Abstract

Collagen XVIII (ColXVIII) is a non-fibrillar collagen and proteoglycan that exists in three isoforms: short, medium and long. The medium and long isoforms contain a unique N-terminal domain of unknown function, DUF959, and our sequence-based secondary structure predictions indicated that DUF959 could be an intrinsically disordered domain. Recombinant DUF959 produced in mammalian cells consisted of ∼50% glycans and had a molecular mass of 63 kDa. Circular dichroism spectroscopy confirmed the disordered character of DUF959, and static light scattering indicated a monomeric state for glycosylated DUF959 in solution. Small-angle X-ray scattering showed DUF959 to be a highly extended, flexible molecule with a maximum dimension of ∼23 nm. Glycosidase treatment demonstrated considerable amounts of O-glycosylation, and expression of DUF959 in HEK293 SimpleCells capable of synthesizing only truncated O-glycans confirmed the presence of N-acetylgalactosamine-type O-glycans. The DUF959 sequence is characterized by numerous Ser and Thr residues, and this accounts for the finding that half of the recombinant protein consists of glycans. Thus, the medium and long ColXVIII isoforms contain at their extreme N-terminus a disordered, elongated and highly O-glycosylated mucin-like domain that is not found in other collagens, and we suggest naming it the Mucin-like domain in ColXVIII (MUCL-C18). As intrinsically disordered regions and their post-translational modifications are often involved in protein interactions, our findings may point towards a role of the flexible mucin-like domain of ColXVIII as an interaction hub affecting cell signaling. Moreover, the MUCL-C18 may also serve as a lubricant at cell–extracellular matrix interfaces.