J Appl Physiol

91: 1017–1028, 2001.

invited review
Energy-sensing and signaling by AMP-activated protein
kinase in skeletal muscle

W. W. WINDER
Department of Zoology, Brigham Young University, Provo, Utah 84602

Winder, W. W. Energy-sensing and signaling by AMP-activated pro-

tein kinase in skeletal muscle. J Appl Physiol 91: 1017–1028, 2001.—
AMP-activated protein kinase (AMPK) is emerging as an important
energy-sensing/signaling system in skeletal muscle. This kinase is acti-
vated allosterically by 5⬘-AMP and inhibited allosterically by creatine
phosphate. Phosphorylation of AMPK by an upstream kinase, AMPK
kinase (also activated allosterically by 5⬘-AMP), results in activation. It
is activated in both rat and human muscle in response to muscle con-
traction, the extent of activation depending on work rate and muscle
glycogen concentration. AMPK can also be activated chemically in rest-
ing muscle with 5-aminoimidazole-4-carboxamide-riboside, which enters
the muscle and is phosphorylated to form ZMP, a nucleotide that mimics
the effect of 5⬘-AMP. Once activated, AMPK is hypothesized to phosphor-
ylate proteins involved in triggering fatty acid oxidation and glucose
uptake. Evidence is also accumulating for a role of AMPK in inducing
some of the adaptations to endurance training, including the increase in
muscle GLUT-4, hexokinase, uncoupling protein 3, and some of the
mitochondrial oxidative enzymes. It thus appears that AMPK has the
capability of monitoring intramuscular energy charge and then acutely
stimulating fat oxidation and glucose uptake to counteract the increased
rates of ATP utilization during muscle contraction. In addition, this
system may have the capability of enhancing capacity for ATP produc-
tion when the muscle is exposed to endurance training.
fatty acid oxidation; GLUT-4; glucose transport; malonyl-CoA; muscle
mitochondria

ONE OF THE FASCINATING PURSUITS in the field of exercise
physiology involves elucidation of the intramuscular
signals that allow matching of the rate of ATP produc-
tion with the increased energy requirements during
exercise. What is the intramuscular change that trig-
gers an enhancement of glucose uptake and fatty acid
oxidation during a single bout of exercise? How is the
availability of energy stores monitored? When a muscle
is exposed repeatedly to endurance exercise, an adap-
tation occurs that allows enhancement of the rate of
ATP production, thus allowing increased endurance at
higher intensities. During the course of an endurance
training program, what are the intramuscular signals
Address for reprint requests and other correspondence: W. W.
Winder, Dept. of Zoology, Brigham Young Univ., Provo, UT 84602
(E-mail: william_winder@byu.edu).
http://www.jap.org 8750-7587/01 $5.00 Copyright © 2001 the American Physiological Society
that trigger accumulation of increased quantities of
mitochondrial oxidative enzymes responsible for oxida-
tion of carbohydrate and fat? Recent studies have pro-
vided initial evidence that at least one of the energy-
sensing/signaling proteins of the muscle is AMP-
activated protein kinase (AMPK) (16, 34–36, 99). This
protein appears to be designed to monitor the energy
charge of the muscle fiber and to initiate responses
that prevent high-energy phosphate depletion. This
signaling system is currently being characterized with
respect to muscle phosphorylation targets. These pro-
tein targets of AMPK appear to be involved in both
regulation of short-term metabolic responses and
chronic adaptation to exercise. In this review, the
AMPK system will be described in terms of its occur-
rence in skeletal muscle, the mechanisms of activation,
processes influenced by AMPK, recently described tar-
gets of the kinase, possible roles in Type 2 diabetes and
1017
1018 INVITED REVIEW
Fig. 1. Model of subunit association of AMP-activated protein kinase
(AMPK) in the absence and presence of 5⬘-AMP (see Ref. 15). Both
the ␣- and the ␥-subunits participate in the allosteric activation of
AMPK by AMP. Both the catalytic domain of this kinase and the
phosphorylation site [threonine 172 (T172)] for AMPK kinase
(AMPKK) are located on the ␣-subunit.
obesity, and feasibility of manipulation of the kinase in
treatment of obesity and diabetes.
STRUCTURE, DISTRIBUTION, AND QUANTITATION
OF AMPK
AMPK is a protein consisting of three subunits des-
ignated ␣, ␤, and ␥ (34, 35) (Fig. 1). The approximate
molecular masses are 63 and 38 kDa for the ␣- and
␤-subunits, respectively (34, 35, 99). The ␥-subunit
molecular mass varies considerably among isoforms
(15). The ␣-subunit is the catalytic subunit containing
the kinase domain, which transfers a phosphate from
ATP to the target protein (34, 35). The ␤- and ␥-sub-
units are considered regulatory components (15, 34,
35). All three subunits are required for expression of
full activity (Fig. 1) (15, 24, 104). Each subunit has two
or three isoforms, designated ␣l, ␣2, ␤1, ␤2, ␥1, ␥2, and
␥3 (9, 15, 35). Information is available on tissue distri-
bution based on immunoprecipitation studies (15,000 g
supernatant) using antibodies to all three subunits and
in combination with AMPK activity measurements
(15). Both ␣1- and ␣2-isoforms are found in skeletal
muscle, but the ␣3-isoform is more abundant, repre-
senting 80% of the total AMPK activity (15, 82, 83, 88,
91). The ␣2-isoform is more sensitive to AMP (81). Both
␤1 and ␤2 are found in extensor digitorum longus (EDL;
predominantly type II fibers), but only ␤1 appears in
the soleus (type I fibers) (13). Data on distribution of all
subunits in the different muscle types are not currently
available. On the basis of immunoprecipitation stud-
ies, the ␥1-isoform predominates in skeletal muscle
(muscle type unspecified), but these data are not con-
sistent with Northern blot analyses, which indicate
J Appl Physiol • VOL 91 • SEPTEMBER 2001 •
high relative abundance of the ␥3-isoform mRNA in
muscle (15). Because Western blots were performed
only on postnuclear supernatants, it is possible that
AMPK containing the ␥3-isoform sedimented with the
cellular particulate fraction. The ␥-subunits appear to
play a role in determining sensitivity to AMP (15).
Activity of AMPK can be measured on ammonium
sulfate precipitates, polyethylene glycol precipitates,
or immunoprecipitates of muscle homogenates (34, 35,
91, 98). The assay measures incorporation of 32P-la-
beled phosphate from [32P]ATP into an artificial pep-
tide substrate resembling an AMPK phosphorylation
target site on acetyl-CoA carboxylase (ACC). Most in-
vestigators have used a 15-amino acid peptide called
“SAMS” peptide as substrate. It should be remembered
that an increase in AMPK activity measured in these
assays represents increases due to phosphorylation of
the enzyme. As will be discussed in detail later, 5⬘-
AMP (AMP) can allosterically activate AMPK and cre-
atine phosphate (CP) can allosterically inhibit AMPK.
The phosphorylated fraction of the enzyme could be
activated allosterically in the intact muscle, and this
increased AMPK activity would not show up in the
assay, since AMP and CP would for the most part be
discarded in the supernatant after collection of the
precipitated fraction. In the assay mix, AMP is added
in maximally effective activating concentrations.
MECHANISMS OF MUSCLE AMPK ACTIVATION
In addition to the glycolytic pathway and oxidative
phosphorylation, two enzymes are responsible for
maintaining relatively stable ATP concentrations in
the muscle as muscle begins to contract and utilize
ATP at increased rates (Fig. 2). Myokinase (adenylate
kinase) transfers a phosphate from one ADP to an-
other, resulting in the products AMP and ATP. An
increase in free ADP concentration resulting from in-
Fig. 2. Creatine phosphate (CP) decreases, creatine (C) increases,
and AMP increases in muscle during exercise as a result of actions of
creatine phosphokinase (CPK) and myokinase (adenylate kinase).
www.jap.org
 INVITED REVIEW
creased rates of ATP utilization allows this reaction to
proceed. This reaction contributes to buffering of ex-
pected contraction-induced ATP concentration changes
in the muscle. As muscle contracts, estimated AMP
concentration increases as a function of work rate (23).
Another enzyme responsible for buffering ATP changes
is creatine phosphokinase. During the course of con-
traction, phosphate may be transferred from CP to
ADP, forming ATP. Thus CP decreases and AMP in-
creases in the muscle fiber during contraction.
AMPK could be considered a protein that monitors
the energy state of the muscle cell and triggers meta-
bolic processes, which when activated are designed to
restore high-energy phosphate concentrations. This
enzyme is regulated by both allosteric and covalent
mechanisms (Figs. 1 and 3). AMPK is activated allos-
terically by an increase in AMP concentration (12, 16,
37). In addition, it is inhibited by CP and ATP (16, 19,
37, 73). The increase in AMP and the decrease in CP as
muscle contracts result in allosteric activation of
AMPK. AMP binds to AMPK, making it a better sub-
strate for the upstream kinase, AMPK kinase
(AMPKK) (37). AMP also directly activates AMPKK,
which phosphorylates the ␣-subunit of AMPK at thre-
onine 172, resulting in activation (37). AMP binding
also makes AMPK a poor substrate for protein phos-
phatase 2C, the phosphatase that appears to be re-
sponsible for removal of the phosphate at threonine
172 and inactivation of the enzyme (19). CP apparently
has no effect on the phosphorylation of AMPK; its
inhibitory effects are entirely allosteric (see Ref. 99).
