Secondary Logo

AMP-Activated Protein Kinase: A Key System Mediating Metabolic Responses to Exercise


Medicine & Science in Sports & Exercise: January 2004 - Volume 36 - Issue 1 - p 28-34
doi: 10.1249/01.MSS.0000106171.38299.64
CLINICAL SCIENCES: Symposium—Protecting Muscle ATP during Fatigue: How Metabolites Contribute to Crisis Management

HARDIE, D. G. AMP-Activated Protein Kinase: A Key System Mediating Metabolic Responses to Exercise. Med. Sci. Sports Exerc., Vol. 36, No. 1, pp. 28–34, 2004. The finding that AMP-activated protein kinase (AMPK) was activated by exercise in skeletal muscle, reported by Winder and Hardie in 1996, provided the first hint that this signaling pathway might represent the elusive and long-sought system that was responsible for the metabolic changes associated with exercise. It triggered an increasing volume of research that has now largely vindicated this hypothesis although, in the usual manner of these things, it is not the whole story. As discussed in this article, it is becoming clear that the AMPK system is partly, but not entirely, responsible for the acute metabolic responses of muscle to acute exercise. It is particularly involved in the switch from the anaerobic metabolism of glycogen to oxidative metabolism of blood glucose and fatty acids. It also appears to be responsible for most, if not all, of the long-term metabolic adaptations to aerobic exercise (i.e., to endurance training), particularly the up-regulation of mitochondrial content and oxidative metabolism. Interestingly, this role is a reflection of the evolutionary origins of the kinase, because the homolog of AMPK in a single-celled eukaryote, the brewer’s yeast Saccharomyces cerevisiae, is also involved in the switch from anaerobic to aerobic metabolism.

Division of Molecular Physiology, Faculty of Life Sciences, University of Dundee, Dundee, Scotland, UNITED KINGDOM

Address for correspondence: Prof. D. Grahame Hardie, Division of Molecular Physiology, School of Life Sciences, Wellcome Trust Biocentre, University of Dundee, Dow Street, Dundee, DD1 5EH, Scotland, United Kingdom; E-mail:

Submitted for publication March 2003.

Accepted for publication June 2003.

A protein kinase is an enzyme that alters the function of another protein by covalently attaching to it a phosphate group derived from ATP. It is now clear that protein kinases (along with the protein phosphatases that remove phosphate groups) are the major agency by which cellular function is altered in eukaryotes in response to external stimuli. There are no less than 500 genes encoding members of the “typical” eukaryotic protein kinase family in the human genome, two of which are alternate isoforms of the catalytic subunit of the AMP-activated protein kinase (AMPK), the subject of this review. It is important not to confuse AMP-activated protein kinase with the more well-known cyclic AMP-dependent protein kinase (PKA). AMPK is activated by 5′-AMP but not cyclic 3′, 5′-AMP, whereas the converse is true of PKA. AMPK is also termed “AMP-activated” and not “AMP-dependent” because (unlike PKA) it has a significant basal activity in the absence of the activating nucleotide.

The AMP-activated protein kinase was given its name by the author in 1988 (47), after our findings that earlier reports of protein factors in rat liver that caused ATP-dependent inactivation of 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR) and acetyl-CoA carboxylase (ACC) could be explained by a single-protein kinase that was activated by AMP (9). HMGR and ACC are the key regulatory enzymes of cholesterol and fatty acid synthesis, respectively, and at the time, it was not clear whether the kinase had a significant role outside of liver or outside of these particular pathways of lipid synthesis. A key development that facilitated its study was our development of the SAMS peptide assay. Previously it has been measured by its ability to inactivate either ACC or HMGR in the presence of ATP, methods that were both complex and difficult to use, particularly in crude cell extracts. Identification of the primary site of phosphorylation on rat liver ACC (Ser-79) introduced the possibility that a synthetic peptide might be used as a substrate. The SAMS peptide substrate was based on the sequence around Ser-79, except that a neighboring serine residue phosphorylated by PKA was changed to alanine, making it more specific for AMPK (15). Although other peptides can be used, the SAMS peptide assay has remained the “gold standard” for studies of AMPK. This assay was used to show that AMPK was present in rat skeletal muscle in 1989 (15), but the role of AMPK in this tissue only became apparent in 1996 with the report by Winder and the author that the kinase was activated by exercise (51). Using the same assay, it was subsequently shown that AMPK was activated in human skeletal muscle during exercise (20,55).

Back to Top | Article Outline


Although the catalytic subunit had been identified earlier, AMPK was not purified to homogeneity until 1994 (16) when it was shown to contain three subunits, now termed α, β, and γ. DNA encoding the catalytic (α) subunit was cloned in the same year (7) and its sequence shown to be related to that of the SNF1 protein kinase from a single-celled eukaryote, the brewer’s yeast Saccharomyces cerevisiae (11). In the presence of abundant glucose, yeast grows exclusively via fermentative (anaerobic) metabolism of glucose to ethanol, even if oxygen is available. However, once glucose starts to run out, the expression of numerous genes required for metabolism of alternative carbon sources (e.g., sucrose) can be switched on. This includes mitochondrial genes involved in oxidative metabolism, which are required for the metabolism of nonfermentable carbon sources such as ethanol. However, studies of mutants showed that growth on sucrose or on nonfermentable carbon sources required a functional SNF1 gene (the acronym SNF stands for sucrose, nonfermentable). This sheds interesting light on the evolution of the AMPK system because, as described in more detail below, AMPK is also involved in mammalian muscle in the switch from anaerobic to aerobic metabolism.

