| After proteins are synthesized in cytoplasma, they should be transported to different subcellular sites where they work, a process defined as protein sorting. Sorting signals embedded in proteins are essential for this process. For example, the secretion of proteins depends on the existence of leader sequence locating on the N terminus of proteins. By this characteristics, biologists developed various methods for the identification of different types of proteins, such as signal sequence trap system (SST) [1,2], nuclear localization signal trap system NLST) [3, 4] and the like.Previously, we established the suc2 signal sequence trap method, and screened a mouse embryonic cDNA library (day 11) with it. The results suggested that it was an efficient method for the identification of secretory proteins [5,6].However, with the coming of the post-genomic era, the development of bioinformatics and implement of cDNA-sequencing project make most of the genes of different model organisms available readily, so it dose not seem to be an economic way to identify secretory proteins by classic biological methods such as suc2-SST alone, although it's more efficient than other SSTs. In this work, wedesigned a new regime for identifying secretory molecules in large scale by recently developed bioinformatic methods combined with suc2-SST. In this method, the candidate molecules from public databases were firstly analysed by computer programs to predict whether there is a potential signal peptide on their N terminus and where will they be in cells, those who were predicted to own signal peptides and locate in secretory pathway (out of cells, on the plasma membrane, in the Golgi apparatus and the endoplasmic reticulum (ER)) were then put to suc2-SST to confirm the existence of a functional signal peptide.In a pilot experiment, we obtained 50 potential secretory proteins from about 1000 mouse hypothetical protein candidates. Out of them, thirteen molecules were randomly selected for further analysis by suc2-SST, and proven to have a functional signal peptide respectively. The subcellular location of two molecules named HYP36 and HYP32 were further determined in mammalian cells by confocol microscope observation, and proven to be in the secretory pathway. The results support that the combined strategy developed by us is a reliable method for rapid identification of secretory proteins on a large scale.In order to achieve the automatic analysis of secretory proteins, we developed a bioinformatic platform for the prediction of secretory proteins together with our colaborators. A systemic analysis of mouse hypothetical proteins in public databases gave 1329 secretory proteins out of 47716 candidates, and secretory proteins are about 2.8% of the total mouse hypothetical proteins analysed in this work.HYP36 and HYP32 were subjected to further analysis in detail. The mouse HYP36 gene, comprised of 11 exons, covering 7.0 kb or so in the C region of 17th chromosome, is transcribed into 4 kinds of transcripts at least. There is a hydrophobic signal peptide (24 amino acids) on the N terminus of HYP36 protein, and domain-motif analysis suggested the existence of 1 M6PR (mannose 6-phosphate receptor) domain and 2 N-linked glycosylation sites. BLAST resultssuggested that HYP36 was the mouse homologue of the human GlcNAc-1-phosphotransferase Y subunit (GenBank accession no. AAR19081). We show that this molecule was unbiquitously expressed in all the tissues examined and located mainly in Golgi apparatus as shown by RT-PCR and subcellular localization experiments. In addition, we found that over-expression of this molecule in NIH-3T3 cells induced an inclusion cells (I-cell) phenotype characterized by intra-cytosol vacuoles as shown by phase-contrast microscopy observation, and dramatic increase of lysosomes as shown by electron microscope. The activity of lysosomal enzymes ( P -Glucuronidase and P -galactosidase) decreased in cytosol and increased in the cultural media, as detected by 4-MUG ( 4-methylumbelliferyl-D-glucuronide )...
|
PDF Full Text Request