Abstract
(Englisch)
|
Many proteins of Saccharomyces and other yeasts as well as humans are anchored to the membrane by means of a glycosylphosphatidylinositol (GPI) anchor. Thus these so called GPI proteins are covalently attached via their C-terminal amino acid to a glycophospholipid. Most of the GPI proteins are found at the cell surface, i.e. at the external side of the plasma membrane. Maybe half of the enzymes involved in elaborating the GPI structure and its attachment to the newly synthesized proteins in the ER have been identified. The lacking enzymes probably are to be found among the many open reading frames of unknown function that were uncovered in part by EUROFAN 1. Our particular interest was directed at finding genes involved in the lipid exchange which occurs shortly after a primary GPI anchor has been attached to newly synthesized proteins in The ER. This lipid exchange, also called lipid remodeling, introduces very long chain fatty acids (C26) into the GPI anchor and can involve several different exchange reactions. While practically all GPI anchored proteins are remodeled, some just exchange their original C16 and C18 fatty acids to incorporate C26. In others, the entire diayclglycerol moiety is exchanged for a ceramide of composition phytosphingosine-C26 (PHS-C26) or PHS-C26-OH. Ceramides with hydroxylated fatty acids (C26-OH) are only introduced in the Golgi while other remodeling reactions also can occur in the ER. At present, no selection procedures for the isolation of 'remodelase'-mutants can be devised, simply because the phenotype of a remodelase-deficient cell is totally unknown. In the framework of Eurofan 2 we thus undertook to screen mutants deleted for single ORF's for there capacity of remodeling. This was done by metabolically yeast cells with 3H-myoinositol, isolating and delipidating their proteins, further delipidating and purifying the glycoproteins by affinity chromatography over Concanavalin A-Sepharose, anchor peptides by protease treatment, finally purifying the anchor peptides over octyl-Sepharose, liberating the labeled lipid moiety by nitrous acid treatment and analysing the liberated lipids by TLC. The separated lipid moieties were finally quantitated by Berchtold radioscanning and visualized by fluorography. As this procedure is quite tedious, we only were able to analyze a selected number of strains which were chosen because they had deletions in genes of unknown function but showing some homology with phospholipases A, B or C, with enzymes synthesizing or utilizing ceramides, with acyltransferases transferring on any type of substrate, or with other proteins binding acyl-CoA. After having established the above described work up of GPI anchors from a purely analytical tool into a reproducible semi-quantitative procedure we thus labeled 52 strains having deletions generated during Eurofan 1 and 20 further strains which were collected from different labs. In all experiments the corresponding parental lines were run along with the examined test strains. Results for the 52 Eurofan strains are resumed in a attached excel file. The results are in form of a table which indicates the relative abundance of the three possible remodeling products, namely IPC/B = PHS-C26, IPC/C = PHS-C26-OH and 'PI' = a C26-containing phosphatidylinositol. Conclusions on the EUROFAN 1 strains can be summarized as follows. While the repeated analysis of the same strain in most cases yielded very reproducible results, there is significant difference between different wild type strains. Choosing as arbitrary criteria to label as elevated or reduced lipid levels which were 40% more or 40% less, respectively, than the mean lipid level in wt cells we found that of the 50 strains analyzed 6 showed elevated, 12 reduced relative amounts of IPC/C, 1 strain showed elevated levels of IPC/B, 1 showed elevated relative amounts of PI and 3 strains had too low levels of PI. No strain was completely deficient in one of the three anchor lipids which are generated by remodeling reactions. Thus, if any remodelase is encoded by the ORFs analyzed in here, there must be other ORFs with overlapping functions. The reduced levels of IPC/C may simply be secondary to a delay of transport of GPI proteins out of the ER, because IPC/C is only added to proteins having reached the Golgi. An attached table contains the abnormal strains with a comment on the reason why they were included in the screen.
|
