A R T I C L E
Ca2+/calmodulin-dependent protein kinase kinase-β acts
upstream of AMP-activated protein kinase in mammalian cells
Angela Woods,1Kristina Dickerson,1Richard Heath,1Seung-Pyo Hong,2Milica Momcilovic,2
Stephen R. Johnstone,1Marian Carlson,2and David Carling1,*
1Cellular Stress Group, MRC Clinical Sciences Centre, Imperial College, Hammersmith Hospital, Du Cane Road, London W12 0NN,
2Institute of Cancer Research, Columbia University, New York, New York 10032
AMP-activated protein kinase (AMPK) is the downstream component of a kinase cascade that plays a pivotal role in
energy homeostasis. Activation of AMPK requires phosphorylation of threonine 172 (T172) within the T loop region of the
catalytic α subunit. Recently, LKB1 was shown to activate AMPK. Here we show that AMPK is also activated by Ca2+/
calmodulin-dependent protein kinase kinase (CaMKK). Overexpression of CaMKKβ in mammalian cells increases AMPK
activity, whereas pharmacological inhibition of CaMKK, or downregulation of CaMKKβ using RNA interference, almost
completely abolishes AMPK activation. CaMKKβ isolated from rat brain or expressed in E. coli phosphorylates and acti-
vates AMPK in vitro. In yeast, CaMKKβ expression rescues a mutant strain lacking the three kinases upstream of Snf1,
the yeast homolog of AMPK. These results demonstrate that AMPK is regulated by at least two upstream kinases and
suggest that AMPK may play a role in Ca2+-mediated signal transduction pathways.
family (Hong et al., 2003; Sutherland et al., 2003). A previous
study had demonstrated that purified CaMKK phosphorylates
and activates AMPK in vitro but much more weakly than a
preparation of AMPK kinase from rat liver (Hawley et al., 1995).
Furthermore, the AMPK kinase activity from rat liver was not
stimulated by Ca2+/calmodulin, unlike CaMKK (Hawley et al.,
1995). Based on this evidence, it was concluded that CaMKKs
were unlikely to play a significant physiological role in the acti-
vation of AMPK. Another closely related kinase to Elm1, Pak1,
and Tos3 is LKB1, a protein kinase that is mutated in Peutz-
Jeghers syndrome, a rare hereditary form of cancer (Hemminki
et al., 1998; Jenne et al., 1998). We and others previously re-
ported that LKB1 phosphorylates and activates AMPK (Hawley
et al., 2003; Hong et al., 2003; Shaw et al., 2004; Woods et al.,
2003) and that LKB1 correlates with the activity of AMPK ki-
nase purified from rat liver (Hawley et al., 2003; Woods et al.,
2003). In addition, although basal activity of AMPK was de-
tected in cells lacking LKB1, it was reported that activation of
AMPK in these cells was almost completely abolished (Hawley
et al., 2003; Shaw et al., 2004). These findings indicate that
LKB1 plays a major role in the activation of AMPK in a number
of cell types. However, yeast has three Snf1-activating kinases,
suggesting that mammals most likely have more than one, and
a recent study reported that AMPK is activated by an AMPK
kinase activity distinct from LKB1 in ischemic heart (Altarejos
et al., 2005).
Here we have revisited the role of CaMKK in phosphorylation
and activation of AMPK. We first show that AMPK can be
markedly activated in cells lacking LKB1, confirming that other
kinases besides LKB1 act upstream of AMPK. Several lines of
evidence suggest that CaMKKβ plays a physiological role in
activating AMPK in mammalian cells. Firstly, increasing intra-
cellular Ca2+leads to activation of AMPK. Secondly, expres-
AMP-activated protein kinase (AMPK) is the downstream com-
ponent of a protein kinase cascade that plays an important role
in maintaining energy balance (Carling, 2004; Hardie et al.,
1998; Kahn et al., 2005). AMPK is activated by a rise in the
AMP:ATP ratio within the cell following depletion of ATP. How-
ever, under some conditions, e.g., hyperosmotic stress, AMPK
is activated without a detectable change in the AMP:ATP ratio
(Fryer et al., 2002). Once activated, AMPK initiates a series of
responses including inhibition of anabolic pathways, stimula-
tion of catabolic pathways, and changes in gene and protein
expression (Carling, 2004; Hardie et al., 1998; Kahn et al.,
2005). A number of studies have indicated that AMPK plays an
important role in whole-body energy metabolism. Leptin and
adiponectin, hormones secreted from adipocytes, activate
AMPK in skeletal muscle, and this results in increased fatty
acid oxidation (Minokoshi et al., 2002; Yamauchi et al., 2002).
In addition, adiponectin activates AMPK in the liver, causing a
decrease in glucose output as well as an increase in glucose
and fatty acid oxidation (Yamauchi et al., 2002). More recently,
AMPK activity in the hypothalamus has been shown to be reg-
ulated by hormones that control feeding, suggesting that it may
also play a role in appetite (Andersson et al., 2004; Kim et al.,
2004a, 2004b; Minokoshi et al., 2004).
Activation of AMPK requires phosphorylation of threonine
172 (T172) within the T loop segment of the catalytic α subunit
(Hawley et al., 1996). In yeast, three protein kinases, Elm1,
Pak1, and Tos3, have been identified that phosphorylate and
activate SNF1, the yeast homolog of AMPK (Hong et al., 2003;
Sutherland et al., 2003). The most closely related mammalian
protein kinases to these yeast kinases are members of the
Ca2+/calmodulin-dependent protein kinase kinase (CaMKK)
CELL METABOLISM : JULY 2005 · VOL. 2 · COPYRIGHT © 2005 ELSEVIER INC.DOI 10.1016/j.cmet.2005.06.005 21
A R T I C L E
sion of CaMKKβ increases AMPK activity, whereas both phar-
macological inhibition of CaMKK activity and downregulation
of CaMKKβ expression, using a small interfering RNA ap-
proach, reduce AMPK activity. In addition, CaMKKβ is an effi-
cient AMPK kinase in vitro. Finally, we show that rat CaMKKβ
rescues a mutant yeast strain lacking Elm1, Pak1, and Tos3,
demonstrating that CaMKKβ is functionally competent as a
Snf1-activating kinase in vivo. Our combined results demon-
strate that AMPK can be regulated by CaMKKβ and indicate
that AMPK is activated in response to alternate signals: in-
creases in the AMP:ATP ratio and elevation of intracellular Ca2+.
AMPK is activated in mutant cells lacking LKB1
We and others previously reported that LKB1 lies directly up-
stream of AMPK (Hawley et al., 2003; Hong et al., 2003; Shaw
et al., 2004; Woods et al., 2003). Two previous studies reported
that, in mouse embryonic fibroblasts (MEFs) isolated from
LKB1−/−embryos, AMPK activity and T172 phosphorylation,
although very low, were detectable. However, it was also re-
ported that activation of AMPK by treatments known to acti-
vate AMPK, i.e., 5-aminoimidazole-4-carboxamide riboside
(AICAR), phenformin, and oxidative stress (hydrogen peroxide),
was almost completely abolished. We decided to reinvestigate
whether LKB1 is essential for AMPK activation by measuring
AMPK activity in cells lacking LKB1. Cultured MEFs isolated
from LKB1−/−embryos (Bardeesy et al., 2002) were subjected
to either oxidative stress, by incubation with 1 mM hydrogen
peroxide, or hyperosmotic stress, by incubation with 0.5 M sor-
bitol. Previous studies have shown that hydrogen peroxide ac-
tivates AMPK concomitant with a decrease in the ATP:AMP
ratio (Choi et al., 2001), whereas hyperosmotic stress activates
AMPK in the absence of a detectable change in this ratio (Fryer
et al., 2002). Figure 1A shows that, in both cases, AMPK activ-
ity was markedly increased in LKB1−/−MEFs relative to un-
treated cells. Western blot analysis of total cell lysates using a
phospho-specific anti-pT172 antibody revealed an increase in
T172 phosphorylation in treated cells compared to control cells
(Figure 1B). These results provide direct evidence that AMPK
is activated by phosphorylation in the absence of LKB1, sup-
porting the existence of another AMPK kinase.
Figure 1. Activation of AMPK in LKB1-deficient cells
Mouse embryonic fibroblasts isolated from LKB−/−mice (Bardeesy et al., 2002)
were treated in the absence (control) or presence of 1 mM hydrogen peroxide
(H2O2) or 0.5 M sorbitol for 15 min prior to harvesting.
A) AMPK was immunoprecipitated from 200 ?g total protein using an anti-
AMPKβ antibody, and activity in the immune complexes was measured using the
SAMS peptide assay. AMPK activity is calculated as pmol/min/mg lysate, and
the results shown are the mean ± SEM of three independent experiments, with
assays carried out in duplicate in each case.
