No abstract

beta1-6 GlcNAc branching, a product of N-acetylglucosaminyltransferase V (GnT-V), is a key structure that is associated with malignant transformations and cancer metastasis. Although a number of reports concerning tumor metastasis-related glycoproteins that contain beta1-6 GlcNAc branching have appeared, the precise function of beta1-6 GlcNAc branching on glycoproteins remains to be elucidated. We previously reported on the importance of beta1-6 GlcNAc branching on matriptase in terms of proteolytic degradation in tumor metastasis. We report here that matriptase purified from GnT-V transfectant (beta1-6 GlcNAc matriptase) binds strongly to L4-PHA, which preferentially recognizes beta1-6 GlcNAc branches of tri- or tetraantennary sugar chains, indicating that the isolated matriptase contains beta1-6 GlcNAc branching. The beta1-6 GlcNAc matriptase was resistant to autodegradation, as well as trypsin digestion, compared with matriptase purified from mock-transfected cells. Furthermore, N-glycosidase-F treatment of beta1-6 GlcNAc matriptase greatly reduced its resistance to degradation. An analysis of matriptase mutants that do not contain potential N-glycosylation sites clearly shows that the beta1-6 GlcNAc branching on N-glycans attached to Asn 772 in the serine protease domain plays a major role in trypsin resistance. This is the first example of a demonstration of a direct relationship between beta1-6 GlcNAc branching and a biological function at the protein level.

UDP-N-acetylglucosamine:alpha-mannoside beta-1,6-N-acetylglucosaminyltransferase (GnT-V) catalyzes the formation of beta-1-6 GlcNAc branches on asparagine-linked oligosaccharides, which is directly linked to tumorigenesis. Our recent studies indicate that the secretion of matriptase from cancer cells is increased via the action of GnT-V, as evidenced by the fact that matriptase-bearing beta-1-6 GlcNAc branching is dramatically inhibited. In this study, we report on an investigation of the expression of GnT-V and matriptase in thyroid neoplasm tissues to determine the clinical significance on the co-expression of these two proteins in thyroid cancer. Although neither GnT-V nor matriptase was expressed in normal thyroid tissue, positive staining for matriptase and GnT-V was observed in 52/68 and 66/68 cases of papillary carcinoma, 3/23 and 10/23 cases of follicular carcinoma, 5/13 and 9/13 cases of follicular adenoma, and 11/28 and 6/28 cases of anaplastic carcinoma, respectively. Immunohistochemistry, as well as western blotting, showed that the expression of matriptase paralleled the expression to GnT-V. However, the expression of matriptase mRNA was not correlated with its protein levels, suggesting that the enhancement in matriptase expression could be regulated by a posttranslational modification such as glycosylation through GnT-V-mediated glycosylation. In papillary carcinoma, the levels of expression of both GnT-V and matriptase were significantly higher in tumors 1 cm or less in size (microcarcinoma) and in those without poorly differentiated lesions, and the two proteins were significantly correlated. In contrast, the prognosis of thyroid carcinoma after surgery was neither correlated with the expression GnT-V nor matriptase, because the levels of their expression were quite low in anaplastic (undifferentiated) carcinomas. These results suggest that prolonged stabilization of matriptase is stabilized by GnT-V-mediated glycosylation in vivo, thus extending its halftime and permitting it to play role in the early phases of papillary carcinoma, but not in its later phase progression.