Báo cáo khoa học: Salt-inducible kinase-1 represses cAMP response element-binding protein activity both in the nucleus and in the cytoplasm pdf

Salt-inducible kinase-1 represses cAMP response element-binding
protein activity both in the nucleus and in the cytoplasm
Yoshiko Katoh
1,
*, Hiroshi Takemori
1
, Li Min
1
, Masaaki Muraoka
1,3
, Junko Doi
4
, Nanao Horike
1
and Mitsuhiro Okamoto
1,2
1
Department of Biochemistry and Molecular Biology, Graduate School of Medicine (H-1) and
2
Laboratories for Biomolecular
Networks, Graduate School of Frontier Biosciences, Osaka University, Japan;
3
ProteinExpress Co., Ltd, Choshi, Chiba, Japan;
4
Department of Food and Nutrition, Senri Kinran University, Osaka, Japan
Salt-inducible kinase-1 (SIK1) is phosphorylated at Ser577
by protein kinase A in adrenocorticotropic hormone-sti-
mulated Y1 cells, and the phospho-SIK1 translocates
from the nucleus to the cytoplasm. The phospho-SIK1 is
dephosphorylated in the cytoplasm and re-enters the nucleus
several hours later. B y using green-fluorescent protein-tag-

ged SIK1 fragments, we found that a peptide region (586–
612) was responsible for the nuclear localization of SIK1.
TheregionwasnamedtheÔRK-rich regionÕ because o f i ts
Arg- and Lys-rich nature. SIK1s mutated in the RK-rich
region were localized mainly in the cytoplasm. Because SIK1
represses c AMP-response e lement (CRE)-mediated tran-
scription of steroidogenic genes, the mutants were examined
for their effect on transcription. To our surprise, the cyto-
plasmic mutants strongly repressed the CRE-binding pro-
tein (CREB) activity, t he extent of repression being similar to
that of SIK1(S577A), a mutant localized exclusively in the
nucleus. Several chimeras were constructed from SIK1 and
from its isoform SIK2, which was localized mainly in the
cytoplasm, and they were examined for i ntracellular l ocal-
ization as well as CREB-repression activity. A SIK1-derived
chimera, where the RK-rich region had been replaced
with the corresponding region of SIK2, was found in the
cytoplasm, its CREB-modulating activity being similar to
that of wild-type SIK1. On the other hand, a SIK2-derived
chimera w ith the RK-rich region of SIK1 was localized in
both the nucleus and the cytoplasm, and had a CREB-
repressing activity similar to that of the wild-type SIK2.
Green fluorescent protein-fused transducer of regulated
CREB activity 2 (TORC2), a CREB-specific co-activator,
was localized in the cytoplasm and nucleus of Y1 cells, and,
after treatment with adrenocorticotropic hormone, cyto-
plasmic TORC2 entered the nucleus, activating CREB. The
SIK1 mutants, having a strong CRE-repressing activity,
completely inhibited the adrenocorticotropic hormone-
induced nuclear entry of green fluorescent protein-fused

TORC2. This suggests that SIK1 may regulate the intra-
cellular movement of TORC2, and as a result modulates the
CREB-dependent transcription a ctivity. Together, these
results indicate that the RK-rich region of SIK1 is important
for determining the nuclear localization and attenuating
CREB-repressing activity, but the degree of the nuclear
localization of SIK1 itself does not necessarily reflect the
degree of SIK1-mediated CREB repression.
Keywords: cAMP; CRE; nuclear localization signal; SIK;
transcription repression.
cAMP response element (CRE)-binding protein (CREB), a
transcription factor involved in numerous physiological
processes, regulates gene expression in a phosphorylation-
dependent manner [1–3]. Ser133, the major transcription
activation site of CREB, is phosphorylated by protein
kinase A (PKA) [4,5], p38 mitogen-activated protein k inase
(MAPK) [6], mitogen- and stress-activated protein kinase 1
[7], pp90
rsk
[8], protein kinase B [9] and calcium-calmodulin-
dependent kinase II/IV [10,11], whereas Ser142, the negative
regulation site, is phosphorylated by calcium-calmodulin-
dependent kinase II [12,13]. The phosphorylation of Ser
residues alters the affinity of CREB to CREB-binding
protein and p300, and results in a change of t he transcrip-
tion efficiency [3,14–17]. Several kinases, such as CPG16 [18]
and leucine zipper protein kinase [19], even though they do
not phosphorylate CREB directly, are known to modulate
CRE/CREB activity.
Salt-inducible kinase-1 (SIK1), a Ser/Thr protein k inase

cloned from high-salt diet-fed rat adrenals [20], and also
from PC12 pheochromocytoma cells induced by membrane
Correspondence to M. Okamoto, Department of Biochemistry and
Molecular Biology, Graduate School of Medicine (H-1), Osaka
University 2-2 Yamadaoka, Suita, Osaka, 565-0871 Japan.
Fax: +81 6 6879 3289, Tel.: +81 6 6879 3280,
E-mail:
Abbreviations: ACTH, adrenocorticotropic hormone; b-ZIP, basic
leucine zipper domain; C, cytoplasm; CRE, cAMP-response element;
CREB, CRE-binding protein; DAPI, 4¢,6-diamidino-2-phenylindole;
FITC, fluorescein-5-isothiocyanate; FRAP, FKBP12-rapamycin-
associated protein kinase; GFP, green fluorescent protein; GST,
glutathione S-transferase; HA, haemagglutinin; MAPK, mitogen-
activated protein kinase; MK5, MAPK-activated protein kinase 5;
N, nucleus; NES, nuclear export signal; NLS, nuclear localization
signal; PKA, protein kinase A; SIK, salt-inducible kinase; TORC,
transducer of regulated CREB activity.
*Note: Yoshiko Katoh is a research fellow of the Japan Society for the
Promotion of Science.
(Received 9 June 2004, revised 26 August 2004,
accepted 21 September 2004)
Eur. J. Biochem. 271, 4307–4319 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04372.x
depolarization [21], is a member of the sucrose nonfer-
menting-1 protein kinas e/AMP-activated protein k inase
family [22–26]. SIK1 was found to be exp ressed in adreno-
cortical cells at an early phase of the adrenocorticotropic
hormone (ACTH) stimulation [20,27]. When overexpressed
in Y1 cells, SIK1 repressed the expression of steroidogenic
genes, such as side-chain cleavage cytochrome P450 and
steroidogenic acute regulatory protein [27,28]. Promoter

analyses of these genes indicated t hat the CREs in these
promoters were the sites for SIK1-mediated transcriptional
repression and that SIK1 might repress C REB a ctivity
[28,29]. Although SIK1 seemed not to phosphorylate CREB
directly, it repressed CREB in a kinase activity-dependent
manner. The results of studies with several Gal4–fusion
CREBs suggested that SIK1 repressed CREB by acting on
the basic leucine zipper (b-ZIP) domain of CREB [29].
SIK1 is found mainly in the nucleus of resting Y1 cells. In
ACTH-treated cells, SIK1 is phosphorylated at Ser577 by
PKA, and the phospho-SIK1 moves to the cytoplasm within
10 min. The phospho-SIK1 is then dephosphorylated, and
SIK1 re-enters t he nucleus several hours later. A period when
SIK1 was not present in the nucleus appeared to coincide
with that when the SIK1-dependent CREB-repression was
not detected. A mutant SIK1 (having Ala577) was exclu-
sively present in the nucleus, and its CREB-repression
activity was higher than that of the wild-type SIK1 [28,30].
Based on these results, we conclude that the nuclear presence
of SIK1 is important for the repression of CREB.
The recent isolation and characterization of an a dipose-
specific isoform, SIK2 [31], forced us to reconsider the
relationship between intracellular localization and CREB-
repression activity of SIKs. SIK2 is localized ma inly in the
cytoplasm of mouse 3T3-L1 preadipocytes but represses
CREB activity, a lthough the degree of repression by SIK2
seems to be lower than that by SIK1. Therefore, we decided,
first, to identify a domain(s) that determined the intracel-
lular localization of SIK1 and, second, to examine its role in
the CREB repression. Our results indicated that a short

peptide stretch (comprising residues 5 86–612), the RK-rich
region, was important for the nuc lear localiz ation of SIK1.
Surprisingly, the RK-rich region-defective SIK1 mutants,
seen mainly in the cytoplasm, were able to repress the
CREB-mediated transcription strongly. Moreover, chime-
ras constructed from SIK1 and SIK2, having their RK-rich
region and the corresponding region exchanged between the
two enzymes, showed no co rrelation between their nuc lear
localization and CREB-repression activities. On the other
hand, both the nuclear and c ytoplasmic SIK1 mutants
inhibited the ACTH-induced nuclear entry of a CREB-
specific co-activator, TORC2 (transducer of regulated
CREB activity 2). We therefore concluded that SIKs, even
present in the cytoplasm, could repress CREB-mediated
gene expression, and the RK-rich region of SIK1 was
important for not only the nuclear localization, but also the
attenuation of CREB-repression activity.
Materials and methods
Cell culture and reporter assay
Y1 cells were maintained in DMEM (Dulbecco’s modified
Eagle’s medium) (Sigma), containing 10% (v/v) fetal bovine
serum and antibiotics, at 37 °C under an atmosphere of 5%
CO
2
/95% air. 3T3-L1 cells, obtained from Japan Health
Sciences Foundation (Osaka, Japan), were maintained in
DMEM, as described previously [31]. 4¢,6-Diamidino-2-
phenylindole (DAPI) dilactate was from Molecular Probes,
and fluorescein isothiocyanate (FITC)-conjugated anti-rat
IgG was from Funakoshi (Tokyo, Japan).

