G lucose stimulation of insulin release involves closure of ATP-sensitive K+ channels (K(+)-ATP channels), depolarization, and Ca2+ influx in B cells. However, by using diazoxide to open K(+)-ATP channels, and 30 mM K to depolarize the membrane, we could demonstrate that another mechanism exists, by which glucose can control insulin release independently from changes in K(+)-ATP channel activity and in membrane potential (Gembal et al. 1992. J. Clin. Invest. 89:1288-1295). A similar approach was followed here to investigate, with mouse islets, the nature of this newly identified mechanism. The membrane potential-independent increase in insulin release produced by glucose required metabolism of the sugar and was mimicked by other metabolized secretagogues. It also required elevated levels of cytoplasmic Cai2+, but was not due to further changes in Cai2+. It could not be ascribed to acceleration of phosphoinositide metabolism, or to activation of protein kinases A or C. Thus, glucose did not increase inositol phosphate levels and hardly affected cAMP levels. Moreover, increasing inositol phosphates by vasopressin or cAMP by forskolin, and activating protein kinase C by phorbol esters did not mimic the action of glucose on release, and down-regulation of protein kinase C did not prevent these effects. On the other hand, it correlated with an increase in the ATP/ADP ratio in islet cells. We suggest that the membrane potential-independent control of insulin release exerted by glucose involves changes in the energy state of B cells.