B) A representative Western blot of total cell lysate (100 ?g) probed with an anti-
phosphospecific T172 antibody is shown. Total AMPK was determined using an
imately 0.5 ?g/ml (Figure 2B). At 10 ?g/ml, a concentration
that has previously been reported to almost completely inhibit
CaMKK activity in HeLa cells (Tokumitsu et al., 2002), STO-609
inhibited AMPK activation by 95%. At this concentration STO-
609 directly inhibits the in vitro activity of AMPK by approxi-
mately 20%, however, this inhibition does not persist through
the immunoprecipitation step used for isolating AMPK (data
not shown) and so does not contribute to the decrease in activ-
ity observed here.
Although ionomycin is well characterized as a Ca2+iono-
phore, we examined whether it had an effect on ATP levels in
HeLa cells. As can be seen from Figure 2C, ionomycin caused
a marked decrease in the intracellular ATP:ADP ratio from
4.2 ± 0.5 (n = 3) in untreated cells to 1.82 ± 0.03 (n = 3), but
hydrogen peroxide treatment caused a larger decrease in this
ratio (1.2 ± 0.1, n = 3). Because ionomycin decreases ATP, we
cannot establish whether activation of AMPK is a result of in-
creased Ca2+per se. However, the relative degree of AMPK acti-
vation in HeLa cells by ionomycin versus hydrogen peroxide does
Evidence that Ca2+regulates AMPK
An obvious candidate for the AMPK kinase activity in cells
lacking LKB1 are the CaMKK’s since it had been reported pre-
viously that purified CaMKK phosphorylates and activates
AMPK in vitro, albeit weakly (Hawley et al., 1995). CaMKK is
activated by Ca2+/calmodulin (Edelman et al., 1996; Lee and
Edelman, 1994; Selbert et al., 1995; Tokumitsu et al., 2001),
but there are no reports of whether Ca2+directly regulates acti-
vation of AMPK. As an initial attempt to address this issue, we
tested the effect of the Ca2+ionophore, ionomycin, on AMPK
activity in HeLa cells. Figure 2A shows that ionomycin treat-
ment activated AMPK over 10-fold relative to untreated cells.
In comparison, hydrogen peroxide treatment activated AMPK
3-fold (Figure 2A). We took advantage of STO-609, a relatively
selective and cell permeable inhibitor of CaMKK (Tokumitsu et
al., 2002), to determine the role of CaMKK on AMPK activation
by ionomycin. STO-609 caused a dose-dependent reduction
in the activation of AMPK by ionomycin, with an IC50of approx-
CELL METABOLISM : JULY 2005
CaMKKβ is an AMPK kinase
which, like HeLa cells, do not express LKB1 (Woods et al.,
2003). Endogenous CaMKKα and CaMKKβ activities were de-
tected in CCL13 cells following immunoprecipitation with the
relevant antibodies (Figure 3A). Expression of CaMKKβ but not
CaMKKα led to a dramatic increase in AMPK activity (Figure
3B). CaMKKβ expression increased AMPK activity in untreated
cells as well as following treatment with hydrogen peroxide and
ionomycin. A kinase-dead form of CaMKKβ, harboring a muta-
tion of aspartic acid residue 329 to alanine, had no effect on
AMPK activity, demonstrating that the catalytic activity of
CaMKKβ is required for AMPK activation. In untreated cells,
expression of CaMKKβ led to an increase in the phosphoryla-
tion of acetyl-CoA carboxylase (ACC), a downstream substrate
of AMPK (Figure 3C). Basal phosphorylation of ACC was de-
tected in cells transfected with CaMKKα or catalytically in-
active CaMKKβ and also in the β-galactosidase transfections
following longer exposure (Figure 3C and data not shown).
However, in these cases, ACC phosphorylation was markedly
lower than in cells transfected with CaMKKβ. Expression of the
CaMKK constructs was confirmed by Western blot analysis of
the cell lysates (Figure 3D). Furthermore, there was a clear in-
crease in both CaMKKα and CaMKKβ activities following
transfection with the corresponding cDNAs (Figure 3E), show-
ing that both CaMKK isoforms are functionally active in this
system. These results demonstrate that high-level expression
of CaMKKβ is sufficient to activate AMPK in the absence of
additional signals, whereas expression of CaMKKα has no sig-
Inhibition of CaMKK reduces AMPK activity
in mammalian cells
To determine the role of CaMKK isoforms on AMPK activation,
we used a small interfering RNA (siRNA) approach to knock
down the expression of CaMKKα and CaMKKβ. To avoid the
complication of interference from activation of AMPK by LKB1,
we used HeLa cells, which do not express LKB1 (Tiainen et
al., 1999). RNA duplexes, designed to specifically target either
human CaMKKα or CaMKKβ, were transfected into HeLa cells.
CaMKK expression was measured 48 hr posttransfection by
Western blot analysis using isoform-specicific antibodies (Fig-
ure 4A). A significant reduction in the level of either CaMKKα
(approximately 50% reduction) or CaMKKβ (approximately
80% reduction) was observed following transfection of the
specific targeted RNA sequence. In a recent study, it was re-
ported that HeLa cells express CaMKKα and a novel CaMKKβ
isoform termed CaMKKβ-3 (Ishikawa et al., 2003). This isoform
lacks exon 16 and has a C-terminal sequence distinct from
CaMKKβ-1 and CaMKKβ-2. We used a siRNA sequence that
is conserved in all three β isoforms and would therefore be
predicted to downregulate all three isoforms. The estimated
molecular mass of the CaMKKβ isoform detected by Western
blotting (w60 kDa) corresponds closely to the predicted mass
of CaMKKβ-3, suggesting that this is the predominant isoform
expressed in HeLa cells, consistent with the findings of a previ-
ous study examining the expression of CaMKK mRNAs in
these cells (Ishikawa et al., 2003). In parallel experiments, we
examined the effect of downregulating CaMKK isoforms on
AMPK activity in cells treated with hydrogen peroxide, hyper-
osmotic stress, or ionomycin (Figure 4B). Downregulation of
CaMKKβ significantly reduced AMPK activation in response to
all treatments. In addition, basal AMPK activity was signifi-
Figure 2. The Ca2+ionophore ionomycin activates AMPK in HeLa cells
A) AMPK activity in immune complexes isolated from HeLa cells (100 ?g lysate)
by immunoprecipitation with an anti-AMPKβ antibody was measured. HeLa cells
were treated with either 1 mM hydrogen peroxide (15 min) or 1 ?M ionomycin (5
min) prior to harvesting.
B) Cells were preincubated for 4 hr with the indicated concentration of STO-609
prior to treatment with ionomycin (1 ?M, 5 min). In both cases, AMPK activity is
plotted as pmol/min/mg lysate and is the mean ± SEM of three independent ex-
C) The ATP:ADP ratio in HeLa cell extracts was determined by separation of
nucleotides by ion-exchange chromatography. Results shown are the mean ±
SEM from three independent measurements.
not correlate with the corresponding changes in the ATP:ADP ra-
tio. Taken together, these results suggest that CaMKK activates
AMPK under conditions in which intracellular Ca2+is increased.
Expression of CaMKKβ increases AMPK activity
in mammalian cells
To assess whether the CaMKK isoforms act as AMPK kinases
in mammalian cells, we first expressed them in CCL13 cells,
CELL METABOLISM : JULY 2005 23
A R T I C L E
Figure 3. Expression of CaMKKβ increases AMPK activity in mammalian cells
A) Endogenous CaMKK in CCL13 cells was immunoprecipitated using anti-CaMKKα or anti-CaMKKβ antibodies and activity present in the immune complexes
determined by phosphorylation of GST-CaMKI(1−293)(10 ?g). As a negative control, an immune complex isolated using a nonspecific antibody was included. A
representative autoradiogram of the phosphorylated GST-CaMKI(1–293)is shown.
B) CCL13 cells were transiently transfected with cDNAs encoding either wild-type CaMKKα, wild-type CaMKKβ, a catalytically inactive mutant of CaMKKβ (harboring
a D329A mutation), or β-galactosidase. Posttransfection (48 hr), cells were either left untreated or incubated with either 1 mM hydrogen peroxide (H2O2) for 15 min or
1 ?M ionomycin for 5 min and AMPK activity determined in anti-AMPKβ immune complexes. Results are the mean ± SEM of three independent experiments.
C) Cell lysates (50 ?g) from transfected, untreated cells were probed with an anti-phosphospecific ACC antibody (pACC). A representative Western blot (in each case,
two samples from independent cell lysates) is shown. In parallel, total ACC expression was determined using horseradish peroxidase-streptavidin conjugate.
D) Following transfection, CaMKK expression was determined by probing the cell lysates (50 ?g) with an anti-CaMKKα or anti-CaMKKβ antibody.