The method of reporter assays was as described previ-
ously [29]. To introduce plasmids into cells, Lipofect-
AMINE 2000 (Invitrogen Corp., Carlsbad, CA, USA) was
used in this study. Luciferase activities were measured by
using the Dual-Luciferase Reporter Assay System (Promega
Corp., Madison, WI, USA). For the CRE-reporter a ssay,
Y1 cells (1 · 10
5
per well) were transfected with the SIK1
expression plasmid (pIRES-SIK1, pIRES-SIK1 mutants or
pIRES empty vector: 0.2 lg), C RE-luciferase reporters
[pTAL-CRE or pTAL (empty reporter) alone: 0.2 lg], the
PKA expression plasmid (pIRES-PKA or pIRES: 0.1 lg)
and pRL-SV40 (internal standard: 0.03 lg). For the CREB
reporter assay, cells were transformed with Gal4 DNA-
binding domain-linked CREB expression vectors [pM-
CREB(F), pM-CREB(S) or pM, empty vector: 0 .15 lg],
expression vectors for SIK1 mutants or empty expression
vectors ( pIRES-SIK1 or pIRES: 0.2 lg), the PKA expres-
sion vector or empty v ector (pIRES-PKA or pIRES:
0.1 lg), and reporter vectors [GAL4-linked luciferase
reporter (pTAL-5x GAL-4: 0.15 lg) and an internal control
(pRL-SV40: 0.03 lg). Transformation efficiencies were
corrected by Renilla luciferase activities. The specific tran-
scriptional activities derived from the CRE and CREB were
expressed as fold-expression of the reporter activity of the
empty vector, pTAL and pM, respectively.
For fluorescence microscopy observations, cells were
cultured on poly-
L

-lysine coated coverslips (18 mm;
Matsunami Co. Ltd, Tokyo , Japan) in a 12-well dish. Cells
(< 1 · 10
4
) were transformed with 0.5 lgofGFP-SIK1
expression vector, incubated for 16 h and stimulated with or
without ACTH (10
)6
M
) for 1 h, fixed with 1 mL of 4%
paraformaldehyde dissolved in NaCl/P
i
for 15 min, stained
with DAPI [1 ngÆmL
)1
in NaCl/P
i
containing 0.01% ( v/v)
Triton X-100] for 5 min, and then washed with NaCl/P
i
four
times. Cells on the coverslip were embedded onto a glass
slide using 50% (v/v) glycerol. On average,  20–30% of the
cells showed detectable GFP-SIK1 signals in independent
triplicate experiments. The majority of localization patte rns
of GFP-SIK1 was classified according to whether it was
found extensively in the nucleus (N), was present at a higher
level in the nucleus (N > C), was evenly distributed between
the nucleus and the cytoplasm (N ¼ C), was present at a
higher level i n the cytoplasm (N < C) or was f ound

extensively in the cytoplasm (C), for more than 200 cells.
To visualize haemagglutinin (HA)-tagged SIK1, rat a nti-HA
tagged IgG (Boehringer Mannheim Biochemicals Inc.,
Mannheim, Germany) and FITC-conjugated anti-rat IgG
were used. The transformation efficiency of GFP-tagged, as
well as HA-tagged, proteins was always constant ( 20%).
Plasmids
cDNA fragments for rat SIK1 [27] and mo use SIK2 [31] were
cloned into the green fluorescent protein (GFP) expression
vector, pEGFP-C [28]. For the r eporter assay, the nontagged
4308 Y. Katoh et al. (Eur. J. Biochem. 271) Ó FEBS 2004
expression vector, pIRES (B-E), was used. The original
pIRESneo1 vector (Clontech Laboratories Inc., Palo Alto,
CA, USA) contained, in its cloning region, EcoRI and
BamHI sites in a reverse order to that of the pEGFP-C
vector. So, BamHI and EcoRIsitesinthisorder(B-E)were
created using oligonucleotides at th e NotI/BamHI s ite of
vector pIRESneo1. The oligonucleotides used were 5¢-GGC
CGGATCCGAATTC and 5¢-GATCGAATTCGGATCC.
cDNA fragments of C-terminal deletion SIK1s (1–740,
1–708, 1–632, 1–572, 1–341) were amplified by PCR using a
common forward primer (5¢-AAA
GGATCCATGGTGA
TCATGTCGGAGTTC: the BamHI site is indicated by
the underlined region) and each s pecific reverse primer
(740R: 5¢-AAA
GAATTCCTGTGGCAGGGGACCAGT
GG; 708R: 5¢-AAA
GAATTCGGGCTGGAGGAGGGG
CGTTG; 632R: 5¢-AAA

GAATTCCGGGGTGTGGAAG
GTACTCA; 572R: 5¢-AAA
GAATTCCTCCTGGAAGC
TGACAGG; 341R: 5¢-AAA
GAATTCGAGCAGGAGG
TAGTAAAT; the EcoRI site is indicated by the underlined
region). Similarly, N-terminal-deleted cDNA fragments were
prepared by a common reverse primer (5 ¢-AAA
GAATTC
TCACTGTACCAGGACGAACGTCC: the EcoRI site is
indicated by the underlined region) and specific forw ard
primers (24F: 5¢-AAA
GGATCCGTGGGCTTTTACGAC
GTGGA; 163F: 5¢-AAAGGATCCATCAAGCTGGCAG
ATTTTGGA; 342F: 5¢-AAA
GGATCCGAGCGCCTCA
GGGAGCATCGA; 571F: 5¢-AAA
GGATCCCAGGA
GGGACGGAGAGCG; 632F: 5 ¢-AAA
GGATCCCCGG
CCCCAAGCTCAGGTCTG; the BamHI site is indicated
by the underlined region). Amplified cDNA fragments were
digested by BamHI/EcoRI and ligated into the BamHI/
EcoRI site of vector pEGFP-C. To prepare internal deletion
mutants, cDNA fragments of the N-terminal regions and of
the C-terminal regions were separately amplified by PCR
and then ligated to these fragments. To amplify frag-
ments c ontaining residues 1–572, 1–585 and 1–612, the
common forward primer and specific reverse primers
(572R: 5¢AAA

GGGCCCCTCCTGGAAGCTGACAGG;
585R: 5¢AAA
GGGCCCCAGCCCTTGAGTGAGAGA
CG;612R:5¢AAA
GGGCCCACGGGCCAATCCTTT
GATCTTGTTCAG; the Bsp120I site is indicated b y the
underlined region) were used. cDNA fragments containing
residues 586–776, 613–776 and 632–776 were amplified
by using the common reverse primer and specific primers
(586F: AAA
GGGCCCAAGGCCTCCCGGCAGCAG;
613F: AAA
GGGCCCCAGGTGTGCCAGTCCTCCATC;
632F: AAA
GGGCCCGCCCCAAGCT CAGGTCTG;
the Bsp120I site is indicated by the underlined region). The
cDNA fragments for the N -terminal regions and t he
C-terminal regions were digested with BamHI/Bsp120I
and b y EcoRI/Bsp120I, respectively, mixed at a ratio of
1 : 1, and ligated into the BamHI/EcoRI site of vector
pEGFP-C.
cDNA fragments encoding the RK-rich region were
amplified by PCR by using the BamHI-linked 586F primer
and EcoRI-linked 612R primer, digested by BamHI/EcoRI
and ligated into the BamHI/EcoRI site of vector pEGFP-C.
GFP expression vectors, having T-antigen nuclear localiza-
tion signal (NLS) and Rev-nuclear export signal (NES)
sequences, were prepared b y direct ligation of o ligonucleo-
tides a t the NotI site of t he pEGFP-C v ector. T-antigen
NLS oligonucleotides were 5¢-GGCCATTCAAAGTA

AAGAAGAAACGTA AGCCGT and 5¢-GGCCACG
GCTTACGTTTCTTCTTTACTTTGAAT; Rev NES
oligonucleotides were 5¢-GGCCTCTGCAGCTCCCGC
CACTGGAACGTCTTACCCTCGACA and 5¢-GGCC
TGTCGAGGGTAAGACGTTCCAGTGGCGGGAG
CTGCAGA. T-antig en NLS- and Rev-NES-inserted
D586-612 SIK1 mutants were prepared by ligation of
the above sets of oligonucleotides in the Bsp120I site of
D586-612.
Site-directed mutagenesis was carried out by using a kit,
GeneEditor (Promega), according to the manufacturer’s
protocol. The following primers were used: SIK-1 (R593A/
K594A), 5¢-TTCCGGCAGCAGCTAGCGGCAAACG
CGAGGACCAAG; SIK-1 (R597A/K599A), 5¢-CTAA
GGAAAAACGCGGCGACCGCGGGGTTCCTGGGA
CTG); S IK-1 (L602A/L604A), 5¢-AGGACCAAGGGG
TTCGCGGGCGCCAACAAGATCAAAGGA); SIK-1
(K606A/K608A), 5¢-TTCCTGGGACTGAACGCGAT
CGCCGGCTTGGCCCGTCAGGTG); SIK-1 (I607A/
L610A), 5¢-CTGGGACTGAACAAGGCCAAAGGCGC
CGCCCGTCAGGTGTGC).
To prepare chimeric SIK1 and SIK2 mutants whose
domain 3Bs were exchanged, the Quick change Site-directed
Mutagenesis Kit (Stratagene) was used. Because the cDNA
fragment of the domain 3B was too long to perform site-
directed mutagenesis, it was separated into 5¢ and 3¢ regions,
and the mutagenesis w as performed o n the r espective
fragments. The 5¢ fragments of the domain 3B from SIK1
and SIK2 were mutated with 5¢-GATACGTCTCTC ACT
CAAGGGATTGTAGCCTTCCGGCAGCATCTACAG

