23Na and 31P NMR spectroscopy were used to follow intracellular [Na+] ([Na+]i) and energy metabolism in isolated, perfused rat hearts. During 30 min of Ca(2+)-free perfusion no significant change in [Na+]i could be detected, but during a subsequent 45 min period of ischemia [Na+]i rose significantly as expected, from 8.6 +/- 2.4 to 36.8 +/- 9.4 mM. In contrast, already during 30 min of Ca(2+)- and Mg(2+)-free perfusion [Na+]i rose significantly from 7.3 +/- 3.7 to 71.3 +/- 15.6 mM. During this period, the Na(+)-K+ ATPase was not limited by depletion of high energy phosphates, decrease of intracellular free Mg2+ or accumulation of inorganic phosphate. During the first 8 min of a subsequent period of ischemia, the rate of rise in [Na+]i even increased, suggesting that during the preceding period of Ca(2+)- and Mg(2+)-free perfusion, the Na(+)-K+ ATPase was indeed operative but apparently not coping with the large Na(+)-influx. Using verapamil, we could demonstrate that this large Na(+)-influx occurs through the L-type Ca2+ channels, and that both Mg2+ and verapamil can block this Na(+)-influx. Previously, we have demonstrated that [Na+]i does not play a role in the origin of the calcium paradox. The notion that an increased [Na+]i is a prerequisite for the calcium paradox to occur apparently results from experimental evidence obtained under conditions of low or absent Mg2+.