We have recently identified a novel amino-terminal fragment of pro-brain natriuretic peptide (NT-proBNP) in the circulation of humans, the concentration of which increases progressively as the left ventricle fails. To clarify the origins of NT-proBNP in experimental animals, we have developed an RIA for NT-proBNP based on residues 52-71 of ovine proBNP-(1103) and used it to study cardiac processing, secretion, and metabolism of BNP in sheep with cardiac overload induced by coronary artery ligation (CAL) or rapid left ventricular pacing (rLVP). The concentration of NT-proBNP in left atrial plasma extracts drawn from normal control sheep was threefold that of mature BNP. Size-exclusion and reverse-phase HPLC analyses of plasma extracts coupled to RIA revealed a single peak of immunoreactive (ir) NT-proBNP [~8,000 relative molecular weight (M_r)], quite distinct from a single peak of ir-mature BNP (~3,000 M_r). In contrast, ovine cardiac tissue contained only a single immunoreactive peak of high-molecular-weight BNP (~11,000 M_r), consistent in size with proBNP-(1103). Sampling from the cardiac coronary sinus in normal control sheep (n = 5) and sheep with CAL (n = 5) revealed that the molar ratio of NT-proBNP to mature BNP was similar. There was a significant gradient of both mature and NT-proBNP across the heart in normal sheep, whereas after CAL the gradient was significant for mature BNP only. In both forms of cardiac overload (CAL and rLVP), left atrial plasma levels of NT-proBNP were significantly increased above normal levels, in contrast with mature BNP levels, which were raised only in the rLVP group of animals. Blockade of natriuretic peptide metabolism in sheep with heart failure (induced by rLVP) raised mature BNP levels threefold but did not affect levels of NT-proBNP. In conclusion, these studies show that NT-proBNP is formed from proBNP stores during secretion and, compared with mature BNP, accumulates in plasma because metabolism of NT-proBNP appears to differ from that of mature BNP. Although its function, if any, remains unclear, plasma NT-proBNP may prove to be a sensitive marker of cardiac overload and/or decompensation.