We investigated the role of p38 mitogen-activated protein kinase (p38 MAPK) in prostaglandin F2 alpha (PGF_2α)-induced contraction of isolated rat small intra-pulmonary arteries (IPA). Male Wistar rats were killed by cervical dislocation in accordance with Home Office Regulations and IPA (internal diameter 200 -500μm) were dissected free of surrounding tissue, mounted on a myograph, bathed in Krebs salt buffer at 37°C and gassed with 5 % CO₂/balance air. Student t-test was used for all statistical comparisons. In Western Blot analysis of isolated rat IPA, 20 μM PGF_2α significantly increased levels of the phosphorylated forms of p38 MAPK (168 ± 20% of control, n = 16, P< 0.01) and of its downstream effector, heat shock protein 27 (HSP27) (278 ± 77% of control, n = 21, P< 0.05). Both increases were reversed by co-incubation with 2 μM SB203580 (77 ± 7% of control for p38 MAPK, n = 5; 71 ± 16% of control for HSP27, n = 8), but not by 2μM SB202474 (194 ± 53% of control for p38 MAPK, n = 7; 276 ± 66% of control for HSP27, n = 8). IPA constricted with 20μM PGF_2α were relaxed by SB203580 with an apparent EC₅₀ of 1.6 ± 0.4 μM (n = 12) and a near maximal effect at 30 μM. The inactive analogue SB202474 was approximately 30-fold less potent (only 2.8 ± 1.4% relaxation at 1 μM and only 39 ± 6% relaxation at 20 μM, n = 10). SB203580 was unable (up to 30 μM) to relax the PGF_2α-induced contraction that remained after Ca²⁺ was removed from and 1 mM EGTA was added to the buffer (n = 6), and was unable (at 2 μM) to relax PGF-induced contractions in α-toxin permeabilised IPA, clamped at pCa 7.0 (n = 7), suggesting that the effect was dependent on Ca²⁺-influx and did not involve Ca²⁺-sensitization. Endothelial denudation caused a large right-ward shift in sensitivity to SB203580 (EC₅₀ 16 ± 2.3 μM, n = 10, P< 0.001 vs. control), such that it was no longer distinguishable from the response to SB202474. The relaxation response to SB202474 however, was not affected by endothelial-denudation. The SB203580 dose response was also significantly shifted to the right by 1 mM L-NAME (EC₅₀ = 12 ± 3 μM, n = 13, P< 0.01 vs. control) and by 1 μM indomethacin (EC₅₀ = 5.6 ± 1 μM, n = 11, P< 0.01), indicating the involvement of nitric oxide and prostacyclin in the relaxation response. Relaxation of endothelium-denuded arteries to the NO donor SNAP were not significantly altered by co-incubation with 1 μM SB203580, suggesting that SB203580 is acting on NO release rather than altering the action of NO on the smooth muscle. These results may suggest a role for the p38 MAPK/HSP27 pathway in mediating agonist-induced contraction and endothelium-dependent relaxation in rat IPA.