We investigated the role of mitochondria (MT) in calcium signaling in a culture of rat aortic smooth muscle cells. Mitochondrial [Ca2+] ([Ca2+]MT) was measured using MT-targeted aequorin (1). Purinergic stimulation with ATP, which initiates a cytoplasmic [Ca2+] transient followed by a plateau, caused a large, transient increase in [Ca2+]MT . This transient was blocked by CPA, an inhibitor of the sarco-endoplasmic reticulum Ca2+-ATPase (SERCA), but was not affected by removal of extra-cellular Ca2+ immediately before the stimulus, indicating SR Ca2+ release as the source for MT Ca2+ uptake. The MT uncoupler FCCP abolished the MT response to ATP, while both the IP3-receptor (IP3R) inhibitor 2-APB and the ryanodine receptor (RyR) blocker procaine induced partial inhibition. Interestingly ATP-mediated [Ca2+]MT transients were synergistically and almost completely inhibited by combined blockade of IP3R and RyR indicating functional interaction between IP3R and RyR and confirming SR Ca2+ release as the source for the MT Ca2+-transient. Prolonged removal of extracellular Ca2+ decreased the MT response to ATP, and by inference SR Ca2+ content by close to 50% (1) . A role for Ca2+ extrusion by the Na+/Ca2+ exchanger (NCX) was suggested by the dependence of SR Ca2+ depletion on extracellular Na+. This confirms the earlier finding in intact vascular smooth muscle indicating that forward NCX unloads SR Ca2+. Furthermore, in the cultured smooth muscle IP3R-associated and RyR-associated SR-elements could be functionally separated, since procaine, but not 2-APB, blocked Ca2+-release from a sub-compartment of the SR that was depleted of Ca2+ via forward-mode NCX. We next investigated whether IP3R and RyR localize to the same or separate SR fractions and how they relate to MT. Confocal microscopy and semi-quantitative computation of co-localization of immuno-labeled RyR and IP3R-1 revealed that the two receptors localize to separate but interlaced reticular lattices (Poburko and van Breemen, unpublished observation). In addition, we found that co-localization of IP3R-1 and RyR was 1.4-times more likely in SR elements adjacent to MT. Using a novel computational model to estimate the likelihood of random co-localization, we found that the sub-cellular heterogeneity of co-localization was more likely due to increased SR density near MT than to RyR and IP3R associating with common sites on the MT. Although SR targeted D1ER revealed a continuous tubular/lamellar membrane network throughout the cytoplasm, the present finding of localization of IP3R-1 and RyR to separate SR elements could explain how IP3R and RyR might differentially influence cytosolic and MT Ca2+-signaling. We showed earlier that reverse-mode NCX mediates SR refilling during agonist-induced smooth muscle Ca2+ oscillations (2). To explore a possible role for MT in this refilling process, we measured [Ca2+]MT and the sub-plasmalemmal space ([Ca2+]subPM) with targeted aequorins and [Ca2+]i with fura-2. Substitution of extracellular Na+ with N-methyl-D-glucamine transiently increased [Ca2+]MT (~2mM) and [Ca2+]subPM (~1.3mM), followed by a decrease to sustained plateaus. In contrast, Na+-substitution caused a delayed and tonic increase in [Ca2+]i (< 100nM). Inhibition of Ca2+-uptake by the SR (30mM CPA) or MT (2mM FCCP or 2mM ruthenium red) enhanced the elevation of [Ca2+]subPM, and abolished the delay in the [Ca2+]i response to 0Na+, while increasing its amplitude (3). Altogether, the above indicates that smooth muscle MT take up Ca2+ released from the SR and Ca2+ influx via reverse NCX. However, if the NCX is to function in the physiological refilling of the SR during activation it also has to redirect Ca2+ towards the SR. To investigate this process we directly measured SR [Ca2+] using the targeted FRET-based indicator D1ER. Inhibiting the mitochondrial NCX with CGP-37157 impaired SR refilling during purinergic stimulation as did inhibition of the PM-NCX with KB-R7943. Over-expression of hFis1 to remove mitochondria from the sub-plasmalemmal space inhibited SR refilling to the same extend suggesting that a subpopulation of peripheral mitochondria funnel Ca2+ back into the SR during maintained smooth muscle stimulation (4). We conclude that mitochondrial Ca2+ signalling derives from Ca2+ uptake from cytoplasmic nanodomains defined by apposing SR and PM, where local [Ca2+] transiently rises well above global [Ca2+]. Furthermore we propose that subplasmalemmal MT ensure efficient SR refilling by cooperating with the plasmalemmal NCX to funnel Ca2+ into the SR.