AATCTCGCGAGGACCAAGGGGTTCCTG/5¢-CAG
GAACCCCTTGGTCCTCGCGAGATTCTGTAGAT
GCTGCCGGAAGGCTACAATCCCTTGAGTGAGA
GACGTATC and 5¢-GATACGTCCCTTACACAAG
GACTTAAGGCATTTAGACAACAGCTTCGGAAG
AATGCTAGAACCAAAGGATTTCTG/5¢- CAGAAA
TCCTTTGGTTCTAGCATTCTTCCGAAGCTGTTG
TCTAAATGCCTTAAGTCCTTGTGTAAGGGACG
TATC, respectively. The 3¢ fragmen ts of domain 3B
from SIK1 and SIK2 were mutated with 5¢-GCGAGG
ACCAAGGGGATTCTGGAGCTGAACAAGGTGC
AATTGTTGTACGAACAGGTGTGCCAGTCCTCC/
5¢-GGAGGACTGGCACACCTGTTCGTACAATAA
TTGCACCTTGTTCAGCTCCAGAATCCCCTTTGGC
CTCGC and 5 ¢-CTTGCTAGAACCAAAGGATTTCT
GGGGTTGAACAAAATAAAAGGGCTGGCTCGGC
AAATGGGATCAAACGCAGAC/5¢-GTCTGCGTTTG
ATCCCATTTGTCGAGC CAGCCCTTTTATTTTGT
TCAACCCCAGAAATCCTTTGGTTCTAGCAAG.
The mammalian expression vector, pCMVspor6, con-
taining full-length mouse TORC2 (Clone ID: 5345301)
cDNA, was purchased from Invitrogen. Site-directed muta-
genesis was used to introduce a BglII site at the 5¢ terminus
of the cDNA. The following primers were used: mTORC2
Bgl II-F, 5-GGCGGGGACGGACGCGGGAGATCTA
TGGCGACGTCAGGG; mTORC2 Bgl II-R, 5-CCC
TGACGTCGCCATAGATCTCCCGCGTCCGTCCCC
GCC. The resultant full-length mouse TORC2 cDNA was
digested with BglII and NotI, and ligated into the BamHI/
NotI site of vector pEGFP-C.

Ó FEBS 2004 Cytoplasmic SIK1 represses CREB (Eur. J. Biochem. 271) 4309
Immunoprecipitation
Immunoprecipitation w as performed a s described previ-
ously [27]. Briefly, cells (5 · 10
5
) plated on a 10 cm dish were
transformed with 3 lg of expression plasmids (pSVL-HA)
for HA-tagged wild-type and mutant SIKs using 8 lLof
LipofectAMINE 2000. After 36 h of incubation, cells were
lysed in 0.7 mL of lysis buffer [50 m
M
Tris/HCl (pH 8.0),
300 m
M
NaCl, 5 m
M
EDTA, 5 m
M
EGTA, 2 m
M
dithio-
threitol, 50 m
M
b-glycerol phosphate, 50 m
M
NaF, 1 m
M
NaVO
4
, 0.5% T riton X-100, and protease inhibitor]. HA-

tagged SIK protein was immunoprecipitated by using anti-
HA-tag IgG (2 lg) and protein G-Sepharose (30 lL). T he
aliquots of immunopurified SIK1 were subjected to Western
blot analyses with anti-SIK1 IgG and in vitro kinase assays.
Purified SIKs were incubated with glutathione S-transferase
(GST)-Syntide2 in the presence 0.5 lCi (18.5 kBq) of
[
32
P]dATP[cP] at 30 °C for 30 min. The kinase reaction
was stopped by adding 3· SDS sample buffer [150 m
M
Tris/
HCl (pH 6.8), 6% (v/v) SDS, 30% (v/v) glycerol, and 0.1%
(v/v) bromophenol blue] and heating at 100 °Cfor5min.
The aliquots were subjected to 15% SDS/PAGE, and
phosphorylated peptides were visual ized by autoradio-
graphy.
Results
Subcellular distribution of truncated SIK1s
By using the GFP-fusion technique, w e previously reported
the phosphorylation-dependent nuclear export of SIK1 in
ACTH-stimulated Y1 cells [28]. When SIK1 was present in
the nucleus, i t was able to repress CREB a ctivity. However,
a cytoplasmic isoform of SIK, SIK2 [31], forced us to
reconsider correlation between the loss of CREB-repressing
activity and the cytoplasmic l ocalization o f SIK1. To
address to this problem, we decided to isolate SIK1
mutants, localized essentially in the cytoplasm, by modify-
ing a region determining the nuclear localization of SIK1.
When GFP-fused full-length SIK1 (i.e. SIK1 with GFP at

its N terminus) was expressed in unstimulated Y1 cells, the
majority of GFP signals formed speckles in the nucleus, and
the minority diffused into t he cytoplasm (the left column in
Fig. 1A). The nuclear GFP–SIK1, however, was completely
translocated to the c ytoplasm after treatment w ith ACTH
(the right column i n Fig. 1 A). When GFP alone was
expressed, the green fluorescence was present more exten-
sively in the nucleus than the cytoplasm, with or without
ACTH. Although the GFP fusion technique would yield
rapid results regarding the distribution of target molecules,
we must consider, as a caveat, that the GFP protein tends to
accumulate in the nucleus. In any case, in the hope of
determining domain(s) important for the intracellular
localization, we first attempted to investigate the intracel-
lular distribution of a number of GFP–SIK1 mutants with
N- and C-terminal deletions. Figure 1B illustrates areas in
the primary sequence of SIK1 t hat altered the intracellular
distribution of GFP–SIK1 fragments.
The distribution of SIK1(1–740) was similar to that of the
full-length SIK1. When the region containing residues 709–
776 was deleted, the resultant SIK1(1–708) was localized
only in the nucleus of Y1 cells and was not translocated to
the cytoplasm, even when the cells were stimulated with
ACTH. Similar results were obtained with SIK1(1–632).
From these results, we surmise that the region containing
residues 708–740 was important for the nuclear export of
SIK1. When the C-terminal deletion reached residue 573,
part of SIK1(1–572) was again found in the cytoplasm.
Similar results were obtained for SIK1(1–341). These results
suggest that the region containing residues 572–632 might

be important for t he nuclear localization of S IK1. Consid-
ering that proteins of < 40–6 0 kDa molecular mass would
enter the nucleus passively [32,33], the nuclear presence of
GFP-SIK1(1–572) and GFP-SIK1(1–341), having mole-
cular masses of 90 kDa and 66 kDa, respectively, suggest
the presence of minor nuclear localization activities in the
N-terminal fragment (1–572).
Next, the subcellular distribution of N-terminus-deleted
mutants was investigated. The distribution of SIK1(24–776)
was similar to that of full-length SIK1. On the other hand,
SIK1(163–776) and SIK1(342–776) were localized only in
the nucleus, and they failed to move to the cytoplasm in
response to ACTH. SIK1(571–776) was localized exclu-
sively in the nucleus, suggesting the presence of an active
nuclear localization signal in the region comprising residues
571–776. When the N-terminal deletion reached residue 631,
the resultant SIK1(632–776) was diffusely distributed all
over the cell. Taken together, these results indicate that
a major nuclear localization signal might exist in re gion
573–631.
The region 573–631 does not contain a cluster of more
than three successive basic residues, a feature often seen in
typical nuclear localization signals, such as th e unipartite
basic cluster KKKRK of SV40 (simian virus 40) T-antigen
[34] and the bipartite cluster RKR-Xn-RKRKR of T-cell
protein tyrosine phosphatase [35]. However, close examina-
tion revealed the presence in the region of a peptide stretch,
Lys586-Ala-Phe-Arg-Gln-Gln-Leu-Arg-Lys-Asn-Ala-Arg-
Thr-Lys-Gly-Phe-Leu-Gly-Leu-Asn-Lys-Ile-Lys-Gly-Leu-
Ala-Arg612 ( basic residues shown in bo ld), in which basic

and hydrophobic residues were interspersed. We named this
stretch the ÔRK-rich region (586–612)Õ. It should be noted
that Ser577, an important residue for the phosphorylation-
dependent nuclear export of SIK1, exists near the RK-rich
region.
Region 586–612, the RK-rich region, determines
the intracellular localization of SIK1
To investigate further the role(s) played by the region 573–
631 for determining the intracellular distribution of SIK1,
we created several mutants w ith deletions in this re gion.
Noting that the RK-rich region (586–612) was positioned
in the centre of 573–631, we decided to s plit the region
573–631 into three parts – an N-terminal region (573–585),
the RK-rich region (586–612), and a C-terminal region
(613–631) – and produced four deletion mutants, shown in
Fig. 2A.
SIK1(D573–631), a mutant lacking the entire region, was
distributed diffusely all over the cell with no response to
ACTH. S IK1(D573–585) was localized specifically in the
nucleus with a weak response to ACTH. However,
SIK1(D586–612) (i.e. the mutant minus the RK-rich region)
was present mainly in the cytoplasm of resting Y1 cells,
4310 Y. Katoh et al. (Eur. J. Biochem. 271) Ó FEBS 2004
although found to a minor extent in the nucleus, and was
translocated completely to the cytoplasm a fter st imulation
with ACTH. S IK1(D613–631) b ehaved similarly to wild-
type SIK1. Whether the RK-rich peptide alone has nuclear
translocation activity was tested by using the GFP-fused
RK-rich peptide. As shown in Fig. 2B, GFP-SIK(586–612)
was localized only in the nucleus. (Note that GFP alone was

