1

Glycogen synthase kinase 3 is activated by cAMP and

plays an active role in the regulation of melanogenesis

Mehdi KHALED, Lionel LARRIBERE, Karine BILLE, Edith ABERDAM, Jean-Paul

ORTONNE, Robert BALLOTTI and Corine BERTOLOTTO*.

INSERM U385, Biologie et Physiopathologie de la peau, IFR 50, 28, avenue de

Valombrose, 06107 NICE Cedex 2, France.

Phone number: (33) 4 93 37 77 90

Fax number: (33) 4 93 81 14 04

Email: bertolot@unice.fr

* Address request to Corine Bertolotto

Running title: Involvement of GSK3β during cAMP-induced melanogenesis

key words: cAMP, AKT, GSK3β, MITF, tyrosinase

Copyright 2002 by The American Society for Biochemistry and Molecular Biology, Inc.

JBC Papers in Press. Published on July 1, 2002 as Manuscript M202939200 by guest on A

ugust 20, 2018http://w

ww

.jbc.org/D

ownloaded from

2

Abstract

In human and mouse, cAMP plays a key role in the control of pigmentation. cAMP,

through the activation of PKA, increases the expression of Microphthalmia-

associated transcription factor (MITF), which in turn stimulates tyrosinase gene

expression, to allow melanin synthesis. Beyond this simplified scheme, cAMP inhibits

the phosphatidylinositol-3-kinase (PI3K) and inhibition of PI3K, by a specific inhibitor,

stimulates melanogenesis. However, the link between the PI3K pathway and

melanogenesis remained to be elucidated. In this report, we showed that cAMP,

through a PKA-independent mechanism, led to inhibition of AKT phosphorylation and

activity. Consistent with the role of AKT in the regulation of GSK3β, cAMP decreased

the phosphorylation of GSK3 β and stimulated its activity. Further, experiments were

performed to investigate the role of GSK3β in the regulation of MITF expression and

function. We observed that GSK3β regulated neither MITF promoter activity nor the

intrinsic transcriptional activity of MITF, but synergized with MITF to activate the

tyrosinase promoter. Additionally, lithium, a GSK3β inhibitor, impaired the response

of the tyrosinase promoter to cAMP and cAMP increased the binding of MITF to the

M-box. Taking into account that GSK3β phosphorylates MITF and increases the

ability of MITF to bind its target sequence, our results indicate that activation of

GSK3β by cAMP facilitates MITF binding to the tyrosinase promoter, thereby leading

to stimulation of melanogenesis.

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

3

Introduction

In mammals, epidermal melanocytes synthesize and transfer melanin pigment to

surrounding keratinocytes to allow skin and hair pigmentation. In human, melanins

play a crucial photo-protective role against the carcinogenic and deleterious effects of

ultraviolet radiation of the solar light. Hence, numerous efforts have been made to

understand the molecular mechanisms that govern pigment production.

In mice as well as in humans, it is now well established that pro-opiomelanocortic

(POMC) peptides, adrenocorticotrophic hormone (ACTH) and α-melanocyte

stimulating hormone (αMSH), play a key role in the control of pigmentation. Indeed,

any disturbance of αMSH or ACTH signalling, due to mutations of the melanocortin

type I receptor (MC1R) (1) or to a decrease of αMSH or ACTH expression, results in

an inhibition of melanin synthesis (2). Conversely, pathologic over-expression of

αMSH (3) or ACTH (4,5), or administration of αMSH analogue to human voluntaries

increase skin pigmentation (6,7). Binding of ACTH or αMSH to the Gαs-coupled

MC1R, leads to adenylate cyclase activation, elevation of intracellular cAMP and

activation of the protein kinase A (PKA) (8). The cAMP pathway plays a pivotal role in

the regulation of skin pigmentation. Indeed, patients with Mac Cune-Albright

syndrome display large hyper-pigmented areas caused by an activating mutation in

the Gαs protein that controls cAMP level (9,10). Further, mutations in the type Iα

regulatory subunit of PKA, leading to a constitutive activation of PKA, have been

described in patients with Carney syndrome characterized by spotty skin

pigmentation (11,12). Finally, pro-pigmenting effects of αMSH can be mimicked, in

vitro, by forskolin that directly binds and activates adenylate cyclase (13,14).

Taken together, these observations clearly demonstrate the meaningful role of the

cAMP pathway in the regulation of melanogenesis and skin pigmentation by

melanocortins.

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

4

cAMP increases melanogenesis mainly through the stimulation of tyrosinase

expression, the enzyme catalyzing the rate limiting reaction of the melanin synthesis

process (15-17). We have shown that cAMP, through activation of PKA and CREB

transcription factor, promotes an increase in the expression of Microphthalmia-

associated transcription factor (MITF) (18), a melanocyte-specific transcription factor

crucial for melanocyte development and differentiation (19,20). As a result, MITF

binds to and activates the tyrosinase promoter, leading thereby to stimulation of

melanogenesis (21-23).

In the course of these studies, we observed that cAMP regulates other signaling

pathways that are also involved in the control of melanogenesis. In melanocytes and

melanoma cells, elevation of the intracellular cAMP content results in the activation of

the Ras/ERK cascade (13). However, activation of Ras/ERK leads to an inhibition of

melanogenesis, and this pathway has been thought to be a feedback mechanism

preventing an excessive production of melanin that would be toxic for cells (24).

Indeed, ERK and RSK, which is activated by ERK, phosphorylate MITF and promote

its degradation thereby leading to an inhibition of tyrosinase expression and of

melanogenesis (25,26). Further, in B16 melanoma cells, we previously showed that

cAMP inhibits the phosphatidylinositol-3 kinase (PI3K) (27). Additionally, the PI3K

specific inhibitor, LY294002, stimulates melanogenesis. These observations suggest

that PI3K pathway might be involved in the regulation of melanin synthesis by cAMP.

