5'-AMP-activated protein kinase (AMPK) has been proposed to be a pivotal factor in cellular responses to both acute exercise and exercise training. To investigate whether protein levels and gene expression of catalytic (alpha(1), alpha(2)) and regulatory (beta(1), beta(2), gamma(1), gamma(2), gamma(3)) AMPK subunits and exercise-induced AMPK activity are influenced by exercise training status, muscle biopsies were obtained from seven endurance exercise-trained and seven sedentary young healthy men. The alpha(1)- and alpha(2)-AMPK mRNA contents in trained subjects were both 117 +/- 2% of that in sedentary subjects (not significant), whereas mRNA for gamma(3) was 61 +/- 1% of that in sedentary subjects (not significant). The level of alpha(1)-AMPK protein in trained subjects was 185 +/- 34% of that in sedentary subjects (P < 0.05), whereas the levels of the remaining subunits (alpha(2), beta(1), beta(2), gamma(1), gamma(2), gamma(3)) were similar in trained and sedentary subjects. At the end of 20 min of cycle exercise at 80% of peak O(2) uptake, the increase in phosphorylation of alpha-AMPK (Thr(172)) was blunted in the trained group (138 +/- 38% above rest) compared with the sedentary group (353 +/- 63% above rest) (P < 0.05). Acetyl CoA-carboxylase beta-phosphorylation (Ser(221)), which is a marker for in vivo AMPK activity, was increased by exercise in both groups but to a lower level in trained subjects (32 +/- 5 arbitrary units) than in sedentary controls (45 +/- 1 arbitrary units) (P < 0.01). In conclusion, trained human skeletal muscle has increased alpha(1)-AMPK protein levels and blunted AMPK activation during exercise.