Mammalian AMPK and yeast SNF1 are highly conserved systems that exist as multi-subunit complexes containing catalytic α subunits and regulatory β and γ subunits. In mammals (including humans), there are two genes encoding alternate isoforms of the α subunit (α1 and α2), two encoding β subunits (β1 and β2), and three encoding γ subunits (γ1, γ2, and γ3). All different combinations of α, β, and γ appear to be able to form a complex, so that there are 12 possible isoform combinations even if alternative splicing of mRNA (which remains a possibility) is excluded. Although all cells probably express all seven subunits to some extent, α1, β1, and γ1 are the major forms in most cells, whereas α2, β2, γ2, and γ3 appear to be expressed at highest levels in cardiac and skeletal muscle. Although the role of the different isoforms remains incompletely understood, it is tempting to suggest that muscle expresses the additional isoforms because it is the tissue where the demand for ATP is subject to the most dramatic fluctuations. The catalytic (α) subunits contain N-terminal domains related to those of all other members of the “typical” eukaryotic protein kinase family. The β subunits contain a conserved C-terminal domain that is required for the formation of the αβγ complex and a central glycogen-binding domain whose exact function is unknown (28,41). The γ subunits appear to be involved in binding of the activator, AMP (13).

As signified by its name, AMPK is allosterically activated (up to 10-fold) by 5′-AMP (9). Although AMPK is a protein kinase and requires micromolar concentrations of ATP for use as substrate, high (mM) concentrations of ATP antagonize activation by AMP, probably by competing for binding at the allosteric site (14). AMPK is also activated by one or more upstream protein kinases that phosphorylate a threonine residue (Thr-172) in the “activation loop” of the kinase domain, equivalent to the region where other protein kinases (such as mitogen-activated protein (MAP) kinases) are phosphorylated and activated by upstream kinases. This phosphorylation causes >100-fold activation, so it is quantitatively more significant than the allosteric activation by AMP. However, AMP also stimulates the phosphorylation, both by binding to AMPK and making it a better substrate, and by activating the upstream kinase. AMP binding to AMPK also inhibits dephosphorylation by protein phosphatases, and the combination of these effects of AMP means that AMPK is an “ultrasensitive” system in that, over the critical range of AMP concentrations, a very small change produces a large change in the final output (23). Binding of high concentrations of ATP to the allosteric sites also inhibits phosphorylation by AMPKK and promotes dephosphorylation by protein phosphatases, thus tending to switch the system off.

Under what conditions does AMP rise in intact cells? Although AMP is produced in several cellular reactions, the most important source in quantitative terms appears to be the adenylate kinase reaction: 2ADP ↔ ATP +AMP. This is a readily reversible reaction and the enzyme (while also known as myokinase due to its abundance in muscle) appears to be very active in all mammalian cells. In healthy, unstressed cells (or in resting muscle) the ATP:ADP ratio is maintained at a high level (analogous to a well-charged battery), and the adenylate kinase reaction therefore runs from right to left, so that AMP is very low. However, if the cell experiences a stress that depletes ATP, the ATP:ADP ratio will fall (analogous to the battery becoming discharged), and adenylate kinase will run from left to right, leading to a large increase in AMP as ATP falls. These are exactly the conditions in which AMPK is activated, by the mechanisms described in the previous paragraph. Treatments that activate AMPK can either be stresses that interfere with ATP production, such as heat shock, metabolic poisons, glucose deprivation, hypoxia, or ischemia (reviewed in (22)) or stresses that increase ATP consumption, such as exercise in skeletal muscle (51). These findings led to the concept that the AMPK system acted as a “fuel gauge” or “cellular energy sensor” (22). This concept was reinforced by findings that AMPK was allosterically inhibited by physiologically relevant concentrations of phosphocreatine (42). Unlike AMP, phosphocreatine does not appear to affect the phosphorylation state of AMPK, and its binding site has not yet been identified.

Back to Top | Article Outline


Activation of AMPK in rat muscle occurs in response to both exercise (51) and electrical stimulation (29,49). Activation is dependent on exercise intensity (44), and activity remains elevated for up to 30 min after the cessation of exercise (43). Although muscle contains both isoforms of the catalytic subunit, the proportion of total AMPK activity contributed by the α1 and α2 isoforms appears to depend on the muscle type. In one study of rat muscle, α2 activity was several times higher than α1 activity in the largely “fast-twitch” red and white quadriceps, although they were almost equal in soleus (18), a “slow-twitch”, postural muscle. In most studies of rat or human muscle, activation of α2 complexes by exercise or contraction can be readily detected, whereas activation of α1 complexes is not. However, this may partly be because the much lower activity of α1 complexes in most muscle types makes measurement of any changes more difficult. Although significant activation of α1 (unlike α2) complexes could not be detected in human quadriceps muscle during moderate exercise (48), it was detected during vigorous sprint exercise (12).

