Glycosyltransferases that use polyisoprenol-linked donor substrates are categorized in the GT-C superfamily. In eukaryotes, they act in the endoplasmic reticulum (ER) lumen and are involved in N-glycosylation, glypiation, O-mannosylation, and C-mannosylation of proteins. We generated a membrane topology model of C-mannosyltransferases (DPY19 family) that concurred perfectly with the 13 transmembrane domains (TMDs) observed in oligosaccharyltransferases (STT3 family) structures. A multiple alignment of family members from diverse organisms highlighted the presence of only a few conserved amino acids between DPY19s and STT3s. Most of these residues were shown to be essential for DPY19 function and are positioned in luminal loops that showed high conservation within the DPY19 family. Multiple alignments of other eukaryotic GT-C families underlined the presence of similar conserved motifs in luminal loops, in all enzymes of the superfamily. Most GT-C enzymes are proposed to have an uneven number of TDMs with 11 (POMT, TMTC, ALG9, ALG12, PIGB, PIGV, and PIGZ) or 13 (DPY19, STT3, and ALG10) membrane-spanning helices. In contrast, PIGM, ALG3, ALG6, and ALG8 have 12 or 14 TMDs and display a C-terminal dilysine ER-retrieval motif oriented towards the cytoplasm. We propose that all members of the GT-C superfamily are evolutionary related enzymes with preserved membrane topology.

Mass spectrometry (MS) easily detects C-mannosylated peptides from purified proteins but not from complex biological samples. Enrichment of specific glycopeptides by lectin affinity prior to MS analysis has been widely applied to support glycopeptide identification but was until now not available for C-mannosylated peptides. Here, we used the α-mannose-specific Burkholderia cenocepacia lectin A (BC2L-A) and show that, in addition to its previously demonstrated high-mannose N-glycan binding capability, this lectin is able to retain C- and O-mannosylated peptides. Besides testing binding abilities to standard peptides, we applied BC2L-A affinity to enrich C-mannosylated peptides from complex samples of tryptic digests of HEK293 and MCF10A whole cell extracts, which led to the identification of novel C-mannosylation sites. In conclusion, BC2L-A enabled specific enrichment of C- and O-mannosylated peptides and might have superior properties over other mannose binding lectins for this purpose.

We identified ADAMTS16 as a target protein for Cmannosylation and showed that this modification is needed for proper secretion of ADAMTS16. Targeting a distinct Cmannosyltransferase by CRISPR-Cas9 in medaka fish embryos caused defects in eye development. We conclude that these developmental defects are caused by reduced secretion of ADAMTS16 when Cmannosylation is missing. Highlights• TSR1 of ADAMTS16 can be C-mannosylated.• Deletion of DPY19L1 or DPY19L3 in hiPSCs caused reduced secretion of ADAMTS16.• Targeting of dpy19l3 in medaka occasionally led to coloboma.