Protein kinases and insulin action in fat cells.

the various guanine nucleotide regulatory proteins that have been fully characterized to date. It has been a matter of some contention as to whether insulin can exert an inhibitory effect upon adenylate cyclase activity. However, based upon our observations on the plasma-membrane cyclic AMP phosphodiesterase, it is apparent that such effects should only be observed at relatively high, but nevertheless physiological, GTP concentrations. Under such conditions we (Heyworth & Houslay, 19836) have indeed observed an inhibition of adenylate cyclase activity using low, physiological concentrations of insulin. As with insulin’s action on the plasmamembrane cyclic AMP phosphodiesterase, no inhibition can be observed in broken membranes obtained from glucagon-treated (‘uncoupled’) hepatocytes where the putative Nin has presumably been modified. Similarly, no effect of insulin on adenylate cyclase activity can be seen in membranes from hepatocytes treated with cholera toxin where, presumably, both N,, and Nin are fully activated. The major ADP-ribosylated band in hepatocyte plasma membranes is N, and this undoubtedly exerts a dominating stirnulatory influence on adenylate cyclase in membranes from cholera toxin-treated cells and in assays containing high glucagon concentrations (Heyworth & Houslay, 19836). These observations on the effect of insulin on adenylate cyclase activity lend support then to our contention that a distinct guanine nucleotide regulatory protein is involved in mediating certain of insulin’s actions. More recently we have been able to show that three distinct treatments block both the ability of glucagon to ‘uncouple’ adenylate cyclase activity and prevent insulin’s activation of the plasma membrane cyclic AMP phosphodiesterase. These are either the pretreatment or preexposure of hepatocytes to adenosine, Ca2 + 4onophore and islet activating protein (IAP) from pertussis toxin. We believe that this emphasizes the inter-relationships between glucagon’s ‘uncoupling’ of adenylate cyclase (Heyworth & Houslay, 1983a) and its ability to block insulin’s activation of the plasma-membrane phosphodiesterase (Heyworth et al., 1983b). Importantly, it shows that glucagon can modify certain of insulin’s actions at the level of the plasma membrane independently of its ability to affect intracellular cyclic AMP concentrations. Clearly the putative Nin cannot mediate all of the effects of insulin that are exerted upon the cell. This is demonstrated by the ‘dense-vesicle’ enzyme whose activity is synergistically increased by treatment of hepatocytes with both glucagon and insulin. We have provided evidence that this enzyme is activated by insulin through a route which is independent from that used to activate the plasma membrane phosphodiesterase. In this case activation appears to ensue as a result of the internalization and subsequent processing of the insulin receptor or insulin itself (Wilson et al., 1983). The concept that insulin might exert widespread effects on cellular functioning by controlling the activity of specific membrane-bound protein kinases has been suggested by a number of investigators (Denton et al., 1981; Houslay, 1981). The attraction of such a hypothesis is that it would provide the necessary degree of amplification and the diversity required to account for insulin’s cellular effects. We envisage here that the putative guanine nucleotide regulatory protein, Nin, may itself mediate insulin’s actions on membrane-bound protein kinases and perhaps even modulate the functioning of the kinase activity which is associated with the receptor itself.