The extent of activation of AMPK by exercise is markedly affected by prior training. In rats, activation of α2 complexes in red quadriceps by treadmill exercise of moderate intensity was greatly reduced in rats that had experienced seven weeks of prior treadmill training, compared with untrained controls with matched food intake (18). Similarly, in a cross-sectional study of well-trained human volunteers (mean O2peak = 66 mL·min−1·kg−1) compared with more sedentary control subjects (mean O2peak = 46 mL·min−1·kg−1), α2 phosphorylation and activation in quadriceps (vastus lateralis) muscle was lower in the trained group, even though in this study the exercise was performed at the same relative (80% of O2peak) rather than absolute intensity (38). This correlated with a smaller increase in phosphorylation of the target protein, ACC, in the trained group.

What is the explanation for the reduced activation of α2 complexes by exercise in endurance-trained animals or humans? No differences in muscle AMP or ATP levels could be detected, but another possibility is that the effect was due to the high muscle glycogen content, which was elevated in the trained group in both the animal (18) and the human (38) study. Consistent with this hypothesis, acute elevation of muscle glycogen content by prior manipulation of exercise and diet was also associated with reduced activation of AMPK triggered by perfusion with the pharmacological agent AICA riboside (see below) in rats (56) or by exercise in humans (54). This behavior makes physiological sense, because elevation of the cellular AMP:ATP ratio activates glycogenolysis and glycolysis via the classical allosteric regulation of phosphorylase and phosphofructokinase (and/or via phosphorylation of the former by phosphorylase kinase) without any requirement for AMPK. The interplay between these mechanisms would ensure that, if muscle glycogen stores were plentiful, glycogenolysis and anaerobic glycolysis would tend to be utilized preferentially for ATP synthesis. Only if or when glycogen stores became depleted would AMPK become activated and the aerobic oxidation of blood glucose and/or fatty acids be utilized instead. The molecular mechanism by which high cellular glycogen represses AMPK activation remains unclear, although it is tempting to suggest that it involves the glycogen-binding domain on the β subunit (28,41) acting as a “glycogen sensor.” However, in humans with McArdle’s disease (who cannot mobilize muscle glycogen due to a hereditary lack of muscle phosphorylase and consequently have muscle weakness), AMPK is hyper-activated by low levels of exercise, despite the high glycogen content (39). Thus, it appears to be the capability to rapidly mobilize glycogen, rather than high glycogen content per se, which somehow represses AMPK activation.

Intriguingly, AMPK is activated by exercise in tissues other than muscle, including liver and adipose tissue (10,40). The mechanism for this effect is not known, although one possibility is that it is an effect of catecholamine release, as AMPK has been reported to be activated by receptors coupled to the G protein Gq, including the α1b-adrenergic receptor (30). Whatever the mechanism, it means that changes in metabolism induced by AMPK activation in other tissues could also contribute to the health benefits of regular exercise.

Back to Top | Article Outline


The classical target for AMPK, acetyl-CoA carboxylase, is now known to occur as two isoforms, i.e., ACC-1 (also known as ACC-α), which is expressed in tissues active in fatty acid synthesis such as liver and adipose tissue, and ACC-2 (ACC-β), which is expressed in tissues active in fatty acid oxidation such as skeletal and cardiac muscle. Experiments with ACC-2 knockout mice (2) suggest that the malonyl-CoA produced by this isoform may be exclusively involved with regulation of fatty acid oxidation, rather than as a precursor for fatty acid synthesis (the likely preserve of ACC-1). Malonyl-CoA inhibits fatty acid oxidation by inhibiting carnitine:palmitoyl-CoA acyl transferase-1 (CPT1) on the outer surface of the inner mitochondrial membrane, thus preventing uptake of long chain fatty acids into mitochondria (34). ACC-2 localizes to mitochondria (1), and the malonyl-CoA it produces may therefore be presented directly to CPT1 and consequently play no role in cytoplasmic fatty acid synthesis. Although the sites phosphorylated by AMPK on ACC-2 (unlike ACC-1) have not been defined in detail, AMPK does phosphorylate and inactivate this isoform (53), and the use of a phospho-specific antibody confirms that the site on ACC-2 homologous with the key regulatory site on ACC-1 (Ser-79) is phosphorylated on human muscle ACC in response to exercise (12).

In order to identify which metabolic responses to exercise, if any, are mediated by AMPK activation, it was necessary to develop a method (other than exercise or contraction itself) to activate the kinase in muscle. An opportunity came with the discovery that 5-aminoimida z ole-4-carboxamide riboside m ono p hosphate (ZMP) activates AMPK in cell-free assays. The nonphosphorylated riboside (AICA riboside) was already known to be taken up by cells and be phosphorylated by cellular adenosine kinase to ZMP. ZMP is an analog of AMP, and it was found to mimic all of the effects of AMP on the AMPK system, including increased phosphorylation and decreased dephosphorylation (14). In 1995, two groups reported that incubation of rat hepatocytes with AICA riboside caused AMPK activation and, as expected (given that ACC-1 and HMGR were known targets in this tissue), inhibition of fatty acid and cholesterol synthesis (14,25).