While PKA/CREB and Ras/ERK pathways have been thoroughly dissected, the

involvement of the PI3K pathway in cAMP-induced melanogenesis remains to be

elucidated.

One of the key effector of PI3K is the serine/threonine kinase AKT. In response to

growth factors and hormones, AKT is activated by binding to the membrane PI3K

phospholipid products and phosphorylation on threonine 308 and serine 473 (28,29).

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

5

Activated AKT phosphorylates, the glycogen synthase kinase 3β (GSK3β) on serine

9, and promotes its inactivation (30). GSK3β is a serine/threonine kinase, first

described to regulate glycogen synthase activity. More recently, GSK3β was

demonstrated to be a key mediator of vertebrate development, tumorigenesis and

cell differentiation (31,32).

Taking into account the role of cAMP in the regulation of melanogenesis, we wished

to evaluate the involvement of AKT and GSK3β in the regulation of melanocyte

differentiation by cAMP. In this report, we clearly demonstrated, in B16 melanoma

cells, that cAMP, by a PKA-independent mechanism, led to inhibition of AKT

phosphorylation and activity, resulting in a dephosphorylation and activation of

GSK3β. Further, we showed that GSK3β regulated neither MITF transcription nor

MITF intrinsic transcriptional activity. However, we found that lithium, a GSK3β

inhibitor, decreased the response to cAMP of the tyrosinase promoter. Additionally,

we showed that GSK3β synergized with MITF to stimulate the tyrosinase promoter.

Finally, short cAMP treatment, that did not up-regulate MITF expression, enhanced

the binding of MITF to the M-box sequence of the tyrosinase promoter. Together, our

results suggest that activation of GSK3β by cAMP, is involved in the regulation of

cAMP-induced melanogenesis by up-regulating the binding of MITF to the tyrosinase

promoter.

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

6

Experimental procedures

Materials.

Forskolin, sodium fluoride, sodium orthovanadate, 4-(2-aminoethyl)-benzene-sulfonyl

fluoride (AEBSF), aprotinin and leupeptin were purchased from Sigma Chem.Co. The

PKA inhibitor H89-dihydrochloride and the PI3K inhibitor LY294002 were from

MERCK Eurolab. Dulbecco's modified Eagle's medium (DMEM), trypsin and

Lipofectamine reagent were from GIBCO and fetal calf serum (FCS) was from

Hyclone. C. sordelii lethal toxin (LT) was a gift from Pr. P. Boquet (Nice, France). The

peptide GRPRTSSFAEG (Crosstide) and the glycogen synthase peptide-2 (GS

peptide-2) (YRRAAVPPSPSLSRHSSPHQ-pSEDEEE) were from Euromedex.

Antibodies.

The polyclonal phospho-specific AKT (S473 and T308), GSK3β (S9), CREB (S133),

p42/44 MAPK antibodies and polyclonal antibody that recognizes AKT regardless of

its phosphorylation state were from Cell Signaling. The monoclonal GSK3β (0011-A)

and monoclonal ERK2 (D-2) antibodies were from Santa Cruz. Monoclonal anti-

hemagglutinin (HA) 12CA5 antibody was from BABCO. The phospho-glycogen

synthase antibody (Ab-1) and the phospho-specific Tau antibody (S396) were

purchased from Oncogene Research products. Horseradish peroxidase-conjugated

anti-rabbit or anti-mouse antibodies were from Dakopatts.

Cell Cultures.

B16/F10 murine melanoma cells were grown at 37°C under 5% CO2 in DMEM

supplemented with 7% FCS and penicillin/streptomycin (100 U/ml/50 µg/ml).

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

7

Western blot assays.

B16 melanoma cells were cultured in 6-well dishes with or without different effectors

for the time indicated in the figure legends. Then, cells were lysed in buffer A

containing 50 mM Tris pH 7.4, 150 mM NaCl, 1% Triton X-100, 10 µM leupeptin, 1

mM AEBSF, 100 U/ml aprotinin, 10 mM NaF and 1 mM Na3VO4. Samples (30 µg)

were resolved by 10% SDS-PAGE, transferred to PVDF membrane and then

exposed to the appropriate antibodies. Proteins were visualized with the ECL system

from Amersham using horseradish peroxidase-conjugated anti-rabbit or anti-mouse

secondary antibody. Western blot assays were representative of at least 3

experiments.

Expression vectors, transfection and luciferase assays.

The luciferase reporter plasmids pTyro and pMITF and the expression vector

encoding MITF were previously described (18,33). The pCDNA3 vector encoding

GSK3β was kindly provided by Dr. T.C Dale (London, UK).

B16 melanoma cells were seeded in 24 well dishes and transient transfections were

performed the following day using 2 µl of lipofectamine and 0.5 µg of total DNA

plasmid. pCMVβGal was transfected with the test plasmids to control the variability in

transfection efficiency. After 48 h, cells were harvested in 50 µl of lysis buffer and

assayed for luciferase and β-galactosidase activities. All transfections were repeated

at least three times.

Immunoprecipitation and in vitro AKT assay.

B16 cells were seeded in 6-well dishes and transient transfections were performed

the following day using 10 µl of lipofectamine and 1.5 µg of HA-tagged AKT or 1 µg

of HA-tagged AKT with 2 µg of empty pCDNA3 or vector encoding RasV12. HA-

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

8

tagged AKT in the mammalian expression vector pCDNA3 was a kind gift from Dr. G.

Baier (Innsbruck, Austria) and the vector encoding RasV12 was provided by Dr. A.

Eychene (Orsay, France). Cells were treated for 1 h with 50 µM forskolin or 15 µM

LY294002 and, then were lysed in buffer A and immunoprecipitated with the anti-HA

antibody coupled to protein G-sepharose, for 3 h at 4 °C. Immunocomplexes were

washed and AKT activity was assayed with Crosstide as a substrate in a reaction

mixture containing 50 mM Tris, 10 mM MgCl2, 1 mM dithiothreitol, 5 µM ATP, 30 µM

Crosstide and 3.3 µCi of [γ-32PATP], for 30 min at 30 °C. The phosphorylation

reaction was stopped by spotting onto Whatman p81 papers and immersing them in

1% orthophosphoric acid. The papers were washed, rinsed in ethanol, air dried and

the radioactivity was determined by Cerenkov counting.

