Contrasting roles for receptor-stimulated inositol lipid metabolism in secretory cells.

the cells in the absence of lysis and without loss of plasmamembrane skeletal protein unless fusion of the nuclear and plasma membranes had occurred, followed by exovesiculation of the resulting expanded surface bilayer. We have further investigated the factors which control this putative membrane fusion event in the hope that an understanding of this process in chicken erythrocytes may give us some pointers to the mechanism of fusion between intracellular granules and plasma membrane in other (more interesting!) cells. It seems quite clear from our studies that raised intracellular Ca2+ exerts a variety of important effects on chicken erythrocytes. As in other cell types (Campbell, 1983) Ca2+ causes a dissolution of microtubules which in the untreated chicken erythrocyte may prevent interaction of the nucleus with plasma membrane. In addition Ca2+ entry causes an increased degradation of sphingomyelin by an endogenous plasma-membrane enzyme which, curiously, does not require Ca2+ for its activity in vitro. More dramatically, Caa+ activates a cytosolic protease that specifically attacks a-spectrin, a major constituent of the plasma membrane skeleton, microtubuleassociated proteins and also a 5OkDa protein, which is probably a constituent of the intermediate filament network holding the nucleus in its normal central position in the cell (Granger et a/., 1982). This protease activity, which appears to require a free sulphydryl group, is essential not only for the fusion of nuclear and plasma membranes but also for the release of microvesicles, since neither of these events occurs when the protease is blocked by iodoacetamide or tosyllysylchloromethylketone. (Thomas et al., 1983). Thus we can understand the process leading to quasiexocytosis of the nucleus and to microvesicle release in terms of a Ca'+-induced removal of those cellular structures which normally prevent interaction and fusion of nuclear and plasma membranes. Of these effects of Ca2+, the most important seems to be activation of the endogenous protease, whose action seems to be first to degrade the intermediate filaments anchoring the nucleus, then to break down the microtubule-associated proteins which may be involved in attachment of microtubules to the plasma membrane and finally to disrupt the membrane skeleton, thus making more likely an interaction between nucleus and plasma membrane bilayer. At this stage however, we cannot exclude the suggestion that some Ca2+-induced changes in the structure of the bilayer itself, occurring as a result either of protease or sphingomyelinase activity, or through a direct interaction of Ca2+ with membrane lipids, may be essential for the completion of the fusion events. It is worth noting that the product of sphingomyelinase activity is ceramide, the sphingolipid analogue of diacylglycerol, a molecule which has received some attention as a possible modulator of fusogenic events and of protein kinase activity (Allan & Michell, 1979; Kishimoto et al., 1980). The above observations in chicken erythrocytes raise the possibility that fusion of intracellular granules with plasma membrane in other cellular systems may also require the activation of a protease. Certainly, Ca2++-sensitive proteases are common in a great variety of different cell types (Murachi, 1983) and in platelets, Ca2+ -dependent degradation of certain proteins accompanies exocytotic release, (Phillips & Jacabova 1977; Fox et al., 1983). Recently Strittmatter and his colleagues have made a determined effort to link membrane fusion events (including secretory processes) with protease activation, (Couch & Strittmatter, 1983; Baxter et al., 1983) but in these experiments it seems that protease activity is dependent on heavy metals since it is inactivated by o-phenanthroline. It is not clear how the activity of such an enzyme could be controlled after a stimulus which raised intracellular Caa+ concentration. It is plain that the evidence in favour of a role for protease activation in the membrane fusion events involved in secretion is not overwhelming. However, it is only recently that the structure of the cytoskeleton of secretory cells has been investigated in detail and, furthermore, it is likely to be difficult to detect changes in individual skeletal polypeptides in such a complex system as the typical secretory cell where, in general, plasma membrane-associated proteins will be a relatively small proportion of the total. More sensitive techniques may be required to convincingly demonstrate the need for protease activation in exocytosis.