The first report of the use of AICA riboside to activate AMPK in muscle was a study of the regulation of ACC-2. Consistent with the idea that ACC-2 is a target for AMPK, perfusion of rat hind-limb muscle with AICA riboside resulted in activation of AMPK, inactivation of ACC-2, reduction in the content of malonyl-CoA, and stimulation of palmitate oxidation (36). This provided strong evidence that AMPK activation during exercise could trigger increased fatty acid oxidation to generate ATP with which to sustain the exercise. At that time, the question of whether malonyl-CoA decreased during exercise in human muscle had been controversial. However, it was subsequently found that exercise did cause inactivation of ACC-2 and lowered malonyl-CoA in humans, although the decrease in malonyl-CoA was only detectable at high exercise intensities (17).

Could AMPK also have a role in regulating carbohydrate metabolism? In the original report of effects of AICA riboside on fatty acid oxidation (36), glucose uptake was also found to be stimulated, providing the first evidence that AMPK might be responsible for the increased glucose uptake observed during exercise or contraction. This study did not distinguish whether the effect was on glucose transport or glucose metabolism, but it was subsequently shown that transport of a nonmetabolizable analog such as 2-deoxyglucose was also stimulated by AICA riboside (24). The effects of AICA riboside on glucose transport, like the effects of contraction, were not blocked by the phosphatidylinositol-3-kinase inhibitor, wortmannin, showing that the mechanism was distinct from that of insulin. However AICA riboside, like insulin, caused translocation of the glucose transporter, GLUT4, to the plasma membrane (31).

Although the actual target for phosphorylation by AMPK that triggers translocation of GLUT4 to the plasma membrane remains unknown, these findings pointed to the conclusion that contraction and AICA riboside were activating glucose transport via a common mechanism. However, although they suggested that activation of AMPK was sufficient for stimulation of glucose uptake, they did not prove that it was necessary for stimulation of glucose uptake during muscle contraction. A more definitive test of this was carried out by Mu and coworkers (37), who constructed transgenic mice expressing an inactive mutant of the AMPK α2 subunit from a muscle creatine kinase promoter. Because the α subunits are unstable in the absence of β and γ, the overexpressed, inactive α2 subunit exerts a “dominant negative” effect by completely replacing the endogenous α1 and α2 subunits in the αβγ complex. Consequently, in these mice, AMPK activity is almost undetectable in muscle, and there is no increase in activity in response to contraction. Stimulation of 2-deoxyglucose uptake by AICA riboside was completely abolished in muscles from these mice, elegantly proving that the effects of this pharmacological agent on glucose transport are mediated by AMPK. However, stimulation of 2-deoxyglucose uptake by contraction was only partly abolished, suggesting that, although AMPK is involved, other mechanisms also play a part in this response. Intriguingly, these mice were intolerant of exercise and took much less voluntary exercise than control mice (37), presumably because they became fatigued more readily.

These results suggest that, although AMPK is involved in exercise-induced glucose uptake, it is not the whole story. An elegant synthesis of the available data has been suggested by Richter and colleagues (45). They suggest that there are at least two parallel mechanisms stimulating GLUT4 translocation in response to contraction in skeletal muscle: (i) a feed-forward mechanism signaling the onset of contraction, possibly mediated by increases in cytosolic Ca2+; and (ii) a feed-back mechanism, mediated by AMPK, that is only utilized when ATP starts to become depleted. Which of these two mechanisms is more important under particular circumstances may depend on a number of factors, such as the muscle type, the duration and intensity of exercise, the glycogen content, and whether or not prior training had occurred.

Along with ACC-2, another good candidate as a target for AMPK in skeletal muscle is glycogen synthase (GS). The muscle isoform of GS is phosphorylated by AMPK at Ser-7 in cell-free assays (8), and this “primes” the phosphorylation at a neighboring site (Ser-10) by the protein kinase CK1, leading to a marked inactivation of GS that is most evident at low concentrations of glucose-6-phosphate (G6P) (19). Perfusion of rat hind-limb muscle with AICA riboside also leads to inactivation of GS at low G6P, and this was associated with a reduced gel mobility that was reversed by protein phosphatase treatment (56). This is consistent with phosphorylation of GS by AMPK, although the actual sites phosphorylated under these conditions require further study. Studies of the effects of exercise on GS activity in vivo are complicated by the fact that glycogen content exerts a strong feed-back regulation on GS activity, so that any inactivation of GS mediated by AMPK might be obscured by activation caused by reduced glycogen content. However, this difficulty could be overcome by studying GS regulation in human subjects with McArdle’s disease, who have a hereditary lack of muscle phosphorylase and therefore cannot mobilize glycogen during exercise. In these subjects, where even the low intensity of exercise that they are able to tolerate caused a large degree of activation of AMPK, GS was inactivated at low G6P concentrations during exercise. This correlated with phosphorylation of ACC-2, a marker of AMPK activation. In normal subjects carrying out the same exercise intensity, AMPK activation was not significant and GS was activated, probably because of a decrease in muscle glycogen content (39). Another group (21) failed to find effects on glycogen synthase activity when isolated muscles were incubated with AICA riboside. However, interpretation of these experiments is complicated by the fact that ZMP, the nucleoside formed from AICA riboside inside the cell, not only activates AMPK (14) but also allosterically activates glycogen phosphorylase (33). Once again, any effects due to phosphorylation of glycogen synthase by AMPK could therefore be obscured by effects of decreased glycogen content. Analysis of the sites phosphorylated on glycogen synthase is needed to resolve these issues.