E) CaMKK activity present in immune complexes from CCL13 cells transfected with the indicated cDNA was determined following immunoprecipitation with either
anti-CaMKKα or anti-CaMKKβ antibodies.
CELL METABOLISM : JULY 2005
CaMKKβ is an AMPK kinase
cantly reduced in cells depleted of CaMKKβ. In contrast,
downregulation of CaMKKα had no significant effect on basal
AMPK activity or activation by either hydrogen peroxide or
hyperosmotic stress. Downregulation of CaMKKα, however,
did cause a significant reduction in AMPK activity in response
to ionomycin, although this was significantly less than that ob-
served following downregulation of CaMKKβ. These results
suggest that CaMKKβ is the predominant isoform regulating
AMPK in HeLa cells but that CaMKKα may also contribute to
AMPK activation in response to ionomycin.
As an initial attempt to determine the relative contribution of
CaMKK versus LKB1 to the activation of AMPK, we examined
the effect of STO-609 on AMPK activity in cells expressing
LKB1 compared with cells lacking LKB1. For this study, we
compared AMPK activity in NIH3T3 fibroblasts, which have
been shown previously to express LKB1 (Tiainen et al., 1999)
and LKB1−/−MEFs. We first confirmed that CaMKKα and
CaMKKβ are expressed in these cell types (Figure 4C). Preincu-
bation of NIH3T3 cells with 10 ?g/ml STO-609 reduced AMPK
activity by between 20% and 50%, depending on the cell treat-
ment (Figure 4D). In LKB1−/−cells, STO-609 had a much
greater effect, reducing AMPK activity by between 50% and
90% (Figure 4D). STO-609 had a dramatic effect on AMPK ac-
tivation by ionomycin in LKB1−/−cells, reducing activity by over
90%. In all cases, the degree of T172 phosphorylation corre-
lated closely with AMPK activity, although this is not a rigor-
ously quantitative correlation due to limitations of Western
blots (Figure 4E). Importantly, at the concentration used in the
cell incubations (10 ?g/ml), STO-609 inhibited LKB1 activity in
vitro by less than 10% (data not shown).
served with the isolated rat brain CaMKKβ could be due to
other coprecipitating kinases. To confirm that CaMKKβ directly
phosphorylates AMPK, we expressed full-length rat CaMKKβ
in E. coli and tested the purified recombinant kinase for its abil-
ity to phosphorylate and activate AMPK. Figure 5D shows a
Coomassie-stained gel and corresponding autoradiogram of
wild-type (α1β1γ1) or kinase-dead (α1[D57A]β1γ1) AMPK com-
plexes after phosphorylation with recombinant CaMKKβ in the
presence of [32P]-ATP. Kinase-dead AMPK, which is unable to
autophosphorylate, was phosphorylated on the α subunit
alone, whereas the wild-type complex was phosphorylated on
both the α and β subunits. The stoichiometry of phosphoryla-
tion was determined by measuring the radioactivity incorpo-
rated into recombinant His-tagged AMPK following isolation on
nickel agarose beads. For the kinase-dead complex, this was
calculated to be 1.1 ± 0.3 (n = 4) mole/mole and 5.5 ± 0.9
(n = 4) mole/mole for the wild-type complex. The increased
incorporation into the wild-type AMPK complex is due to auto-
phosphorylation, which led to a noticeable shift in the mobility
of the α and β subunits on SDS-PAGE (Figure 5D). Analysis
of the autoradiogram by densitometry indicated approximately
equal levels of incorporation into the α and β subunits of the
wild-type AMPK complex. In parallel, we determined the effect
of CaMKKβ phosphorylation on AMPK activity (Figure 5E).
AMPK activity increased in response to increasing concentra-
tions of bacterially expressed CaMKKβ, reaching a maximum
activity of 0.2 ?mol/min/mg. A similar degree of activation was
achieved following phosphorylation of AMPK by purified LKB1
(data not shown). We went on to determine that activation of
AMPK by CaMKKβ was due to phosphorylation of T172, since
this had not been demonstrated previously. Western blot analy-
sis using anti-pT172-specific antibodies revealed that CaMKKβ
phosphorylates T172 within AMPK (Figure 5F).
CaMKK isoforms phosphorylate and activate
AMPK in vitro
To assess the ability of CaMKKα and CaMKKβ to phosphory-
late AMPK in vitro, we used isoform-specific CaMKK antibod-
ies to immunoprecipitate them from a rat brain extract. The
resulting immune complexes were assayed for their ability to
phosphorylate kinase-dead AMPK (α1(D157A)β1γ1) and a trun-
cated, kinase-dead form of CaMKI (CaMKI[1–293]) that lacks the
regulatory region required for binding of calmodulin (Tokumitsu
et al., 2000). The phosphorylated products were resolved by
SDS-PAGE and detected by autoradiography (Figure 5A) or an-
alyzed using a phosphoimager in order to quantitate the rela-
tive degree of phosphorylation (Figure 5B). In the presence
of Ca2+/calmodulin, both CaMKK isoforms phosphorylated
CaMKI with approximately equal efficiency. In contrast, when
AMPK was used as a substrate, CaMKKβ was much more ef-
fective than CaMKKα. There was no detectable phosphoryla-
tion of either CaMKI or AMPK following incubation with a con-
trol immune complex isolated using an irrelevant antibody (data
not shown). To examine whether phosphorylation by CaMKK
caused activation of AMPK, bacterially expressed AMPK was
incubated with CaMKKα or CaMKKβ immune complexes in the
presence of MgATP and the resulting AMPK activity measured
(Figure 5C). In the presence of Ca2+/calmodulin, CaMKKα
caused a small but barely detectable increase in AMPK activity.
In contrast, CaMKKβ caused a marked activation of AMPK,
even in the absence of Ca2+/calmodulin, and this was in-
creased approximately 4-fold by the presence of Ca2+/calmod-
ulin in the activation assay (Figure 5C).
It remained possible that the activation of AMPK we ob-
AMP does not directly activate CaMKKβ
In addition to direct allosteric activation of AMPK, it has been
reported that AMP also promotes the phosphorylation of T172.
A previous study of the effect of AMP on phosphorylation of
AMPK by CaMKK purified from pig brain suggested that AMP
exerts a substrate mediated effect on AMPK but no direct ef-
fect on CaMKK (Hawley et al., 1995). Because we are not able
to establish which isoform of CaMKK was used in the previous
study, we examined the effect of AMP on CaMKKβ-mediated
activation of AMPK. Figure 6A shows that there was a small
but statistically significant increase in the activation of α2 com-
plexes by AMP (1.27 ± 0.07, n = 6), whereas AMP had no sig-
nificant effect on the activation of α1 complexes (1.16-fold ±
0.11, n = 6). In comparison, AMP stimulated LKB1-mediated
activation of α1 by 1.34-fold ± 0.09 (n = 4) and α2 by 1.77-
fold ± 0.13 (n = 4). In parallel, we determined the effect of AMP
directly on AMPK by carrying out the AMPK assay (after activa-
tion) in the presence or absence of 0.2 mM AMP. AMP signifi-
cantly stimulated α1 complexes and, to a greater extent, α2
complexes, consistent with previous reports (Salt et al., 1998;
Stein et al., 2000). We were unable to determine the effect of
AMP on CaMKKα-mediated activation of AMPK using this
method because the activation was too low to measure reli-
ably. However, using CaMKI(1–293)as a substrate, AMP did not
directly activate either CaMKKα or CaMKKβ (Figure 6B).
CELL METABOLISM : JULY 200525
A R T I C L E
Figure 4. Inhibiting CaMKK activity in HeLa cells decreases AMPK activity
HeLa cells were incubated for 48 hr in the presence of synthetic RNA duplexes (0.06 ?M) targeted to human CaMKKα, human CaMKKβ, or a control RNA duplex
containing an unrelated sequence.
A) A representative Western blot of cell lysates probed with antibodies against either CaMKKα or CaMKKβ is shown. Lysates from two independent transfections with
CaMKKα siRNA are shown for the CaMKKα blot and from two independent transfections with CaMKKβ siRNA for the CaMKKβ blot.
B) Following RNA transfection, cells were treated in the presence or absence of 1 mM hydrogen peroxide (H2O2) for 15 min, 0.5 M sorbitol for 15 min, or 1 ?M
ionomycin for 5 min and AMPK activity determined in anti-AMPKβ immune complexes isolated from 100 ?g cell lysate. Results are the mean ± SEM for three
independent experiments. For each treatment, activities that are statistically significant from the control RNA incubation are denoted by * (p < 0.05, unpaired t test).