distributed diffusely all over the cell, as shown in Fig. 1 A.)
In the control experiments, the GFP-linked SV40 T-antigen
nuclear localization signal (NLS), KKKRK (basic residues in
bold) [34] was found only in the nucleus, while the GFP-
linked HIV-1 Rev protein nuclear export s ignal (NES),
LGLPPLERLTLD ( hydrophobic residues in bold) [36],
was found only in the cytoplasm. These results indicated
that the RK-rich region might be important for the nuclear
localization of SIK1.
Basic residues in the RK-rich region (586–612) were
replaced with Ala, and the intracellular distribution of
mutants was examined (Fig. 3A). SIK1(R593A/K594A),
SIK1(R597A/K599A) and SIK1( K606A/K608A) were
found mainly in the cytoplasm of resting Y1 cells. When
the cells were treated with ACTH, those mutants were
translocated to the cytoplasm. These results suggested that
the basic residues in the RK-rich region might be important
for the nuclear localization of SIK1. Next, bulky
hydrophobic residues, such as Leu and Ile, were replaced
with Ala, and the intracellular distribution of mutants was
investigated (Fig. 3A). SIK1(L602A/L604A) was distri-
buted diffusely both to the nucleus and to the cytoplasm
of resting Y1 cells. On the other hand, SIK1(I607A/L610A)
was localized mainly in the c ytoplasm. These findings
suggest that these hydrophobic residues, too, might have an
important role for the nuclear localization.
Based on the above findings, we attempted to prepare
SIK1 mutants that would be predicted to localize exclusively
either in the nucleus or in the cytoplasm. Hence, either
canonical SV40 T-antigen NLS or Rev NES was inserted

into the deleted part of SIK1(D586–612) (Fig. 3B).
SIK1(D586–612 + NLS) was accumulated more in the
nucleus than the parent S IK1(D586–612), and, when the
cells were stimulated with ACTH, SIK1(D586–
612 + NLS) was translocated to the cytoplasm (Fig. 3A).
A
B
Fig. 1. Intracellular distribution of green
fluorescent protein (GFP)-fusion salt-inducible
kinase 1 (SIK1) mutants. (A) Y1 cells, cultured
on coverslips and transformed with over-
expression vectors for GFP-tagged SIK1
protein, were t reated with or without
adrenocorticotro pic hormone (ACTH)
(10
)6
M
) for 1 h and fixed for fluorocyto-
chemical an aly ses. G reen fluorescent signals o f
GFP-SIK1 (upper), and blue fluorescent sig-
nals representing nuclear staining with 4¢,6-
diamidino-2-phenylindole (DAPI) (middle),
are shown. The intracellular localization of
GFP alone is shown in the bo ttom panels. T he
patterns of the intracellular distribution of
green fluorescent signals were classified into
five groups (N, N > C, N ¼ C, N < C and
C), as described in the Materials and methods,
and representative pictures are shown.
(B) C-terminal- and N-terminal-deleted SIK1

mutants were expressed in Y1 cells. The wild-
type SIK1 (Full) contains amino acids 1–776.
The kinase domain (27–278) is in the
N-terminal half, whereas the region essential
for nuclear localization (573–631) is in the
C-terminal half. The Arg/Lys-rich region
(RK-rich region) is shown as a black box.
Ó FEBS 2004 Cytoplasmic SIK1 represses CREB (Eur. J. Biochem. 271) 4311
By contrast, SIK1(D586–612 + NES) was found exclu-
sively in the cytoplasm of both resting and ACTH-treated
Y1 cells. Thus, the T-antigen NLS could b e rep laced with
the RK-rich region as the nuclear localization signal, while
the Rev NES, if inserte d into SIK1( D586–612), c ould
function as the nuclear export signal.
We also produced an unpho sphorylatable SIK1(D586–
612) mutant, in which Ser577 was replaced w ith Ala. As
expected, the nuclear accu mulation of SI K1(D586–
612 + S577A) was higher than that of parent mutant
SIK1(D586–612) and was influenced a little by ACTH
treatment. On the other hand, SIK1(S577A), the nuclear
export-defective mutant, was localized only in the nucleus,
as reported p reviously [28]. A comparison between the
intracellular distribution of SIK1(D586–612/S577A) and
that of SIK1(S577A) again highlights the importance of the
RK-rich region as the nuclear localization signal.
The RK-rich region of SIK1 and the corresponding region
of SIK2 determine the intracellular localization
The above results suggest that the intact RK-rich region is
an important determinant for the nuclear localization of
SIK1. Before proceeding to investigate the CREB-repress-

ing activity of these cytoplasmic SIK mutants, we focused
our attention on S IK2, an adipose-specific isoform of
SIK1. SIK2, having a s imilar, but distinct, amino acid
sequence in a region corresponding to the RK-rich region
(Fig. 4 A,B), was localized mainly in the cytoplasm of 3T3-
L1 mouse preadipocytes [31,37]. Alignment of the two
isoforms indicates that they have three highly conserved
domains (Fig. 4A). Protein kinase structures are present in
domain 1, and the function of domain 2 remains to be
explored. D omain 3 shares 73% identical amino a cid
residues between the two isoforms. Domain 3A, the
N-terminal half, contains a PKA-dependent phosphory-
latable Ser (Ser577 in SIK1 and Ser587 in SIK2,
respectively), whereas domain 3B, the C -terminal half,
corresponds to the RK-rich region and its equivalent
(Fig. 4 B). T he simila rity of domain 3B between the two
isofoms i s lower than that of domain 3A. Notably, five of
nine basic residues present in the R K-rich region are
replaced with other residues in the corresponding region
of SIK2. To examine whether or not the lower similarity
in domain 3 B between the two isoforms contributes to the
difference in their intracellular distribution, two chimeras,
each having its domain 3B replaced with that of the other
isoform, were constructed a nd expressed in Y 1 cells
(Fig. 4 C). Chimera 1, SIK1, in which RK-rich region had
been replaced with domain 3B of SIK2, was localized less
inthenucleusthanthewild-typeSIK1(seethedifference
between the first and second panels in Fig. 4C). On the
other h and, Chimera 2, SIK2, in which domain 3B had
been replaced with the RK-rich region, was localized more

inthenucleusthanthewild-typeSIK2(seethedifference
between panels 4 and 5 in Fig. 4C). The treatment of Y1
cells with ACTH seemed to induce the nuclear export of
these chimeras. When Ser577 of Chimera 1 was mutated
to Ala, a substantial amount of the mutant was retained
in the nucleus after treatment with ACTH (see the
difference between panels 2 and 3 in Fig. 4C). In contrast,
the similar SerfiAla mutation in Chimera 2 seemed not to
influence the intracellular distribution of Chimera 2 (see
the difference between panels 5 a nd 6 in Fig. 4C). These
results indicate that the RK-rich region, domain 3B of
SIK1, plays an important role in the nuclear localization
of SIK1, but not for domain 3B of SIK2.
Although several mutants, such as D573–585 (Fig. 2A),
I607A/L610A (Fig. 3A), D586–612 + S577A (Fig. 3B),
Chimera 1 + S577A (Fig. 4C) and Chimera 2 + S 587A
(Fig. 4 C), did not give unequivocal results as to the
intracellular l ocalization as well as the ACTH-dependent
B
A
Fig. 2. The RK-rich region is essential for the
nuclear localization of salt-inducible kinase 1
(SIK1). (A) An overall structure o f green
fluorescent protein (GFP)-SIK1 is shown at
the top. Ser577 is the residue that is phos-
phorylated by protein kinase A (PKA) when
the cells are stimulated. The intracellular
distribution of SIK1 mutants containing the
deletion in the RK-rich region were investi-
gated as described in the legend to Fig. 1 and

in the Materials and methods. (B) The GFP-
fused RK-rich region nuclear localization sig-
nal(NLS)(KKKRK)ofSV40(Simianvirus
40) T-antigen, and nuclear export signal
(NES) (LGLPPLERL TLD) of t he H IV-1 Rev
protein were expressed, and their intracellular
distribution were investigated.
4312 Y. Katoh et al. (Eur. J. Biochem. 271) Ó FEBS 2004
intracellular translocation, we decided to investigate the
relationship between the intracellular localization and
CREB repression activity of SIK1 mutants.
Even SIK1 mutants present in the cytoplasm repress
PKA-induced CRE activity
As reported previously, SIK1(S577A) was retained in
the nucleus, even after the activation of PKA, and its
transcriptional repression activity was higher than wild-type
SIK1, indicating t hat SIK1 p resent in the nucleus could
repress CREB activity [28]. We decided to test the
transcriptional repression activity of SIK1(1–708), another
mutant localizing exclusively i n the nucleus (shown in
Fig. 1B). The SIK1(1–708) expression plasmid, pIRES-
SIK1(1–708), was co-introduced with the CRE–reporter
construct into Y1 cells in the p resence or absence of the
PKA expression vector, and the SIK1-dependent repression
of CRE activity was examined. As expecte d, the extent of
CRE repression by SIK1(1–708) was greater t han that
induced by wild-type SIK1 (Fig. 5A). The protein kinase
activity was required for SIK1(1–708) to exert this repres-
sion, because t he repression was not seen by SIK1(K56M/
1–708). We further t ested the CRE-repression activities of a