Another proposed target of AMPK in muscle is malonyl-CoA decarboxylase (MCD), suggested by reports that the enzyme is activated by contraction or by treatment of isolated muscle with AICA riboside (46). This is a potentially attractive mechanism because it would mean that a fall in malonyl-CoA during exercise could be produced both by inactivation of the enzyme that produces it (ACC-2) and by activation of the enzyme that degrades it (MCD). However, others have failed to reproduce this finding and also found that MCD was not a substrate for AMPK in cell-free assays (21).

Back to Top | Article Outline


The finding that AMPK was activated by exercise in muscle (51) suggested that it might also be responsible for some of the long-term adaptations to endurance exercise (such as the up-regulation of mitochondrial content and respiratory capacity (26)), in addition to the acute effects on metabolism already discussed. The first evidence in favor of this hypothesis came from studies by Holmes et al. (27) and Winder et al. (52) involving repeated dosing of rats with AICA riboside, which caused increased expression of GLUT4, hexokinase and mitochondrial enzymes. The effect on GLUT4 was due to increased transcription, as demonstrated using transgenic mice in which expression of a reporter gene was driven by the GLUT4 promoter. The actual target for AMPK responsible for the increased transcription is not known, although these workers obtained evidence for increased binding of the transcription factor, myocyte enhancer factor-2 (MEF2) to the promoter (58). The α2 isoform of the catalytic subunit is enriched in the nucleus of muscle cells (3), where it could phosphorylate nuclear factors, and a small increase in the amount of nuclear α2 has been reported after exercise (35). There is evidence that AMPK phosphorylates nuclear factors involved in regulating transcription in tissues other than muscle. The transcriptional coactivator, p300, has been found to be a target for AMPK both in cell-free assays and in intact cells, and phosphorylation decreased its interaction with transcription factors of the nuclear hormone receptor family such as PPAR-α (57). In liver cells, AMPK activation has been found to down-regulate the expression of several important transcription factors, including HNF-4α (32) and SREBP-1c (59).

The best evidence that AMPK is involved in the long-term adaptation to endurance exercise comes from studies with the transgenic mice lacking active AMPK in muscle, which were described earlier (37). Unfortunately, because these mice are intolerant of exercise the effect of endurance training itself could not be studied. However, AMPK can be activated in muscle by treatment in vivo with β-guanidinoproprionic acid (GPA), a creatine analog that replaces muscle creatine but is a poor substrate for creatine kinase and consequently causes depletion of phosphocreatine and ATP, and elevation of AMP. GPA treatment of normal mice caused many of the responses seen during endurance training, i.e., up-regulation of mitochondrial content and increased expression of the mitochondrial markers cytochrome c and δ-aminolevulinate synthase (ALS), but these effects were completely abolished in mice lacking muscle AMPK (60). Two other candidates as signaling proteins involved in up-regulation of mitochondrial function (based on studies in which they were overexpressed in muscles of transgenic mice) are calmodulin-dependent protein kinase IV (CaMKIV) and the transcriptional coactivator, peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α). Intriguingly, the expression of both of these was up-regulated by GPA in control mice but not in the mice lacking muscle AMPK (60). These results suggest that, if CaMKIV and PGC-1α are involved in up-regulating mitochondrial function, they may act downstream of AMPK.

Back to Top | Article Outline


Although much undoubtedly remains to be learned, the research reviewed above makes it clear that activation of AMPK is responsible for many of the acute metabolic changes that occur in muscle during exercise or contraction, particularly the increased glucose uptake, increased fatty acid oxidation, and decreased glycogen synthesis. Activation of AMPK occurs in response to decreased phosphocreatine and ATP and increased AMP, and acts as a feed-back mechanism that adjusts the supply of ATP to the demand. The effects of exercise and contraction on glucose transport are not entirely mediated by AMPK and there appears to be an additional mechanism, possibly mediated by calcium ions, although the details remain unclear. However, initial results using transgenic mice lacking functional AMPK in muscle (60) suggest that activation of AMPK may also be responsible for many of the long-term adaptations of muscle to sustained exercise (i.e., to endurance training), especially the increases in mitochondrial content and oxidative capacity. Intriguingly, this is similar to the role of the yeast homolog of AMPK (the SNF1 complex), which is required for the switch from anaerobic to aerobic metabolism, and for up-regulation of genes required for oxidative metabolism.

AMPK activation therefore stimulates, both acutely and chronically, the uptake and oxidation of glucose and fatty acids by skeletal muscle. This led to the suggestion, first made in a review by Winder and Hardie (50), that the beneficial effects of exercise on the development of Type 2 diabetes and the metabolic syndrome may be mediated by AMPK, and conversely that AMPK activators may be good candidates as novel drugs aimed at treatment of these conditions. In support of this, treatment with AICA riboside ameliorates insulin resistance and other metabolic defects of the metabolic syndrome in various animal models, e.g., genetically obese rats (4–6). As final proof of this concept, it has recently been reported that metformin, currently the most widely used drug for treatment of Type 2 diabetes, may function by activating AMPK (59).

