A novel regulator of the synthesis of complex and hybrid N-glycans

Abstract

Protein N-glycosylation affects many biological functions of glycoconjugates in development and disease. It occurs in the endoplasmic reticulum (ER) and Golgi compartments, and is catalyzed by resident glycosyltransferases and glycosidases. The overall goal of my thesis was to identify a novel activity which affects glycosylation in Chinese hamster ovary (CHO) cells and in the mouse. In order to facilitate my search, the first aim was to characterize the glycome of CHO cells and CHO glycosylation mutants for use as hosts in identifying glycosylation changes induced by a novel activity. The second part was to characterize a novel regulator of N-glycosylation, NM_026233 , which was identified through database analyses. An N-terminal-tagged version of the encoded protein stably expressed in CHO cells specifically inhibited N-acetylglucosaminyltransferase I (GlcNAcT-I) activity and phenocopied the Lec1 CHO glycosylation mutant. Therefore, the new activity was termed GlcNAcT-I inhibitory protein (GnT1IP). GnT1IP transcripts are detected primarily in testis and their levels are highly regulated during spermatogenesis. A smaller transcript, GnT1IP-S, encodes a glycoprotein with a cleaved signal sequence and a transcript 132 nucleotides longer, GnT1IP-L, encodes a type II transmembrane protein. Membrane-bound forms of GnT1IP inhibit GlcNAcT-I and also cause GlcNAcT-I to mislocalize from the medial -Golgi to the ER, ERGIC and cis-Golgi. However, mislocalization of GlcNAcT-I with C-terminal KDEL to the ER did not inhibit its activity. CHO cells overexpressing GnT1IP-L were shown to exhibit markedly increase adhesion to TM4 Sertoli cells. Therefore GnTIIP may regulate germ cell N-glycosylation and affect their interaction with Sertoli cells during spermatogenesis. The third aim was to generate a mouse defective in GnT1IP to test this hypothesis. Bacterial artificial chromosome (BAC) and recommbineering-based techniques were applied in an effort to generate conditional GnTI IP mutant mice in order to understand GnT1IP functions in vivo. A number of targeted embryonic stem (ES) cells were isolated but homologous recombinants were not obtained. An alternative strategy would seem necessary to generate the desired mouse mutant.