The distribution, excretion and hepatic metabolism of [3H]microcystin-LR (sublethal i.v.) were measured in mice. Plasma elimination was biexponential with alpha- and beta-phase half-lives of 0.8 and 6.9 min, respectively. At 60 min, liver contained 67 +/- 4% of dose. Through the 6-day study the amount of hepatic radioactivity did not change whereas 23.7 +/- 1.7% of the dose was excreted; 9.2 +/- 1.0% in urine and 14.5 +/- 1.1% in feces. Approximately 60% of the urine and fecal radiolabel 6 and 12 hr postinjection was the parent toxin. Hepatic cytosol, which contained 70 +/- 2% of the hepatic radiolabel (1 hr through 6 days), was prepared for high-performance liquid chromatography analysis by heat denaturation, pronase digestion and C18 Sep Pak extraction. At 1 hr, 35 +/- 2% of the radiolabel was insoluble or C18 Sep Pak-bound; 43 +/- 3% was associated with a peak of retention time (rt) 6.6 min, and 16 +/- 3% with the parent toxin (rt 9.4 min). After 6 days, 8 +/- 1% was C18 Sep Pak-bound or insoluble; 5 +/- 0% occurred at rt 6.6 min, 17 +/- 1% with parent and 60 +/- 2% was associated with rt 8.1 min. Two other peaks, rt 4.9 and 5.6 min, appeared transiently. Analysis of hepatic cytosol by desalting chromatography under nondenaturing and denaturing conditions revealed that all of the radiolabel was associated with cytosolic components, and 83 +/- 5% was bound covalently through 1 day. By day 6 the amount of covalently bound isotope decreased to 42 +/- 11%. This is the first study to describe the long-term hepatic retention of microcystin toxin and documents putative detoxication products.