Challenges in O-glycan engineering of plants

Abstract

Plants are attractive alternative expression hosts for the production of recombinant proteins. Many therapeutic proteins are glycosylated with N- and O-glycosylation being the most prevalent forms of protein glycosylation. While N-glycans have already been modified in plants toward the formation of homogenous mammalian-type glycoforms with equal or improved biological function compared to mammalian-cell culture produced glycoproteins little attention has been paid to the modification of O-linked glycans. Recently, the first step of mammalian O-glycan biosynthesis has been accomplished in plants. However, as outlined in this short review there are important issues that have to be addressed in the future. These include: (i) elimination of potentially immunogenic or allergenic carbohydrate epitopes containing arabinosides or arabinogalactans, (ii) a detailed investigation of the interplay between engineered N- and O-glycosylation pathways to avoid competition for common metabolites like UDP-GlcNAc, and (iii) a deeper understanding of signals and mechanisms for distribution of glycan processing enzymes, which is a prerequisite for complete and homogenous glycosylation of recombinant proteins.

Keywords: glycoprotein therapeutics; glycosylation; metabolic engineering; molecular farming; secretory pathway.

FIGURE 1

O-glycosylation engineering steps required to produce the disialylated core 1 structure which is, for example, the major O-glycan on Ser-126 of human EPO (Tsuda et al., 1990) In addition to providing CMP-NeuAc in the Golgi and initiation of O-glycosylation by transfer of a GalNAc residue the biosynthesis requires the expression of a mammalian core 1 β1,3-galactosyltransferase together with its specific chaperone Cosmc as well as α2,3- and α2,6-sialyltransferases (Gill et al., 2011; Ju et al., 2011). These steps are depicted in blue and have not been engineered into plants so far. The generation of O-GalNAc residues has been shown (Daskalova et al., 2010; Yang et al., 2012) and the successful expression of the enzymes (GNE, UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase; NANS, N-acetylneuraminic acid phosphate synthase; CMAS, CMP-N-acetylneuraminic acid synthetase; NANP, N-acetylneuraminate-9-phosphate phosphatase, this dephosphorylation step is catalyzed by an endogenous plant enzyme) for CMP-sialic acid synthesis has been demonstrated in A. thaliana and N. benthamiana (Castilho et al., 2008, 2010). In total, the efficient generation of disialylated core 1 structures requires the expression of at least nine non-plant proteins. According to a recent publication the ectopic expression of a UDP-GalNAc transporter is not essential for efficient O-glycosylation initiation in N. benthamiana (Yang et al., 2012). The requirement of the specific chaperone Cosmc for core 1 β1,3-galactosyltransferase folding and activity needs to be tested in plants.