In yeast, there are at least two vesicle populations upon ER (endoplasmic reticulum) exit, one containing Gap1p (general aminoacid permease) and a glycosylated α-factor, gpαF (glycosylated proα-factor), and the other containing GPI (glycosylphosphatidylinositol)-anchored proteins, Gas1p (glycophospholipid-anchored surface protein) and Yps1p. We attempted to identify sorting determinants for this protein sorting event in the ER. We found that mutant Gas1 proteins that lack a GPI anchor and/or S/T region (serine- and threonine-rich region), two common characteristic features conserved among yeast GPI-anchored proteins, were still sorted away from Gap1p-containing vesicles. Furthermore, a mutant glycosylated α-factor, gpαGPI, which contains both the GPI anchor and S/T region from Gas1p, still entered Gap1p-containing vesicles, demonstrating that these conserved characteristics do not prevent proteins from entering Gap1p-containing vesicles. gpαF showed severely reduced budding efficiency in the absence of its ER exit receptor Erv29p, and this residual budding product no longer entered Gap1p-containing vesicles. These results suggest that the interaction of gpαF with Erv29p is essential for sorting into Gap1p-containing vesicles. We compared the detergent solubility of Gas1p and the gpαGPI in the ER with that in ER-derived vesicles. Both GPI-anchored proteins similarly partitioned into the DRM (detergent-resistant membrane) in the ER. Based on the fact that they entered different ER-derived vesicles, we conclude that DRM partitioning of GPI-anchored proteins is not the dominant determinant of protein sorting upon ER exit. Interestingly, upon incorporation into the ER-derived vesicles, gpαGPI was no longer detergent-insoluble, in contrast with the persistent detergent insolubility of Gas1p in the ER-derived vesicles. We present different explanations for the different behaviours of GPI-anchored proteins in distinct ER-derived vesicle populations.

You do not currently have access to this content.