C) CaMKK activity present in anti-CaMKKα and anti-CaMKKβ immune complexes isolated from NIH3T3 or LKB−/−cells was determined by phosphorylation of GST-
CaMKI(1−293)(10 ?g). No further activity was recovered in a subsequent immunoprecipitation assay using the same antibody, whereas activity was recovered if an
antibody to the alternate isoform was used in the subsequent immunoprecipitation.
CELL METABOLISM : JULY 2005
CaMKKβ is an AMPK kinase
Rat CaMKKβ activates Snf1 in yeast cells
Previously, we identified three protein kinases that lie directly
upstream of Snf1 in yeast (Hong et al., 2003). Since the mam-
malian AMPK and yeast Snf1 pathways are highly conserved,
yeast cells provide a convenient model system to assess the
in vivo function of CaMKK in these pathways. We previously
showed that in yeast deletion of the Snf1 upstream kinases
Elm1, Pak1, and Tos3 resulted in a Snf phenotype (Hong et al.,
2003). To test the ability of CaMKK to activate Snf1 in yeast,
we expressed rat CaMKKβ in the pak1? tos3? elm1? triple
mutant yeast strain, which lacks all three native Snf1-activating
kinases. The triple mutant grows on glucose but is unable to
grow on carbon sources whose utilization requires Snf1 protein
kinase function. Expression of CaMKKβ complemented this
defect, conferring growth on raffinose and glycerol-ethanol
(Figure 7A). In control experiments, expression of Pak1 also
allowed growth on these carbon sources, as expected,
whereas the vector alone did not.
To confirm that CaMKKβ remedied the growth defect by acti-
vating Snf1 protein kinase, we assayed Snf1 catalytic activity.
Triple mutant cells expressing CaMKKβ were grown in high glu-
cose and then subjected to glucose limitation, which results in
activation of Snf1 protein kinase in wild-type yeast cells. Snf1
was partially purified from cell extracts and was assayed by
phosphorylation of the SAMS peptide substrate. CaMKKβ re-
stored Snf1 catalytic activity in the triple mutant, as did Pak1
(Figure 7B). Together, these results indicate that rat CaMKKβ
functions as an upstream Snf1-activating kinase in yeast cells
and suggest a physiological role for CaMKK isoforms as AMPK
kinases in mammalian cells.
nous Snf1 upstream kinases (Hong et al., 2005). These results
raise new possibilities for both the physiological regulation of
AMPK and for the role of CaMKK in vivo.
LKB1 was recently identified as a major upstream kinase in
the AMPK cascade (Hawley et al., 2003; Hong et al., 2003;
Shaw et al., 2004; Woods et al., 2003). Moreover, rat liver
AMPK kinase activity was found to copurify with LKB1, and
this activity was immunoprecipitated using anti-LKB1 antibod-
ies (Hawley et al., 2003; Woods et al., 2003). These studies
suggest that the major AMPK kinase activity in rat liver corres-
ponds to LKB1, but they do not rule out the possibility that
other AMPK kinase activities exist in different tissues. Previous
studies (Hawley et al., 2003; Shaw et al., 2004) reported low
but detectable AMPK activity in MEFs isolated from LKB1−/−
embryos. However, these same studies also reported that
AMPK activation by AICAR (Hawley et al., 2003; Shaw et al.,
2004), phenformin (Hawley et al., 2003), or oxidative stress
(Shaw et al., 2004) was almost completely abolished in these
cells. The significance of the residual AMPK activity, or the
identity of the upstream kinase(s) responsible for AMPK phos-
phorylation, in LKB−/−cells was not investigated further in
these studies. We reexamined the regulation of AMPK by phos-
phorylation in cells lacking LKB1 and found that there was a
marked activation of AMPK following oxidative stress (hy-
drogen peroxide) and hyperosmotic stress (sorbitol). These re-
sults clearly demonstrate that AMPK is activated in cells lack-
ing LKB1. Interestingly, both stresses have been reported to
increase intracellular Ca2+(see below) (Qin et al., 1997).
Based on the results obtained from siRNA-mediated knock-
down of CaMKK expression, we conclude that CaMKKβ is the
major AMPK kinase activity present in HeLa cells, although
CaMKKα may play a role in the activation of AMPK in response
to ionomycin. Two separate lines of evidence also lead us to
conclude that CaMKKβ has the major role over CaMKKα in
the activation of AMPK. Firstly, in CCL13 cells, expression of
CaMKKβ but not CaMKKα caused a marked increase in AMPK
activity. Secondly, CaMKKβ phosphorylated and activated
AMPK much more efficiently than CaMKKα in vitro. The finding
that CaMKKα is a very poor AMPK kinase may also provide an
explanation for a much earlier observation. Almost 10 years
ago, a preparation of CaMKK purified from pig brain was
shown to phosphorylate and activate AMPK in vitro. However,
it was concluded that CaMKK was unlikely to be a physiologi-
cally relevant AMPK kinase, in part because AMPK was a poor
substrate for CaMKK relative to CaMKI (Hawley et al., 1995). A
potential explanation for the apparent discrepancy between
our study and the earlier study is that the pig brain preparation
of CaMKK contained predominantly the CaMKKα isoform, al-
though we are unable to test this hypothesis directly.
CaMKKβ expression increased AMPK activity under basal
conditions, indicating that it is at least partially active at resting
levels of Ca2+. Substantial CaMKKβ activity in unstimulated
Jurkat T cells has been reported previously (Anderson et al.,
We show here that CaMKKβ is an upstream kinase in the
AMPK cascade. In cells lacking LKB1, a recently identified
AMPK kinase, AMPK was activated by oxidative stress and
hyperosmotic stress. Pharmacological inhibition of CaMKK in
cells lacking LKB1 almost completely abolished this activation.
Furthermore, using siRNA we show that downregulation of
CaMKKβ rather than CaMKKα activity markedly reduced
AMPK activation in HeLa cells. Expression of CaMKKβ but not
CaMKKα in CCL13 cells caused a dramatic increase in AMPK
activity. CaMKKβ purified from rat brain, or recombinant
CaMKKβ expressed in E. coli, phosphorylated and activated
AMPK efficiently in vitro, whereas, in comparison, rat brain
CaMKKα is a very poor AMPK kinase. Taken together, these
findings suggest that CaMKKβ, rather than CaMKKα, is the
major CaMKK activity directed against AMPK, at least in the
systems that we have studied. In addition, CaMKKβ was func-
tionally active in yeast in vivo and rescued a mutant yeast strain
lacking the three endogenous Snf1 upstream kinases. While
this manuscript was under revision, it was reported that
CaMKKα can rescue a mutant yeast lacking the three endoge-
D) AMPK activity in NIH3T3 cells and LKB1−/−cells following treatment with either 1 mM hydrogen peroxide (H2O2), 1 mM AICAR, 0.5 M sorbitol (all for 15 min), or 1
?M ionomycin (5 min) in the presence or absence of STO-609 (10 ?g/ml) was measured. Results shown are the mean ± SEM from four independent experiments and
are plotted as pmol/min/mg lysate. In every case, STO-609 caused a significant decrease in AMPK activity (* denotes p < 0.05, ** denotes p < 0.005, unpaired t test).
E) AMPK phosphorylation was determined by Western blot analysis of total cell lysates (50 ?g) probed with an anti-phosphospecific T172 antibody. Total AMPK was
determined using an anti-panβ antibody.
CELL METABOLISM : JULY 2005 27
A R T I C L E
Figure 5. CaMKK isoforms phosphorylate AMPK in vitro
A and B) CaMKKα and CaMKKβ were immunoprecipitated from 125 ?g rat brain lysate using isofom-specific CaMKK antibodies bound to protein A-Sepharose. The
immune complexes were used to phosphorylate GST-CaMKI(1−293)(10 ?g) or kinase-dead AMPK (α1β1γ1, 2 ?g) in the presence or absence of 2 mM Ca2+and 2 ?M
calmodulin. After 30 min incubation, the immune complexes were removed by brief centrifugation and the supernatant fraction containing the substrate was resolved
by SDS-PAGE. (A) Phosphorylated products were detected by autoradiography (1 hr for CaMKI, 12 hr for AMPK). A representative autoradiogram (showing duplicate
reactions in each case) is shown. (B) The gel was subjected to phosphoimage analysis and the relative degree of phosphorylation of the two substrates quantified. In
each case, the results are plotted as arbitrary units from the phospho-image analysis per nmol substrate used.
C) Bacterially expressed wild-type AMPK was incubated in the presence of MgATP with either CaMKKα or CaMKKβ immune complexes (isolated from 25 ?g rat brain
lysate) and in the presence (shaded bars) or absence (open bars) of 2 mM Ca2+and 2 ?M calmodulin. After 10 min, the immune complexes were removed by
centrifugation and AMPK activity in the supernatant fraction measured using the SAMS peptide assay. Results shown are the mean ± SEM for three independent
experiments, with duplicate assays in each case.