AMPKK and AMPK can also be activated artificially
by use of the chemical 5-aminoimidazole-4-carboxam-
ide-riboside (AICAR) (17, 41, 42, 60, 79, 87). AICAR
was reported to be taken up by the cells and phosphor-
ylated to form an AMP analog, termed ZMP, which
then activated AMPKK and AMPK similarly to AMP.
Fig. 3. Activation of AMPK in muscle in response to muscle contrac-
tion. AMPK is activated as a result of allosteric stimulatory effects of
AMP, the removal of allosteric inhibitory effects by CP, and the
phosphorylation by AMPK kinase in the T172 position. ⫺, CP allo-
sterically inhibits AMPK; ⫹, AMP allosterically activates both
AMPKK and AMPK.
J Appl Physiol • VOL 91 • SEPTEMBER 2001 •
1019
AICAR has since been determined to be effective in
activating AMPK in perfused hindlimb muscle, incu-
bated muscle of the rat, cultured mouse muscle cells,
and muscle of injected rats in vivo (29, 38, 50, 60, 101).
CONTRACTION-INDUCED ACTIVATION OF AMPK
IN MUSCLE
The first indication that AMPK could be involved in
regulation of muscle metabolism during exercise came
from a study reported in 1996 (98). Rats were run on a
treadmill at 21 m/min up a 15% grade for periods up to
30 min. The deep red region of the quadriceps muscle
was quickly removed and frozen for AMPK isolation
and analyses. AMPK activity was found to increase
two- to threefold within 5 min of the beginning of
exercise and remained elevated for as long as the rat
continued to run. In rats running on the treadmill, the
increase in AMPK activity was found to be dependent
on work rate (75), and the activity remained elevated
for several minutes following 5- and 30-min exercise
bouts (74). When rat gastrocnemius muscle was stim-
ulated in situ at a frequency of once per second, 10-ms
duration, the estimated free AMP concentration in-
creased, CP decreased, and AMPK activity increased
(53). Interestingly, the decline in ACC activity (phos-
phorylation target for AMPK, a reporter for AMPK
activity) correlated more closely with the decline in CP
than with the rise in free AMP or with the increase in
measurable AMPK activity. This illustrates the impor-
tance of the allosteric mechanisms. AMPK activity also
increased in incubated epitrochlearis muscles in re-
sponse to contraction and AICAR treatment (38)
Vavvas et al. (91) published the first information on
the activation of specific isoforms of AMPK in response
to muscle contraction. The gastrocnemius muscle was
stimulated via the sciatic nerve (5 pulses/s, 100-ms
trains, 50 Hz, 10-ms duration) to contract in situ for
periods up to 5 min. The ␣1- and ␣2-isoforms were
precipitated from muscle homogenates using specific
antisera. The ␣2-isoform was maximally activated (3-
to 4-fold) within 30 s of the beginning of stimulation,
but no significant increase occurred in the ␣1-isoform
(91). In isolated epitrochlearis muscle, both the ␣1- and
␣2-isoforms were reported to be activated in response
to a number of stimuli that decreased CP and glycogen,
including contraction, hypoxia, dinitrophenol, rote-
none, and sorbitol (hyperosmotic stress) (29, 39). Only
the ␣2-isoform was activated in epitrochlearis of rats
running on the treadmill (65).
Studies published during the last year indicate that
the glycogen content of muscle may modulate the
AMPK response to contraction (21, 55). Derave et al.
(21) subjected rats to 2 h of swimming, followed by
feeding with a 100% fat diet overnight or with normal
chow ⫹ 20% glucose in drinking water. The two proto-
cols produced rats with low and high muscle glycogen,
respectively. Hindlimbs of these rats were perfused
with cell-free medium. After a 5-min washout period,
one sciatic nerve was stimulated for 10 min with
100-ms trains (100 Hz) at 2-s intervals. In fast-twitch
www.jap.org
1020 INVITED REVIEW
white fibers from the gastrocnemius, AMPK activity
measured in postnuclear supernatants of stimulated
muscles was increased 1.8-fold over resting muscles in
the high-glycogen group compared with 4.6-fold over
resting muscles in the low-glycogen group. In slow-
twitch soleus muscle, a contraction-stimulated activa-
tion of AMPK occurred in the muscles with low glyco-
gen but not in those with high glycogen. The activity of
ACC declined during contraction, however, indicating
that the AMPK may have been activated allosterically.
This would not have been detected in the final AMPK
assay procedure. Studies using the incubated epitroch-
learis muscle (fast twitch) have also demonstrated an
attenuation of the AMPK response in muscles with
high glycogen (55). It is unclear whether there is a
direct effect of glycogen on AMPK activity or whether
the attenuation is due to blunting of changes in ATP,
AMP, and CP. ATP and CP responses to stimulation
did not appear to be significantly altered in the soleus
regardless of glycogen content, although the time
course of changes were not studied (21).
Recent studies have clearly demonstrated AMPK to
be activated in human muscle during exercise.
Wojtaszewski et al. (103) exercised (cycle ergometer)
human subjects at 50% of maximal oxygen consump-
tion for 90 min or at 55 min at 75% followed by 5 min
at 90% of maximal oxygen consumption. They reported
no change in AMPK activity in needle biopsies from the
vastus lateralis in response to the lower intensity work
bout but a three- to fourfold increase in activity of the
␣2-isoform in response to the higher intensity work
bout. The activity of the ␣1-isoform was not changed in
response to either work rate. The activity of the ␣2-
isoform had returned to baseline by 3 h postexercise.
Fugii et al. (30) earlier the same year reported no
change in either isoform of AMPK in the vastus late-
ralis of human subjects in response to 20 min of work
at 50% of maximal oxygen consumption, whereas ac-
tivity of the ␣2-isoform increased after 20 and 60 min of
work at 70% of maximal oxygen consumption on the
cycle ergometer. Muscle CP and glycogen were signif-
icantly decreased in response to subjects working at
70% but not at 50% of maximal oxygen consumption.
One recent report indicates that the activity of both ␣1-
and ␣2-isoforms may increase with very high-intensity
work. Chen et al. (14) had subjects do a 30-s sprint
(peak power ⫽ 834 W) on a cycle ergometer and found
the activity of both isoforms to be increased. Richter et
al. (76a) recently obtained data from human subjects
indicating that elevated glycogen prevents the increase
in AMPK activity in response to working at 70% of
maximal oxygen consumption for 60 min, confirming
the relationship between glycogen content and AMPK
activity first demonstrated in animal studies (personal
communication).
In summary, from the currently available data, it
appears that the muscle AMPK increases in response
to muscle contraction. The increase is dependent on
work rate. In prolonged exercise at work rates exceed-
ing 70% of maximal oxygen consumption, only activity
J Appl Physiol • VOL
of the ␣2-isoform increases. In 30-s maximal sprints,
activities of both isoforms increase. The presence of
high glycogen in muscle appears to attenuate the in-
crease in AMPK during prolonged exercise.
REGULATION OF MUSCLE FATTY ACID OXIDATION
BY AMPK
The rate of fatty acid oxidation in muscle is con-
trolled in part at the level of carnitine palmitoyltrans-
ferase-1 (CPT-1) (Fig. 4). This enzyme is subject to
inhibition by malonyl-CoA (59, 78). Malonyl-CoA de-
clines rapidly in the red region of the quadriceps mus-
cle of rats during exercise, presumably relieving inhi-
bition and allowing fatty acid oxidation to increase as
long-chain acyl-CoA becomes available (74, 75, 97).
Malonyl-CoA is synthesized by the muscle isoform of
ACC, which has a molecular mass of ⬃272 kDa (89).
The ACC isoform isolated from rat hindlimb muscle
can be phosphorylated in vitro by both cAMP-depen-
dent protein kinase (PKA) and by AMPK, but only
phosphorylation by AMPK inactivates the enzyme,
thus differing from principal isoforms in both liver and
heart, which respond to phosphorylation by PKA with
a decline in activity (98, 102).
The muscle isoform of ACC is activated by citrate
over a range of 0–20 mM (89, 91, 98). The shape of this
activation curve is dependent on the phosphorylation
state of the enzyme (91, 98). Phosphorylation by AMPK
causes the curve to shift to the right, and the maximal
activation at high-citrate concentrations (Vmax) is re-
duced. The activation constant (Ka) for citrate in-
creases as a consequence of phosphorylation. Although
the determination of the citrate activation curve (ac-
tivity of ACC with citrate varied between 0 and 20 mM)
is important in characterizing ACC and its phosphor-
ylation state, muscle levels of citrate are in the range of
0.2 mM. The phosphorylated ACC is essentially inac-
tive at physiological citrate concentrations. Neverthe-
less, measurement of the citrate activation curve for
ACC isolated from muscle is a convenient way to assess
Fig. 4. Malonyl-CoA inhibits fatty acid oxidation by inhibiting car-
nitine palmitoyltransferase-1 (CPT-1), the enzyme necessary for
entrance of long-chain fatty acids into mitochondria for oxidation.
