The crystal structure of a sugar-transferring enzyme offers insight into the mechanism of a ubiquitous protein-modification reaction, and solves the mystery of how the enzyme recognizes certain sequences in proteins. See Article p.350
One of the most common protein-modification reactions in the cells of eukaryotes (organisms that include plants, animals and fungi) is N-linked glycosylation, in which sugars are attached to the side chain of the amino acid asparagine. This reaction has diverse roles in protein folding and stability, intracellular trafficking and cell–cell interactions, and is catalysed by the enzyme oligosaccharyltransferase (OST). More specifically, OST mediates the transfer of an oligosaccharide from a donor substrate onto acceptor asparagine residues in newly synthesized proteins. In eukaryotes, such glycosylation occurs only at asparagine residues located within asparagine-X-threonine/serine amino-acid sequences (N–X–T/S sequences), where X can be any amino acid except proline. It has been unclear how OST recognizes the acceptor sites and the associated amino-acid sequences, and how it activates the normally unreactive nitrogen in asparagine's side chain to take part in glycosylation. But on page 350 of this issue, Lizak et al.1 now provide remarkable insight into these issues with their report of the X-ray crystal structure of PglB, an OST from the bacterium Campylobacter lari.
Although often described as post-translational modifications, most N-linked oligosaccharides in eukaryotic cells are added co-translationally to a nascent polypeptide as it is threaded through the protein-translocation channel in the endoplasmic reticulum (ER) membrane2(Fig. 1). The preassembled oligosaccharide is anchored by a lipid to the luminal surface of the ER. Consequently, the OST has to scan nascent polypeptides for acceptor sites that are moving past the enzyme at the rate of protein synthesis (about 6–8 residues per second for eukaryotes), while simultaneously positioning the lipid-linked oligosaccharide (LLO) in the enzyme's active site.
Unlike the oligomeric OST complexes that are present in eukaryotes3, PglB is monomeric, making it a more suitable candidate for protein crystallography. PglB consists of an amino-terminal domain that binds to the inner bacterial membrane, and a carboxy-terminal domain that resides in the periplasm (the space between the inner and outer bacterial membranes). Proteins are glycosylated by PglB as they thread through the plasma membrane. The enzyme's periplasmic domain includes a short, evolutionarily conserved amino-acid sequence (tryptophan–tryptophan–aspartic acid, abbreviated as WWD) that is essential in all OST catalytic subunits4,5. The structure of this domain in PglB from the archaeon Pyrococcus furiosus has been solved previously6, but the structure of the membrane-bound domain was unknown, not least because crystallizing integral membrane proteins is always technically challenging. Lizak et al.1 surmounted this challenge by crystallizing the complete Campylobacter PglB protein in the presence of its peptide substrate.
The authors' structure reveals that two large loops (EL1 and EL5) of the N-terminal domain extend into the periplasm to form a platform that supports the periplasmic domain, and contain residues required for peptide binding and catalysis. Remarkably, the peptide-binding cleft is formed by the interface between the membrane-binding and periplasmic domains, and is located on the opposite face of PglB from the cleft that harbours the catalytic site and the LLO-binding site. This architecture explains how the large LLO and peptide substrates can independently enter the enzyme's active site. The structure is consistent with biochemical studies7 that indicated that acceptor sequences must be located within flexible or unfolded segments of polypeptides. It also shows that the critical WWD motif binds to the threonine side chain of the peptide substrate, thereby explaining why serine or threonine must be located two amino acids after asparagine in the N–X–T/S acceptor sequence.
Another intriguing feature of the structure1 is that the asparagine side chain in the peptide projects through a narrow 'porthole' of PglB that connects the peptide-binding cleft to the catalytic site. This site is formed by a cluster of charged residues that bind a magnesium ion (Mg2+). Although OST activity has long been known to be dependent on a divalent cation (either Mg2+ or the manganese ion, Mn2+), the catalytic-site residues had not been identified. Lizak et al. report that three of the active-site residues have acidic side chains (they are aspartic acid or glutamic acid residues). When the authors replaced any of these residues with alanine, which has a non-acidic, methyl side chain, the glycosylation activity of the resulting enzymes was reduced by 50–90% compared with the wild-type enzyme.