Immunoprecipitation and in vitro GSK3 assay.

B16 cells were seeded in 6-well dishes, exposed to 50 µM forskolin or 15 µM

LY294002 for the time indicated in the figure legend, lysed as previously described

for the AKT assay and, GSK3β was immunoprecipitated with an anti-GSK3β antibody

for 3 h at 4°C. The activity of GSK3β was assayed with GS peptide-2 as a substrate

in buffer containing 20 mM Tris, 10 mM MgCl2, 5 mM DTT, 20 µM glycogen synthase

peptide-2 as substrate, 10 µM ATP and 3.3 µCi of [γ-32PATP] for 10 min at 30 °C. The

following steps were as described for AKT assay.

Nuclear extracts and gel mobility shift assay.

Nuclear extracts from control cells or cells incubated with forskolin for 1 h were

prepared as previously described (34). Double-stranded synthetic M-box, 5'-

GAAAAAGTCATGTGCTTTGCAGAAGA-3' was γ32P end-labeled using T4

polynucleotide kinase. 5 µg of nuclear proteins were preincubated in a binding buffer

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

9

containing 10 mM Tris pH 7.5, 100 mM NaCl, 1 mM DTT, 1 mM EDTA, 4 % glycerol,

80 µg/ml of salmon sperm DNA, 0.1 µg poly (dIdC), 10% FCS, 2 mM MgCl2 and 2

mM spermidine for 15 min on ice. Then, 30,000-50,000 cpm of 32

P labeled probe

were added to the binding reaction for 10 min at room temperature. DNA-protein

complexes were resolved by electrophoresis on a 4 % polyacrylamide gel (37.5:1

Acrylamide-Bisacrylamide) in TBE buffer (22.5 mM Tris-borate, 0.5 mM EDTA, pH 8)

for 1 h 30 at 100 V. For supershift assays, 0.3 µl of pre-immune serum or anti-MITF

antibody (33) was preincubated with nuclear extracts in the binding reaction buffer

before adding the labeled probe.

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

10

Results

In B16 melanoma cells, cAMP leads to an inhibition of AKT phosphorylation

and activity.

First, we investigated, in B16 melanoma cells, the effects of two cAMP-elevating

agents, αMSH and forskolin on AKT. Immunoblotting of cell lysates, with phospho-

specific antibodies to AKT, revealed that αMSH, a physiologic melanocyte

differentiating agent, as well as forskolin, induced a strong inhibition of AKT

phosphorylation on both threonine 308 and serine 473 compared to control cells (Fig.

1A). As expected, LY294002, a specific pharmacological inhibitor of PI3K, also

abolished AKT phosphorylation on both residues. Increase in cAMP content led to a

complete dephosphorylation of AKT after 30 min (Fig. 1B). The effect of LY294002

was even more rapid since no phosphorylation persisted after 15 min of treatment.

Further, B16 cells, transfected with a vector encoding a HA-tagged AKT were left

untreated, exposed to forskolin or to LY294002 for 30 min. Then, HA-AKT was

immunoprecipitated to perform an in vitro kinase assay. As shown in Fig. 1C,

forskolin and LY294002 reduced the activity of AKT to about 50 and 70%

respectively indicating that the decrease in AKT phosphorylation correlated with an

inhibition of its activity. In each experiment, detection of AKT regardless of its

phosphorylation state ensured even loading of each lane (Fig. 1).

To investigate whether cAMP mediates its effect through a phosphatidylinositol-3-

kinase (PI3K)-dependent mechanism, B16 cells were infected with an adenovirus

encoding a constitutively active (p110CAAX) or a kinase-dead (p110KD) form of the

p110 sub-unit of PI3K. B16 cells, were then exposed or not to forskolin and

phosphorylation of AKT was examined. p110CAAX stimulated the phosphorylation of

AKT over the basal and markedly reduced the effect of cAMP compared to cells

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

11

infected with a control adenovirus (Fig.2A). The kinase dead-mutant, had no effect on

AKT regulation by cAMP. Therefore, in B16 melanoma cells, cAMP leads to an

inhibition of AKT phosphorylation and activity that appears to be mediated through a

PI3K-dependent mechanism.

To better define the site of cAMP action, experiments were designed to evaluate

whether cAMP acted upstream of p21Ras, that is described to activate PI3K. B16

cells were co-transfected with vectors encoding a HA-tagged AKT and a

constitutively active mutant of p21Ras, RasV12. Cells were treated or not with

forskolin and HA-AKT was immunoprecipitated to perform an in vitro kinase assay.

Consistent with fig. 1C, forskolin reduced the activity of AKT to about 50% in

presence of a pCDNA3 empty plasmid (Fig. 2B). Further, co-expression of HA-AKT

with RasV12 increased the basal AKT activity that was also inhibited by forskolin to

about 50%. These results indicate that cAMP acts independently of Ras to inhibit

AKT.

In B16 cells, cAMP decreases GSK3 phosphorylation and promotes its

activation.

Next, we wanted to determine if AKT inactivation by cAMP had an impact on

downstream events in B16 melanoma cells. In this aim, we focused our attention on

the glycogen synthase kinase 3β (GSK3β), a cellular substrate of AKT.

