Fura-2 is the most common dye to measure cytosolic Ca2+ concentrations ([Ca2+]i). To facilitate simultaneous imaging of many cells while preserving their cytosolic environment, fura-2 is often loaded into the cytosol in its membrane-permeant ester form. It has been reported that small amounts of fura-2 accumulate in intracellular compartments, an effect that is usually neglected. We show that either focal or non-focal stimulation methods induce large [Ca2+]i gradients in T-lymphocytes during both, Ca2+ release and Ca2+ influx across the plasma membrane. Interfering with mitochondrial Ca2+ homeostasis and by labeling mitochondria with MitoTracker, we demonstrate that [Ca2+]i gradients co-localize with mitochondria and are attributable to mitochondrial fura-2 sequestration. Gradients could not be avoided by different loading protocols, compromising measurements of "real" [Ca2+]i gradients following T-cell stimulation. They were observed in human blood and lamina propria lymphocytes, Jurkat T-cells, mast cells, but not to the same extent in HEK-293 cells. Finally, we show that T-lymphocytes can be efficiently loaded with the membrane-impermeant fura-2 salt by electroporation and by osmotic lysis of pinocytic vesicles, which result in the loss of [Ca2+]i gradients. These methods are therefore suitable to study localized Ca2+ signals in large populations of T-cells while preserving their cytosolic integrity.