N termini of hOGT constructs are indicated by arrows. Secondary structure elements (cylinders/arrows) are shown, a brown bar marks the active site lid. coloured bars indicate the TPRs (yellow=A helix; orange=B helix), the TLRs (green), the connecting loop (light green), the N-terminal GT-B subdomain (cyan), the intersubdomain linker (blue) and the C-terminal GT-B subdomain (magenta). Blue triangles give the location of the previously described (triangle up) and proposed additional (triangle down) PIP3-interacting residues, orange and red circles locate point mutations described previously and in the present work, respectively. Labels represent hOGT/XcOGT residues.
	Figure 2. Structure of OGT. (A) Cartoon view of XcOGT in complex with the UDP-GlcNAc analogue, using the domain colours as described in Figure 1. Beyond the XcOGT N-terminal TPRs, the TPR superhelix is continued by the superimposed hOGT TPR structure (PDB id, 1W3B, semitransparent). The phosphoinositide-binding site is identified by an IP4 model obtained by superposition as explained in the text. Mutated and PIP3-binding side chains (colours as in Figure 1) are shown as sticks and labelled with XcOGT residue numbers, the equivalent hOGT residue numbers can be obtained from Figure 1. The UDP-GlcNAc analogue is shown with green carbons. The OGT middle domain is indicated by the ‘OMD' label. (B) Surface representation of the XcOGT/hOGT superposition, coloured by metazoan OGT sequence conservation (blue (100% identity) to white (<50%)).
	Figure 3. Mutational analysis and substrate specificity of OGT. (A) hOGT (overexpressed and purified from baculovirus) was incubated with the protein substrate TAB1 in the absence of magnesium and in the presence of 1 mM EGTA, and the reaction mixture was probed by anti-O-GlcNAc western blotting. (B) Single-point hOGT mutants (overexpressed and purified from E. coli) localised in the putative peptide and UDP-GlcNAc-binding site assayed against TAB1 and studied by anti-O-GlcNAc western blotting. (C, D, E) Activity assay of the truncated baculovirally expressed hOGT constructs against different acceptor substrates. (C) Immunoblot detecting O-glycosylated TAB1, (D) O-glycosylated HEK293 lysates and (E) radioactivity assay against the CKII peptide. (F) Anti-O-GlcNAc immunoblot of Arabidopsis thaliana lysate incubated with XcOGT and UDP-GlcNAc shows one single protein substrate around 130 kDa.
	Figure 4. OGT active site and schematics of the TPR interactions. (A) Stereo representation of the active site of XcOGT, with the UDP-GlcNAc analogue (green carbons) and an omit map (blue, 2.25σ). The protein backbone is shown as coil in domain colours as above, amino acids of interest are shown as sticks, coloured red/orange if they correspond to described point mutants in hOGT; labels as in Figure 2A. Black dotted lines indicate potential hydrogen bonds. The ‘active site lid' is shown in both its apo- (brown) and ligand-bound conformation (cyan). A black arrow indicates the proposed direction of nucleophilic attack compatible with an inverting mechanism. (B) Model of how hOGT and protein kinase B (an OGT substrate) could colocalise in a PIP3-dependent manner on the plasma membrane, as suggested by a recent study (Yang et al, 2008). The OGT active site is identified by a sticks model of the sugar donor. (C) Cartoon representation of the OGT catalytic domain (colours as in (A, B)) and the different possible models of how the TPRs (rods) might contribute to the regulation of OGT activity, as seen in other TPR proteins, such as the widely used ‘substrate localisation mode' seen in most TPR proteins (D'Andrea and Regan, 2003) and the ‘autoinhibition mode' seen in protein phosphatase 5 (Yang et al, 2005). The crystal structure appears to be most compatible with the ‘substrate-binding mode', with the TPRs tightly associated with, but not occluding, the active site.

References [+] :

Chang, Synthesis of the C1-phosphonate analog of UDP-GlcNAc. 2006, Pubmed