The side chain of asparagine contains an amide group (CONH2) in which the nitrogen atom is a 'weak nucleophile' — which means that it shouldn't react readily with the oligosaccharide of the LLO. So how does the enzyme activate the amide so that it can react? Previous models for OST catalysis envisaged that activation occurred through the formation of hydrogen bonds between the asparagine side chain and the threonine (or serine) residue in the peptide substrate8,9. But Lizak and colleagues' finding1 that the asparagine residue projects through the porthole in the active site rules out this possibility.
Instead, the authors propose that two active-site residues form hydrogen bonds to the amide hydrogens of asparagine's side chain, and in so doing increase the ability of the amide to react with the oligosaccharide. When the researchers replaced these catalytic residues with others to alter the hydrogen bonding to the asparagine side chain, the resulting mutants lost their catalytic activity, thus providing strong support for the proposed scheme. This activation mechanism could also explain why a glutamine-containing pseudo-acceptor sequence (Q–X–T/S, where Q is glutamine) can bind to the OST active site, yet not be glycosylated — the glutamine side chain cannot form properly positioned hydrogen bonds to the active site10.
Following glycosylation, the asparagine side chain on the polypeptide substrate is covalently linked to a bulky oligosaccharide that cannot pass through the narrow active-site porthole. So how can the reaction product leave the active site? Lizak and colleagues' structure1 reveals that the porthole is formed by the packing of the flexible, partially disordered EL5 loop against the periplasmic domain. They therefore propose that product formation induces a conformational change that promotes disengagement of EL5 from the periplasmic domain, followed by release of the glycosylated product and the lipid anchor.
Lizak and colleagues' crystal structure of PglB in complex with a peptide substrate provides invaluable information about PglB–peptide binding and the enzyme's catalytic mechanism, but higher-resolution structures are needed to provide detailed insight into the mechanism. What's more, the location of the LLO binding site in PglB cannot be precisely known until the structure of a PglB–LLO complex is solved. Such a structure should also reveal why N-acetylglucosamine (a sugar that contains an NHCOCH3 group in place of one OH group) is present at the reducing end of all eukaryotic LLO donors — that is, at the end that connects the oligosaccharide to the lipid11. Finally, the structure of an enzyme–product complex would provide insight into how the product dissociates from the enzyme.
Lizak, C., Gerber, S., Numao, S., Aebi, M. & Locher, K. P. Nature 474, 350–354 (2011).
Ruiz-Canada, C., Kelleher, D. J. & Gilmore, R. Cell 136, 272–283 (2009).
Kelleher, D. J. & Gilmore, R. Glycobiology 16, 47R–62R (2006).
Wacker, M. et al. Science 298, 1790–1793 (2002).
Yan, Q. & Lennarz, W. J. J. Biol. Chem. 277, 47692–47700 (2002).
Igura, M. et al. Embo J. 27, 234–243 (2008).
Kowarik, M. et al. Science 314, 1148–1150 (2006).
Imperiali, B., Shannon, K. L., Unno, M. & Rickert, K. W. J. Am. Chem. Soc. 114, 7944–7945 (1992).
Bause, E. & Legler, G. Biochem. J. 195, 639–644 (1981).
Nilsson, I. et al. J. Cell Biol. 161, 715–725 (2003).
Wacker, M. et al. Proc. Natl Acad. Sci. USA 103, 7088–7093 (2006).
About this article
Biochimica et Biophysica Acta (BBA) - General Subjects (2015)
Crystallographic and NMR Evidence for Flexibility in Oligosaccharyltransferases and Its Catalytic Significance
Crystal structures of an archaeal oligosaccharyltransferase provide insights into the catalytic cycle of N-linked protein glycosylation
Proceedings of the National Academy of Sciences (2013)
Proceedings of the National Academy of Sciences (2012)
Journal of Theoretical Biology (2012)