Immunoblotting experiments of cells exposed to forskolin or to LY294002, with

phospho-specific antibody revealed a decrease of GSK3β phosphorylation on serine

9 compared to control cells (Fig. 3A). Next, we investigated the effect of this

dephosphorylation on GSK3β activity. After treatment of the cells with forskolin or

LY294002, GSK3β has been immunoprecipitated and subjected to an in vitro kinase

assay. Results presented in Fig. 3B showed that forskolin and LY294002 stimulated

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

12

GSK3β activity. Immunodetection of total GSK3β indicated that these differences

were not due to variations in the level of protein expression. To confirm the activation

of GSK3β, the effect of forskolin and LY294002 were then analyzed, in intact cells,

toward the phosphorylation of endogenous substrates of GSK3β such as the

glycogen synthase and the microtubule-associated protein, Tau. Using phospho-

specific antibodies, we observed that forskolin and LY294002 increased

phosphorylation of both glycogen synthase and Tau protein compared to cells in

basal condition (Fig. 3C). The blot was also incubated with an anti-ERK2 antibody to

control for the equal loading of the gel. In conclusion, these results clearly

demonstrate that, in B16 melanoma cells, GSK3β is activated in response to cAMP.

The effect of cAMP on AKT and GSK3 is not mediated by PKA and Ras/ERK

pathways.

Although PKA is the major cellular target of cAMP, this nucleotide appears to have

some PKA-independent actions (35-37). Thus, to define whether PKA was involved

in the regulation of AKT and GSK3β following cAMP treatment, we used H89, a cell-

permeable and selective inhibitor of PKA. Treatment of melanoma cells with H89 had

no effect on forskolin-induced inhibition of AKT phosphorylation on threonine 308 and

serine 473 (Fig. 4A). Immunoblotting also indicated that H89 did not prevent the

inhibition of GSK3β phosphorylation on serine 9 induced by forskolin (Fig. 4A). The

functionality of H89 was assessed by its ability to block forskolin-induced

phosphorylation of CREB at serine 133. Detection of ERK2 ensured that each lane

was even loaded. These results clearly indicate that, in B16 melanoma cells, PKA is

not involved in the regulation of AKT and GSK3β by cAMP.

Interestingly, we have recently reported that cAMP stimulates the Ras/ERK pathway

via a PKA independent mechanism (35). Therefore, we wished to explore the

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

13

possibility that the Ras/ERK pathway was involved in the effect of cAMP on AKT and

GSK3β. To this purpose, we took advantage of C. sordelii lethal toxin (LT), which is

an inhibitor of p21Ras, and thus allows the inhibition of the downstream cascade.

Immunoblotting with phospho-specific antibodies demonstrated that LT failed to

abolish the inhibition of AKT and GSK3β phosphorylation induced by forskolin

exposure (Fig. 4B). The functionality of LT was demonstrated by its ability to block

forskolin-induced phosphorylation and activation of ERK1/ERK2. Together, in B16

melanoma cells, cAMP regulates the AKT/GSK3β cascade via a mechanism that

involves neither PKA nor Ras/ERK signaling pathways.

Involvement of GSK3 in the effect of cAMP on the tyrosinase promoter.

As MITF, through the regulation of tyrosinase expression is a key actor in cAMP-

induced melanogenesis, we explored the possibility that GSK3β regulates MITF

expression through a stimulation of the transcriptional activity of its promoter. As

shown in Fig. 5A, GSK3β did not alter basal or cAMP-induced MITF promoter

activity demonstrating that GSK3β did not affect MITF transcription. We also

investigated the possible involvement of GSK3β in the stimulation of MITF activity.

Using a mammalian one hybrid system, we found that GSK3β did not up-regulate the

intrinsic transcriptional activity of MITF (Fig. 5B). Finally, we wished to evaluate the

potential role of GSK3β in the regulation of tyrosinase expression, the enzyme that

controls melanin synthesis. To do so, B16 cells were transfected with the tyrosinase

promoter and then exposed to forskolin in presence or in absence of lithium, an

inhibitor of GSK3β. The results presented on Fig. 6A showed that lithium reduced by

50% the response to cAMP of the tyrosinase promoter. Additionally, we showed that

GSK3β by itself did not significantly affect the activity of the tyrosinase promoter but

synergized with MITF to stimulate the tyrosinase promoter (Fig. 6B). Together, these

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

14

data suggest that GSK3β, through the control of MITF function, is involved in the

regulation of tyrosinase transcription by cAMP. Finally, gel shift assays were

performed with nuclear extracts from control cells or cells that have been exposed to

cAMP for 1 h. We observed an increased binding of nuclear proteins from cAMP-

treated cells to the M-box probe compared to control cells (Fig. 6C, lower panel).

The complexes were shifted by specific MITF antibody, indicating that cAMP

stimulated the formation of MITF/M-box complexes. Since, cAMP did not stimulate

MITF expression at 1 h (Fig. 6C, upper panel), we concluded that cAMP increased

the ability of MITF to bind its target sequence. Together, these results suggest that

GSK3β plays a key role in cAMP-induced melanogenesis by enhancing the binding of

MITF to the tyrosinase promoter.

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

15

Discussion

In the course of investigating the molecular mechanisms involved in the regulation of

melanocyte differentiation, we have previously observed, in B16 cells, an inhibition of

the phosphatidylinositol-3 kinase (PI3K) during cAMP-induced melanogenesis (27).

Since the inhibition of PI3K stimulates melanogenesis, we wished to dissect, in the

present report, the molecular events that connected the PI3K pathway to melanin

synthesis. We first focused our attention on AKT, a well-characterized kinase that

functions downstream of PI3K. Depending on the cell type, cAMP has been reported

to inhibit or stimulate AKT (38,39). Hence, we verified the effect of cAMP on AKT in

B16 melanoma cells. Our results showed that cAMP potently inhibited the

phosphorylation of AKT at threonine 308 and serine 473 and led to an inhibition of

AKT activity. Additionally, cAMP effect on AKT was abolished in presence of a

constitutively active mutant of PI3K, demonstrating that cAMP promotes an inhibition

of AKT through a PI3K-dependent mechanism. However, over-expression of a

constitutively active mutant of p21Ras did not block the inhibition of AKT evoked by

cAMP, indicating that the inhibition of AKT by cAMP is mediated through a p21Ras

independent pathway. It should be noted that we have previously reported an

activation of p21Ras by cAMP in melanocytes and melanoma cells (35). Thus, in the

conditions in which we observed an inhibition of AKT, p21Ras is activated

demonstrating that the activation of p21Ras cannot prevent the inhibition of AKT.