91 • SEPTEMBER 2001 • www.jap.org
 INVITED REVIEW
whether AMPK is activated (either allosterically or by
phosphorylation). ACC activity thus serves as a re-
porter for in vivo activation of AMPK.
ACC is inactivated concurrently with AMPK activa-
tion in muscle of rats running on the treadmill (75, 98)
and in muscle stimulated electrically (53, 91). It is also
clear that ACC can be inactivated early during the
time course of prolonged submaximal muscle contrac-
tion with no detectable increase in AMPK activity (53),
implying that the AMPK may be activated allosteri-
cally or that other mechanisms exist for inactivation of
ACC. When AMPK is activated using AICAR in resting
perfused muscle of rats, ACC activity declines, malo-
nyl-CoA declines, and the rate of radiolabeled palmi-
tate oxidation increases (60, 61). These effects are also
observed in incubated soleus (3) and in cultured C2C12
myocytes and myotubes (63). The rate of palmitate
oxidation is dependent on palmitate concentration, but
inclusion of AICAR in the medium increased the rate of
oxidation regardless of palmitate concentration (61).
The processes regulating the substrate oxidation mix
(carbohydrate vs. fat) in muscle are poorly understood.
Although a number of factors may be important in
determining this mix (including availability of glucose,
concentration of insulin, muscle glycogen, availability
of fatty acids, availability of carnitine and CoA, and so
forth; see Ref. 94), it is likely that AMPK plays a role.
Addition of insulin to medium during rat hindlimb
perfusion increases glucose uptake and tends to reduce
the rate of palmitate oxidation in perfused muscles
(100). If AICAR is also included in the medium to
activate AMPK, inactivate ACC, and reduce malonyl-
CoA, the presence of insulin, even at very high concen-
trations, does not prevent an acceleration of the rate of
palmitate oxidation. With the use of this method, a
wide range of malonyl-CoA concentrations were gener-
ated in perfused rat hindlimbs. A curvilinear relation-
ship between malonyl-CoA and palmitate oxidation
was observed, with a half-maximal effect inhibition
seen at ⬃0.6 nmol malonyl-CoA/g of muscle. This is
much higher than would have been predicted from data
generated from malonyl-CoA inhibition curves in iso-
lated mitochondria (59, 78).
Although a reduction of ACC activity would be ex-
pected to reduce the rate of malonyl-CoA synthesis
during muscle contraction, there remained the ques-
tion of what actually caused the reduction in malonyl-
CoA concentration. This question was approached by
Saha et al. (80), who studied muscle responses of ma-
lonyl-CoA decarboxylase (MCD) activity to electrical
stimulation and AICAR treatment. MCD is thought to
catalyze the major pathway for degradation of malo-
nyl-CoA in muscle, removing one carbon and producing
acetyl-CoA. MCD activity increased approximately
twofold in the gastrocnemius muscle in response to
electrical stimulation. Treatment of muscle extracts
with protein phosphatase 2A reversed this stimulation
of activity, providing evidence that phosphorylation is
involved. In addition, MCD could be activated in the
J Appl Physiol • VOL
1021
EDL by incubation with AICAR. Phosphorylation of
purified MCD with purified AMPK in vitro has not yet
been reported. The AICAR data provide initial evi-
dence that AMPK regulates not only synthesis but also
degradation of malonyl-CoA (Fig. 5).
The role of the AMPK system in control of fatty acid
oxidation in human muscle is less well defined. All
components of the signaling system are present in
human muscle, including AMPK, ACC, and malonyl-
CoA (14, 20). Evidence has been presented above for
activation of AMPK in muscle of human subjects dur-
ing exercise. Data indicate that ACC is phosphorylated
(14) and inactivated (14, 20) in muscle during exercise.
A small decline in malonyl-CoA has been reported by
one group (20) but not so in the majority of studies
(68–70). The concentration of malonyl-CoA is much
lower (⬃0.1) in human muscle than in rat muscle,
making it more difficult to measure, particularly in
small muscle biopsy samples (7, 20, 68–70, 77). It is
also more difficult to obtain muscle samples that are
homogeneous with respect to fiber type, as can be done
in rat studies. In one of the first rat studies, malonyl-
CoA was found to decline rapidly in type IIA fibers but
not in IIB fibers (97). If a mixture of contracting and
noncontracting fibers was sampled in the biopsy of
human muscle, any change in contracting fibers would
be diluted. One possibility is that ACC in human mus-
cle is arranged in close proximity to CPT-1 so that local
concentration of malonyl-CoA is responsible for regu-
lation. In this case, total tissue malonyl-CoA may not
change significantly, yet the concentration near CPT-1
could vary markedly depending on the ACC phosphor-
ylation state. This is entirely speculative, and it is
possible that malonyl-CoA-independent mechanisms
(see Refs. 68–70, 84, 94) are more important in control
of fatty acid oxidation in human muscle. The fact
remains, however, that ACC is present in the muscle.
It is phosphorylated at the AMPK target site, and its
activity decreases during exercise.
Fig. 5. Malonyl-CoA content of muscle is controlled by the relative
rate of synthesis by acetyl-CoA carboxylase (ACC) and the rate of
degradation by malonyl-CoA decarboxylase (MCD). AMPK phos-
phorylates (P) and inactivates ACC. On the basis of effects of 5-ami-
noimidazole-4-carboxamide-riboside (AICAR) on activation of MCD
in incubated extensor digitorum longus, AMPK is hypothesized to
phosphorylate and activate MCD (Ref. 80). These changes could
result in a decline in malonyl-CoA (rat muscle studies).
91 • SEPTEMBER 2001 • www.jap.org
1022 INVITED REVIEW
ROLE IN GLUCOSE UPTAKE
During the first hindlimb perfusion studies on the
role of AMPK in regulation of fatty acid oxidation in
muscle, it was noted that AICAR not only activated
AMPK and increased palmitate oxidation but also it
stimulated an increase in glucose uptake (60). It was
hypothesized at the time that AMPK activated by mus-
cle contraction stimulated both an increase in fatty
acid oxidation and an increase in glucose uptake to
meet the energy demands of working muscle (60).
The insulin-like effect of muscle contraction has been
extensively studied (31, 32, 40, 47). The insulin-sensi-
tive glucose transporter, GLUT-4, resides in two loca-
tions in the muscle (31, 32, 40, 47). In the basal state
(low insulin), the GLUT-4 transporters reside in mi-
crovesicles beneath the sarcolemma and T-tubule
membranes. In response to insulin or contraction,
these transporters are translocated and inserted into
membranes of the T tubules and sarcolemma, thus
increasing the capacity for moving glucose into the
muscle fiber (Fig. 6). The signaling pathways for the
two systems differ. The insulin pathway involves occu-
pied receptor-catalyzed phosphorylation of insulin re-
ceptor substrate 1 followed by activation of phosphati-
dylinositol 3-kinase, whereas the contraction pathway
does not involve either of these signaling proteins (31,
32, 40, 47, 105). The maximal effect of contraction on
translocation of GLUT-4 is additive with the effects of
high insulin. Although careful and extensive efforts
have been expended by many laboratories in attempts
to elucidate initial steps in the contraction-mediated
GLUT-4 translocation, the signaling steps have been
elusive. The possibility that AMPK activation may be
involved has therefore generated considerable interest.
Using the incubated epitrochlearis, it was deter-
mined that AICAR would increase 3-methylglucose
(3-MG) transport with characteristics similar to the
Fig. 6. Both insulin and muscle contraction trigger translocation of
GLUT-4 to the membranes of the sarcolemma and T tubules, result-
ing in enhancement of glucose uptake. These processes occur by
different signaling pathways. In the case of the contraction-initiated
pathway, other signals besides AMPK activation are possible. PI3K,
phosphatidylinositol 3-kinase; IR, insulin receptor; IRS-1, insulin
receptor substrate 1.
J Appl Physiol • VOL 91 • SEPTEMBER 2001 •
contraction effect (38). Wortmannin, an inhibitor of
phosphatidylinositol 3-kinase, inhibited insulin-stimu-
lation of 3-MG transport but not that caused by con-
traction or AICAR. Stimulatory effects of AICAR were
additive with those of insulin but not with those of
contraction. In perfused hindlimbs, AICAR was found
to stimulate glucose uptake in the absence of insulin
and to cause GLUT-4 translocation to the sarcolemmal
fraction from the microvesicle fraction (56). Infusion of
AICAR into live rats resulted in increased labeled
hexose uptake into all three muscle fiber types (8).
More recent studies have demonstrated that several
factors that decrease CP and glycogen in incubated
epitrochlearis (contraction, hypoxia, dinitrophenol, ro-
tenone, and sorbitol) also activate AMPK and stimu-
late glucose transport (29, 39). These data provide
evidence that AMPK activation mediates the effect of
contraction on stimulation of glucose transport.
Two studies (29, 39) have now demonstrated that
both the contraction-induced increase in AMPK activ-
ity and the contraction stimulation of glucose uptake
are decreased in muscles composed of type II fibers
when glycogen content of the muscle is elevated.
