Nuclear magnetic resonance (NMR) spectroscopy detects only free, unbound metabolites. We have therefore compared the free high-energy phosphate content of isolated perfused rat hearts (determined by 31P-NMR) with the total high-energy phosphates of the same hearts (determined by chemical analysis) to determine the fractions, if any, that are NMR invisible. Aerobic perfusion (40 min at 37 degrees C, Pi-free Krebs buffer) was followed by 10, 14, or 18 min total global ischemia and 30 min reperfusion (n = 6 in each group). Fully relaxed 31P-NMR spectra (40 scans using 90 degrees pulses at 15-s intervals) were collected at various times throughout the protocol, and the signal intensities of the beta-phosphate of ATP, phosphocreatine (PCr), and Pi were quantified using methylenediphosphonate as an external standard. Hearts were freeze clamped either before ischemia or at the end of reperfusion and were chemically assayed for ATP, PCr, and Pi. After 40 min of normoxia, the ATP and PCr contents determined by NMR were almost identical to the values determined by chemical analysis. However, only 39 +/- 8% of the total Pi was NMR visible. After reperfusion, after 14 or 18 min of ischemia, the proportion of NMR-visible ATP had decreased to 64 +/- 9% (P less than 0.005). After reperfusion after 18 min ischemia, the proportion of NMR-visible Pi had increased to 76 +/- 10% (P less than 0.05). In conclusion, whereas the total cellular content of PCr is always NMR visible, ischemia-reperfusion can alter the fraction of NMR-visible ATP and Pi.