Cell type differences in Golgi retention signals for transmembrane proteins.

The transmembrane domain of Golgi resident proteins such as beta-galactoside alpha 2,6-sialyltransferase (ST) and N-acetylglucosaminyltransferase 1 (NT) contain a Golgi retention signal which confers Golgi retention to reporter proteins appended to them in the appropriate context. Thus, chimeras of the cell surface protein dipeptidyl peptidase IV containing the transmembrane domain of ST and NT are retained in the Golgi apparatus in MDCK and COS cells, as assessed by indirect immunofluorescence microscopy. Transfection of these chimeric constructs into CHO cells, however, results in their transport to vesicular structures which do not colocalize with that of an endogenous Golgi marker, mannosidase II. Furthermore, the staining pattern of these structures are not affected by brefeldin A. Biochemical analysis of the transgene products in pulse-chase experiments revealed that the chimeric proteins eventually become resistant to endoglycosidase H, suggesting that they are transported beyond the medial Golgi and therefore the vesicular structures are likely to be post-Golgi. The vesicular structures colocalized well with a lysosomal marker, cathepsin D, and also with internalized FITC-dextran chased into the lysosomal compartment. Monitoring the cell surface appearance of the chimeric protein suggests that the majority is transported directly to the lysosomal compartment. Golgi retention can be completely restored for ST and improved for NT by the inclusion of sequences flanking the transmembrane domain. Our results reflect cell type differences in the interpretation of the transmembrane domain Golgi retention signal, established that general Golgi retention of type II glycosyltransferases requires the hydrophilic flanking sequence as well as the transmembrane domain, and demonstrate that proteins which escape Golgi retention may be channeled to the lysosomal pathway.