Previous studies have proposed that caffeine-induced activation of glucose transport in skeletal muscle is independent of AMP-activated protein kinase (AMPK) because alpha-AMPK Thr172 phosphorylation was not increased by caffeine. However, our previous studies, as well as the present, show that AMPK phosphorylation measured in whole muscle lysate is not a good indicator of AMPK activation in rodent skeletal muscle. In lysates from incubated rat soleus muscle, a predominant model in previous caffeine-studies, both acetyl-CoA carboxylase-beta (ACCbeta) Ser221 and immunoprecipitated alpha(1)-AMPK activity increased with caffeine incubation, without changes in AMPK phosphorylation or immunoprecipitated alpha(2)-AMPK activity. This pattern was also observed in mouse soleus muscle, where only ACCbeta and alpha(1)-AMPK phosphorylation were increased following caffeine treatment. Preincubation with the selective CaMKK inhibitor STO-609 (5 microM), the CaM-competitive inhibitor KN-93 (10 microM), or the SR Ca(2+) release blocking agent dantrolene (10 microM) all inhibited ACCbeta phosphorylation and alpha(1)-AMPK phosphorylation, suggesting that SR Ca(2+) release may work through a CaMKK-AMPK pathway. Caffeine-stimulated 2-deoxyglucose (2DG) uptake reflected the AMPK activation pattern, being increased with caffeine and inhibited by STO-609, KN-93, or dantrolene. The inhibition of 2DG uptake is likely causally linked to AMPK activation, since muscle-specific expression of a kinase-dead AMPK construct greatly reduced caffeine-stimulated 2DG uptake in mouse soleus. We conclude that a SR Ca(2+)-activated CaMKK may control alpha(1)-AMPK activation and be necessary for caffeine-stimulated glucose uptake in mouse soleus muscle.