AKT phosphorylates GSK3β at serine 9, leading to GSK3β inactivation. Therefore,

the inhibition of AKT by cAMP should lead to the activation of GSK3β. However, it

has been also reported that PKA directly phosphorylates GSK3β at serine 9 and

inhibits its activity (39,40). In B16 cells, cAMP decreased the phosphorylation of

GSK3β at serine 9, and led to the stimulation of its kinase activity. Consistent with

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

16

the effect of AKT on GSK3β and the effect of cAMP on AKT, we demonstrate for the

first time that cAMP promotes an activation of GSK3β.

As PKA is the major intracellular target of cAMP, it was tempting to propose that

effect of cAMP on AKT and GSK3β involved the PKA. However, a pharmacological

inhibitor of PKA did not impair the effect of cAMP on AKT and GSK3β. As mentioned

above, in B16 cells, cAMP also stimulates the Ras/ERK pathway via a PKA

independent mechanism (35). Thus, we hypothesized that the activation of the

Ras/ERK pathway could mediate the effect of cAMP on AKT and GSK3β. However,

inhibition of p21Ras, and consequentely of ERK, by C. sordelii lethal toxin, did not

alter the effect of cAMP on AKT and GSK3β. Hence, cAMP leads to an inhibition of

AKT and a stimulation of GSK3β through PKA and p21Ras/ERK independent

pathways. Recently, it has been also reported that cAMP inhibits AKT through a

mechanism that does not involved PKA (41).

GSK3β has been widely implicated in cell homeostasis by its ability to phosphorylate

a broad range of substrates including the glycogen synthase, the microtubule-

associated protein Tau and β-catenin (42). Interestingly, cAMP-induced activation of

GSK3β, stimulated phosphorylation of the microtubule-associated protein, Tau (Fig.

2C). Considering the role of Tau in the organization of microtubule and actin

cytosqueleton (43,44), as well as in the regulation of organelles traffic (45), our

results point out to a potential role of GSK3β and Tau in the regulation of cAMP-

induced dendritogenesis and melanosome transport, two key parameters of

melanocyte differentiation.

Further, phosphorylation of β-catenin by GSK3β promotes its degradation and

prevents its translocation to the nucleus, where β-catenin acts in concert with the

lymphoid enhancer factor-1/T-cell factor (LEF-1/TCF) family of transcription factors to

transactivate target genes. The promoter of MITF contains a LEF-1 binding site and

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

17

β-catenin stimulates the transcriptional activity of the MITF promoter (46). Thus, the

activation of GSK3β following cAMP exposure should lead to an inhibition of the

MITF promoter activity due to β-catenin degradation. However, in B16 cells, elevation

of the cAMP content did not induce the phosphorylation of β-catenin and did not

affect the activity of a TOPFlash reporter plasmid bearing LEF-1 binding sites (data

not shown). Cotransfection of MITF promoter reporter construct with GSK3β confirms

that GSK3β did not regulate the transcriptional activity of the MITF promoter (Fig.

5A).

Recently, Takeda et al., have demonstrated that GSK3β phosphorylates MITF on

serine 298 thereby enhancing its binding to the tyrosinase promoter (47).

Interestingly, mutation of the MITF serine 298 has been identified in patients with

Waardenburg syndrome type II which are characterized by pigmentary disorders,

emphasing the physiopathologic importance to understand the role of GSK3β in the

control of pigmentation. Therefore, since GSK3β stimulates neither the MITF

promoter activity nor the intrinsic transcriptional activity of MITF, we hypothesized

that cAMP, through GSK3β activation, could increase the ability of MITF to bind its

target sequence. Consistent with the work from Takeda et al., we demonstrated that

GSK3β stimulated the action of MITF on the tyrosinase promoter. Additionally,

lithium, a GSK3β inhibitor, decreased cAMP-induced stimulation of the tyrosinase

promoter. Finally, we demonstrated that cAMP increases the binding of MITF to the

M-box sequence. To demonstrate the key role of the serine 298 phosphorylation, we

studied the effect of a MITF mutant, MITF S298A. This mutant cannot be

phosphorylated by GSK3β and thus the effect of this mutant should not be increased

by GSK3β or by cAMP. However, in agreement with the previous observation of

Takeda et al., this mutant is devoid of any transcriptional activity (data not shown).

Further, we constructed two other mutants in which the serine 298 was replaced

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

18

either by an aspartate or glutamate. The positive charge of these amino acids was

expected to mimic the charge brought by the phosphate group. Unexpectedly, these

two mutants were also unable to transactivate the tyrosinase promoter (data not

shown).

Together, these results show that, in parallel to the PKA/CREB pathway which

stimulates MITF expression, cAMP, through the activation of GSK3β, increases the

ability of MITF to bind to the tyrosinase promoter. These two pathways cooperate,

allowing cAMP elevating agents to efficiently stimulate tyrosinase transcription and

thereby melanogenesis. In conclusion, elevation of intracellular cAMP leads to the

activation of a complex network of signaling pathways that converge to MITF to

control melanin synthesis and melanocyte differentiation (Fig. 7). Our findings,

demonstrating that GSK3β plays an active role in cAMP-induced melanogenesis and

in melanocyte differentiation, bring information of paramount importance on the

molecular mechanisms that control melanin synthesis and skin pigmentation.

Acknowledgments

We are grateful to S. Tartare-Deckert and G. Ponzio for critical reading of the

manuscript. This work was supported by INSERM, The Ligue Nationale contre le

Cancer and the Association pour la Recherche sur le Cancer grant 5808.