variety of other mutants, whose intracellular localizations
were examined in Fig. 3B. To our surprise, all the mutants
seemed to inhibit CRE activity more strongly than wild-type
SIK1 (Fig. 5B). SIK1(D586–612 + NES), which w as
localized exclusively in the c ytoplasm ( Fig. 3B), could also
repress CRE activity as strongly as the nuclear resident
SIK1(S577A). These results suggest t hat the nuclear
localization of SIK1 might not be a prerequisite for its
CRE repress ion.
By using Gal4–CREB reporter systems, w e previously
demonstrated that SIK1 repressed PKA-induced CREB
activation by acting on the b-ZIP domain of CREB [29],
and the nuclear resident S IK1(S577A) repressed CREB
more strongly than the wild-type SIK1 [28]. Similar
experiments were designed to test whether the b-ZIP
domainofCREBwasthetargetsitefortherepressive
action of the cytoplasmic SIK1. As shown in the left panel
of Fig. 5C, the cytoplasmic SIK1(D586–612 + NES) and
the nuclear re sident SIK1(S577A) similarly r epressed the
full-length CREB-dependent reporter activity. On the
contrary, neither mutant repressed the reporter activities
of b-ZIP-less CREB (the right panel of Fig. 5C). The other
cytoplasmic SIK1 mutants, such as SIK1( D586–612) and
SIK1(D573–631), were also able to repress the full-length
CREB-dependent reporter activity (data not shown). To
verify the above results, the expression levels of exogenous
mutant SIK1 protein and their kinase activity were exam-
ined (Fig. 5D). The results showed no significant difference
in their expression levels. These results suggest that SIK1,
when present in the cytoplasm, is able to repress the

A
B
Fig. 3. Effe ct of alteration of the RK-rich
region on the intracellular distribution of
salt-inducible kinase 1 (SIK1). (A) Effect of
site-directed mutagenesis in the RK-rich re-
gion on the intracellular distribution of SIK1.
Basic and hydrophobic residues, indicated by
underlines and asterisks, respectively, were
mutated to Ala. The intracellular distribution
of mutant SIKs was investigated as detailed in
the legend to Fig. 1. (B) The RK-rich region
(586–612) was replaced with a nuclear local-
ization signal (NLS) of T-antigen or with a
nuclear export signal (NES) of the Rev pro-
tein. To disrupt the nuclear export, mutation
S577A was introduced into wild-type SIK1
and SIK1(D586–612). These mutants were
expressed as GFP-fusion proteins, and their
intracellular distribution was examined as
detailed in the legend of Fig. 1.
Ó FEBS 2004 Cytoplasmic SIK1 represses CREB (Eur. J. Biochem. 271) 4313
PKA-induced CREB-activation through the b-ZIP domain
of CREB. We also confirmed that the size of the tag
(GFP or HA) did not influence the intracellular distribution
of SIK1 and i ts mutants. The intracellular distribution
of HA-tagged SIK1 and its mutants, S577A (nucleus) or
D586–612 + NES (cytoplasm), showed similar features to
those of their GFP-tagged counterparts (Fig. 3B).
The RK-rich region and the corresponding region

of SIK2 modulate the CRE-repression activity
The chimeras produced from the two SIK isoforms were
tested for their CRE repression activity in Y1 cells. Chimera
1, localizing mainly in the cytoplasm (Fig. 4C), repressed
CRE ac tivity to an extent similar to that of wild-type SIK1
(Fig. 6A). On the other hand, Chimera 1, having an S577A
mutation, was a strong repressor. To our surprise, both
Chimera 2 and wild-type SIK2 elevated, not repressed, the
PKA-dependent enhanced CRE activity (Fig. 6B), whereas
Chimera 2, having an S587A mutation, was a strong
repressor. Because we previously reported that w ild-type
SIK2 repressed, not elevat ed, CRE activity in 3T3-L1 cells
[31], t he experiments with Chimera 2 and wild-type SIK 2
were repeated by using 3T3-L1 cells. Both Chimera 2 and
wild-type SIK2 repressed the CRE activity to a moderate
extent in 3T3-L1 cells.
The kinase activities of SIKs were essential for their CRE-
repression activities, a nd some kinase-defective mutant
SIKs [SIK1 ( K56M) and SIK2 (K49M)] h ave been shown
to influence the CRE a ctivity in a domi nant-negative
manner [31]. To exclude the possibility that the enhance-
ment of CRE activity by SIK2 in Y1 cells was a re sult of
lower kinase activity, GST-tagged SIK1 and SIK2 were
expressed in Y1 cells, purified using glutathione columns
and subjected to an in vitro kinase assay. The specific kinase
activity of SIK2 in Y1 cells was similar to that of SIK1.
Although we cannot presently discuss the molecular mech-
anism underlying the up-regulation of CRE activity by
SIK2 in Y1 cells, the RK-rich region of SIK1, and the
corresponding region of SIK2, may play a critical role in

modulating the CREB-repression activity of SIKs.
Both nuclear and cytoplasmic SIK1s inhibited
the ACTH-induced nuclear entry of TORC2
TORCs, ubiquitously expressed in a variety of cells, are
CREB-specific co-activators, essential for bo th basal and
cAMP-induced CREB activity [38,39]. Recently, cAMP-
induced nuclear import of TORC2 was found in insulinoma
cells [40], suggesting that the nucleo-cytoplasmic shuttling
of TORC2 was important for the regulation of the
co-activation function of TORC2. We also showed that
SIK2 phosphorylated TORC2 at S er171 and that the
phospho-TORC2 could not enter the nuclei in forskolin-
stimulated cells [40]. These findings prompted us to examine
whether or not the nuclear import of TORC2 could occur in
the SIK1 mutant-expressing Y1 cells.
As shown in Fig. 7 , GFP-fused TORC2 was localized
both in the nucleus and in the cytoplasm of resting Y1 cells,
ACTH(–). When the cells were treated with ACTH,
A
B
C
Fig. 4. Difference between salt-inducible kinase
1 (SIK1) and SIK2 with special reference to
domain 3. (A) Alignment of the primary
sequence and the subcellular distribution of
SIK1 (blue bar) and SIK2 (pink bar). Three
highly conserved regions, domains 1–3, were
depicted. (B) Amin o acid sequ ences of do main
3 are shown. The p rotein kinase A (PKA)-
dependent phosphorylation sites, Ser577 in

SIK1 and Ser587 in SIK2, are indicated in
bold. The basic residues in domain 3B are
indicated by underlines. The similarity
between SIK1 and SIK2 in domain 3A, the
N-terminal half of domain 3, is higher than
that in domain 3B, the C-terminal half.
(C) The intracellular distribution of chimeras
in Y1 cells.
4314 Y. Katoh et al. (Eur. J. Biochem. 271) Ó FEBS 2004
ACTH(+), GFP-TORC2 was found mostly in the nucleus,
indicating that the ACTH signalling induced the nuclear
import of TORC2, resulting in a TORC2-dependent CREB
activation. When HA-tagged SIK1 was co-expressed with
GFP-TORC2 in these cells, most GFP-TORC2 signals were
localized in the cytoplasm. In the ACTH-treated cells,
TORC2 seemed to move into the nucleus, although the
extent of the nuclear import seemed less than that in the
SIK1-nonexpressing control cells (–). When either the n uc-
lear resident mutant SIK1, S577A, or the cytoplasmic
mutant SIK1, D586–612 + NES, was co-expressed,
moreTORC2seemedtoberetainedinthecytoplasm,
corroborating that these SIK1 mutants could constitutively
repress CREB activity (Fig. 5B).
Discussion
We reported previously that SIK1 could repress the
CREB-mediated transcription activation in cultured cells,
and the extent of CREB repression apparently correlated
with the amount of SIK1 present in the nucleus [28].
Based on the assumption that a certain region of the SIK1
A