D) AMPK (3 ?g) wild-type (wt) or catalytically inactive (D157A) complex was incubated with 0.5 ?g bacterially expressed CaMKKβ in the presence of32P-ATP. Products
were resolved by SDS-PAGE and visualized by staining the gel with Coomassie (top panel), and radiolabeled products were detected by autoradiography (bottom panel).
CELL METABOLISM : JULY 2005
CaMKKβ is an AMPK kinase
nomycin treatment also led to a significant decrease in the
ATP:ADP ratio, raising the possibility that AMP may be involved
in the signaling pathway. However, whereas AMP was almost
ineffective in directly stimulating CaMKKβ activity, Ca2+/cal-
modulin stimulated its activity over 4-fold. The simplest inter-
pretation of these results is that ionomycin exerts its effect on
AMPK predominantly via a Ca2+-CaMKK-mediated pathway,
although further studies will be required to rule out other possi-
Numerous studies have shown that AMPK is activated in re-
sponse to depletion of ATP and a concomitant rise in the
AMP:ATP ratio (Carling, 2004; Hardie et al., 1998). Our results
now provide strong evidence that intracellular Ca2+acts as a
second signaling pathway to activate AMPK. A previous study
reported that AMP promotes the phosphorylation of AMPK by
LKB1, probably via a substrate (AMPK)-mediated effect (Haw-
ley et al., 2003). In our current study, AMP caused a modest
stimulation of the activation of AMPK by LKB1 and a smaller
effect on CaMKKβ-mediated activation. Similar to that reported
for LKB1 (Hawley et al., 2003), the effect of AMP on CaMKKβ-
mediated phosphorylation of AMPK is most likely a substrate-
mediated one, since AMP had no effect on the phosphorylation
of CaMKI by CaMKKβ. It is interesting to note that an earlier
study reported that AMP promoted AMPK activation by
CaMKIK but had no effect on activation of CaMKI by CaMKIK
(Hawley et al., 1995). On the other hand, Ca2+/calmodulin stim-
ulated CaMKKβ activity over 4-fold but has no significant effect
on LKB1 activity (our unpublished data). Consistent with our
results, Hawley et al. reported previously that Ca2+/calmodulin
did not directly activate a rat liver preparation of AMPKK (Haw-
ley et al., 1995), which, although not known at the time, was
most likely to be LKB1. Taken together, an attractive hypothe-
sis would be that CaMKKβ (and possibly CaMKKα) mediates
the activation of AMPK primarily in response to Ca2+, whereas
LKB1 mediates activation primarily in response to an increase
in the AMP:ATP ratio. However, our findings suggest that
CaMKKβ (or CaMKKα) may also play a role in the activation of
AMPK in response to a change in the AMP:ATP ratio, since the
effect of AMP appears to be mediated via AMPK. Furthermore,
distinguishing between the two signaling pathways may be dif-
ficult because under some circumstances both pathways are
likely to be affected in parallel. Many situations that lead to an
increase in Ca2+would also be predicted to deplete ATP
through activation of Ca2+pumps (Carafoli, 2004). Conversely,
many of the conditions that activate AMPK concomitant with
an increase in the AMP:ATP ratio also increase intracellular
Ca2+, e.g., exercise in muscle, oxidative stress, and uncoupling
of mitochondria (Patel et al., 2001). It is possible, therefore,
that in many cases AMPK activation is a consequence of both
nucleotide- and Ca2+-mediated effects and may involve LKB1
In order to investigate the relative contribution of CaMKKs
and LKB1 to activate AMPK in response to different stimuli, we
investigated the effect of STO-609 on AMPK activity in NIH3T3
Figure 6. Effect of AMP on CaMKKβ activity
A) Rat liver AMPK was immunoprecipitated with anti-α1 or anti-α2 antibodies
and treated with recombinant protein phosphatase 2C for 30 min at 30°C. After
extensive washing, the immune complexes were incubated with either recombi-
nant CaMKKβ or recombinant LKB1 and MgATP in the presence or absence of
0.2 mM AMP. The activity of CaMKKβ and LKB1 was assessed by determining
the resulting activation of AMPK using the SAMS peptide assay (measured in the
presence of 0.2 mM AMP). The direct effect of AMP on AMPK was measured by
assaying active AMPK (in immune complexes) in the presence or absence of 0.2
mM AMP. In each case, the results are plotted as ratios of the activities with and
without AMP and are the mean ± SEM of six (CaMKKβ) or four (LKB1) indepen-
dent immunoprecipitations with assays in duplicate in each case. Statistically
significant differences between activities in the presence and absence of AMP
are denoted by * (p < 0.05, unpaired t test) and the dotted line denotes activity
in the absence of AMP.
B) CaMKKα and CaMKKβ, immunoprecipitated from rat brain, were assayed in
the presence of 2 mM Ca2+and 2 ?M calmodulin and in the presence or absence
of 0.2 mM AMP using GST-CaMKI as substrate. Phosphorylated products were
detected by autoradiography, and a representative autoradiogram (with duplicate
reactions in each case) is shown.
1998). Whether this represents a physiologically relevant activ-
ity or is a consequence of the overexpression systems used
remains to be determined. As a first step to address whether
increasing intracellular Ca2+activates AMPK, we used iono-
mycin, a Ca2+ionophore. Ionomycin caused a significant acti-
vation of AMPK in all cell lines we studied, and in HeLa cells
this activation was almost completely blocked by STO-609,
suggesting that activation is dependent on CaMKK activity. Io-
E) The activity of wild-type AMPK (0.05 ?g) following phosphorylation with varying amounts of bacterially expressed CaMKKβ was determined using the SAMS peptide
assay. Results are the mean ± SEM from four independent experiments and are calculated as ?mol phosphate incorporated per minute per mg protein.
F) In separate incubations, the degree of T172 phosphorylation was determined by Western blot analysis of recombinant AMPK (0.5 ?g) with an anti-phosphospecific
CELL METABOLISM : JULY 2005 29
A R T I C L E
Figure 7. Rat CaMKKβ activates Snf1 protein kinase
HA-tagged yeast Pak1 and rat CaMKKβ were ex-
pressed from plasmids in the pak1? tos3? elm1?
triple mutant yeast strain. Control cells expressed
HA alone from the empty vector.
A) Cells were spotted with serial 5-fold dilutions on
synthetic complete medium lacking uracil (to select
for plasmids) and containing either 2% glucose, 2%
raffinose with 1 mg/l of the respiratory inhibitor anti-
mycin A, or 2% glycerol plus 3% ethanol. Plates
were incubated at 30° and photographed after 2
days (glucose) or 6 days (raffinose and glycerol-eth-
anol). Two independent yeast transformants ex-
pressing CaMKKβ were spotted.
B) Cells were grown to midlog phase in selective
synthetic medium containing 2% glucose, collected
by filtration, resuspended in medium containing
0.05% glucose for 30 min, and collected by filtra-
tion. Extracts were prepared, and Snf1 protein ki-
nase was partially purified. Snf1 catalytic activity
was assayed by measuring phosphorylation of the
SAMS peptide substrate, as described previously
(Hong et al., 2003). In each case, extracts were pre-
pared from two independent cultures, and values
are averages of six assays from the two experi-
ments. Kinase activity is expressed as nmol of
phosphate incorporated into the peptide per min
per mg of protein.
fibroblasts expressing CaMKKs and LKB1. Although these
cells are not a perfect control for the LKB1−/−MEFs, which are
a heterogeneous population of fibroblasts isolated from an en-
tire embryo, they do express both CaMKK isoforms and LKB1.
