The most sensitive nonradiometric routine assay for angiotensin-converting enzyme (ACE) activity uses fluorometry to detect His-Leu released from Hip-His-Leu. Our results indicate that, in contrast to human serum, rat serum and plasma contain large and variable amounts of dipeptidase activity that lead to a subestimation of the ACE activity measured in 0.1 M potassium phosphate buffer, pH 8.3, containing 0.3 M NaCl, the most commonly used assay for human serum and tissue ACE. We describe and validate an assay for 1 to 10 microL rat and human serum or plasma using 5 mM Hip-His-Leu in 500 microL of 0.4 M sodium borate buffer, pH 8.3, containing 0.9 M NaC1 at 37 degrees C that reduced the subestimation error to less than or equal to 3% (rat serum) and less than or equal to 0.1% (human serum) and increased the ACE activity twofold to threefold. The Km and Vmax are reported for rat serum ACE (Hip-His-Leu) and dipeptidase (His-Leu) in borate buffer and phosphate buffer. Rat serum ACE hydrolysis of Hip-His-Leu measured by fluorometry correlated (r = 0.99, p less than 0.05) with the hydrolysis of angiotensin I measured by high-performance liquid chromatography. A direct method based on amino acid analysis is described for evaluating the dipeptidase error of complex mixtures such as tissue extracts and other physiological fluids. We have found that the assay can be used to measure ACE activity in 25 samples (in duplicate) in 2 hours with small intraassay (2.2%) and interassay (3.9%) coefficients of variation.(ABSTRACT TRUNCATED AT 250 WORDS)