Human CaCo-2 cells were incubated with [C-14]linoleic (18:2(n - 6)), [C-14]linolenic (18:3(n - 3)) and [H-3]eicosapentaenoic acid (20:5(n - 3)), and the interconversion of the radioactive fatty acids to higher homologues and their acylation into triacylglycerols (TG) and phospholipids were examined. An active conversion of [C-14]18:3 to [C-14]20: 5 and [C-14]docosapentoenoic acid (22:5(n - 3)) and of [H-3]20:5 to [H-3]22:5, but not to [H-3]docosahexaenoic acid (22:6(n - 3)) was observed. In relation to the amounts that had been incorporated into cellular phospholipids and TG, the interconversion of [C-14]18:3 clearly exceeded that of [C-14]18:2. Addition of 10-100 muM 18:2 or 10-50 muM arachidonic acid (20:4(n - 6)) increased the percent interconversion of [C-14]18:2 to [C-14]20:4. E.g., addition of 50 muM 20:4 increased the formation of [C-14]20:4 from 4.4+/-0.1% to 5.9+/-0.8%, decreased the incorporation into phospholipids from 64.8 +/- 6.3% to 31.4 +/- 1.2% and increased the incorporation into TG from 8.8 +/- 0.4% to 28.8 +/- 1.1%. In contrast, addition of 10-100 muM 18:3 or 20:5 significantly decreased the interconversion of both [C-14]18:2 and [C-14]18:3. E.g., addition of 50 muM 20:5 decreased the formation of [C-14]20:4 from [C-14]18:2 from 4.4 +/- 0.1% to 0.9 +/- 0.1%, whereas the effects on the acylation reactions were very similar to those of 20:4. 20:5 also decreased the formation of interconversion products from [C-14]18:3. 18:2 and 20:4 caused a smaller decrease in the formation of [C-14]20:5 and actually increased percent conversion to [C-14]22:5. The percent conversion of [H-3]20:5 to [H-3]22:5 was also increased by the addition of 50-100 muM unlabeled 20:5. [C-14]18:2 and [C-14]18:3 were predominantly incorporated into phosphatidylcholine (PC) whereas more of the radioactive 20:4, 20:5 and 22:5 was incorporated into phosphatidylethanolamine (PE). An active fatty acid interconversion catalyzed by DELTA6 and DELTA5 desaturases thus occurs in the human CaCo-2 cell line, whereas conversion of 20:5(n - 3) to 22:6(n - 3) could not be demonstrated. The desaturation-elongation pathway has a preference for 18:3(n - 3) and is subjected to an efficient feedback regulation by 20:5(n - 3). Formation of 22:5 increases with available 20:5 mass and by the presence of other polyunsaturated fatty acids competing with 20:5 for acylation into phospholipids.