Kawanaka et al. (55) hypothesized that the smaller
degree of AMPK activation was responsible for the
decreased contraction stimulated translocation of
GLUT-4 in glycogen supercompensated epitrochlearis
muscles. Derave et al. (21) reported a complete absence
of AMPK activation in perfused supercompensated so-
leus muscles (composed of type I fibers) in response to
electrical stimulation, but glucose transport was not
diminished. On the basis of this apparent dissociation
between the two processes in the soleus, the authors
concluded that AMPK activation is not an obligatory
signaling step in contraction activation of glucose
transport. There was some evidence of allosteric acti-
vation of AMPK in the soleus, however, because ACC
activity at 0.2 mM citrate was lower after stimulation.
Nevertheless, it is entirely possible that redundant
pathways are present for the insulin-like effect of con-
traction and that the AMPK-independent pathway
predominates in type I fibers.
Part of the difficulty in identifying AMPK-dependent
processes in the muscle has been due to the lack of
specific inhibitors of AMPK. In isolated epitrochlearis,
iodotubercidin, Ara-A, and 8-bromo-AMP inhibited ef-
fects of AICAR on 3-MG uptake and AMPK ␣2 activa-
tion but had no or little effect on contraction-stimu-
lated AMPK activity and 3-MG uptake (65). It will be
important to identify more specific inhibitors of AMPK
to clearly link control of cellular processes to the AMPK
signaling pathway.
Another approach to linking AMPK activation with
contraction-stimulated glucose uptake is that of pro-
duction of transgenic mice lacking functional forms of
AMPK. This approach is complicated by the occurrence
of multiple isoforms of the three subunits of AMPK. Mu
and Birnbaum (62) found that muscle from mice ex-
pressing a kinase-inactive ␣2-catalytic subunit of
www.jap.org
 INVITED REVIEW
AMPK resulted in a depression of endogenous AMPK
activity in both the resting and contracting muscle.
Stimulation of hexose transport was completely
blocked in AICAR-treated and hypoxic muscles but
was only partially blocked in electrically stimulated
contracting muscles. The authors concluded that a
redundant AMPK-independent pathway must in part
mediate the effect of contraction on glucose uptake in
muscle. It is unclear whether residual endogenous
AMPK could have been activated allosterically to me-
diate contraction-stimulated glucose transport in these
transgenic mice. No mention was made of measure-
ment of ACC activity as a reporter of possible allosteric
activation.
Not all effects of AMPK on glucose transport may be
due to effects on translocation of GLUT-4. Abbud et al.
(1) studied effects of AICAR on glucose transport in
C2C12 myoblasts, which rely on GLUT-1 to transport
glucose. In these cells, a two- to threefold increase in
glucose transport was observed in response to incuba-
tion with AICAR. Thus enhancement of the activity of
other isoforms of glucose transporters may also help
explain the effect of AICAR on glucose transport in
muscle.
Some attention has been given to identifying the
specific target protein(s) for mediating the effect of
AMPK on GLUT-4 translocation and glucose trans-
port. Previous studies have demonstrated a possible
role of nitric oxide (NO) in contraction stimulation of
glucose transport (5, 6). Fryer et al. (29) recently re-
ported nitric oxide synthase (NOS) activity to be in-
creased along with glucose transport in H-2K myo-
tubes in response to activation of AMPK with AICAR.
NOS inhibitors blocked the increase in glucose trans-
port after AMPK activation. Both the neuronal NOS
and endothelial NOS isoforms can be phosphorylated
by AMPK in vitro. NO activates guanylate cyclase,
which then catalyzes synthesis of cGMP. Treatment of
these cells with an inhibitor of guanylate cyclase pre-
vents the increase in glucose transport wrought by
AICAR. In soleus and EDL muscle strips, the stimula-
tory effect of AICAR on glucose transport was also
blocked by the NOS inhibitor, NG-nitro-L-arginine
methyl ester. These results provide evidence that an
obligatory target of AMPK in inducing an increase in
glucose transport is NOS.
A more recent study using isolated EDL and soleus
muscles provides evidence against NO being a media-
tor of the contraction-induced activation of glucose
uptake (43). Incubation of muscles with sodium nitro-
prusside (SNP ⫽ NO donor) resulted in stimulation of
2-deoxy-D-glucose uptake. SNP was also reported to
activate ␣1- but not the ␣2-isoform of AMPK. Inclusion
of a NOS inhibitor in the incubation medium failed to
block the effect of contraction on 3-MG transport. The
authors proposed that NO activates glucose uptake by
a pathway separate from the insulin- and contraction-
induced signaling pathways. The reason for the appar-
ently contradicting results in these two studies is un-
clear at this time.
J Appl Physiol • VOL
1023
ROLE IN CONTROL OF GLUT-4 AND HEXOKINASE
One of the benefits of performing regular endurance
exercise is that muscles tend to become more sensitive
to the action of insulin (4, 22, 26, 47–49, 54, 67, 76, 90,
93). This is thought to be due in part to the increased
amounts of GLUT-4 that accumulate in muscle in re-
sponse to repetitive bouts of exercise over a several-day
period (26, 52, 54, 76). With daily exposure of the
muscle to exercise, there is an increase in the amount
of GLUT-4 mRNA that accumulates in muscle after
each bout (66, 76). The GLUT-4 gene has a regulatory
region that appears to have what is termed an “exer-
cise response element” that mediates the effect of reg-
ular bouts of contraction on GLUT-4 transcription (27,
66). Although the presence of this regulatory element
has been demonstrated, it is unclear what transcrip-
tion factor(s) binds to this region and how this factor
(or factors) is influenced by the contractile process.
AMPK appears to be a candidate for controlling the
putative transcription factors, thereby mediating
adaptive responses to exercise. AMPK would be ex-
pected to be activated with each bout of exercise. Could
it then phosphorylate a target transcription factor,
which would then be more effective in turning on the
GLUT-4 gene? Another possibility is that activated
AMPK could trigger synthesis of relevant transcription
factors? AMPK has previously been reported to be
responsible for negative regulation of several genes
involved in gluconeogenesis and fatty acid metabolism
in liver and in pancreatic islet ␤-cell metabolism (18,
28, 57, 58).
If AMPK activation is involved in inducing an in-
crease in GLUT-4 transcription in response to training,
it should be possible to mimic the effect of training with
chronic chemical activation of AMPK in resting mus-
cle. Injection of a high dose (1 mg/g body wt) of AICAR
into resting rats was found to produce an increase in
muscle ZMP, induce activation of AMPK, and induce a
decrease in ACC activity and in malonyl-CoA content
(50). The AMPK was activated in muscle for at least 2 h
in response to a single injection. When muscle AMPK
was activated daily with AICAR in resting rats for a
5-day period, the total amount of GLUT-4 in epitroch-
learis and gastrocnemius muscle was increased by
100% and 60%, respectively. Hexokinase activity was
also significantly increased (⬃2.5-fold) in both muscles.
The AICAR-treated rats exhibited glycogen supercom-
pensation the morning after injection, similarly to
trained rats after an overnight resting period. This
increase in GLUT-4 was maintained for at least 4 wk
with daily treatment of rats with AICAR (101). The
increase in muscle GLUT-4 induced by AICAR injec-
tion into rats has recently been confirmed and ex-
tended (11, 107). The increase in GLUT-4 protein is
accompanied by increases in mRNA (11, 107) and in-
creases in glucose uptake of isolated epitrochlearis in
response to insulin (11). Ojuka et al. (71) reported that
incubation of epitrochlearis muscles with 0.5 mM
AICAR for 18 h induced an increase in both GLUT-4
protein and hexokinase activity, implying that the in-
91 • SEPTEMBER 2001 • www.jap.org
1024 INVITED REVIEW
crease in these proteins in muscle was due to direct
effects of AMPK activation and not to some indirect or
humoral change (92, 107).
Transgenic mice have been produced with regions of
the 5⬘-flanking region of the human GLUT-4 gene
controlling expression of the bacterial chloramphenicol
acetyltransferase (CAT) reporter gene. These trans-
genics have been utilized to determine which region of
the promoter is responsible for the AICAR induction of
mRNA synthesis (107). The transgenic mice with ei-
ther 1,154 or 895 bp of the promoter responded to
AICAR with an increase in CAT mRNA expression, but
the transgenic with 730 bp was not responsive. This
implies that the region between ⫺730 and ⫺895 bp
from the GLUT-4 transcription start site contains an
element essential for induction of GLUT-4 mRNA syn-
thesis by AMPK activation. Evidence was also ob-
tained from gel-shift assays for possible roles of
MEF-2A and/or MEF-2D transcription factors in this
enhancement of transcription (107). A recent prelimi-
nary report indicates hexokinase II transcription (us-
ing nuclear run-on analysis) can be enhanced 6.5- and
10-fold, respectively, in red and white gastrocnemius
locally infused with AICAR (86).