Current studies in the author’s laboratory are funded by a Program Grant from the Wellcome Trust, by an RTD Contract (QLG1-CT-2001–01488) from the European Commission, and by Project Grants from Diabetes UK and the Medical Research Council.

Back to Top | Article Outline


1. Abu-Elheiga, L., W. R. Brinkley, L. Zhong, S. S. Chirala, G. Woldegiorgis, and S. J. Wakil. The subcellular localization of acetyl-CoA carboxylase 2. Proc. Natl. Acad. Sci. USA 97: 1444–1449, 2000.
2. Abu-Elheiga, L., M. M. Matzuk, K. A. Abo-Hashema, and S. J. Wakil. Continuous fatty acid oxidation and reduced fat storage in mice lacking acetyl-CoA carboxylase 2. Science 291: 2613–2616, 2001.
3. Ai, H., J. Ihlemann, Y. Hellsten, et al. Effect of fiber type and nutritional state on AICAR- and contraction-stimulated glucose transport in rat muscle. Am. J. Physiol. 282: E1291–E1300, 2002.
4. Bergeron, R., S. F. Previs, G. W. Cline, et al. Effect of 5-aminoimidazole-4-carboxamide-1-beta-D-ribofuranoside infusion on in vivo glucose and lipid metabolism in lean and obese Zucker rats. Diabetes 50: 1076–1082, 2001.
5. Buhl, E. S., N. Jessen, R. Pold, et al. Long-term AICAR administration reduces metabolic disturbances and lowers blood pressure in rats displaying features of the insulin resistance syndrome. Diabetes 51: 2199–2206, 2002.
6. Buhl, E. S., N. Jessen, O. Schmitz, et al. Chronic treatment with 5-aminoimidazole-4-carboxamide-1-beta-D-ribofuranoside increases insulin-stimulated glucose uptake and GLUT4 translocation in rat skeletal muscles in a fiber type-specific manner. Diabetes 50: 12–17, 2001.
7. Carling, D., K. Aguan, A. Woods, et al. Mammalian AMP-activated protein kinase is homologous to yeast and plant protein kinases involved in the regulation of carbon metabolism. J. Biol. Chem. 269: 11442–11448, 1994.
8. Carling, D., and D. G. Hardie. The substrate and sequence specificity of the AMP-activated protein kinase: phosphorylation of glycogen synthase and phosphorylase kinase. Biochim. Biophys. Acta 1012: 81–86, 1989.
9. Carling, D., V. A. Zammit, and D. G. Hardie. A common bicyclic protein kinase cascade inactivates the regulatory enzymes of fatty acid and cholesterol biosynthesis. FEBS Lett. 223: 217–222, 1987.
10. Carlson, C. L., and W. W. Winder. Liver AMP-activated protein kinase and acetyl-CoA carboxylase during and after exercise. J. Appl. Physiol. 86: 669–674, 1999.
11. Celenza, J. L., and M. Carlson. A yeast gene that is essential for release from glucose repression encodes a protein kinase. Science 233: 1175–1180, 1986.
12. Chen, Z. P., G. K. Mcconell, B. J. Michell, R. J. Snow, B. J. Canny, and B. E. Kemp. AMPK signaling in contracting human skeletal muscle: acetyl-CoA carboxylase and NO synthase phosphorylation. Am. J. Physiol. 279: E1202–E1206., 2000.
13. Cheung, P. C. F., I. P. Salt, S. P. Davies, D. G. Hardie, and D. Carling. Characterization of AMP-activated protein kinase g subunit isoforms and their role in AMP binding. Biochem. J. 346: 659–669, 2000.
14. Corton, J. M., J. G. Gillespie, S. A. Hawley, and D. G. Hardie. 5-Aminoimidazole-4-carboxamide ribonucleoside: a specific method for activating AMP-activated protein kinase in intact cells? Eur. J. Biochem. 229: 558–565, 1995.
15. Davies, S. P., D. Carling, and D. G. Hardie. Tissue distribution of the AMP-activated protein kinase, and lack of activation by cyclic AMP-dependent protein kinase, studied using a specific and sensitive peptide assay. Eur. J. Biochem. 186: 123–128, 1989.
16. Davies, S. P., S. A. Hawley, A. Woods, D. Carling, T. A. J. Haystead, and D. G. Hardie. Purification of the AMP-activated protein kinase on ATP-g-Sepharose and analysis of its subunit structure. Eur. J. Biochem. 223: 351–357, 1994.
17. Dean, D., J. R. Daugaard, M. E. Young, et al. Exercise diminishes the activity of acetyl-CoA carboxylase in human muscle. Diabetes 49: 1295–1300, 2000.
18. Durante, P. E., K. J. Mustard, S. H. Park, W. W. Winder, and D. G. Hardie. Effects of endurance training on activity and expression of AMP-activated protein kinase isoforms in rat muscles. Am. J. Physiol. 283: E178–E186, 2002.
19. Flotow, H., and P. J. Roach. Synergistic phosphorylation of rabbit muscle glycogen synthase by cyclic AMP-dependent protein kinase and casein kinase I. Implications for hormonal regulation of glycogen synthase. J. Biol. Chem. 264: 9126–9128, 1989.