C
D
B
Fig. 5. Bo th nuclear and cy toplas mic salt-inducible kinase 1 (SIK 1) mutants repress cAMP-response element (CRE)/CRE-binding protein (CREB)
activities. (A) The protein kinase A (PKA)-induced CRE activity, and its repression by C-terminal-deleted nuclear SIK1, DC(708), and its kinase-
defective mutant, K56M/ DC(708), were investigated by using Y1 c ells. Y1 cells (1 · 10
5
/well) were transfected with SIK1 expression plasmid, CRE-
luciferase reporter, PKA expression plasmid and pRL-SV40 (internal standard), as described in the Materials and methods. After 15 h of
incubation, cells were harvested for luciferase assay. The specific transcriptional activities derived from the CRE were expressed as fold-expression
of the reporter activity of the empty vector, pTAL. Mean values and SD are indicated (n ¼ 4). (B) The CRE activity of the RK-rich region mu tants
and S577A mutants (0.1 or 0.3 lg of SIK1-expression plasmid was used), as given in Fig. 3B, were examined as described above. Mean values and
SD are indicated (n ¼ 3). (C) The transactivation activities of full-length CREB [CREB (F): (left panel)] and basic leucine zipper domain (b-ZIP)
minus CREB [CREB (S): (right panel)] were investigated by using the Gal4 DNA-binding domain-linked assay. Cells were transformed with the
Gal4 DNA-binding domain-linked CREB expression vectors, described above, the expression vectors for SIK1, and PKA and reporters (GAL4-
linked luciferase reporter pTAL-5x G AL4 and internal standard pRL-SV40). The specific transcriptional activities o f C REB w ere expressed a s fold-
expression of the empty Gal4 vector, pM. Mean values and SD are indicate d (n ¼ 3). Grey bars indicate CREB activities in Y1 cells overexpressing
kinase-defective SI K1s. Cells were transformed w ith expression plasmids for h aemagglutin in (H A)-tagged wild -type an d m utant S IKs. Aft er 36 h of
incubation, cells were lysed, and HA-tagged SIK protein was immunoprecipitated by using anti-HA-tag IgG and protein G–Sepharose, as
described in the Materials and methods. The aliquots of immunopurified SIK1 were subjected to Western blot analyses with anti-SIK1 IgG (upper
panel) and in vitro kinase assays (lower panel). GST-Syntide2 was used for a substrate. The results were representative of experiments carried out in
duplicate.
Ó FEBS 2004 Cytoplasmic SIK1 represses CREB (Eur. J. Biochem. 271) 4315
peptide is responsible for the nuclear localization of SIK1
and that this region may play a crucial role in the
regulation of the SIK1-mediated transcription repression,
we first tried to identify a nuclear localization signal of
SIK1. Second, by utilizing SIK1 mutants with a defective
nuclear localization signal, we examined the relationship
between the nuclear localization and the transcription

repression activity of SIK1.
The RK-rich region (586–612), containing nine basic
residues, was shown to act as a nuclear localization signal
(Figs 1–4). This region, however, was not typical of nuclear
localization signals [34,35,41–45]. The importance of the
basic residues in the RK-rich region for the nuclear
localization of S IK1 w as also demonstrated by using
chimeras of SIK1 and SIK2 (Fig. 4). Hence, domain 3B
of SIK2, the region corresponding to the RK-rich region,
has only four basic residues, and SIK2 was localized mainly
in the cytoplasm. On the other hand, the SIK2-derived
chimera, in w hich domain 3B was replaced with the
RK-rich region, accumulated in t he nucleus. Conversely,
the SIK1-derived chimera, containing domain 3B of SIK2,
was localized to a significant extent in the cytoplasm.
The RK-rich region contains several h ydrophobic resi-
dues that intersperse in the peptide stretch, a f eature typical
of the nuclear export signal (L-X
2,3
-L-X
1,2
-L-X-L) (hydro-
phobic residues are shown in bold) [46–49]. The experiments
using SIK1 mutants with disrupted hydrophobic residues,
especially those of SIK1(I607A/L610A), indicated the
importance of these residues f or the nuclear distribution
of SIK1.
Using a variety of SIK1 mutants localized in the
cytoplasm, we next examined CRE/CREB-repression activ-
ity of these mutants. The results indicated t hat the mutants,

although localized mainly in the cytoplasm, could repress
CRE-reporter activity more strongly than wild-type SIK1
(Fig. 5 ). Moreover, SIK1(D586–612 + NES), SIK1 having
its RK-rich re gion replaced with a strong nuclear export
signal, was localized exclusively in the cytoplasm, and
nonetheless i t behaved as a strong CREB-repressor. The
extent of repression by SIK1(D586–612 + NES) was
similar to that induced by the nuclear SIK1(S577A). The
experiments using chimeras of SIK1 and SIK2 also
supported no correlation between the in tracellular distribu-
tion and the CREB-repressing activity of SIKs (Fig. 6).
Fig. 7. Salt-inducible kinase 1 (SIK1) inhibits
the nuclear translocation of TORC2 (transdu-
cer of regulated CREB activity). The effects of
SIK1 and its mutants on the subcellular
localization of green fl uorescent protein
(GFP)-TORC2 were investigated as described
inthelegendtoFig.1andintheMaterials
and methods. An e xpression plasmid for
GFP-TORC2 was co-transfected without (–)
or with wild-type (WT) SIK1 or mutant SIKs
(nuclear: S577A or cytoplasmic: D586-612 +
NES) to Y1 cells, and cells were treated
with (+) or without (–) adrenocorticotropic
hormone (ACTH) for 30 min.
Fig. 6. Eff ect of chimeric salt-inducible kinases (SIKs) on cAMP-
response elem ent (CRE)/CRE-binding protein (CREB) activities. The
CRE-repressing activities of SIK1 and chimera 1 (A) and of SIK2
and chimera 2 (B) were examined by using Y 1 cells, as described in
thelegendtoFig.5.ToexpressSIK1andSIK2,weusedpIRES

(B-E) p lasmid and pTarget plasmid, respectively. The means and
SD values are indicated (n ¼ 3).
4316 Y. Katoh et al. (Eur. J. Biochem. 271) Ó FEBS 2004
Many intracellular shuttling molecules that chan ge their
intracellular l ocalizations under the influence of cellular
metabolism play their distinct roles in the respective cellular
compartments. Some molecules, however, when forced to
remain in only one compartment, behave similarly to when
they are retained in the other c ompartment. For example,
serum and glucocorticoid-induced kinase (Sgk), is known to
change its intracellular l ocalization during the cell cycle in
mammary tumour cells [50,51]. When ectopically expressed,
the wild-type Sgk was retained only in the cytoplasmic
compartment and induced growth arrest of transformed
cells. On the other hand, a Sgk mutant with the SV40
T-antigen NLS at its C terminus was retained in the nucleus,
but was also found to inhibit cell growth [51]. FKBP12-
rapamycin-associated protein (FRAP; also called mTOR or
RAF), another shuttling kinase, is involved in the regulation
of translation initiation by activating p70 S6 kinase and
phosphorylating eIF-4E-binding protein [52]. When FRAP
was restricted to the nucleus by adding an exogenous
nuclear localiz ation signal, the a ctivation o f p 70 S6 kinase
and the phosphorylation of e IF-4E-binding protein were
enhanced. In contrast, the addition of an exogenous nuclear
export signal to FRAP resulted in the inhibition of FRAP-
dependent cascades. A translational-initiation reporter
assay, however, indicated that the forced retention of FRAP
not only in the cytoplasm, but also in the nucleus, inhibited
rapamycin-sensitive translational in itiation [52]. These

examples suggest that, in these cases, the nucleo-cytoplasmic
shuttling, rather than the intracellular distribution, may be
important for regulating the cellular functions.
Some shuttling molecules involved in transcription regu-
lation(s) have their intracellular localization signals over-
lapped with the regions involved in transcription regulation.
Smad proteins transduce transforming g rowth factor-b
(TGF-b) family cytokine signals from the plasma mem-
brane to the nucleus. The disruption of an intrinsic nuclear
localization signal, or a nucle ar export signal of Smad
proteins, r esulted in the accumulation of Smads in the
cytoplasm or nucleus, respectively [53]. Because a domain
having these signals is overlapped with regions regulating
transcriptional activity [54] and accepting upstr eam signals
[55], the disruption of these signals resulted i n the same
phenotype, the impaired transcription activity of Smads.
Similarly, an intrinsic nuclear localization signal of MAPK-
activated protein kinase 5 (MK5) w as shown to overlap
with a binding site for the upstream kinase p38 [56]. Thus, it
was proposed that importin, a receptor for the nuclear
localization signal and p38 might compete for their binding
sites in MK5, and, as the result, the nucleo-cytoplasmic
shuttling and the activation of MK5 occurred simulta-
neously in a coupled manner.
Taken together, we conclude that SIK1, present in the
cytoplasm, could well repress the CREB-mediated tran-
scription. Therefore, the SIK1-mediated CREB-repression
may not occur through a direct physical interaction between
SIK1 and the CREB-transcriptional machinery complex in
the nucleus. Then, what mechanism underlies the SIK1-

dependent repression? A possible answer to this question
would be that one of CREB co-activators, that when
present in the nucleus interacts with CREB and activates the
CREB-dependent transcription, is itself a phosphorylation-
dependent shuttling molecule. This co-activator could be
phosphorylated by SIK1, either in t he nucleus or in the
cytoplasm, and the phosphorylated co-activator would
move out of the nucleus or be tethered to the cytoplasm. We
found that TORCs were qualified as such co-activators
[34,35].TORCswereshowntobindtotheb-ZIPdomainof
CREB, the binding being i ndependent of the Ser133–
phosphorylation of C REB, and a ctivate CREB-mediated
transcription. On the other hand, SIK1 was shown to
repress CREB activity v ia the b-ZIP domain but fail to
directly phosphorylate CREB. The close examination
revealed that TORCs indeed have SIKs-phosphorylation
motifs. These findings prompted us to test the possibility
that SIK1 might regulate the TORCs activity in a
phosphorylation-dependent manner. Thus, we have shown
that SIK2, when b ound to TORC2, phosphorylated and
inhibited the PKA-induced nuclear entry of TORC2 in
insulinoma cells [36], and, in this study, SIK1 inhibited the
ACTH-induced nuclear import of GFP-TORC2 (Fig. 7).
The inhibition of nuclear entry of TORC2 occurred only in
the case using SIK1 lacking the phosphorylatable Ser577 or
domain 3B, and seemed to occur independently of the
intracellular localization of SIK1. These results strongly
indicate that TORCs are the phosphorylation targets of
SIKs-mediated CREB repression.
Figure 8 illustrates d omains of SIK1, w ith special refer-

