Objective: The aim was to examine whether mitochondrial Ca2+ fluxes are high enough to change mitochondrial and cytosolic calcium concentration during the contraction cycle. Methods: Isolated guinea pig ventricular myocytes were stimulated with paired voltage clamp pulses until contractions were maximal (2 mM [Ca2+]o, 36-degrees-C). At defined times of diastole or systole, the cells were shock frozen. Electronprobe microanalysis measured the concentration of total calcium in mitochondria (SIGMACa(mito)) and surrounding cytosol (SIGMACa(c)). Other experiments were performed to evaluate DNP sensitive mitochondrial Ca2+ uptake from depolarisation induced [Ca2+]c transients (K5indo-1 fluorescence). Results: At end of diastole, SIGMACa(mito) was 446 mumol-litre-1. During systole, SIGMACa(mito) increased with a 20 ms delay. A peak SIGMACa(mito) of 1050 mumol.litre-1 was measured 40 ms after start of systole, while 95 ms after start of systole SIGMACa(mito) had fallen to 530 mumol.litre-1. From the changes in SIGMACa(mito) the rates of net mitochondrial Ca2+ flux were estimated at 100 nmol.s-1.mg-1 protein for Ca2+ influx and 36 nmol.s-1.mg-1 protein for Ca2+ egress. Decay of SIGMACa(mito) was coupled to a rise in SIGMANa(mito). SIGMACl(mito) and SIGMAK(mito) rose and fell in parallel with SIGMACa(mito), suggesting Ca2+ activation of mitochondrial anion and cation channels. Activation of the non-specific permeability can be excluded. Block of mitochondrial Ca2+ uptake with DNP (100 muM) or FCCP (10 muM) increased the amplitude of the [Ca2+]c transients tor 1-3 min by about 50%; evaluation of mitochondrial Ca2+ uptake from DNP sensitive difference signals, however, was hampered by sequestration of mitochondrial Ca2+ into the sarcoplasmic reticulum. Conclusions: Mitochondrial calcium content changes during each individual contraction cycle; a substantial amount of calcium is taken up during the systole and released during later systole and diastole.