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

19

References

1. Valverde, P., Healy, E., Jackson, I., Rees, J. L., and Thody, A. J. (1995) Nat Genet 11, 328-

330

2. Krude, H., Biebermann, H., Luck, W., Horn, R., Brabant, G., and Gruters, A. (1998) Nat Genet

19, 155-157

3. Pears, J. S., Jung, R. T., Bartlett, W., Browning, M. C., Kenicer, K., and Thody, A. J. (1992) Br

J Dermatol 126, 286-289.

4. Lamerson, C. L., and Nordlund, J. J. (1998) Nordlund, JJ, Boissy, RE, Hearing, VJ, King, RA,

Ortonne, JP, Oxford University Press, Oxford 120, 1695-1708

5. Sowers, J. R., and Lippman, H. R. (1985) Cutis 36, 351-352, 354

6. Levine, N., Sheftel, S. N., Eytan, T., Dorr, R. T., Hadley, M. E., Weinrach, J. C., Ertl, G. A.,

Toth, K., McGee, D. L., and Hruby, V. J. (1991) Jama 266, 2730-2736

7. Lerner, A. B., and S., M. J. (1961) Nature 189, 176-179

8. Lalli, E., and Sassone-Corsi, P. (1994) J Biol Chem 269, 17359-17362

9. Schwindinger, W. F., Francomano, C. A., and Levine, M. A. (1992) Proc Natl Acad Sci U S A

89, 5152-5156

10. Weinstein, L. S., Shenker, A., Gejman, P. V., Merino, M. J., Friedman, E., and Spiegel, A. M.

(1991) N Engl J Med 325, 1688-1695

11. Casey, M., Vaughan, C. J., He, J., Hatcher, C. J., Winter, J. M., Weremowicz, S.,

Montgomery, K., Kucherlapati, R., Morton, C. C., and Basson, C. T. (2000) J Clin Invest 106,

R31-38

12. Kirschner, L. S., Carney, J. A., Pack, S. D., Taymans, S. E., Giatzakis, C., Cho, Y. S., Cho-

Chung, Y. S., and Stratakis, C. A. (2000) Nat Genet 26, 89-92.

13. Englaro, W., Rezzonico, R., Durand-Clement, M., Lallemand, D., Ortonne, J. P., and Ballotti,

R. (1995) J Biol Chem 270, 24315-24320.

14. Hunt, G., Todd, C., Cresswell, J. E., and Thody, A. J. (1994) J Cell Sci 107 ( Pt 1), 205-211

15. Hearing, V. J., Jr. (1987) Methods Enzymol 142, 154-165

16. Korner, A., and Pawelek, J. (1982) Science 217, 1163-1165

17. Prota, G. (1988) Prog Clin Biol Res 256, 101-124

18. Bertolotto, C., Abbe, P., Hemesath, T. J., Bille, K., Fisher, D. E., Ortonne, J. P., and Ballotti, R.

(1998) J Cell Biol 142, 827-835.

19. Hodgkinson, C. A., Moore, K. J., Nakayama, A., Steingrimsson, E., Copeland, N. G., Jenkins,

N. A., and Arnheiter, H. (1993) Cell 74, 395-404.

20. Steingrimsson, E., Moore, K. J., Lamoreux, M. L., Ferre-D'Amare, A. R., Burley, S. K., Zimring,

D. C., Skow, L. C., Hodgkinson, C. A., Arnheiter, H., Nakayama, N. G., Copeland, N. G., and

Jenkins, N. A. (1994) Nat Genet 8, 256-263

21. Goding, C. R., and Fisher, D. E. (1997) Cell Growth Differ 8, 935-940

22. Sato, S., Roberts, K., Gambino, G., Cook, A., Kouzarides, T., and Goding, C. R. (1997)

Oncogene 14, 3083-3092

23. Yasumoto, K., Yokoyama, K., Takahashi, K., Tomita, Y., and Shibahara, S. (1997) J Biol

Chem 272, 503-509

24. Englaro, W., Bertolotto, C., Busca, R., Brunet, A., Pages, G., Ortonne, J. P., and Ballotti, R.

(1998) J Biol Chem 273, 9966-9970.

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

20

25. Hemesath, T. J., Price, E. R., Takemoto, C., Badalian, T., and Fisher, D. E. (1998) Nature

391, 298-301.

26. Wu, M., Hemesath, T. J., Takemoto, C. M., Horstmann, M. A., Wells, A. G., Price, E. R.,

Fisher, D. Z., and Fisher, D. E. (2000) Genes Dev 14, 301-312.

27. Busca, R., Bertolotto, C., Ortonne, J. P., and Ballotti, R. (1996) J Biol Chem 271, 31824-

31830.

28. Bellacosa, A., Testa, J. R., Staal, S. P., and Tsichlis, P. N. (1991) Science 254, 274-277

29. Jones, P. F., Jakubowicz, T., Pitossi, F. J., Maurer, F., and Hemmings, B. A. (1991) Proc Natl

Acad Sci U S A 88, 4171-4175

30. Cross, D. A., Alessi, D. R., Cohen, P., Andjelkovich, M., and Hemmings, B. A. (1995) Nature

378, 785-789.

31. Kim, L., and Kimmel, A. R. (2000) Curr Opin Genet Dev 10, 508-514

32. Plyte, S. E., Hughes, K., Nikolakaki, E., Pulverer, B. J., and Woodgett, J. R. (1992) Biochim

Biophys Acta 1114, 147-162

33. Bertolotto, C., Bille, K., Ortonne, J. P., and Ballotti, R. (1996) J Cell Biol 134, 747-755.

34. Bertolotto, C., Bille, K., Ortonne, J. P., and Ballotti, R. (1998) Oncogene 16, 1665-1670.

35. Busca, R., Abbe, P., Mantoux, F., Aberdam, E., Peyssonnaux, C., Eychene, A., Ortonne, J. P.,

and Ballotti, R. (2000) Embo J 19, 2900-2910.