20. Fujii, N., T. Hayashi, M. F. Hirshman, et al. Exercise induces isoform-specific increase in 5′AMP-activated protein kinase activity in human skeletal muscle. Biochem. Biophys. Res. Commun. 273: 1150–1155, 2000.
21. Habinowski, S. A., M. Hirshman, K. Sakamoto, et al. Malonyl-C oa decarboxylase is not a substrate of AMP-activated protein kinase in rat fast-twitch skeletal muscle or an islet cell line. Arch. Biochem. Biophys. 396: 71–79, 2001.
22. Hardie, D. G., and S. A. Hawley. AMP-activated protein kinase: the energy charge hypothesis revisited. BioEssays 23: 1112–1119, 2001.
23. Hardie, D. G., I. P. Salt, S. A. Hawley, and S. P. Davies. AMP-activated protein kinase: an ultrasensitive system for monitoring cellular energy charge. Biochem. J. 338: 717–722, 1999.
24. Hayashi, T., M. F. Hirshman, E. J. Kurth, W. W. Winder, and L. J. Goodyear. Evidence for 5′ AMP-activated protein kinase mediation of the effect of muscle contraction on glucose transport. Diabetes 47: 1369–1373, 1998.
25. Henin, N., M. F. Vincent, H. E. Gruber, and G. Van den Berghe. Inhibition of fatty acid and cholesterol synthesis by stimulation of AMP-activated protein kinase. FASEB J. 9: 541–546, 1995.
26. Holloszy, J. O., and E. F. Coyle. Adaptations of skeletal muscle to endurance exercise and their metabolic consequences. J. Appl. Physiol. 56: 831–838, 1984.
27. Holmes, B. F., E. J. Kurth-Kraczek, and W. W. Winder. Chronic activation of 5′-AMP-activated protein kinase increases GLUT-4, hexokinase, and glycogen in muscle. J. Appl. Physiol. 87: 1990–1995, 1999.
28. Hudson, E. R., D. A. Pan, J. James, et al. A novel domain in AMP-activated protein kinase causes glycogen storage bodies similar to those seen in hereditary cardiac arrhythmias. Curr. Biol. 13: 861–866, 2003.
29. Hutber, C. A., D. G. Hardie, and W. W. Winder. Electrical stimulation inactivates muscle acetyl-CoA carboxylase and increases AMP-activated protein kinase activity. Am. J. Physiol. 272: E262–E266, 1997.
30. Kishi, K., T. Yuasa, A. Minami, et al. AMP-activated protein kinase is activated by the stimulations of G(q)-coupled receptors. Biochem. Biophys. Res. Commun. 276: 16–22, 2000.
31. Kurth-Kraczek, E. J., M. F. Hirshman, L. J. Goodyear, and W. W. Winder. 5′ AMP-activated protein kinase activation causes GLUT4 translocation in skeletal muscle. Diabetes 48: 1667–1671, 1999.
32. Leclerc, I., C. Lenzner, L. Gourdon, S. Vaulont, A. Kahn, and B. Viollet. Hepatocyte nuclear factor-4a involved in type 1 maturity-onset diabetes of the young is a novel target of AMP-activated protein kinase. Diabetes 50: 1515–1521, 2001.
33. Longnus, S. L., R. B. Wambolt, H. L. Parsons, R. W. Brownsey, and M. F. Allard. 5-Aminoimidazole-4-carboxamide 1-beta-D-ribofuranoside (AICAR) stimulates myocardial glycogenolysis by allosteric mechanisms. Am. J. Physiol. 284: R936–R944, 2003.
34. Mcgarry, J. D., and N. F. Brown. The mitochondrial carnitine palmitoyltransferase system: from concept to molecular analysis. Eur. J. Biochem. 244: 1–14, 1997.
35. Mcgee, S. L., K. F. Howlett, R. L. Starkie, D. Cameron-Smith, B. E. Kemp, and M. Hargreaves. Exercise increases nuclear AMPK alpha-2 in human skeletal muscle. Diabetes 52: 926–928, 2003.
36. Merrill, G. M., E. Kurth, D. G. Hardie, and W. W. Winder. AICAR decreases malonyl-CoA and increases fatty acid oxidation in skeletal muscle of the rat. Am. J. Physiol. 273: E1107–E1112, 1997.
37. Mu, J., J. T. Brozinick, O. Valladares, M. Bucan, and M. J. Birnbaum. A role for AMP-activated protein kinase in contraction- and hypoxia-regulated glucose transport in skeletal muscle. Mol. Cell 7: 1085–1094, 2001.
38. Nielsen, J. N., K. J. Mustard, D. A. Graham, et al. 5′-AMP-activated protein kinase activity and subunit expression in exercise-trained human skeletal muscle. J. Appl. Physiol. 94: 631–641, 2003.
39. Nielsen, J. N., J. F. P. Wojtaszewski, R. G. Haller, et al. Role of 5′-AMP activated protein kinase in exercise regulation of glucose utilization and glycogen synthase activity in skeletal muscle from patients with McArdle’s disease. J. Physiol. 541: 979–989, 2002.
40. Park, H., V. K. Kaushik, S. Constant, et al. Coordinate regulation of malonyl-CoA decarboxylase, sn-glycerol-3-phosphate acyltransferase, and acetyl-CoA carboxylase by AMP-activated protein kinase in rat tissues in response to exercise. J. Biol. Chem. 277: 32571–32577, 2002.