Comparison of the activation of AMPK in these cells versus
LKB1−/−MEFs provides an indication of the contribution of the
different upstream kinases. For all of the conditions we exam-
ined, including incubation with AICAR, STO-609 caused a sig-
nificant reduction in AMPK activity, although the degree of inhi-
bition varied between 20% and 50%, depending on the cell
treatment. These results suggest that, in cells expressing
CaMKKs and LKB1, inhibition of CaMKK is not sufficient to
completely block AMPK activation. In contrast, in cells lacking
LKB1, STO-609 almost completely abolished AMPK activity in
response to all the treatments we used. The simplest inter-
pretation of these results is that AMPK activation involves
CaMKKs and LKB1, but in the absence of LKB1, CaMKKs con-
tribute the major role in activation. A caveat to this hypothesis
is that, in addition to CaMKKs, other AMPK kinases that are
also sensitive to inhibition by STO-609 are present in cells. Al-
though we cannot rule out this possibility, we think it is unlikely
to have a major bearing on our conclusions for the following
reasons. In vitro, STO-609 inhibits CaMKKα with an IC50of 120
ng/ml and CaMKKβ with an IC50of 40 ng/ml, concentrations
that are at least two orders of magnitude lower than the IC50
values reported for several other protein kinases (Tokumitsu et
al., 2002), suggesting that STO-609 is a selective CaMKK in-
hibitor. Secondly, in our cell-based studies, STO-609 caused a
dose-dependent reduction in the activation of AMPK with an
IC50value of approximately 0.5 ?g/ml, remarkably similar to
the IC50value previously reported for inhibition of CaMKIV acti-
vation in HeLa cells (Tokumitsu et al., 2002). Finally, downregu-
lation of CaMKKβ using siRNA also significantly reduced
AMPK activity in cells lacking LKB1, suggesting that, if other
AMPK kinases do exist, they are unlikely to play a major role
in AMPK activation, at least under the conditions we have
A key goal that stems from our study will be to determine the
physiological roles of CaMKKs and LKB1 in regulating AMPK
activity. CaMKKα and CaMKKβ are closely related, and the rat
isoforms share approximately 65% amino acid sequence iden-
tity (Anderson et al., 1998). Northern blot analysis shows that
both CaMKKα and CaMKKβ mRNAs are expressed predomi-
nantly in the brain (Hsu et al., 1998; Tokumitsu et al., 1995).
Based on these findings, brain represents a likely tissue where
CaMKK may play a significant role in activation of AMPK. In
addition, a number of stimuli predicted to increase Ca2+have
been reported to lead to activation of AMPK, e.g., Gq-coupled
receptor agonists (Kishi et al., 2000), or thrombin and histamine
in endothelial cells (Thors et al., 2004). Furthermore, it was re-
ported recently that increasing intracellular Ca2+by K+-induced
depolarization in MIN6 pancreatic β cells led to AMPK activa-
tion, despite elevation of ATP levels in these cells (Leclerc and
Rutter, 2004). It will be important to determine the role of
CaMKK in these and other responses that activate AMPK con-
comitant with an increase in Ca2+signaling. Clearly, it will be
important to establish the relevant contribution of the individual
CaMKK isoforms and LKB1 to activation of AMPK in different
tissues and in response to different stimuli.
STO-609 was from Tocris (Ellisville, Missouri). Ionomycin was from Sigma
(Poole, Dorset, UK). LKB−/−cells were a generous gift from Dr. R. DePinho
(Harvard Medical School). cDNAs encoding CaMKI and CaMKK isoforms
were a generous gift from Dr. H. Tokumitsu (Kagawa Medical University,
Japan). Recombinant LKB1 and phospho-specific antibodies against ACC
were from Upstate.
CELL METABOLISM : JULY 2005
CaMKKβ is an AMPK kinase
Preparation and assay of AMPK
Recombinant AMPK was expressed in E. coli and purified as previously
described (Neumann et al., 2003). Rat liver AMPK was purified up to DEAE-
Sepharose step as described previously (Carling et al., 1989). AMPK was
immunoprecipitated using a rabbit anti-pan β antibody (Woods et al., 1996a)
bound to protein A-Sepharose, or sheep anti-α1 or anti-α2 antibodies
(Woods et al., 1996b) bound to protein G-Sepharose. In some cases, AMPK
was inactivated by incubation with 16 U/ml recombinant protein phospha-
tase 2C (Marley et al., 1996) for 30 min 37°C. AMPK activity was determined
by phosphorylation of the synthetic SAMS peptide (Davies et al., 1989) in
the presence of 5 mM MgCl2, 0.2 mM [γ-32P]ATP (specific radioactivity ap-
proximately 200 cpm/pmol). Unless stated otherwise, reactions were car-
ried out in the presence 0.2 mM AMP.
quent analysis. CaMKK activity was measured in one of two ways. In one
method, GST-CaMKI (1–293, K49E) was used as a substrate, as previously
described (Tokumitsu et al., 2000). Briefly, purified GST-CaMKI (10 ?g) was
incubated with CaMKK in 50 mM HEPES (pH 7.4) containing 0.2 mM
[γ-32P]ATP, 5 mM MgCl2, for 30 min at 37°C. Unless stated otherwise, incu-
bations also contained 2 mM CaCl2and 2 ?M bovine brain calmodulin
(Calbiochem). The reaction was stopped by addition of SDS sample buffer
and the samples resolved by SDS-PAGE on 12% gels. The gels were
stained with Coomassie blue, dried, and subjected to autoradiography or
phosphoimage analysis (Storm 820 using ImageQuant Software, Amer-
sham). In the second method, purified recombinant AMPK complexes (0.2
?g) were incubated with CaMKK in 50 mM HEPES (pH 7.4) containing 0.2
mM ATP, 5 mM MgCl2, for 20 min at 37°C. Unless stated otherwise, incub-
ations also contained 2 mM CaCl2and 2 ?M bovine brain calmodulin (Cal-
biochem) and 0.2 mM AMP. At the end of the incubation period, an aliquot
containing 0.05 ?g AMPK was removed and AMPK activity measured using
the SAMS peptide assay.
Expression of CaMKK isoforms
cDNA encoding rat CaMKKβ was cloned into the pET DUET-1 vector (Nova-
gen) so that the recombinant protein encoded a hexahistidine tag at the N
terminus. CaMKKβ was expressed in E. coli (BL21 cells) following induction
by 1 mM isopropyl-β-thiogalactopyranoside for 4 hr at 37°C. Cells were
sonicated and insoluble material was removed by centrifugation. CaMKKβ
present in the soluble fraction was purified on Ni-NTA resin (Qiagen). Rat
CaMKI (residues 1–293, harboring a mutation of lysine 49 to glutamic acid,
K49E) was expressed as a fusion with glutathione S-transferase (GST) in
pGEX-4T3 (a gift from H. Tokumitsu) and purified as described elsewhere
(Tokumitsu et al., 2000). cDNAs encoding rat CaMKKα and rat CaMKKβ
were cloned into the mammalian expression vector pCDNA3 (Invitrogen). A
catalytically inactive mutant of CaMKKβ was made by introducing a muta-
tion of aspartatic acid 329 in the ATP binding region to alanine (D329A) and
cloned into pCDNA3. All constructs were validated by DNA sequencing of
Cells were rinsed briefly in phosphate-buffered saline and lysed by the di-
rect addition of perchloric acid to a final concentration of 7% (v/v). Insoluble
material was removed by centrifugation, and nucleotides in the supernatant
fraction were analyzed as described previously (Fryer et al., 2002).
Western blot analysis
CaMKKα was detected using rabbit anti-CaMKKα antibody (R-73, Santa
Cruz) and CaMKKβ using goat anti CaMKKβ (a mixture of C-20 [sc 1540]
and L-19 [sc 9629], both Santa Cruz). T172 phosphorylation was deter-
mined using rabbit anti-pT172 antibody (Cell Signaling). Total AMPK was
detected by blotting with a rabbit anti-panβ antibody (Woods et al., 1996a).
Rabbit phospho-specific ACC antibodies were from Upstate. Blots were
developed with horseradish-peroxidase-linked secondary antibodies and
visualized by enhanced chemiluminescence (Pierce SuperSignal West Dura
or Femto kit).
Transient transfection of CCL13 cells
Plasmid DNA was prepared using a QIAGEN maxiprep kit according to the
manufacturer’s instructions. CCL13 cells were transfected with 30 ?g plas-
mid encoding rat CaMKKα, rat CaMKKβ, or a catalytically inactive mutant
of rat CaMKKβ by Ca2+phosphate precipitation (Chen and Okayama, 1987).
As a control, cells were transfected with pCDNA3 encoding β-galactosi-
dase. In each case, cells were harvested 48 hr posttransfection.
HA-tagged yeast Pak1 and rat CaMKKβ were expressed from the yeast
ADH1 promoter from plasmids pRH104 (Hedbacker et al., 2004) and
pRH134, respectively, which are derivatives of vector pWS93 (Song and
Carlson, 1998). pRH134 was constructed by cloning a PCR product ampli-
fied from the rat CaMKKβ cDNA into pWS93 at the BamHI site. Expression
of HA-tagged CaMKKβ was confirmed by immunoblot analysis. Plasmids
were used to transform Saccharomyces cerevisiae strain MCY5138 (MATa
pak1D::kanMX4 tos3D::kanMX4 elm1D::ADE2 ura3 trp1 ade2 his3 can1
Mammalian cell culture
Cells were maintained in DMEM supplemented with 10% fetal calf serum
and 2 mM L-glutamine. Cells were transferred into serum-free media for 4
hr prior to treatment with 1 mM hydrogen peroxide or 0.5 M sorbitol (both
for 15 min) or 1 ?M ionomycin (for 5 min). Cells were rinsed briefly in phos-
phate-buffered saline (PBS) before lysis in 1 ml of ice-cold buffer (50 mM
HEPES [pH 7.5], 50 mM NaF, 5 mM sodium pyrophosphate, 1 mM EDTA, 1
mM dithiothreitol, 0.1 mM phenylmethylsulphonyl fluoride, 157 ?g/ml ben-
zamidine, 4 ?g/ml trypsin inhibitor, 10% [v/v] glycerol, 1% [v/v] Triton X-100).
