Substitute sweeteners: diverse bacterial oligosaccharyltransferases with unique N-glycosylation site preferences

Abstract

The central enzyme in the Campylobacter jejuni asparagine-linked glycosylation pathway is the oligosaccharyltransferase (OST), PglB, which transfers preassembled glycans to specific asparagine residues in target proteins. While C. jejuni PglB (CjPglB) can transfer many diverse glycan structures, the acceptor sites that it recognizes are restricted predominantly to those having a negatively charged residue in the −2 position relative to the asparagine. Here, we investigated the acceptor-site preferences for 23 homologs with natural sequence variation compared to CjPglB. Using an ectopic trans-complementation assay for CjPglB function in glycosylation-competent Escherichia coli, we demonstrated in vivo activity for 16 of the candidate OSTs. Interestingly, the OSTs from Campylobacter coli, Campylobacter upsaliensis, Desulfovibrio desulfuricans, Desulfovibrio gigas, and Desulfovibrio vulgaris, exhibited significantly relaxed specificity towards the −2 position compared to CjPglB. These enzymes glycosylated minimal N-X-T motifs in multiple targets and each followed unique, as yet unknown, rules governing acceptor-site preferences. One notable example is D. gigas PglB, which was the only bacterial OST to glycosylate the Fc domain of human immunoglobulin G at its native ‘QYNST’ sequon. Overall, we find that a subset of bacterial OSTs follow their own rules for acceptor-site specificity, thereby expanding the glycoengineering toolbox with previously unavailable biocatalytic diversity.

A cell-free platform for rapid synthesis and testing of active oligosaccharyltransferases

“…Desulfovibrio gigas (DgPglB), were expressed in CFPS containing 1 mg/mL nanodiscs using the optimized conditions from above. These OSTs have previously been demonstrated by Ollis et al to glycosylate scFv13-R4 DQNAT with the C. jejuni glycan in vivo (Ollis et al, 2015). Each of these OSTs was produced with approximately 100% solubility as determined by 14 C-leucine incorporation (Figure 3) and the soluble yields were comparable to that achieved for CjPglB.…”

“…In the future, we anticipate the use of our method for the synthesis and study of other single-subunit OSTs, especially given the speed, simplicity, and consistency of this CFPS method. An expression-based system for characterizing OSTs is needed, especially because of the significant challenges that exist in determining meaningful structure-function relationships (Maita et al, 2010 OSTs analyzed in that study were deemed inactive (Ollis et al, 2015). However, it is difficult to determine whether this was due to poor in vivo expression, or due to poor affinity of the enzymes for the given substrates.…”

“…The goal of this work was to develop a CFPS platform for the expression and characterization of OSTs. As a model OST, we chose to synthesize CjPglB because its lipid-linked oligosaccharide and amino acid sequon specificities are well-studied (Chen et al, 2007;Ollis et al, 2015). Additionally, CjPglB functions as a single subunit, making its study more tractable than multi-subunit OSTs found in higher organisms.…”

“…Because a key part of the catalytic cycle of the OST is to physically bind to the sequon before glycan transfer, simple modifications in the 5-amino acid sequon dictate the ability of the OST to exhibit glycosylation activity (e.g., modification of DQNAT to AQNAT changes the -2 sequon position and eliminates activity of the CjPglB on an otherwise identical protein) (Lizak et al, 2011;Ollis et al, 2015). Consistent with earlier in vivo experiments (Ollis et al, 2015), only CFPS-derived DgPglB was able to glycosylate purified scFv13-R4 AQNAT in vitro. It is notable that when IVG reactions were run with CFPS reaction product containing no OST, model glycoproteins were not glycosylated (Figure 4).…”

On-demand, cell-free biomanufacturing of conjugate vaccines at the point-of-care

“…Protein synthesis yields and glycosylation activity were reproducible across different batches of lysate at both small and large scale. To generate CjPglB-enriched lysate, CLM24 cells carrying plasmid pSF-CjPglB (Ollis et al, 2015) was used as the source strain. To generate FtO-PS-enriched lysates, CLM24 carrying plasmid pGAB2 (Cuccui et al, 2013) was used as the source strain.…”