ROLE IN CONTROL OF MITOCHONDRIAL PROTEINS
When muscle is exposed to repeated bouts of exercise
over several days and weeks, adaptations occur that
result in an increase in the maximal capacity to gen-
erate ATP. These adaptations include increases in mi-
tochondrial enzymes of the citric acid cycle and fatty
acid oxidation, proteins of the electron transport chain,
and myoglobin (10, 25, 44–46, 51). Although consider-
able work has been done in characterizing these adap-
tations, the intramuscular signaling mechanisms re-
sponsible for the accumulation of these proteins are not
clearly defined (see Ref. 51). Recently reported data
provide evidence of AMPK involvement in inducing
these adaptations (101). When muscle AMPK is acti-
vated by AICAR injection into nonexercising rats for
periods up to 4 wk, citrate synthase, succinate dehy-
drogenase, malate dehydrogenase, cytochrome c,
␦-aminolevulinate synthetase, and hexokinase in-
crease in a fiber-type-specific manner. Significant in-
creases (in the range of 40–50%) were seen in the white
quadriceps (type IIb fibers) and soleus (type I fibers)
but not in the red region of the quadriceps (type IIa
fibers). However, the muscle AMPK response to AICAR
injection was less at the end of 4 wk than at the
beginning. When rats were given intermittent injec-
tions of AICAR rather than daily injections (to mini-
mize the downregulation), a significant increase in
citrate synthase in the red quadriceps was observed
after 2-wk treatment (101). Enzymes of fatty acid oxi-
dation were not significantly influenced by AICAR
treatment. A more recent study, however, provides
evidence of a twofold increase in the rate of CPT-1 gene
transcription in response to local infusion of AICAR
into muscle of conscious rats in vivo (86). An increase
in transcription of CPT-1 along with other metabolic
J Appl Physiol • VOL
genes has also been reported for human muscle in the
postexercise period (72).
It should be clear that AICAR injection does not
mimic all changes in high-energy phosphate concentra-
tions that occur during muscle contraction. ATP and
CP are not markedly influenced. ZMP increases to
mimic the effects of AMP. ZTP was also found to
increase, and this change may interfere allosterically
with full AMPK activation (101). These differences
may explain why changes in mitochondrial enzymes
were not of the magnitude seen with daily 2-h bouts of
exercise (44, 101). The decline in CP may be essential
for full manifestation of the effect of contraction on
these adaptations. Nevertheless, these data suggest
that the regular activation of AMPK during daily bouts
of exercise may be responsible for inducing increases in
some, but not all, mitochondrial oxidative enzymes.
The mitochondrial uncoupling protein 3 (UCP3) has
also recently been discovered to be under the influence
of AMPK (108). Treatment of incubated muscles with
AICAR resulted in an increased UCP3 mRNA expres-
sion within 30 min. These findings are supported by
recent work indicating a marked stimulation of skele-
tal muscle UCP3 gene transcription after in vivo infu-
sion of AICAR into muscle of conscious rats (86). Al-
though the function of UCP3 in muscle is not entirely
clear, if it is indeed functioning as an uncoupler of
oxidative phosphorylation, more energy substrate
would be consumed for producing a given amount of
ATP when mitochondrial UCP3 is increased. Is it pos-
sible that transient increases in UCP3 after exercise
bouts during training could contribute to an increase in
oxygen consumption and caloric expenditure during
each postexercise period? Teleologically speaking, it is
advantageous for those involved in regular endurance
exercise to have a lower body mass. Could a transient
increase in UCP3 after daily bouts of exercise contrib-
ute to the weight loss and low-fat content of those who
train by running, swimming, or cycling many hours per
week?
CLINICAL CONSIDERATIONS
Two very exciting reports this year emphasize the
importance of continued research into this exciting
field and possible application of the basic research
findings to treatment or prevention of metabolic dis-
eases. Musi et al. (64) reported that the ␣2-isoform of
AMPK is activated in muscle of patients with Type 2
diabetes in response to 20 and 45 min of exercise. They
also demonstrated with Western blots that both ␣-iso-
forms are expressed normally in muscles of these pa-
tients. It has been previously suggested that some
forms of Type 2 diabetes could possibly be due to
deficiencies in the AMPK signaling pathway (98). This
may still prove to be the case, since subjects used in the
above-mentioned study were not obese and were capa-
ble of exercising at work rates required to activate
AMPK. A recent preliminary report indicates that
transgenic mice deficient in muscle AMPK activity
have reduced exercise tolerance (62). When the numer-
91 • SEPTEMBER 2001 • www.jap.org
 INVITED REVIEW
ous actions of AMPK are considered (activation of glu-
cose transport in muscle, control of GLUT-4 expression
in muscle, activation of muscle fatty acid oxidation,
inhibition of hepatic cholesterol synthesis and lipogen-
esis, stimulation of hepatic fatty acid oxidation, and
ketogenesis), it is not difficult to imagine that a defi-
ciency in AMPK could lead to insulin resistance and
obesity. Nevertheless, in those individuals with intact
AMPK signaling, exercise of sufficient duration and
intensity will most likely be beneficial in prevention
and possibly treatment of Type 2 diabetes.
Another major finding this year was from a study in
mice lacking the skeletal muscle isoform of ACC (2).
The consequent low malonyl-CoA in skeletal muscle
was postulated to result in chronically increased fatty
acid oxidation in muscle. Labeled palmitate oxidation
was 30% higher in incubated soleus muscle from these
mice compared with that of wild-type mice. In addition,
insulin suppression of fatty acid oxidation was not
observed in soleus muscles from the ACC-deficient
mice. These mice accumulated less fat in epididymal
fat pads while eating normally. This confirms the stud-
ies in rats that demonstrated that daily AICAR injec-
tions (which results in activation of AMPK, inactiva-
tion of ACC, and decreased malonyl-CoA in muscle)
reduced fat pad size compared with pair-fed controls
(101).
It has been suggested previously that development of
pharmaceutical activators of AMPK may be a feasible
approach for treatment of obesity and Type 2 diabetes
(50, 96, 99–101). AICAR probably will not be the drug
of choice, since large quantities are required for induc-
tion of AMPK activation in vivo. Liver hypertrophy
develops when large doses of AICAR are given for a
prolonged period (101). Much more potent and specific
AMPK activators must be developed.
SUMMARY AND FUTURE DIRECTIONS
Evidence has been presented to support the hypoth-
esis that AMPK, activated in response to muscle con-
traction, is involved in regulation of fatty acid oxida-
tion and stimulation of glucose uptake during exercise
(Fig. 7). The phosphorylation targets for AMPK include
ACC and possibly MCD, which then induce a decline in
malonyl-CoA in rat muscle (but not in human muscle)
and allow an increased rate of entrance of long-chain
fatty acyl-CoA into the mitochondria for oxidation. In
AMPK stimulation of glucose uptake, NOS has been
hypothesized to be a phosphorylation target, but data
have been presented that are inconsistent with the
hypothesis that NOS activation mediates the effect of
muscle contraction. Certainly, it will be important to
focus research efforts on this action of AMPK to discern
the downstream signaling mechanisms responsible for
GLUT-4 translocation and stimulation of glucose up-
take. Development of specific inhibitors of AMPK and
the use of transgenic models will be important. Both of
these rapid actions in muscle would have the effects of
enhancing the rate of ATP production during acute
bouts of exercise.
J Appl Physiol • VOL
1025
Fig. 7. Postulated consequences of acute activation of AMPK on
metabolism during single bouts of exercise (fatty acid oxidation and
glucose transport) and in response to repetitive activation of AMPK
as would be the case during endurance training [effects on GLUT-4,
hexokinase (HK), uncoupling protein 3 (UCP3), and oxidative en-
zymes of the citric acid cycle and electron transport chain]. Possible
phosphorylation target proteins ACC, MCD, and nitric oxide syn-
thetase (NOS) are indicated. ?, Phosphorylation targets in these
actions are uncertain or completely unknown.
On a broader scale, evidence is accumulating for a
role of chronic AMPK activation in increasing the max-
imal capacity for ATP production. This includes en-
hancement of the rate of transcription of the GLUT-4
gene, allowing increased amounts of GLUT-4 to accu-
mulate in the muscle in response to training. This
would allow increases in maximal glucose uptake dur-
ing very heavy exercise. The increase in mitochondrial
oxidative enzymes, hypothesized to be due in part to
the chronic activation of AMPK, would allow increased
maximal rates of pyruvate oxidation and ATP produc-
tion as the glucose is metabolized. Thus the AMPK
energy-sensing and signaling system appears to be
involved in counteracting high-energy phosphate de-
pletion during single bouts of exercise and, in the long
term, in inducing adaptations that counteract high-
energy phosphate depletion during prolonged heavy
exercise. The phosphorylation targets of AMPK re-
sponsible for mediating increases in GLUT-4 and mi-
tochondrial proteins are yet to be identified.
The basic research on the AMPK signaling pathway
provides additional strong rationale for the use of ex-
ercise in prevention and treatment of Type 2 diabetes
and obesity (96). Regular activation of this system
appears to increase fatty acid oxidation, glucose up-
take, and responsiveness of muscle to insulin (by in-
creasing GLUT-4). It may be also feasible to develop
suitable pharmaceutical AMPK activators, which
would mimic many of the effects of exercise. Obviously,
exercise will always be the most economical and most
effective way to activate this kinase and produce the
beneficial effects, but, for those who are unable to
exercise, the development of AMPK activators may
have clinical application (96).