ence to their roles for determining the intracellular localiza-
tion and CREB-modulat ing activity. The RK-rich region
(586–612) (domain 3 B) may act as the major nuclear
localization signal. Because the PKA-dependent phosphory-
lation of Ser577 induced the translocation of SIK1 to the
cytoplasm, the phospho-Ser577 in domain 3 A seems to be a
key regulator for the nuclear export of SIK1. As for the
major determinant for CREB-repression activity, the intact
kinase domain is important. SIK1(S577A) stro ngly repres-
ses the CREB activation, indicating that the phosphoryla-
tion in domain 3 A disrupts the CREB-repression activity of
SIK1. The loss of the RK-rich domain, SIK1(D586–612),
Fig. 8. The regions important for the intracellular distribution and
cAMP-response element-binding protein (CREB)-modulating activity of
salt-inducible kinase 1 (SIK1). The kinase domain is indicated as KD.
The region 567–585 may be important for nuclear export of SIK1
when Ser577 is phosphorylated. The region 586–612 (RK-rich region)
may act as the major nuclear import signal of SIK1. The combination
of phospho-Ser577 and the intact struct ure of the RK-rich region is
required for attenuating the CREB re pression ac tivity of the S IK1
kinase domain.
Ó FEBS 2004 Cytoplasmic SIK1 represses CREB (Eur. J. Biochem. 271) 4317
also enhances the CREB-repression activity of SIK1,
indicating that the complete c ombination of phospho-
domain 3A and intact domain 3B is important for the
modulation of CREB-repression activities of SIK1. An
intact structure of domain 3 of SIK may regulate the
interaction between SIK and TORC. The precise mechan-
ism of SIK-dependent CREB repression is now under
investigation.

Acknowledgements
This research wa s supported by Grants-in-Aid for Scientific Research
from the Ministry of E duc ation, Culture, Sport s, Science and
Technology and M inistry of Health, Labor and Welfare Japan , a
grant ÔTechnology Research Grant Program in 03Õ from the New
Energy and Ind ustrial Technology D evelopment Organization
(NEDO) of Japan, and grants from the Tokyo Biochemical Researc h
Foundation, the ONO Medical Research Foundation and the Suzuken
Memorial Foundation.
References
1. Brindle, P., Linke, S. & Montminy, M. (1993) Protein-kinase-
A-dependent activator in transcription factor CREB rev eals new
role for CREM repressors. Nature 364, 821–824.
2. Quinn, P.G. (1993) Distinct activation domains within cAMP
response element-binding protein (CREB) mediate basal and
cAMP-stimulated transcription. J. Biol. Chem. 268, 16999–17009.
3. Mayr, B. & M ontminy, M. (2001) Transcriptional regulation by
the phosphorylation-dependent factor CREB. Nat. Rev. Mol. Cell
Biol. 2, 599–609.
4. Gonzalez,G.A.&Montminy,M.R.(1989)CyclicAMPstimu-
lates somatostatin gene tran scription by phosphorylation of
CREB at serine 133. Cell 59, 675–680.
5. Boshart, M., Weih, F., Nichols, M. & Schutz, G. (1991) The tis-
sue-specific extinguisher locus TSE1 encodes a regulatory subunit
of cAMP-dependent protein kinase. Cell 66, 849–859.
6. Xing, J., Korn hauser, J.M., Xia, Z., Thiele, E.A. & G reen berg,
M.E. (1998) Nerve growth factor activates extracellular signal-
regulated kinase and p38 mitogen-activated protein kinase path-
ways to stimulate CREB serine 133 phosphorylation. Mol. Cell.
Biol. 18, 1946–1955.

7. Arthur, J.S. & Cohen, P. (2000) MSK1 is required for CREB
phosphorylation in response to mitogens in mouse embryonic
stem cells. FEBS Lett. 482, 44–48.
8. Xing, J., Ginty, D.D. & Greenberg, M.E. (1996) Coupling of the
RAS-MAPKpathwaytogeneactivationbyRSK2,agrowth
factor-regulated CREB kinase. Science 273, 959–963.
9. Du, K. & Montminy, M. (1998) CREB is a regulatory target for
the protein kinase Akt/PKB. J. Biol. Chem. 273, 32377–323779.
10. Sun, P., Enslen, H., Myung, P.S.&Maurer,R.A.(1994)Differ-
ential activation of CREB by Ca
2+/
calmodulin-dependent protei n
kinases type II and type IV involves phosphorylation of a site that
negatively regulates activity. Genes Dev. 8, 2527–2539.
11. Matthews, R.P., Guthrie, C.R., Wailes, L.M., Zhao, X., Means,
A.R. & McKnight, G.S. (1994) Calciu m/calmod ulin-dep endent
protein kinase types II and IV differentially regulate CREB-
dependent gene expression. Mol. Cell. Biol. 14 , 6107–6116.
12. Shimomura, A., Ogawa, Y., Kitani, T., Fujisawa, H. & Hagiwara,
M. (1996) Calmodulin-dependent protein kinase I I potentiates
transcriptional activation through activating tran scription factor 1
but not cAMP resp onse element-b inding protei n. J. Biol. Chem.
271, 17957–17960.
13. Wu, X . & McMurray, C.T. (2001) Calmodulin kinase II
attenuation of gene transcription by preventing cAMP response
element-binding protein (CREB) dimerization and binding of the
CREB-bind ing protein . J. Biol. Chem. 276, 1735–1741.
14. Chrivia, J.C., Kwok, R.P., Lamb, N., Hagiwara, M., Montminy,
M.R. & Goodman, R.H. (1993) Phosphorylated CREB binds
specifically to the nuclear protein CBP. Nature 365, 855–885.

15. Kwok, R.P., Lundblad, J.R., Chrivia, J.C., Richards, J.P.,
Bachinger, H.P., Brennan, R.G., Roberts, S.G., Green, M.R. &
Goodman, R.H. (1994) Nuclear protein CBP is a coactivator for
the transcription factor CREB. Nature 370, 223–226.
16. Arias, J., Alberts, A.S., Brindle, P., Claret, F.X., Smeal, T., Karin,
M., Feramisco, J. & Montminy, M. (1994) Activation of cAMP
and mitogen responsive genes relies on a common nuclear factor.
Nature 370, 226–229.
17. Lundblad, J.R., Kwok, R.P., Laurance, M.E., Harter, M.L. &
Goodman, R.H. (1995) Adenoviral E1A-associated protein p300
as a f unctional homologue of th e t ranscrip tional co-activator
CBP. Nature 374, 85–88.
18. Silverman, M.A., Benard, O., Jaaro, H., Rattner, A., Citri, Y. &
Seger, R. (1999) CPG16, a novel protein serine/threonine kinase
downstream of c AMP-de pendent protein kinase. J. B iol. Chem.
274, 2631–2636.
19. Reddy, U.R., Basu, A., Bannerman,P.,Ikegaki,N.,Reddy,C.D.
& Pleasure, D. (1999) ZPK inhibits PKA ind uced transcriptional
activation by CREB and b locks retinoic acid induced neuronal
differentiation. Oncogene 18, 4474–4484.
20.Wang,Z.,Takemori,H.,Halder,S.K.,Nonaka,Y.&
Okamoto, M. (1999) Cloning of a novel kinase (SIK) of the
SNF1/AMPK family from high salt diet-treated rat adrenal.
FEBS Lett. 453, 135–139.
21. Feldman, J.D., Vician, L., Crispino, M., Hoe, W., Baudry, M.
& Herschman, H.R. (2000) T he salt-inducible kinase, SIK,
is induced by depolarization in b rain. J. Neurochem. 74 , 2227–
2238.
22. Hardie, D.G. & Carling, D. (1997) The AMP-activated protein
kinase – fuel gauge of the mammalian cell? Eur. J. Biochem. 246,

259–273.
23. Drewes, G., Ebneth, A., Preuss,U.,Mandelkow,E.M.&
Mandelkow, E. (1997) MARK, a novel family of protein kinases
that ph osphorylate microtubule-associated prote ins and triggers
microtubule disruption. Cell 89, 297–308.
24. Lanjuin, A. & Sengupta, P. (2002) Regulation of c hemo sensory
receptor expression and sensory signalingbytheKIN-29Ser/Thr
kinase. Neuron 33, 369–381.
25. da Silva Xavier, G., Leclerc, I., Varadi, A., Tsuboi, T., Moule,
S.K. & R utter, G.A. (2003) Role for AMP-activated protein
kinase in glucose-stimulated insulin secretion and preproinsulin
gene expression. Biochem. J. 371, 761–774.
26. Shaw,R.J.,Kosmatka,M.,Bardeesy,N.,Hurley,R.L.,Witters,
L.A., DePinho, R.A. & Cantley, L.C. (2004) LKB1 is a master
kinase that activates 13 k inases of th e AMPK su bfamily, including
MARK/PAR-1. EMBO J. 23, 833–843.
27. Lin, X., Takemori, H., Katoh, Y., Doi, J., Horike, N., Makino,
A., Nonaka, Y. & Okamoto, M . (2001) Salt-inducible kinase i s
involved in the ACTH/cAMP-dependent protein kinase signaling
in Y1 mouse adrenocortical t umor cells. Mol. Endocrinol. 15,
1264–1276.
28. Takemori,H.,Katoh,Y.,Horike,N.,Doi,J.&Okamoto,M.
(2002) ACTH-induced nucleocytoplasmic translocation of salt-
inducible kinase. Implication in the protein kinase A-activated
gene transcrip tion i n m ouse adrenocortical tumor cells. J. Bi ol.
Chem. 277, 42334–42343.
29. Doi, J., T akemori, H., Lin, X Z., Horike, N., Katoh, Y. &
Okamoto, M. (2002) Salt-inducible kinase represses PKA-medi-
ated activation of human cholesterol side chain cleavage cyto-
chrome promoter through the CREB basic leucine zipper do main.

