Human platelet lipids were enriched in vitro with different amounts of either docosahexaenoic acid (22:6n-3), eicosapentaenoic acid (20:5n-3) or linoleic acid (18:2n-6). Of the total fatty acid incorporated, between 82 and 95% was associated with the phospholipid (PL) fraction, with the remainder as either neutral lipid or hydroxy fatty acid. Within the PL fraction, the majority (64% of total) of each fatty acid was incorporated into phosphatidylcholine. It was found that platelet aggregation induced by the thromboxane A2/prostaglandin H2 mimetic (15S)-hydroxy-11,9-(epoxymethano)prosta-5Z,13E-dienoic acid (U46619) was inhibited after PL enrichment with 22:6n-3 or 20:5n-3, but not after 18:2n-6 enrichment. The specificity of 22:6n-3 and 20:5n-3 for U46619 activation was demonstrated by the finding that neither fatty acid significantly inhibited thromboxane A2/prostaglandin H2-independent aggregation induced by A23187 or thrombin. Furthermore, enrichment with 22:6n-3 or 20:5n-3 resulted in inhibition of [3H]U46619 specific binding, while enrichment with 18:2n-6 did not affect binding. Scatchard analysis revealed that thromboxane A2/prostaglandin H2 receptor affinity for [3H]U46619 decreased 4.8-fold following 22:6n-3 incorporation. These results demonstrate that platelet phospholipid enrichment with 22:6n-3 or 20:5n-3 results in a selective inhibition of thromboxane A2/prostaglandin H2 receptor function.