A Golgi retention signal in a membrane-spanning domain of coronavirus E1 protein

One current hypothesis involves the idea that proteins destined for constitutive secretion or insertion at the plasma membrane are transported by default with the bulk flow of lipid (35) . Proteins destined for lysosomes or secretory granules (in cells which perform regulated secretion) are directed by specific signals once they have traversed the Golgi complex . This hypothesis requires that resident proteins of the ER and Golgi complex have specific signals that cause their retention in the appropriate compartment . Evidence is accumulating to support this idea. A tripeptide which is presumed to lack any signals for transport is secreted rapidly from cells and defines the rate of"bulk flow" (47) . Retention signals for both soluble and membranebound ER proteins have been identified (14, 30, 31) . The mannose-6-phosphate modification on lysosomal hydrolases is recognized by a receptor in the Golgi complex which targets these proteins to lysosomes (20 their single membrane-spanning domains were replaced with the first membrane-spanning domain from El. Single amino acid substitutions in this sequence released retention of the chimeric G protein, as well as a mutant El protein which lacks the second and third membrane-spanning domains. The important feature of the retention sequence appears to be the uncharged polar residues which line one face of a predicted alpha helix . This is the first retention signal to be defined for a resident Golgi protein . The fact that it is present in a membrane-spanning domain suggests a novel mechanism of retention in which the membrane composition of the Golgi complex plays an instrumental role in retaining its resident proteins.