91 • SEPTEMBER 2001 • www.jap.org
1026 INVITED REVIEW
Work in the author’s laboratory was supported by National Insti-
tute of Arthritis and Musculoskeletal and Skin Diseases Grant
AR-41438.
REFERENCES
1. Abbud W, Habinowski S, Zhang JZ, Kendrew J, Elkairi
FS, Kemp BE, Witters LA, and Ismail-Beigi F. Stimulation
of AMP-activated protein kinase (AMPK) is associated with
enhancement of GLUT1-mediated glucose transport. Arch Bio-
chem Biophys 380: 347–352, 2000.
2. Abu-Etheiga L, Matzuk MM, Abo-Hashema KAH, and
Wakil SJ. Continuous fatty acid oxidation and reduced fat
storage in mice lacking acetyl-CoA carboxylase 2. Science 291:
2613–2616, 2001.
3. Alam N and Saggerson ED. Malonyl-CoA and the regulation
of fatty acid oxidation in soleus muscle. Biochem J 334: 233–
241, 1998.
4. American Diabetes Association. Diabetes mellitus and ex-
ercise. Diabetes Care 21: S40–S44, 1998.
5. Balon TW. Integrative biology of nitric oxide and exercise.
Exerc Sport Sci Rev 27: 219–253, 1999.
6. Balon TW. Role of nitric oxide in contraction induced glucose
transport. Adv Exp Med Biol 441: 87–95, 1998.
7. Bavenholm PN, Pigon J, Saha AK, Ruderman NB, and
Efendic S. Fatty acid oxidation and the regulation of malonyl-
CoA in human muscle. Diabetes 49: 1078–1083, 2000.
8. Bergeron R, Russell RR, Young LH, Ren JM, Marchucci
M, Lee A, and Shulman GI. Effect of AMPK activation on
muscle glucose metabolism in conscious rats. Am J Physiol
Endocrinol Metab 276: E938–E944, 1999.
9. Beri RK, Marley AE, See CG, Sopwith WF, Aguan K,
Carling D, and Carey F. Molecular cloning, expression and
chromosomal localisation of human AMP-activated protein ki-
nase. FEBS Lett 356: 117–121, 1994.
10. Booth FW and Baldwin KM. Muscle plasticity: energy de-
mand and supply processes. In: Handbook of Physiology. Exer-
cise: Regulation and Integration of Multiple Systems. Bethesda,
MD: Am. Physiol. Soc., 1996, sect. 12, chapt. 24, p. 1075–1123.
11. Buhl ES, Jessen N, Schmitz O, Pedersen SB, Pedersen O,
Holman GD, and Lund S. Chronic treatment with 5-amino-
imidazole-4-carboxamide-1-␤-D-ribofuranoside increases insu-
lin-stimulated glucose uptake and GLUT4 translocation in rat
skeletal muscles in fiber type-specific manner. Diabetes 50:
12–17, 2001.
12. Carling D, Clarke PR, Zammit VA, and Hardie DG. Puri-
fication and characterization of the AMP-activated protein ki-
nase. Copurification of acetyl-CoA carboxylase kinase and 3-hy-
droxy-3-methylglutaryl-CoA reductase kinase activities. Eur
J Biochem 186: 129–136, 1989.
13. Chen ZP, Heierhorst J, Mann RJ, Mitchelhill KI, Michell
BJ, Witters LA, Lynch GS, Kemp BE, and Stapleton D.
Expression of the AMP-activated protein kinase ␤1 and ␤2
subunits in skeletal muscle. FEBS Lett 460: 343–348, 1999.
14. Chen ZP, McConnell GK, Michell BJ, Snow RJ, Canny
BJ, and Kemp BE. AMPK Signaling in contracting human
skeletal muscle: acetyl-CoA carboxylase and NO synthase phos-
phorylation. Am J Physiol Endocrinol Metab 279: E1202–
E1206, 2000.
15. Cheung PCF, Salt IP, Davies SP, Hardie DG, and Carling
D. Characterization of AMP-activated protein kinase ␥-subunit
isoforms and their role in AMP binding. Biochem J 346: 659–
669, 2000.
16. Corton JM, Gillespie JG, and Hardie DG. Role of the
AMP-activated protein kinase in the cellular stress response.
Curr Biol 4: 315–324, 1994.
17. Corton JM, Gillespie JG, Hawley SA, and Hardie DG.
5-Aminoimidazole-4-carboxamide ribonucleoside: a specific
method for activating AMP-activated protein kinase in intact
cells? Eur J Biochem 229: 558–565, 1995.
18. Da Silva XG, Leclerc I, Salt IP, Doiron B, Hardie DG,
Kahn A, and Rutter GA. Role of AMP-activated protein ki-
nase in regulation by glucose of islet beta cell gene expression.
Proc Natl Acad Sci USA 97: 4023–4028, 2000.
J Appl Physiol • VOL 91 • SEPTEMBER 2001 •
19. Davies SP, Helps NR, Cohen PTW, and Hardie DG. 5⬘-
AMP inhibits dephosphorylation, as well as promoting phos-
phorylation, of the AMP-activated protein kinase. Studies using
bacterially expressed human protein phosphatase-2C␣ and na-
tive bovine protein phosphatase-2C␣. FEBS Lett 377: 421–425,
1995.
20. Dean D, Daugaard JR, Young ME, Saha A, Vavvas D, Asp
S, Kiens B, Kim KH, Witters L, Richter EA, and Ruder-
man N. Exercise diminishes the activity of acetyl-CoA carbox-
ylase in human muscle. Diabetes 49: 1295–1300, 2000.
21. Derave W, Ai H, Ihlemann J, Witters LA, Kristiansen S,
Richter EA, and Ploug T. Dissociation of AMP-activated
protein kinase activation and glucose transport in contracting
slow-twitch muscle. Diabetes 49: 1281–1287, 2000.
22. Devlin JT, Hirshman M, Horton ED, and Horton ES.
Enhanced peripheral and splanchnic insulin sensitivity in
NIDDM men after a single bout of exercise. Diabetes 36: 434–
439, 1987.
23. Dudley GA, Tullson PC, and Terjung RL. Influence of
mitochondrial content on sensitivity of respiratory control.
J Biol Chem 262: 9109–9114, 1987.
24. Dyck JRB, Gao G, Widmer J, Stapleton D, Fernandez CS,
Kemp BE, and Witters LA. Regulation of the 5⬘-activated
protein kinase activity by the non-catalytic ␤ and ␥ subunits.
J Biol Chem 271: 17798–17803, 1996.
25. Essig DA. Contractile activity-induced mitochondrial biogene-
sis in skeletal muscle. Exerc Sport Sci Rev 24: 289–319, 1996.
26. Etgen GJ, Jensen J, Wilson CM, Hunt DG, Cushman SW,
and Ivy JL. Exercise training reverses insulin resistance in
muscle by enhanced recruitment of GLUT-4 to the cell surface.
Am J Physiol Endocrinol Metab 272: E864–E869, 1997.
27. Ezaki O. Regulatory elements in the insulin-responsive glu-
cose transporter (GLUT4) gene. Biochem Biophys Res Commun
241: 1–6, 1997.
28. Foretz M, Carling D, Guichard C, Ferr P, and Foufelle F.
AMP-activated protein kinase inhibits the glucose-activated
expression of fatty acid synthase gene in rat hepatocytes. J Biol
Chem 273: 14767–14771, 1998.
29. Fryer LGD, Hajduch E, Rencurel F, Salt IP, Hundal HS,
Hardie DG, and Carling D. Activation of glucose transport by
AMP-activated protein kinase via stimulation of nitric oxide
synthetase. Diabetes 49: 1978–1985, 2000.
30. Fugii N, Hayashi T, Hirshman MF, Smith JT, Habinowski
SA, Kaijser L, Mu J, Ljungqvist O, Birnbaum MJ, Witters
LA, Thorell A, and Goodyear LJ. Exercise induces isoform-
specific increase in 5⬘AMP-activated protein kinase activity in
human skeletal muscle. Biochem Biophys Res Commun 273:
1150–1155, 2000.
31. Goodyear LJ. AMP-activated protein kinase: a critical signal-
ing intermediary for exercise-stimulated glucose transport? Ex-
erc Sport Sci Rev 28: 113–116, 2000.
32. Goodyear LJ and Kahn BB. Exercise, glucose transport, and
insulin sensitivity. Annu Rev Med 49: 235–261, 1998.
33. Gulve EA and Spina RJ. Effect of 7–10 days of cycle ergome-
ter exercise on skeletal muscle GLUT-4 protein content. J Appl
Physiol 79: 1562–1566, 1995.
34. Hardie DG and Carling D. The AMP-activated protein ki-
nase. Fuel gauge of the mammalian cell? Eur J Biochem 246:
259–273, 1997.
35. Hardie DG, Carling D, and Carlson M. The AMP-activated/
SNF1 protein kinase subfamily: metabolic sensors of the eu-
karyotic cell? Annu Rev Biochem 67: 821–855, 1998.
