GA2/GM2/GD2 synthase localizes to the trans-golgi network of CHO-K1 cells

UDP-GalNAc:lactosylceramide/GM3/GD3 beta-1,4-N-acetylgalactosaminyltransferase (GalNAc-T) transforms its acceptors into the gangliosides GA2, GM2 and GD2. It is well established that it is a Golgi-located glycosyltransferase, but its sub-Golgi localization is still unclear. We addressed this question in Chinese hamster ovary K1 cell clones stably transfected with a c-myc-tagged version of GalNAc-T which express the enzyme at different levels of activity. In these cell clones we examined the effect of brefeldin A (BFA) on the synthesis of glycolipids (in metabolic-labelling experiments) and on the sub-Golgi localization of the GalNAc-T (by immunocytochemistry). We found that in cell clones expressing moderate levels of activity, GalNAc-T immunoreactivity behaved as the trans-Golgi network (TGN) marker mannose-6-P receptor (M6PR) both in BFA-treated and untreated cells, and that BFA completely blocked the synthesis of GM2, GM1 and GD1a. On the other hand, in cell clones expressing high levels of activity and treated with BFA, most GalNAc-T immunoreactivity redistributed to the endoplasmic reticulum, as did the medial-Golgi marker mannosidase II, and the synthesis of GM2, GM1 and GD1a was not completely blocked. These results indicate that GalNAc-T is a TGN-located enzyme and that the mechanism that localizes it to this compartment involves steps that, when saturated, lead to its mislocalization to the cis-, medial- or trans-Golgi. Changes of Golgi membrane properties by modification of local glycolipid composition due to the activity of the expressed enzyme were not the main cause of mislocalization, since it persists when glycolipid synthesis is inhibited with d, l-threo-1-phenyl-2-hexadecanoylamino-3-pyrrolidino-1-propanol-HCl.