23 Genes, 23 years later

George Palade's pioneering work in the 1960s had established the secretory pathway and its connection to the biogenesis of organelles as a major theme in eukary-otic cell biology (Palade, 1975). His disciples, including David Sabatini and Gunter Blobel, extended Palade's morphological approaches to the next level of analysis, involving subcellular fractionation and biochemical re-constitution to understand the mechanism of secretory polypeptide synthesis and vectorial translocation (Red-man and Sabatini, 1966; Dobberstein and Blobel, 1975). However, as of the end of the 1970s, little progress had been made in identifying proteins required foi" the organization and execution of the secretory pathway. Beisson and colleagues developed a genetic approach to the discovery of proteins required for secretion to investigate mucocyst discharge in Paramecium (Beisson et al., 1976). Unfortunately, rudimentary and cumbersome genetic and molecular cloning techniques limited the utility of this system. However, Saccharo-myces cerevisiae was known to secrete glycoproteins (Van Rijn et al., 1972), and some intracellular organelles characteristic of a mammalian style secretory system had been visualized (Linnemans et al., 1977). Thus, a genetic approach in yeast, analogous to that used to dissect the cell division cycle (Hartwell, 1974), seemed appropriate to begin to investigate molecular aspects of the secretory pathway. We assumed that a clever genetic selection procedure would be necessary to identify rare, secretion-selective mutations. Unfortunately, neither of us knew much genetics , but our initial effort suggested that such mutants would not be so rare. In 1979, we published our first report on the isolation of a pleiotropic secretion mutant secl (Novick and Schekman, 1979). secl mutant cells accumulated secretory proteins and vesicles at the expense of cell wall and plasma membrane expansion. However, the procedure used in the isolation of this mutant, from a relatively small collection of temperature-sensitive growth mutants, was not exactly as we had described in that paper. The seemingly clever idea was to select against cells expressing an inducible plasma membrane permease, the SO42-permease, which was known to actively concentrate chromate, a toxic mimic of sulfate. Mutagenized cells were incubated at a restrictive temperature, induced for sulfate permease, and then incubated ovemight in the presence of chromate in an effort to kill cells competent in the cell surface expression of the permease, secl emerged from this selection, but the procedure was abandoned when we found, in a reconstruction experiment, that sec I mutant cells were actually killed more efficiently than wild-type cells during …