36. Hardie DG, Salt IP, Hawley SA, and Davies SP. AMP-
activated protein kinase: an ultrasensitive system for monitor-
ing cellular energy charge. Biochem J 338: 717–722, 1999.
37. Hawley SA, Davison Woods A, Davies SP, Beri RK,
Carling D, and Hardie DG. Characterization of the AMP-
activated protein kinase kinase from rat liver, and identifica-
tion of threonine-172 as the major site at which it phosphory-
lates and activates AMP-activated protein kinase. J Biol Chem
271: 27879–27887, 1996.
38. Hayashi T, Hirshman MF, Kurth EJ, Winder WW, and
Goodyear LJ. Evidence for 5⬘-AMP-activated protein kinase
www.jap.org
 INVITED REVIEW
mediation of the effect of muscle contraction on glucose trans-
port. Diabetes 47: 1369–1373, 1998.
39. Hayashi T, Hirshman MF, Fugii N, Habinowski SA, Wit-
ters LA, and Goodyear LJ. Metabolic stress and altered
glucose transport. Activation of AMP-activated protein kinase
as a unifying coupling mechanism. Diabetes 49: 527–531, 2000.
40. Hayashi T, Wojtaszewski JF, and Goodyear LJ. Exercise
regulation of glucose transport in skeletal muscle. Am J Physiol
Endocrinol Metab 273: E1039–E1051, 1997.
41. Henin N, Vincent MF, Gruber HE, and Van den Berghe G.
Inhibition of fatty acid and cholesterol synthesis by stimulation
of AMP-activated protein kinase. FASEB J 9: 541–546, 1995.
42. Henin N, Vincent MF, and Van den Berghe G. Stimulation
of rat liver AMP-activated protein kinase by AMP analogues.
Biochim Biophys Acta 1290: 197–203, 1996.
43. Higaki Y, Hirshman MF, Fujii N, and Goodyear LJ. Nitric
oxide increases glucose uptake through a mechanism that is
distinct from the insulin and contraction pathways in rat skel-
etal muscle. Diabetes 50: 241–247, 2001.
44. Holloszy JO. Biochemical adaptations in muscle. Effects of
exercise on mitochondrial oxygen uptake and respiratory en-
zyme activity in skeletal muscle. J Biol Chem 242: 2278–2282,
1967.
45. Holloszy JO and Booth FW. Biochemical adaptations to
endurance exercise in muscle. Annu Rev Physiol 38: 273–291,
1976.
46. Holloszy JO and Coyle EF. Adaptations of skeletal muscle to
endurance exercise and their metabolic consequences. J Appl
Physiol 56: 831–838, 1984.
47. Holloszy JO and Hansen PA. Regulation of glucose transport
into skeletal muscle. Rev Physiol Biochem Pharmacol 128:
99–193, 1996.
48. Holloszy JO and Kohrt WM. Regulation of carbohydrate and
fat metabolism during and after exercise. Annu Rev Nutr 16:
121–138, 1996.
49. Holloszy JO, Schultz J, Kusnierkiewicz J, Hagberg JM,
and Ehsani AA. Effects of exercise on glucose tolerance and
insulin resistance. Acta Med Scand Suppl 711: 55–65, 1986.
50. Holmes BF, Kurth-Kraczek EJ, and Winder WW. Chronic
activation of 5⬘-AMP-activated protein kinase increases
GLUT-4, hexokinase, and glycogen in muscle. J Appl Physiol
87: 1990–1995, 1999.
51. Hood DA. Contractile activity-induced mitochondrial biogene-
sis in skeletal muscle. J Appl Physiol 90: 1137–1157, 2001.
52. Houmard JA, Egan PC, Neufer PD, Friedman JE,
Wheeler WS, Israel RG, and Dohm GL. Elevated skeletal
muscle glucose transporter levels in exercise-trained, middle-
aged men. Am J Physiol Endocrinol Metab 261: E437–E443,
1991.
53. Hutber CA, Hardie DG, and Winder WW. Electrical stimu-
lation inactivates muscle acetyl-CoA carboxylase and increases
AMP-activated protein kinase. Am J Physiol Endocrinol Metab
272: E262–E266, 1997.
54. Ivy JL. Role of exercise training in the prevention and treat-
ment of insulin resistance and non-insulin-dependent diabetes
mellitus. Sports Med 24: 321–336, 1997.
55. Kawanaka K, Nolte LA, Han DH, Hansen PA, and Hol-
loszy JO. Mechanisms underlying impaired GLUT-4 translo-
cation in glycogen supercompensated muscles of exercised rats.
Am J Physiol Endocrinol Metab 279: E1311–E1318, 2000.
56. Kurth-Kraczek EJ, Hirshman MF, Goodyear LJ, and
Winder WW. 5⬘-AMP activated protein kinase activation
causes GLUT4 translocation in skeletal muscle. Diabetes 48:
1667–1671, 1999.
57. Leclerc I, Kahn A, and Doiron B. The 5⬘-AMP-activated
protein kinase inhibits the transcriptional stimulation by glu-
cose in liver cells, acting through the glucose response complex.
FEBS Lett 431: 180–4, 1998.
58. Lochhead PA, Salt IP, Walker KS, Hardie DG, and Suth-
erland C. 5-Amino-4-carboxamide riboside mimics the effects
of insulin on the expression of the 2 key gluconeogenic genes
PEPCK and glucose-6-phosphatase. Diabetes 49: 896–903,
2000.
J Appl Physiol • VOL 91 • SEPTEMBER 2001 •
1027
59. McGarry JD and Brown NF. The mitochondrial carnitine
palmitoyltransferase system from concept to molecular analy-
sis. Eur J Biochem 244: 1–14, 1997.
60. Merrill GF, Kurth EJ, Hardie DG, and Winder WW. AICA
riboside increases AMP-activated protein kinase, fatty acid
oxidation, and glucose uptake in rat muscle. Am J Physiol
Endocrinol Metab 273: E1107–E1112, 1997.
61. Merrill GF, Kurth EJ, Rasmussen BB, and Winder WW.
Influence of malonyl-CoA and palmitate concentration on rate
of palmitate oxidation in rat muscle. J Appl Physiol 85: 1909–
1914, 1998.
62. Mu J and Birnbaum MJ. A role for AMP-activated protein
kinase in contraction and hypoxia regulated glucose transport
in skeletal muscle (Abstract). FASEB J 15: A1164, 2001.
63. Muoio DM, Seefeld K, Witters LA, and Coleman RA. AMP-
activated kinase reciprocally regulates triacylglycerol synthesis
and fatty acid oxidation in liver and muscle: evidence that
sn-glycerol-3-phosphate acyltransferase is a novel target. Bio-
chem J 338: 783–791, 1999.
64. Musi N, Fugii N, Hirshman MF, Ekberg I, Froberg S,
Ljungqvist O, Thorell A, and Goodyear LJ. AMP-activated
protein kinase (AMPK) is activated in muscle of subjects with
type 2 diabetes during exercise. Diabetes 50: 921–927, 2001.
65. Musi N, Hayashi T, Fugii N, Hirshman MF, Witters LA,
and Goodyear LJ. AMP-activated protein kinase activity and
glucose uptake in rat skeletal muscle. Am J Physiol Endocrinol
Metab 280: E677–E684, 2001.
66. Neufer PD and Dohm GL. Exercise induces a transient
increase in transcription of the GLUT-4 gene in skeletal mus-
cle. Am J Physiol Cell Physiol 265: C1597–C1603, 1993.
67. Oakes ND, Bell KS, Furler SM, Camilleri S, Saha AK,
Ruderman NB, Chisholm DJ, and Kraegen EW. Diet-in-
duced muscle insulin resistance in rats is ameliorated by acute
dietary lipid withdrawal or a single bout of exercise. Diabetes
46: 2022–2028, 1997.
68. Odland LM, Heigenhauser GJ, and Spriet LL. Effects of
high fat provision on muscle PDH activation and malonyl-CoA
content in moderate exercise. J Appl Physiol 89: 2352–2358,
2000.
69. Odland LM, Heigenhauser GJ, Lopaschuk GD, and
Spriet LL. Human skeletal muscle malonyl-CoA at rest and
during prolonged submaximal exercise. Am J Physiol Endocri-
nol Metab 270: E541–E544, 1996.
70. Odland LM, Howlett RA, Heigenhauser GJ, Hultman E,
and Spriet LL. Skeletal muscle malonyl-CoA content at the
onset of exercise at varying power outputs in humans. Am J
Physiol Endocrinol Metab 274: E1080–E1085, 1998.
71. Ojuka EO, Nolte LA, and Holloszy JO. Increased expression
of GLUT-4 and hexokinase in rat epitrochlearis muscles ex-
posed to AICAR in vitro. J Appl Physiol 88: 1072–1075, 2000.
72. Pilegaard H, Ordway GA, Saltin B, and Neufer PD. Tran-
scriptional regulation of gene expression in human skeletal
muscle during recovery from exercise. Am J Physiol Endocrinol
Metab 279: E806–E814, 2000.
73. Ponticos M, Lu QL, Morgan JE, Hardie DG, Partridge TA,
and Carling D. 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.