Insoluble material was removed by brief centrifugation and the supernatant
fraction used for subsequent analysis.
Results are expressed as means ± SEM. Analysis was carried out using a
two-tailed, unpaired Student’s t test.
Synthetic RNA duplexes, based on human cDNAs encoding either for
CaMKKα or CaMKKβ, were designed and used to transfect HeLa cells. The
oligonucleotides used for CaMKKα were the following: sense, r(GCGUUAU
CUGGAAAGUGGA)dTdT; and antisense, r(UCCACUUUCCAGAUAACGC)
dCdG. The following were used for CaMKKβ: sense, r(AGCUGAUUGUU
GUGGUCAA)dTdT; and antisense, r(UUGACCACAACAAUCAGCU)dAdG.
As a control, cells were transfected with a nonspecific RNA duplex (negative
control siRNA, Ambion). Cells were plated in 6-well plates 18 hr prior to
transfection. Transfection was performed using Oligofectamine (Invitrogen)
according to manufacturer’s instructions.
We are grateful to Dr. H. Tokumitsu for cDNA clones encoding CaMKI,
CaMKKα, and CaMKKβ. Mouse embryonic fibroblasts derived from LKB1−/−
mice were a generous gift from Dr. R DePinho. This work was supported
by the Medical Research Council UK, Diabetes UK (RD02/0002383), and
an RTD grant (QLG1-CT-2001-01488) from the European Commission (to
D.C.). M.C. was supported by NIH grant GM34095.
Received: February 22, 2005
Revised: May 2, 2005
Accepted: June 9, 2005
Published: July 19, 2005
CaMKK immunoprecipitation and assay
CaMKK isoforms were immunoprecipitated from rat brain extracts or cul-
tured cell extracts using a rabbit anti-CaMKKα antibody (R-73 [sc11370]
Santa Cruz) bound to protein A-Sepharose or a goat anti-CaMKKβ antibod-
ies (a mixture of C-20 [sc 1540] and L-19 [sc 9629], Santa Cruz) bound
to protein G-Sepharose. The resulting immune complexes were washed
extensively with 50 mM HEPES (pH 7.4), 1 mM EGTA, and used for subse-
Altarejos, J.Y., Taniguchi, M., Clanachan, A.S., and Lopaschuk, G.D. (2005).
Myocardial ischemia differentially regulates LKB1 and an alternate 5#-AMP-
activated protein kinase kinase. J. Biol. Chem. 280, 183–190.
CELL METABOLISM : JULY 2005 31
A R T I C L E
Anderson, K.A., Means, R.L., Huang, Q.H., Kemp, B.E., Goldstein, E.G.,
Selbert, M.A., Edelman, A.M., Fremeau, R.T., and Means, A.R. (1998). Com-
ponents of a calmodulin-dependent protein kinase cascade. Molecular
cloning, functional characterization and cellular localization of Ca2+/cal-
modulin-dependent protein kinase kinase beta. J. Biol. Chem. 273,
Hong, S.P., Momcilovic, M., and Carlson, M. (2005). Function of mammalian
LKB1 and Ca2+/calmodulin-dependent protein kinase kinase alpha as
Snf1-activating kinases in yeast. J. Biol. Chem. 280, 21804–21809.
Hsu, L.S., Tsou, A.P., Chi, C.W., Lee, C.H., and Chen, J.Y. (1998). Cloning,
expression and chromosomal localization of human Ca2+/calmodulin-
dependent protein kinase kinase. J. Biomed. Sci. 5, 141–149.
Andersson, U., Filipsson, K., Abbott, C.R., Woods, A., Smith, K., Bloom,
S.R., Carling, D., and Small, C.J. (2004). AMP-activated protein kinase plays
a role in the control of food intake. J. Biol. Chem. 279, 12005–12008.
Ishikawa, Y., Tokumitsu, H., Inuzuka, H., Murata-Hori, M., Hosoya, H., and
Kobayashi, R. (2003). Identification and characterization of novel compo-
nents of a Ca2+/calmodulin-dependent protein kinase cascade in HeLa
cells. FEBS Lett. 550, 57–63. Bardeesy, N., Sinha, M., Hezel, A.F., Signoretti, S., Hathaway, N.A.,
Sharpless, N.E., Loda, M., Carrasco, D.R., and DePinho, R.A. (2002). Loss
of the Lkb1 tumour suppressor provokes intestinal polyposis but resistance
to transformation. Nature 419, 162–167.
Jenne, D.E., Reimann, H., Nezu, J.I., Friedel, W., Loff, S., Jeschke, R.,
Müller, O., Back, W., and Zimmer, M. (1998). Peutz-Jeghers syndrome is
caused by mutations in a novel serine threonine kinase. Nat. Genet. 18,
38–43.Carafoli, E. (2004). Calcium-mediated cellular signals: a story of failures.
Trends Biochem. Sci. 29, 371–379.
Kahn, B.B., Alquier, T., Carling, D., and Hardie, D.G. (2005). AMP-activated
protein kinase: ancient energy gauge provides clues to modern understand-
ing of metabolism. Cell Metab. 1, 15–25.
Carling, D. (2004). The AMP-activated protein kinase cascade—a unifying
system for energy control. Trends Biochem. Sci. 29, 18–24.
Kim, E.K., Miller, I., Aja, S., Landree, L.E., Pinn, M., McFadden, J., Kuhajda,
F.P., Moran, T.H., and Ronnett, G.V. (2004a). C75, a fatty acid synthase in-
hibitor, reduces food intake via hypothalamic AMP-activated protein kinase.
J. Biol. Chem. 279, 19970–19976.
Carling, D., Clarke, P.R., Zammit, V.A., and Hardie, D.G. (1989). Purification
and characterisation of the AMP-activated protein kinase. Eur. J. Biochem.
Chen, C., and Okayama, H. (1987). High efficiency transformation of mam-
malian cells by plasmid DNA. Mol. Cell. Biol. 7, 2745–2752.Kim, M.S., Park, J.Y., Namkoong, C., Jang, P.G., Ryu, J.W., Song, H.S.,
Yun, J.Y., Namgoong, I.S., Ha, J., Park, I.S., et al. (2004b). Anti-obesity
effects of alpha-lipoic acid mediated by suppression of hypothalamic AMP-
activated protein kinase. Nat. Med. 10, 727–733.
Choi, S.L., Kim, S.J., Lee, K.-T., Kim, J., Mu, J., Birnbaum, M.J., Soo Kim,
S.S., and Ha, J. (2001). The regulation of AMP-activated protein kinase by
H2O2. Biochem. Biophys. Res. Commun. 287, 92–97.
Kishi, K., Yuasa, T., Minami, A., Yamada, M., Hagi, A., Hayashi, H., Kemp,
B.E., Witters, L.A., and Ebina, Y. (2000). AMP-activated protein kinase is
activated by the stimulations of G(q)-coupled receptors. Biochem. Biophys.
Res. Commun. 276, 16–22.
Davies, S.P., Carling, D., and Hardie, D.G. (1989). Tissue distribution of
AMP-activated protein kinase, and lack of activation by cyclic AMP-depen-
dent protein kinase, studied using a specific and sensitive peptide assay.
Eur. J. Biochem. 186, 123–128.
Leclerc, I., and Rutter, G.A. (2004). AMP-activated protein kinase: a new
beta-cell glucose sensor? Regulation by amino acids and calcium ions. Dia-
betes Suppl. 3 53, S67–S74.
Edelman, A.M., Mitchelhill, K.I., Selbert, M.A., Anderson, K.A., Hook, S.S.,
Stapleton, D., Goldstein, E.G., Means, A.R., and Kemp, B.E. (1996). Multiple
Ca(2+)-calmodulin-dependent protein kinase kinases from rat brain. Purifi-
cation, regulation by Ca(2+)-calmodulin, and partial amino acid sequence.
J. Biol. Chem. 271, 10806–10810.
Lee, J.C., and Edelman, A.M. (1994). A protein activator of Ca(2+)-calmod-
ulin-dependent protein kinase Ia. J. Biol. Chem. 269, 2158–2164.
Fryer, L.G., Parbu-Patel, A., and Carling, D. (2002). The anti-diabetic drugs
rosiglitazone and metformin stimulate AMP-activated protein kinase
through distinct signaling pathways. J. Biol. Chem. 277, 25226–25232.