36. Gray, P. C., Scott, J. D., and Catterall, W. A. (1998) Curr Opin Neurobiol 8, 330-334

37. Matthews, G. (1991) Trends Pharmacol Sci 12, 245-247

38. Filippa, N., Sable, C. L., Filloux, C., Hemmings, B., and Van Obberghen, E. (1999) Mol Cell

Biol 19, 4989-5000.

39. Li, M., Wang, X., Meintzer, M. K., Laessig, T., Birnbaum, M. J., and Heidenreich, K. A. (2000)

Mol Cell Biol 20, 9356-9363.

40. Fang, X., Yu, S. X., Lu, Y., Bast, R. C., Jr., Woodgett, J. R., and Mills, G. B. (2000) Proc Natl

Acad Sci U S A 97, 11960-11965.

41. Wang, L., Liu, F., and Adamo, M. L. (2001) J Biol Chem 30, 30

42. Frame, S., and Cohen, P. (2001) Biochem J 359, 1-16

43. Garcia, M. L., and Cleveland, D. W. (2001) Curr Opin Cell Biol 13, 41-48

44. Gundersen, G. G., and Cook, T. A. (1999) Curr Opin Cell Biol 11, 81-94

45. Stamer, K., Vogel, R., Thies, E., Mandelkow, E., and Mandelkow, E. M. (2002) J Cell Biol 156,

1051-1063

46. Takeda, K., Yasumoto, K., Takada, R., Takada, S., Watanabe, K., Udono, T., Saito, H.,

Takahashi, K., and Shibahara, S. (2000) J Biol Chem 275, 14013-14016

47. Takeda, K., Takemoto, C., Kobayashi, I., Watanabe, A., Nobukuni, Y., Fisher, D. E., and

Tachibana, M. (2000) Hum Mol Genet 9, 125-132

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

21

Figure legends

FIG.1: cAMP inhibits the activity of AKT.

(A) B16 melanoma cells were left untreated (C) or were incubated with αMSH 1 µM

(αM), forskolin 50 µM (Fsk) or LY294002 15 µM (LY) for 30 min. Cell lysates was

analyzed by western blot with phospho-specific antibodies to AKT (PT308-AKT and

PS473-AKT). (B) Extracts from control B16 cells or cells exposed to forskolin 50 µM

or LY294002 15 µM for the indicated time were submitted to western blot

experiments as described in (A). (C) B16 cells were transiently transfected with an

HA-tagged AKT and then incubated or not with forskolin 50 µM or LY294002 15 µM

for 1 h. Then, AKT was immunoprecipitated and its kinase activity was evaluated

against Crosstide. Data are expressed as percentage of AKT activity in control cell

and are means ± S.E of 3 experiments performed in triplicate. (A-C) Lower panels

show western blot probed with antibody to AKT to control protein loading. Molecular

masses, indicated on the left of western blots, are expressed in Kilodaltons.

FIG.2: cAMP inhibits AKT through a PI3K-dependent, p21Ras-independent

pathway.

(A) B16 cells were infected with an adenovirus encoding either a constitutively active

(p110CAAX) or a kinase-dead (p110KD) form of the p110 subunit of PI3K. Then,

cells were exposed for 1 h to forskolin 50 µM and AKT phosphorylation was analyzed

as described in A. (B) B16 cells were transfected with an HA-tagged AKT in presence

of an empty pCDNA3 plasmid or a vector encoding RasV12 and then were incubated

or not with forskolin 50 µM for 1 h. The activity of AKT was next tested in vitro against

crosstide. Data are expressed as percentage of AKT activity in control cell and are

means ± S.E of 3 experiments performed in triplicate. Detection of total AKT ensured

that each lane was even loading.

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

22

FIG. 3: cAMP activates GSK3 .

(A) B16 melanoma cells were treated as described in Fig. 1B and immunoblotting of

cell lysates was done with phospho-specific GSK3β antibody (PS9-GSK3β).

Detection of total GSK3β was used to ensure that each lane was even loaded. (B)

Endogenous GSK3β, from control B16 cells or cells exposed to forskolin 50 µM or

LY294002 15 µM for the indicated time, was immunoprecipitated and its activity was

measured with the GS peptide-2 as substrate. The results are expressed as

percentage of GSK3β activity in control cells and are means ± S.E of 3 experiments

performed in triplicate. The level of GSK3β in immunoprecipitate is shown below the

graph. (C) Cells were incubated for 1 h with forskolin 50 µM or LY294002 15 µM and

then submitted to immunoblotting using a phospho specific glycogen synthase (pGS)

antibody or an antibody that recognizes Tau phosphorylated on S396 (P-TAU).

Detection of ERK2 ensured even loading of each lane. Molecular masses, indicated

on the left of western blots, are expressed in Kilodaltons.

FIG. 4: PKA and Ras/ERK are not involved in the regulation of AKT and GSK3

by cAMP.

B16 cells were left untreated or (A) were incubated with H89 dihydrochloride 5 µM for

30 min or (B) with C. sordelii lethal toxin (LT) 70 ng/ml for 4 h, before adding forskolin

50 µM for 1 h. Western blots were next performed with phospho specific antibodies

described in figure 1 and 3. P-CREB antibody recognizes specifically CREB when

phosphorylated on serine 133. P-ERK antibody recognizes ERK1 and ERK2 when

phosphorylated and activated. Detection of ERK2 showed that each lane was loaded

with equal amounts of protein. Molecular masses, indicated on the left of western

blots, are expressed in Kilodaltons.

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

23

FIG. 5: GSK3 regulates neither MITF transcription nor the intrinsic MITF

transcriptional activity.

