AMP-activated protein kinase (AMPK) is a multisubstrate enzyme activated by increases in AMP during metabolic stress caused by exercise, hypoxia, lack of cell nutrients, as well as hormones, including adiponectin and leptin. Furthermore, metformin and rosiglitazone, frontline drugs used for the treatment of type II diabetes, activate AMPK. Mammalian AMPK is an alphabetagamma heterotrimer with multiple isoforms of each subunit comprising alpha1, alpha2, beta1, beta2, gamma1, gamma2, and gamma3, which have varying tissue and subcellular expression. Mutations in the AMPK gamma subunit cause glycogen storage disease in humans, but the molecular relationship between glycogen and the AMPK/Snf1p kinase subfamily has not been apparent. We show that the AMPK beta subunit contains a functional glycogen binding domain (beta-GBD) that is most closely related to isoamylase domains found in glycogen and starch branching enzymes. Mutation of key glycogen binding residues, predicted by molecular modeling, completely abolished beta-GBD binding to glycogen. AMPK binds to glycogen but retains full activity. Overexpressed AMPK beta1 localized to specific mammalian subcellular structures that corresponded with the expression pattern of glycogen phosphorylase. Glycogen binding provides an architectural link between AMPK and a major cellular energy store and juxtaposes AMPK to glycogen bound phosphatases.

The AMP-activated protein kinase (AMPK) is a metabolic-stress-sensing protein kinase that regulates metabolism in response to energy demand and supply by directly phosphorylating rate-limiting enzymes in metabolic pathways as well as controlling gene expression.

The effect of exercise intensity on skeletal muscle AMP-activated protein kinase (AMPK) signaling and substrate metabolism was examined in eight men cycling for 20 min at each of three sequential intensities: low (40 ؎ 2% VO 2 peak), medium (59 ؎ 1% VO 2 peak), and high (79 ؎ 1% VO 2 peak). Muscle free AMP/ATP ratio only increased at the two higher exercise intensities (P < 0.05). AMPK ␣1 (1.5-fold) and AMPK ␣2 (5-fold) activities increased from low to medium intensity, with AMPK ␣2 activity increasing further from medium to high intensity. The upstream AMPK kinase activity was substantial at rest and only increased 50% with exercise, indicating that, initially, signaling through AMPK did not require AMPK kinase posttranslational modification. Acetyl-CoA carboxylase (ACC)-␤ phosphorylation was sensitive to exercise, increasing threefold from rest to low intensity, whereas neuronal NO synthase (nNOS) phosphorylation was only observed at the higher exercise intensities. Glucose disappearance (tracer) did not increase from rest to low intensity, but increased sequentially from low to medium to high intensity. Calculated fat oxidation increased from rest to low intensity in parallel with ACC␤ phosphorylation, then declined during high intensity. These results indicate that ACC␤ phosphorylation is especially sensitive to exercise and tightly coupled to AMPK signaling and that AMPK activation does not depend on AMPK kinase activation during exercise.