Marley, A.E., Sullivan, J.E., Carling, D., Abbott, W.M., Smith, G.J., Taylor,
I.W.F., Carey, F., and Beri, R.K. (1996). Biochemical characterization and
deletion analysis of recombinant human protein phosphatase 2C alpha. Bio-
chem. J. 320, 801–806.
Hardie, D.G., Carling, D., and Carlson, M. (1998). The AMP-activated/SNF1
protein kinase subfamily: metabolic sensors of the eukaryotic cell? Annu.
Rev. Biochem. 67, 821–855.
Minokoshi, Y., Kim, Y.B., Peroni, O.D., Fryer, L.G., Muller, C., Carling, D.,
and Kahn, B.B. (2002). Leptin stimulates fatty-acid oxidation by activating
AMP-activated protein kinase. Nature 415, 339–343.
Hawley, S.A., Selbert, M.A., Goldstein, E.G., Edelman, A.M., Carling, D.,
and Hardie, D.G. (1995). 5#-AMP activates the AMP-activated protein ki-
nase cascade and Ca2+/calmodulin activates the calmodulin-dependent
protein kinase-I cascade, via 3 independent mechanisms. J. Biol. Chem.
Minokoshi, Y., Alquier, T., Furukawa, N., Kim, Y.B., Lee, A., Xue, B., Mu, J.,
Foufelle, F., Ferre, P., Birnbaum, M.J., et al. (2004). AMP-kinase regulates
food intake by responding to hormonal and nutrient signals inthe hypothala-
mus. Nature 428, 569–574.
Hawley, S.A., Davison, M.D., Woods, A., Davies, S.P., Beri, R.K., Carling,
D., and Hardie, D.G. (1996). Characterisation of the AMP-activated protein
kinase kinase from rat liver and identification of threonine-172 as the major
site at which it phosphorylates AMP-activated protein kinase. J. Biol. Chem.
Neumann, D., Woods, A., Carling, D., Wallimann, T., and Schlattner, U.
(2003). Mammalian AMP-activated protein kinase: functional, heterotrimeric
complexes by co-expression of subunits in Escherichia coli. Protein Expr.
Purif. 30, 230–237.
Patel, N., Khayat, Z.A., Ruderman, N.B., and Klip, A. (2001). Dissociation of
5# AMP-activated protein kinase activation and glucose uptake stimulation
by mitochondrial uncoupling and hyperosmolar stress: differential sensitivi-
ties to intracellular Ca2+ and protein kinase C inhibition. Biochem. Biophys.
Res. Commun. 285, 1066–1070.
Hawley, S.A., Boudeau, J., Reid, J.L., Mustard, K.J., Udd, L., Makela, T.P.,
Alessi, D.R., and Hardie, D.G. (2003). Complexes between the LKB1 tumor
suppressor, STRADα/β and MO25α/β are upstream kinases in the AMP-
activated protein kinase cascade. J. Biol. 2, 28.
Qin, S., Minami, Y., Hibi, M., Kurosaki, T., and Yamamura, H. (1997). Syk-
dependent and -independent signaling cascades in B cells elicited by os-
motic and oxidative stress. J. Biol. Chem. 272, 2098–2103.
Hedbacker, K., Hong, S.-P., and Carlson, M. (2004). Pak1 protein kinase
regulates activation and nuclear localization of Snf1-Gal83 protein kinase.
Mol. Cell. Biol. 24, 8255–8263.
Hemminki, A., Markie, D., Tomlinson, I., Avizienyte, E., Roth, S., Loukola,
A., Bignell, G., Warren, W., Aminoff, M., Hoglund, P., et al. (1998). A serine/
threonine kinase gene defective in Peutz-Jeghers syndrome. Nature 391,
Salt, I.P., Celler, J.W., Hawley, S.A., Prescott, A., Woods, A., Carling, D., and
Hardie, D.G. (1998). AMP-activated protein kinase: greater AMP depen-
dence, and preferential nuclear localization, of complexes containing the α2
isoform. Biochem. J. 334, 177–187.
Hong, S.P., Leiper, F.C., Woods, A., Carling, D., and Carlson, M. (2003).
Activation of yeast Snf1 and mammalian AMP-activated protein kinase by
upstream kinases. Proc. Natl. Acad. Sci. USA 100, 8839–8843.
Selbert, M.A., Anderson, K.A., Huang, Q.H., Goldstein, E.G., Means, A.R.,
and Edelman, A.M. (1995). Phosphorylation and activation of Ca(2+)-cal-
modulin-dependent protein kinase IV by Ca(2+)-calmodulin-dependent pro-
CELL METABOLISM : JULY 2005
CaMKKβ is an AMPK kinase Download full-text
tein kinase Ia kinase. Phosphorylation of threonine 196 is essential for acti-
vation. J. Biol. Chem. 270, 17616–17621.
and expression of calcium/calmodulin-dependent protein kinase kinase. J.
Biol. Chem. 270, 19320–19324.
Tokumitsu, H., Muramatsu, M., Ikura, M., and Koboyashi, R. (2000). Regula-
tory mechanism of Ca2+/calmodulin-dependent protein kinase kinase. J.
Biol. Chem. 275, 20090–20095.
Shaw, R.J., Kosmatka, M., Bardeesy, N., Hurley, R.L., Witters, L.A., De-
Pinho, R.A., and Cantley, L.C. (2004). The tumor suppressor LKB1 kinase
directly activates AMP-activated kinase and regulates apoptosis in re-
sponse to energy stress. Proc. Natl. Acad. Sci. USA 101, 3329–3335.
Tokumitsu, H., Iwabu, M., Ishikawa, Y., and Kobayashi, R. (2001). Differential
regulatory mechanism of Ca2+/calmodulin-dependent protein kinase ki-
nase isoforms. Biochemistry 40, 13925–13932.
Song, W., and Carlson, M. (1998). Srb/mediator proteins interact function-
ally and physically with transcriptional repressor Sfl1. EMBO J. 17, 5757–
Tokumitsu, H., Inuzuka, H., Ishikawa, Y., Ikeda, M., Saji, I., and Kobayashi,
R. (2002). STO-609, a specific inhibitor of the Ca(2+)/calmodulin-dependent
protein kinase kinase. J. Biol. Chem. 277, 15813–15818.
Stein, S.C., Woods, A., Jones, N.A., Davison, M.D., and Carling, D. (2000).
The regulation of AMP-activated protein kinase by phosphorylation. Bio-
chem. J. 345, 437–443.
Woods, A., Cheung, P.C.F., Smith, F.C., Davison, M.D., Scott, J., Beri, R.K.,
and Carling, D. (1996a). Characterization of AMP-activated protein kinase β
subunit and γ subunit—assembly of the heterotrimeric complex in vitro. J.
Biol. Chem. 271, 10282–10290.
Sutherland, C.M., Hawley, S.A., McCartney, R.R., Leech, A., Stark, M.J.R.,
Schmidt, M.C., and Hardie, D.G. (2003). Elm1p is one of three upstream
kinases for the Saccharomyces cerevisiae SNF1 complex. Curr. Biol. 13,
Woods, A., Salt, I., Scott, J., Hardie, D.G., and Carling, D. (1996b). The α1
and α2 isoforms of the AMP-activated protein kinase have similar activities
in rat liver but exhibit differences in substrate specificity in vitro. FEBS Lett.
Thors, B., Halldorsson, H., and Thorgeirsson, G. (2004). Thrombin and his-
tamine stimulate endothelial nitric-oxide synthase phosphorylation at
Ser1177 via an AMPK mediated pathway independent of PI3K-Akt. FEBS
Lett. 573, 175–180.
Woods, A., Johnstone, S.R., Dickerson, K., Leiper, F.C., Fryer, L.G., Neu-
mann, D., Schlattner, U., Wallimann, T., Carlson, M., and Carling, D. (2003).
LKB1 is the upstream kinase in the AMP-activated protein kinase cascade.
Curr. Biol. 13, 2004–2008.
Tiainen, M., Ylikorkala, A., and Makela, T.P. (1999). Growth suppression by
Lkb1 is mediated by a G1 cell cycle arrest. Proc. Natl. Acad. Sci. USA 96,
Yamauchi, T., Kamon, J., Minokoshi, Y., Ito, Y., Waki, H., Uchida, S., Yama-
shita, S., Noda, M., Kita, S., Ueki, K., et al. (2002). Adiponectin stimulates
glucose utilization and fatty-acid oxidation by activating AMP-activated pro-
tein kinase. Nat. Med. 8, 1288–1295.
Tokumitsu, H., Enslen, H., and Soderling, T.R. (1995). Characterization of a
Ca2+/calmodulin-dependent protein kinase cascade. Molecular cloning
CELL METABOLISM : JULY 2005 33