To dissect mechanisms of arachidonic acid (20:4) metabolism in pulmonary alveolar macrophages (PAM), two distinct cell populations were investigated, resident and BCG-activated rabbit alveolar macrophages. After purified resident PAM were labeled overnight with [3H]20:4, radioactivity was localized primarily within lyso(bis)phosphatidic acid (L(bis)PA) (13.1% +/- 1.7), phosphatidylethanolamine (PE) (22.8% +/- 0.8), and phosphatidylcholine (PC) (26.7% +/- 1.7), with lesser amounts recovered in phosphatidylserine plus phosphatidylinositol (PS/PI) (9.2 +/- 0.8%). By contrast, analysis of the phospholipid classes from prelabeled BCG-activated PAM revealed that the amount of [3H]20:4 contained in L(bis)PA was profoundly decreased (4.7% +/- 0.4), p less than 0.003), whereas [3H]20:4 contained within other BCG phospholipids remained unchanged. Moreover, L(bis)PA, which composed 18.6% +/- 1.2 of the total phospholipid phosphorus of resident PAM, was reduced to 4.1% +/- 0.1 in BCG-activated macrophages (p less than 0.01). Phospholipase A2 from snake venom or from pancreas failed to release 20:4 from L(bis)PA, and lipase (phospholipase A1) from Rhizopus delmar liberated no more than one-third of this arachidonate. These results suggest that much of the arachidonate is not mobilized by classical phospholipases A1 and A2. When [3H]20:4-labeled PAM were stimulated with 1 microM 12-O-tetradecanoyl-phorbol-13-acetate (TPA), a loss of [3H]20:4 was observed from L(bis)PA, PE, PC, and PS/PI, with a concomitant increase in the synthesis of Hete and leukotriene C4. BCG-activated PAM exposed to either TPA or 3.8 microM calcium ionophore A23187 liberated [3H]20:4 solely from PE and PC, with diminished 20:4 oxidative metabolism. Analysis of the specific radioactivities of phospholipids obtained from resident PAM prelabeled with [3H]20:4 or [32P]i demonstrated that the specific activity of [32P]L(bis)PA was negligible, whereas that of [3H]20:4 was quite high. In addition, L(bis)PA deacylation induced by TPA in resident PAM was always accompanied by a corresponding loss of [3H]20:4 from phosphatidylinositol (PI), suggesting that metabolism of this novel phospholipid proceeded by a deacylation-reacylation reaction rather than by de novo synthesis. BCG-activated PAM, which exhibited depressed eicosanoid formation, consistently failed to deacylate [3H]20:4 from L(bis)PA or PI. These studies demonstrate that, unlike 20:4 derived from PE and PC by BCG-activated PAM, L(bis)PA may indeed provide a novel source of 20:4 that is tightly coupled to the lipoxygenase pathway.