J. Biol. Chem. 277, 15629–15637.
4318 Y. Katoh et al. (Eur. J. Biochem. 271) Ó FEBS 2004
30. Takemori, H., Doi, J., Horike, N., Katoh, Y., Min, L., Lin, X.Z.,
Wang, Z.N., Muraoka, M. & Okamoto, M. (2003) Salt-inducible
kinase-mediated regulation of steroidogenesis at the early stage of
ACTH-stimulation. J. Steroid Biochem. Mol. Biol. 85, 397–400.
31. Horike,N.,Takemori,H.,Katoh,Y.,Doi,J.,Min,L.,Asano,T.,
Sun, X.J., Yamamoto, H., Kasayama, S., Muraoka, M., Nonaka,
Y. & Okamoto, M. (2003) Adipose-spec ific expression, phos-
phorylation of Se r794 in insulin rece ptor substrate-1, and act iva-
tion in diabetic animals of salt-inducible kinase-2. J. Biol. Chem.
278, 18440–18447.
32. Gorlich, D. (1998) Transport into and out of the cell nucleus.
EMBO J. 17, 2721–2727.
33. Nigg, E.A. (1997) Nucleocytoplasmic transport: signals, mecha-
nisms and regulation. Nature 386, 779–787.
34. Kalderon, D., Roberts, B.L., Richardson, W.D. & Smith, A.E.
(1984) A short amino acid sequence able to specify nuclear loca-
tion. Cell 39, 499–509.
35. Tiganis, T., Flint, A.J., Adam, S.A. & Tonks, N.K. (1997) Asso-
ciation of the T-cell protein tyrosine phosphatase with nuclear
import factor p97. J. Biol. Chem. 272, 21548–21557.
36. Fischer, U., Huber, J., Boelens, W.C., Mattaj, I.W. & Luhrmann,
R. (1995) The HIV-1 Rev activation domain is a nuclear export
signal that accesses a n e xport p athway used by specific cellular
RNAs. Cell 82, 475–483.
37. Okamoto, M., Takemori, H. & Katoh, Y. (2004) Salt-inducible
kinase in stero idogenesis and adipogenesis. Trends Endocrinol.
Metab. 15, 21–26.
38. Conkright, M.D., Canettieri, G., Screaton, R., Guzman, E.,

Miraglia, L., Hogene sch, J.B. & Montminy, M. (2003) TORCs:
transducers of regulated CREB activity. Mol. Cell 12, 413–423.
39. Iourgenko, V., Zhang, W., Mickanin, C., Daly, I., Jiang, C.,
Hexham, J.M., Orth, A.P., Miraglia, L., Meltzer, J., G arza, D.,
Chirn, G.W., McWhinnie, E., Cohen, D., Ske lton, J., Terry, R.,
Yu, Y., Bodian, D ., Buxton, F.P., Zhu, J., Song, C. & L abow,
M.A. (2003) Identification of a family of cAMP response element-
binding protein coactivators b y genome-scale functional analysis
in mammalian cells. Proc. Natl Acad. Sci. USA 100, 12147–12152.
40. Screaton, R.A., Conkright, M.D., Katoh, Y., Best, J., Canettieri,
G., Jeffries, S., Guzman, E., Niessen, S., Yates, J.R. III, Takemori,
H.,Okamoto,M.&Montminy,M.R.(2004)TheCREBCoacti-
vator TORC2 Functions as a Calcium- and cAMP-Sensitive
Coincidence Detector. Cell 119, 61–74.
41. Robbins, J., Dilworth, S.M., Laskey, R.A. & Dingwall, C. (1991)
Two interdependent basic d omains in nucleoplasmin nuclear
targeting seq uenc e: iden tification o f a class o f bipar tite nuclear
targeting sequence. Cell 64, 615–623.
42. Siomi, H. & Dreyfuss, G. (1995) A nuclear localization domain in
the hnRNP A1 protein. J. Cell Biol. 129, 551–560.
43. Jakel, S. & Gorlich, D. (1998) Importin beta, transportin, RanBP5
and RanBP7 mediate nuclear import of ribosomal proteins in
mammalian cells. EMBO J. 17, 4491–4502.
44. Mo, Y.Y., Wang, C. & Beck, W.T. (2000) A novel nuclear loca-
lization sign al in human DNA topoisomerase I. J. Biol. Chem.
275, 41107–41113.
45. Jans, D.A., Xiao, C.Y. & Lam, M.H. (2000) Nu clear targeting
signal recognition: a key control point in nuclear transport?
Bioessays 22, 532–544.
46. Fukuda, M., Asano, S., Nakamura, T., Adachi, M., Yoshida, M.,

Yanagida, M. & Nishida, E. (1997) CRM1 is responsible f or
intracellular transport med iated b y the nuc lear expo rt signal.
Nature 390, 308–311.
47. Ducret, C., Maira, S.M., Dierich, A. & Wasylyk, B. (1999) The net
repressor is regulated by nuclear export in response to a nisomycin,
UV, and heat shock. Mol. Cell Biol. 19, 7076–7087.
48. Graves, P.R., Lovly, C.M., Uy, G.L. & Piwnica Worms, H.
(2001) Localization of human Cdc25C is regulated both by
nuclear export a nd 14-3-3 protein binding. Oncogene 20 , 1839–
1851.
49. Kuersten, S., Ohno, M. & Mattaj, I.W. (2001) Nucleocytoplasmic
transport: Ran, beta and beyond. Trends Cell Biol. 11, 497–503.
50. Webster, M.K., Goya, L. & Firestone, G.L. (1993) Immediate-
early transcriptional regulation and rapid mRNA turnover of a
putative serine/threonine protein kinase. J. Biol. Chem. 268,
11482–11485.
51. Buse, P., Tran, S.H., Luther, E., Phu, P.T., Aponte, G.W. &
Firestone, G.L. ( 1999) Ce ll cycle and hormonal control of
nuclear-cytoplasmic localization of the serum- and glucocorticoid-
inducible pro tein kinase, Sgk, in mammary tumor cells. A novel
convergence point of anti-proliferative a nd proliferative cell sig-
naling pathways. J. Biol. Chem. 274, 7253–7263.
52. Kim, J.E. & Chen, J. (2000) Cytoplasmic-nuclear shuttling of
FKBP12-rapamycin-associated protein is involved in rapamycin-
sensitive signaling and translation initiation. Proc. Natl Acad. Sci.
USA 97, 14340–14345.
53. Xiao, Z., Watson, N., Rodriguez, C. & Lodish, H.F. (2001)
Nucleocytoplasmi c shu ttling of S mad1 conferred by its nuclea r
localization and nuclear export signals. J. Biol. Chem. 276, 39404–
39410.

54. Xiao, Z., Latek, R. & Lodish, H.F. (2003) An extended bipartite
nuclear localization signal in Smad4 is required for its nuclear
import and transcriptional activity. Oncogene 22, 1057–1069.
55. Xiao, Z., Brownawell, A.M., Macara, I.G. & Lodish, H.F. (2003)
A novel e xport signal in Smad1 is essential for its signaling activity.
J. Biol. Chem. 278, 34245–34252.
56. Seternes, O.M., Johansen, B., Hegge, B., Johannessen, M., Keyse,
S.M. & Moens, U. (2002) Both binding and activation of p38
mitogen-activated protein kinase (MAPK) play essential ro les
in regulation of the nucleocytoplasmic distribution of MAPK-
activated protein kinase 5 b y cellular stress. Mol. Cell Biol. 22,
6931–6945.
Supplementary material
The following material is available from http://www.
blackwellpublishing.com/products/journals/suppmat/EJB/
EJB4372/EJB4372sm.htm
Fig. S1. Immunocytochemical analysis of the intracellular
distribution of HA-SIK1 and HA-SIK1 mutants (S577A,
D586–612 + NES) in resting and ACTH treated Y1 cells.
After fixation with 4% paraformaldehyde, cells were
incubated with rat anti-HA monoclonal antibody followed
by FITC-labeled goat anti-rat IgG. Cells expressing
HA-SIK1 were observed by fluorescent microscopy (green).
Nuclear staining was indicated by DAPI (blue).
Fig. S2. (A) 3T3-L1 cells were used for CRE-r eporter assay
as described in Fig. 6. (B) The modulation of CRE-reporter
activities by various kinase defective SIK1 (left) a nd SIK2
(right) were examined a s described in the legends for Fig. 5.
Means ± SD are indicated (n ¼ 3). (C) The level of kinase
activity of SIK2 in Y1 cells was almost same as that of

SIK1. Y1 cells (5 · 10
5
) were transformed with 3 lgof
pEBG-SIK1 or pEBG-SIK2 using 10 ll of Lipofect
AMINE 2000. After the 36 h-incubation, GST- tagged
SIK protein was purified using a glutathione-sepharose
column (CP). The aliquots of GST-SIK protein were
subjected to western blot analyses (WB) with anti-GST IgG
(lower panel) and in vitro kinase assays using [
32
P]ATP[cP]
and G ST-Syntide2 as a substrate (upper panel). The results
are representatives of experiments performed in duplicate.
Ó FEBS 2004 Cytoplasmic SIK1 represses CREB (Eur. J. Biochem. 271) 4319