Nuclear magnetic resonance (NMR) spectroscopy does not always detect the total metabolite content in isolated perfused rat hearts. Alterations in NMR peak areas therefore could be caused by a change either in the metabolite content per se or in its NMR visibility. We have therefore compared the ATP, phosphocreatine (PCr), and Pi content of hearts, as determined by 31P-NMR spectroscopy, with the total, chemically determined ATP, PCr, and Pi contents of the same hearts to determine the fractions, if any, that are NMR invisible under different perfusion conditions. The three perfusion buffers used contained 1) glucose, 2) glucose plus high K+, and 3) pyruvate. Fully relaxed 31P-NMR spectra were collected during the final 10 min of perfusion in each group, and the ATP, PCr, and Pi contents were quantified using methylene diphosphonate as an external standard. The hearts were then freeze-clamped and chemically assayed for ATP, PCr, and Pi. Under all three conditions, the NMR-determined ATP and PCr contents were almost identical to the chemically determined values. However, only a portion of the chemically determined Pi was NMR visible. During perfusion with glucose-containing buffer, 39 +/- 8% of the total Pi was NMR visible, and this decreased to 12 +/- 2% (P less than 0.01) during K+ arrest and to 9 +/- 5% (P less than 0.01) during perfusion with pyruvate-containing buffer. In conclusion, whereas the total cellular content of ATP and PCr is always NMR visible during normoxic perfusion, alterations in substrate and contractile status can change the fraction of NMR-visible Pi.
