Divergent secretory behavior of the opposite ends of aggrecan

Date of Completion

January 1996


Biology, Molecular|Biology, Genetics|Biology, Cell




Secreted proteoglycans probably undergo surveillance by molecular chaperones in the endoplasmic reticulum (ER); the chaperones determine if proper conformation is present, serving as gatekeepers for proteoglycan core protein entry into the Golgi stacks. Such entry requires proper conformation of at least one core protein globular domain. Meanwhile, non-globular structures may not be subject to chaperone surveillance. We postulate that the C-terminal G3 domain of proteoglycans is a major conformer for quality control surveillance by chaperones. This postulate is based upon the following considerations: (1) a subdomain of G3 is a common motif in secreted and cell surface proteins; (2) lack of G3 causes proteoglycan core protein retention in the ER and its non-entry into Golgi stacks; (3) G3 seems expendable following proteoglycan secretion. We compared the secretory behavior of the globular ends of aggrecan, namely G1 (N-terminus) and G3 (C-terminus), in separate constructs containing endogenous consensus sites for glycosaminoglycan (GAG) chain addition. The consensus sites served as reporter groups for Golgi transit. Each construct was separately transfected into Chinese hamster ovary K1 (CHO) cells. The transfection system had been previously optimized using chloramphenicol acetyltransferase instead of G1 or G3; expression was $>$500 times basal levels per unit of cellular protein. Transfected G3 was readily expressed as shown by Northern analysis, immunocytology, and secretion of a neoproteoglycan. G1 expression was detected by its appearance as a new protein in cells and spent medium, G3 appeared in the medium as sulfated neoproteoglycan whose GAG chains were largely digestible by chondroitinase ABC. In contrast, G1 was released as a protein without any apparent GAG chains. The core proteins of G1 and G3 each migrated on SDS/PAGE at a larger size than predicted. This was confirmed by transfection into mutant CHO cells which were deficient in the initial two steps of GAG chain addition. Both G1 and G3 were detectable in wild type and mutant cells but differed greatly in their rates of secretion. G3 was rapidly secreted whereas G1 was retained within the cells. Furthermore, a construct containing consensus sites alone was readily modified with GAG chains and secreted, indicating that chaperone surveillance of globular structures dominates the trafficking of non-globular GAG sites from the ER to the Golgi. These composite studies indicate that: (1) G1 and G3 markedly differ in their secretory behavior; (2) G1 either moves through the Golgi without GAG chain addition or is released via a non-Golgi pathway and; (3) G3 secretion is not coupled to GAG chain initiation; (4) non-globular GAG sites follow the default secretory pathway and may not interact with molecular chaperones. ^