(A) B16 cells were transfected with 0.3 µg of the luciferase reporter plasmid pMITF

plus 0.1 µg of empty pCDNA3 or 0.1 µg of pCDNA3 encoding GSK3β. Cells were left

untreated or were exposed to forskolin 50 µM for 48 h. (B) B16 cells were transfected

with 0.3 µg of a Gal4 luciferase reporter plasmid and 0.05 µg of plasmid encoding the

Gal4 DNA binding domain (Gal4) or the Gal4 DNA binding domain fused to MITF

(Gal4-MITF). When indicated, 0.1 µg of GSK3β expression plasmid was added to

the reaction. The total DNA was maintained constant at 0.5 µg per condition by

addition of empty pCDNA3 vector.

Luciferase activity was measured and normalized to β-galactosidase activity. Results

are expressed as fold stimulation of the basal luciferase activity from each

experiment. Data are means ± standard error of three experiments performed in

triplicate.

FIG. 6: GSK3 mediates the effect of cAMP on the tyrosinase promoter.

(A) B16 cells were transfected with 0.3 µg of pTyro. After transfection, cells were

incubated with forskolin 50 µM and/or lithium chloride (LiCl) 15 mM for 48 h.

(B) B16 cells were transfected with 0.3 µg of the luciferase reporter plasmid pTyro

and 0.05 µg of pCDNA3 encoding MITF and/or 0.1 µg of pCDNA3 encoding GSK3β.

The total DNA was maintained constant at 0.5 µg per point by addition of empty

pCDNA3 vector. Luciferase activity was measured and normalized to β-galactosidase

activity. Results are expressed as fold stimulation of the basal luciferase activity from

each experiment. Data are means ± standard error of three experiments performed in

triplicate. (C) Gel shift assay was performed using the 32P labeled M-box and B16

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

24

nuclear extracts from control cells (C) or cells treated for 1 h with forskolin (Fsk).

Binding reactions were performed in presence of 0.3 µl of pre-immune serum (PI) or

with 0.3 µl of a specific anti-MITF serum (I). Western blot experiment in the lower

panel shows the detection of total MITF in each condition.

FIG. 7: Model of signaling pathways involved in cAMP-induced melanogenesis.

Elevation of intracellular cAMP content leads to PKA activation and stimulation of

MITF transcription resulting in stimulation of tyrosinase expression. cAMP,

independently of PKA, also activates the Ras, B-Raf, MEKK, ERK and Rsk-1

cascade. Phosphorylation of MITF on serine 73 and serine 409 by ERK and Rsk-1

respectively promotes its degradation constituing a retrocontrol mechanism that

prevents an excessive production of melanin synthesis. Finally, cAMP, via a PKA-

independent mechanism, inhibits PI3K and AKT and promotes an activation of

GSK3β. GSK3β, by phosphorylating MITF on serine 298, increases its binding to the

M-box of the tyrosinase promoter leading to stimulation of tyrosinase expression.

Taken into account the role of Tau in cytosqueleton organization, we hypothesized a

role of Tau in dendritogenesis.

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

C 15 30 60Fsk

Time, min

63

63

63

C 15 30 60LY

C M Fsk LY

63

63

AKT

PS473-AKT

63

PT308-AKT

B

AKT

C LYFsk

0

25

50

75

100

63

A

C

Khaled et al. Figure 1

Emptyadenovirus

Fsk - - - +++

p110CAAX

p110KD

63

63

AKT

PS473-AKT

63

PT308-AKT

A

Khaled et al. Figure 2

0

25

50

75

100

125

150

C Fsk C Fsk

pCDNA3 RasV12

AKT63

AK

T a

ctiv

ity,

(%

of

con

tro

l)

B

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

A

0

100

200

300

400

C Fsk30 LY60Fsk60

GS

K3

act

ivit

y, (

% o

f co

ntr

ol)

B

GSK346

p-TAU60

p-GS80

C Fsk LY

C

ERK236

46

C 15 30 60LY

Time, min

46

PS9-GSK3

GSK3

46

46

C 15 30 60

Fsk

Time, min

Khaled et al. Figure 3

A

C Fsk H89 Fsk+H89

36

49

46

63

63

Khaled et al. Figure 4

B

C LT Fsk Fsk+LT

63 PT308-AKT

PS473-AKT63

P-ERK36

36 ERK2

46 PS9-GSK3

PT308-AKT

PS473-AKT

P-CREB

ERK2

PS9-GSK3

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

A B

pMITF Gal4 Luc

Gal4-MITF

Gal40

10

40

50

Rel

ativ

e L

uci

fera

se A

ctiv

ity,

Fo

ld s

tim

ula

tio

n o

ver

bas

al

20

0

1

2

4

5

Rel

ativ

e L

uci

fera

se A

ctiv

ity,

Fo

ld s

tim

ula

tio

n o

ver

bas

al

3

Fsk - + - +

30

Khaled et al. Figure 5

BA

pTyro

0

10

20

30

40

50

60

70

80

pCD

NA 3

MIT

F

GSK

3

MIT

F+G

SK3

pTyro

Rel

ativ

e L

uci

fera

se A

ctiv

ity,

Fo

ld s

tim

ula

tio

n o

ver

bas

al

Cont Fsk LiCl Fsk+LiCl

0

2

4

6

8

10

12

14

C Fsk

PI I PI I

MITF

MITF

C

MITF

Khaled et al. Figure 6

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

cAMP

MC1R

MSH

PKA

MITFCREB

MITF Promoter

+

MITF

Tyrosinase

S73 S409P

Degradation

Ras

B-Raf

ERK

P

Rsk-1

MEKK

PI3K

AKT

GSK3

PS298

Tau

Dendrite outgrowth and melanosome transport

?

Khaled et al. Figure 7

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Robert Ballotti and Corine BertolottoMehdi Khaled, Lionel Larribere, Karine Bille, Edith Aberdam, Jean Paul Ortonne,

regulation of melanogenesisGlycogen synthase kinase 3b is activated by cAMP and plays an active role in the

published online July 1, 2002J. Biol. Chem. 

  10.1074/jbc.M202939200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

by guest on August 20, 2018

http://ww

w.jbc.org/

Dow

nloaded from