41. Polekhina, G., A. Gupta, B. J. Michell, et al. AMPK b-subunit targets metabolic stress-sensing to glycogen. Curr. Biol. 13: 867–871, 2003.
42. Ponticos, M., Q. L. Lu, J. E. Morgan, D. G. Hardie, T. A. Partridge, and D. Carling. Dual regulation of the AMP-activated protein kinase provides a novel mechanism for the control of creatine kinase in skeletal muscle. EMBO J. 17: 1688–1699, 1998.
43. Rasmussen, B. B., C. R. Hancock, and W. W. Winder. Postexercise recovery of skeletal muscle malonyl-CoA, acetyl-CoA carboxylase, and AMP-activated protein kinase. J. Appl. Physiol. 85: 1629–1634, 1998.
44. Rasmussen, B. B., and W. W. Winder. Effect of exercise intensity on skeletal muscle malonyl-CoA and acetyl-CoA carboxylase. J. Appl. Physiol. 83: 1104–1109, 1997.
45. Richter, E. A., W. Derave, and J. F. Wojtaszewski. Glucose, exercise and insulin: emerging concepts. J. Physiol. 535: 313–322, 2001.
46. Saha, A. K., A. J. Schwarsin, R. Roduit, et al. Activation of malonyl-CoA decarboxylase in rat skeletal muscle by contraction and the AMP-activated protein kinase activator 5-aminoimidazole-4-carboxamide-1-beta-D-ribofuranoside [In Process Citation]. J. Biol. Chem. 275: 24279–24283, 2000.
47. Sim, A. T. R., and D. G. Hardie. The low activity of acetyl-CoA carboxylase in basal and glucagon-stimulated hepatocytes is due to phosphorylation by the AMP-activated protein kinase and not cyclic AMP-dependent protein kinase. FEBS Lett. 233: 294–298, 1988.
48. Stephens, T. J., Z. P. Chen, B. J. Canny, B. J. Michell, B. E. Kemp, and G. K. Mcconell. Progressive increase in human skeletal muscle AMPK a2 activity and ACC phosphorylation during exercise. Am. J. Physiol. 282: E688–E694, 2002.
49. Vavvas, D., A. Apazidis, A. K. Saha, et al. Contraction-induced changes in acetyl-CoA carboxylase and 5′-AMP-activated kinase in skeletal muscle. J. Biol. Chem. 272: 13255–13261, 1997.
50. Winder, W. W., and D. G. Hardie. The AMP-activated protein kinase, a metabolic master switch: possible roles in Type 2 diabetes. Am. J. Physiol. 277: E1–E10, 1999.
51. Winder, W. W., and D. G. Hardie. Inactivation of acetyl-CoA carboxylase and activation of AMP-activated protein kinase in muscle during exercise. Am. J. Physiol. 270: E299–E304, 1996.
52. Winder, W. W., B. F. Holmes, D. S. Rubink, E. B. Jensen, M. Chen, and J. O. Holloszy. Activation of AMP-activated protein kinase increases mitochondrial enzymes in skeletal muscle. J. Appl. Physiol. 88: 2219–2226, 2000.
53. Winder, W. W., H. A. Wilson, D. G. Hardie, et al. Phosphorylation of rat muscle acetyl-CoA carboxylase by AMP-activated protein kinase and cAMP-dependent protein kinase. J. Appl. Physiol. 82: 219–225, 1997.
54. Wojtaszewski, J. F., C. Macdonald, J. N. Nielsen, et al. Regulation of 5′AMP-activated protein kinase activity and substrate utilization in exercising human skeletal muscle. Am. J. Physiol. Endocrinol. Metab. 284: E813–E822, 2003.
55. Wojtaszewski, J. F., P. Nielsen, B. F. Hansen, E. A. Richter, and B. Kiens. Isoform-specific and exercise intensity-dependent activation of 5′-AMP-activated protein kinase in human skeletal muscle. J. Physiol. 528: 221–226, 2000.
56. Wojtaszewski, J. F. P., S. B. Jørgensen, Y. Hellsten, D. G. Hardie, and E. A. Richter. Glycogen-dependent effects of AICA riboside on AMP-activated protein kinase and glycogen synthase activities in rat skeletal muscle. Diabetes 51: 284–292, 2002.
57. Yang, W., Y. H. Hong, X. Q. Shen, C. Frankowski, H. S. Camp, and T. Leff. Regulation of transcription by AMP-activated Protein Kinase. Phosphorylation of p300 blocks its interaction with nuclear receptors. J. Biol. Chem. 276: 38341–38344, 2001.
58. Zheng, D., P. S. Maclean, S. C. Pohnert, et al. Regulation of muscle GLUT-4 transcription by AMP-activated protein kinase. J. Appl. Physiol. 91: 1073–1083, 2001.
59. Zhou, G., R. Myers, Y. Li, et al. Role of AMP-activated protein kinase in mechanism of metformin action. J. Clin. Invest. 108: 1167–1174, 2001.
60. Zong, H., J. M. Ren, L. H. Young, et al. AMP kinase is required for mitochondrial biogenesis in skeletal muscle in response to chronic energy deprivation. Proc. Natl. Acad. Sci. USA 99: 15983–15987, 2002.


©2004The American College of Sports Medicine