The mitochondrial inhibitors NaN(3) and carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone (FCCP) were used to study the role of mitochondria in pancreatic B-cell Ca2+ homeostasis. In glucose-stimulated B-cells NaN(3) and FCCP both increased the K(ATP) current and thus hyperpolarized the cell membrane potential, as expected for agents depleting cellular ATP. NaN(3) and FCCP stopped the glucose-induced oscillations in the cytosolic free Ca2+ concentration ([Ca2+](c)) and elicited a biphasic response. After a first rapid and transient increase, [Ca2+](c) rose in a second slow phase to a sustained level. In cells pretreated with thapsigargin the first inhibitor-induced rise in [Ca2+](c) was absent, suggesting that it may be due to Ca2+ mobilization from intracellular stores. The glucose-induced oscillations were terminated again by NaN(3) and FCCP, respectively, but the slow increase in [Ca2+](c)of the second phase was still present. A minute increase in [Ca2+](c)elicited by NaN(3) or FCCP was even visible after the removal of extracellular Ca2+, suggesting that the inhibitors also mobilize Ca2+ from mitochondria. NaN(3) and FCCP induced Ca2+ influx into B-cells treated with low glucose concentrations whose voltage-dependent Ca2+ channels are closed. Experiments with thapsigargin-preincubated cells indicate that disturbance of mitochondrial function stimulates Ca2+ influx through voltage-independent Ca2+ pathways. During the NaN(3)-induced increase in [Ca2+](c), K+-elicited depolarizations of the cells did not further augment [Ca2+](c). Evidently, this is due to a direct inhibitory effect of azide on L-type Ca2+ channels. The data demonstrate that disturbing the mitochondrial function affects cellular Ca2+ homeostasis in B-cells at several sites. Thus, it is concluded that intact mitochondrial function is a prerequisite for regular Ca2+ handling in B-cells.