The alpha- and beta-subunits are required for expression of catalytic activity in the hetero-dimeric glucosidase II complex from human liver.

The alpha- and beta-subunits of the hetero-dimeric glucosidase II complex from human liver were cloned and expressed in COS-1 cells. The 4106 bp full-length cDNA for the alpha-subunit contained a 2835 bp ORF encoding a 107 kDa polypeptide. The 2095 bp cDNA for the beta-subunit encodes a approximately 60 kDa protein in a continuous 1605 bp ORF. The alpha- and beta-subunits each contain two potential Asn-Xaa-Thr/Ser acceptor sites, with only one site in the alpha-subunit (Asn97) being glycosylated. Additional lambda-clones were isolated for each subunit containing in-frame insertions/deletions within the coding region, indicating alternative splicing. Analysis of different human tissues revealed approximately 4.4 kb and approximately 2.4 kb transcripts for alpha- and beta-subunit, respectively, consistent with their full-length cDNA. Coexpression of the alpha- and beta-subunits in COS-1 cells resulted in >4-fold increase of glucosidase II activity. An inactive protein was obtained, however, after transfection with the alpha-subunit alone, showing that both subunits are essential for expression of active glucosidase II. The observation that the enzyme, previously purified from pig liver and lacking the beta-subunit, was catalytically active indicates that the beta-subunit is involved in alpha-subunit maturation rather than being required for enzymatic activity once the alpha-subunit has acquired its mature form. The alpha-subunit is expressed in COS-1 cells as an ER-located protein, whether inactive or part of a catalytically active complex. This suggests that ER-localization of the alpha-subunit, when associated with the dimeric enzyme complex, is mediated by the C-terminal HDEL-signal in the beta-subunit, whereas the apparently incompletely folded form of the inactive alpha-subunit could be retained in the ER by the putative "glycoprotein-specific quality control machinery. "

Man9-mannosidase from pig liver is a type-II membrane protein that resides in the endoplasmic reticulum. cDNA cloning and expression of the enzyme in COS 1 cells.

Man9-mannosidase, one of three different alpha 1,2-exo-mannosidases known to be involved in N-linked oligosaccharide processing, has been cloned in lambda gt10, using a mixed-primed pig liver cDNA library. Three clones were isolated which allowed the reconstruction of a 2731-bp full-length cDNA. The cDNA construct contained a single open reading frame of 1977 bp, encoding a 659-residue polypeptide with a molecular mass of approximately 73 kDa. The Man9-mannosidase specificity of the cDNA construct was verified by the observation that all peptide sequences derived from a previously purified, catalytically active 49-kDa fragment were found within the coding region. The N-terminus of the 49-kDa fragment aligns with amino acid 175 of the translated cDNA, indicating that the catalytic activity is associated with the C-terminus. Transfection of COS 1 cells with the Man9-mannosidase cDNA gave rise to a > 30-fold over-expression of a 73-kDa protein whose catalytic properties, including substrate specificity, susceptibility towards alpha-mannosidase inhibitors and metal ion requirements, were similar to those of the 49-kDa enzyme fragment. Thus deletion of 174 N-terminal amino acids in the 73-kDa protein appears to have only marginal influence on the catalytic properties. Structural and hydrophobicity analysis of the coding region, as well as the results from tryptic degradation studies, point to pig liver Man9-mannosidase being a non-glycosylated type-II transmembrane protein. This protein contains a 48-residue cytosolic tail followed by a 22-residue membrane anchor (which probably functions as internal and non-cleavable signal sequence), a lumenal approximately 100-residue-stem region and a large 49-kDa C-terminal catalytic domain. As shown by immuno-fluorescence microscopy, the pig liver enzyme expressed in COS 1 cells, is resident in the endoplasmic reticulum, in contrast to COS 1 Man9-mannosidase from human kidney which is Golgi-located [Bieberich, E. & Bause, E. (1995) Eur. J. Biochem. 233, 644-649]. Localization of the porcine enzyme in the endoplasmic reticulum is consistent with immuno-electron-microscopic studies using pig hepatocytes. The different intracellular distribution of pig liver and human kidney Man9-mannosidase is, therefore, enzyme-specific rather than a COS-1-cell-typical phenomenon. Since we observe approximately 81% sequence similarity between the two alpha-mannosidases, we deduce that the localization in either endoplasmic reticulum or Golgi is likely to be sequence-dependent.