74. Rasmussen BB, Hancock CR, and Winder WW. Postexer-
cise recovery of skeletal muscle malonyl-CoA, acetyl-CoA car-
boxylase, and AMP-activated protein kinase. J Appl Physiol 85:
1629–1634, 1998.
75. Rasmussen BB and Winder WW. Effect of exercise intensity
on skeletal muscle malonyl-CoA and acetyl-CoA carboxylase.
J Appl Physiol 83: 1104–1109, 1997.
76. Ren JM, Semenkovich CF, Gulve EA, Gao J, and Holloszy
JO. Exercise induces rapid increases in GLUT4 expression,
glucose transport capacity, and insulin-stimulated glycogen
storage in muscle. J Biol Chem 269: 14396–14401, 1994.
76a.Richter EA, MacDonald C, Kı̀ens B, Hardie G, and
Wojtaszewski JFP. Dissociation of 5⬘-AMP-activated protein
kinase activity and glucose clearance in human skeletal muscle
during exercise (Abstract). Diabetes 50, Suppl 2: A62, 2001.
www.jap.org
1028 INVITED REVIEW
77. Ruderman NB, Saha AK, Vavvas D, Kurowski T, Laybutt
DR, Schmitz-Peiffer C, Biden T, and Kraegen EW. Malo-
nyl-CoA as a metabolic switch and regulator of insulin sensi-
tivity. Adv Exp Med Biol 441: 263–270, 1998.
78. Ruderman NB, Saha AK, Vavvas D, and Witters LA. Ma-
lonyl-CoA, fuel sensing, and insulin resistance. Am J Physiol
Endocrinol Metab 276: E1–E18, 1999.
79. Sabina RL, Patterson D, and Holmes EW. 5-Amino-4-imi-
dazole carboxamide riboside (Z-riboside) metabolism in eukary-
otic cells. J Biol Chem 260: 6107–6114, 1991.
80. Saha AK, Schwarsin AJ, Roduit R, Massé F, Kaushik V,
Tornheim K, Prentki M, and Ruderman NB. Activation of
malonyl-CoA decarboxylase in rat skeletal muscle by contrac-
tion and the AMP-activated protein kinase activator aminoimi-
dazole-4-carboxamide-1-D-ribofuranoside. J Biol Chem 275:
24279–24283, 2000.
81. Salt IP, Celler JW, Hawley SA, Prescott A, Woods A,
Carling D, and Hardie DG. AMP-activated protein kinase—
greater AMP dependence, and preferential nuclear localization,
of complexes containing the ␣2 isoform. Biochem J 334: 177–
187, 1998.
82. Stapleton D, Mitchelhill KI, Gao G, Widmer G, Mitchell
BJ, The T, House CM, Fernandez CS, Cox T, Witters LA,
and Kemp BE. Mammalian AMP-activated protein kinase
subfamily. J Biol Chem 271: 611–614, 1996.
83. Stapleton D, Woollatt E, Mitchelhill KI, Nicholl JK, Fer-
nandez CS, Michell BJ, Witters LA, Power DA, Suther-
land GR, and Kemp BE. AMP-activated protein kinase isoen-
zyme family: subunit structure and chromosomal location.
FEBS Lett 409: 452–456, 1997.
84. Starritt EC, Howlett RA, Heigenhauser GJ, and Spriet
LL. Sensitivity of CPT I to malonyl-CoA in trained and un-
trained human skeletal muscle. Am J Physiol Endocrinol
Metab 278: E462–E468, 2000.
85. Stein SC, Woods A, Jones NA, Davison MD, and Carling
D. The regulation of AMP-activated protein kinase by phos-
phorylation. Biochem J 345: 437–443, 2000.
86. Stoppani J, Hildebrandt AL, and Neufer PD. Activation of
AMP-activated protein kinase increases transcription of meta-
bolic genes in rat gastrocnemius muscle (Abstract). FASEB J
15: A415, 2001.
87. Sullivan JE, Brocklehurst KJ, Marley AE, Carey F,
Carling D, and Beri RK. Inhibition of lipolysis and lipogene-
sis in isolated rat hepatocytes with AICAR, a cell permeable
activator of AMP-activated protein kinase. FEBS Lett 353:
33–36, 1994.
88. Thornton C, Snowden MA, and Carling D. Identification of
a novel AMP-activated protein kinase subunit isoform that is
highly expressed in skeletal muscle. J Biol Chem 273: 12443–
12450, 1998.
89. Trumble GE, Smith MA, and Winder WW. Purification and
characterization of rat skeletal muscle acetyl-CoA carboxylase.
Eur J Biochem 231: 192–198, 1995.
90. US Department of Health and Human Services. The ef-
fects of physical activity on health and disease. In: Physical
Activity and Health: A Report of the Surgeon General. Centers
for Disease Prevention and Control, National Center for
Chronic Disease Prevention and Health Promotion, Washing-
ton, DC: GPO, 1996.
91. Vavvas D, Apazidis A, Saha AK, Gamble J, Patel A, Kemp
BE, Witters LA, and Ruderman NB. Contraction-induced
changes in acetyl-CoA carboxylase and 5⬘-AMP-activated ki-
nase in skeletal muscle. J Biol Chem 272: 13256–13261, 1997.
J Appl Physiol • VOL 91 • SEPTEMBER 2001 •
92. Vincent MF, Marangos P, Gruber HE, and Van den
Berghe G. AICA riboside inhibits gluconeogenesis in isolated
rat hepatocytes. Adv Exp Med Biol 309B: 1991–1995, 1991.
93. Wallberg-Henriksson H, Rincon J, and Zierath JR. Exer-
cise in the management of non-insulin-dependent diabetes mel-
litus. Sports Med 25: 25–35, 1998.
94. Winder WW. Intramuscular mechanisms regulating fatty acid
oxidation during exercise. Adv Exp Med Biol 441: 239–248,
1998.
95. Winder WW. Malonyl-CoA—regulator of fatty acid oxidation in
muscle during exercise. Exerc Sport Sci Rev 26: 117–132, 1998.
96. Winder WW. AMP-activated protein kinase: possible target for
treatment of type 2 diabetes. Diabetes Technol Therapeut 2:
441–447, 2000.
97. Winder WW, Arogyasami J, Elayan IM, and Cartmill D.
Time course of the exercise-induced decline in malonyl-CoA in
different muscle types. Am J Physiol Endocrinol Metab 259:
E266–E271, 1990.
98. Winder WW and Hardie DG. Inactivation of acetyl-CoA car-
boxylase and activation of AMP-activated protein kinase in
muscle during exercise. Am J Physiol Endocrinol Metab 270:
E299–E304, 1996.
99. Winder WW and Hardie DG. AMP-activated protein kinase,
a metabolic master switch: possible roles in Type 2 diabetes.
Am J Physiol Endocrinol Metab 277: E1–E10, 1999.
100. Winder WW and Holmes BF. Insulin stimulation of glucose
uptake fails to decrease palmitate oxidation in muscle if AMPK
is activated. J Appl Physiol 89: 2430–2437, 2000.
101. Winder WW, Holmes BF, Rubink DS, Jensen EB, Chen M,
and Holloszy JO. Activation of AMP-activated protein kinase
increases mitochondrial enzymes in skeletal muscle. J Appl
Physiol 88: 2219–2226, 2000.
102. Winder WW, Wilson HA, Hardie DG, Rasmussen BB, Hut-
ber CA, Call GB, Clayton RD, Conley LM, Yoon S, and
Zhou B. Phosphorylation of rat muscle acetyl-CoA carboxylase
by AMP-activated protein kinase and protein kinase A. J Appl
Physiol 82: 219–225, 1997.
103. Wojtaszewski JF, Nielsen P, Hansen BF, Richter EA, and
Kiens B. Isoform-specific and exercise intensity-dependent ac-
tivation of 5⬘-AMP-activated protein kinase in human skeletal
muscle. J Physiol (Lond) 528: 221–226, 2000.
104. Woods A, Salt I, Scott J, Hardie DG, and Carling D. The ␣1
and ␣2 isoforms of the AMP-activated protein kinase have
similar activities in rat liver but exhibit differences in substrate
specificity in vitro. FEBS Lett 397: 347–351, 1996.
105. Yeh JI, Gulve EA, Remeh L, and Birnbaum MJ. The effects
of wortmannin on rat skeletal muscle. Dissociation of signaling
pathways for insulin- and contraction-activated hexose trans-
port. J Biol Chem 270: 2107–2111, 1995.
106. Young ME, Radda GK, and Leighton B. Activation of gly-
cogen phosphorylase and glycogenolysis in rat skeletal muscle
by AICAR: an activator of AMP-activated protein kinase. FEBS
Lett 382: 43–47, 1996.
107. Zheng D, MacLean PS, Pohnert SC, Knight JB, Olson AL,
Winder WW, and Dohm GL. Regulation of muscle GLUT-4
transcription by AMP-activated protein kinase. J Appl Physiol
In press.
108. Zhou M, Lin BZ, Coughlin S, Valega G, and Pilch PF.
UCP-3 expression in skeletal muscle: effects of exercise, hyp-
oxia, and AMP-activated protein kinase. Am J Physiol Endo-
crinol Metab 279: E622–E629, 2000.
www.jap.org