Báo cáo khoa học: A novel splicing variant form suppresses the activity of full-length signal transducer and activator of transcription 5A pot

A novel splicing variant form suppresses the activity of
full-length signal transducer and activator of
transcription 5A
Yoshihisa Watanabe
1
, Masaya Ikegawa
2
, Yoshihisa Naruse
3
and Masaki Tanaka
1
1 Department of Cell Biology, Research Institute for Neurological Diseases and Geriatrics, Kyoto Prefectural University of Medicine, Japan
2 Department of Genomic Medical Sciences, Kyoto Prefectural University of Medicine, Japan
3 Department of Anatomy, Medical Education and Research Center, Meiji University of Integrative Medicine, Kyoto, Japan
Introduction
Signal transducers and activators of transcription
(STATs) are cytoplasmic transcriptional factors that
respond to cytokines, growth factors, and peptide hor-
mones [1,2]. In mammals, seven members of the STAT
family (STAT1–4, STAT5A, STAT5B, and STAT6)
have been identified. STATs are activated through the
phosphorylation of a tyrosine residue located between
a Src homology 2 (SH2) domain and a transactivation
domain. Phosphorylated STATs form homodimers,
heterodimers, or tetramers, and translocate into the
nucleus, where they act as transcription activators
[3–6]. In addition to the involvement of STATs in
immunological intracellular signal transduction, hema-
topoiesis, mammary gland development, and lactogen-
esis [7], some reports have demonstrated that STAT3
and STAT5 also play important roles in the central

Keywords
brainstem; coaggregation; STAT5A splicing
variant; suppression of STAT5A activity
Correspondence
M. Tanaka, Department of Cell Biology,
Research Institute for Neurological Diseases
and Geriatrics, Kyoto Prefectural University
of Medicine, Kawaramachi-Hirokoji,
Kamikyo-ku, Kyoto 602-0841, Japan
Tel ⁄ Fax: +81 75 251 5797
E-mail:
Database
The nucleotide sequence for the mouse
STAT5A_DE18 cDNA has been submitted to
the GenBank database under the accession
number EU249543
(Received 4 July 2009, revised 24 August
2009, accepted 1 September 2009)
doi:10.1111/j.1742-4658.2009.07339.x
Signal transducers and activators of transcription (STATs) regulate a vari-
ety of cellular functions, including differentiation and proliferation. STAT3
and STAT5 are known to play important roles in brain processes, such as
energy homeostasis and neuronal development. We isolated a novel splicing
variant of STAT5A from a cDNA library of the mouse brainstem. This
variant, STAT5A_DE18, lacked exon 18 and caused a frameshift in the
C-terminus, resulting in deletion of a tyrosine phosphorylation site and a
transactivation domain. Although the frameshift region had no characteris-
tic motifs, it was highly serine ⁄ threonine-rich and contained a short
proline-rich sequence. Expression of STAT5A_DE18 was detected in the
mouse brainstem, lung and thymus, but not in the mouse cerebrum or cere-

bellum. We developed a specific antibody against STAT5A_DE18 and
investigated the intracellular localization of this variant. STAT5A_DE18
showed dot-like structures in the cytoplasm and could not translocate into
the nucleus after prolactin treatment. STAT5A_DE18 showed a strong ten-
dency to aggregate, which led to coaggregation with STAT5A_full-length.
This coaggregation inhibited the nuclear transport of STAT5A and
suppressed prolactin-induced activation of STAT5A.
Abbreviations
DAPI, 4¢,6-diamidino-2-phenylindole; EGFP, enhanced green fluorescent protein; GST, glutathione S-transferase; MTT, 3-(4,5-dimethylthiazol-
2-yl)-2,5-diphenyl-tetrazolium bromide; PRL, prolactin; SH2, Src homology 2; STAT, signal transducer and activator of transcription;
STAT5A_FL, signal transducer and activator of transcription 5A_full-length.
6312 FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS
nervous system [8–10]. Morphological analyses of
STAT5 knockout mice revealed a reduction in the
number of cortical interneurons in the marginal zone,
abnormalities of corticofugal axon projection, and
defective axon guidance [8]. Furthermore, neuronal
cell-specific suppressor cytokine signaling-3 (SOCS3)
knockout mice exhibit higher leptin-induced phosphor-
ylation of hypothalamic STAT3, loss of body weight,
and suppression of food intake [9]. STAT5 was also
reported to be involved in energy homeostasis. Neu-
ron-specific STAT5A and STAT5B knockout mice
develop severe obesity with hyperphagia, impaired
thermal regulation in response to cold, hyperleptin-
emia, and insulin resistance [10]. These reports indicate
that STAT3 and STAT5 regulate food intake and
energy utilization in the brain.
We previously analyzed the functions and expression
of relaxin 3, which is involved in stress responses and

hyperphagia [11–13]. Relaxin 3 is expressed in neurons
of the nucleus incertus of the median dorsal tegmental
pons, and its expression is regulated by corticotropin-
releasing factor [11,14]. Relaxin 3-expressing neurons
project into the septum, hippocampus and feeding-
associated regions such as the lateral hypothalamus
area and arcuate nucleus [11]. Recently, we determined
the promoter region of the relaxin 3 gene, and found
that it contained a putative binding site for STATs
[14]. In an attempt to analyze this transcriptional regu-
lation, we isolated a novel splicing variant of STAT5A,
STAT5A_DE18, from a cDNA library of the mouse
brainstem. The variant protein was generated by a
frameshift in the C-terminal region, resulting in dele-
tion of the tyrosine phosphorylation site and transacti-
vation domain. Here, we report that this variant is
predominantly expressed in the brainstem and coag-
gregates with STAT5A_full length (STAT5A_FL).
Furthermore, the expression of this variant suppresses
the activity of STAT5A.
Results
Isolation of a novel STAT5A splicing variant
A cDNA encoding STAT5A was cloned from the
mouse brainstem to investigate transcriptional regula-
tion by STAT5 in the brain. During the process of
STAT5A cDNA cloning, we found a novel splicing
variant of STAT5A by nucleotide sequencing analysis.
The mouse STAT5A gene is composed of 20 exons
that encode a 793 amino acid polypeptide with a calcu-
lated molecular mass of 91 kDa (Fig. 1A, STA-

T5A_FL). The novel splicing variant, termed
STAT5A_DE18, lacked the sequence corresponding to
exon 18 of STAT5A. The deletion of exon 18 caused a
frameshift at Ala688, which led to premature termina-
tion at the codon for amino acid 798 (Fig. 1A, STA-
T5A_DE18). The C-terminal region of the variant
lacked the transactivation domain and the tyrosine res-
idue (Tyr694) phosphorylated by Janus protein tyro-
sine kinase or other tyrosine kinases, although the
DNA-binding domain and SH2 domain remained
intact. The frameshift region of STAT5A_DE18
(amino acids 688–797) had no characteristic motifs.
However, it was highly serine ⁄ threonine-rich (25.5%)
and had a short proline-rich sequence [PQMPE-
PAPP(693–701)].
RT-PCR analysis of STAT5A_DE18
To determine the expression of STAT5A_DE18 in
mouse tissues, RT-PCR analyses were conducted
using specific primers designed within exons 16 and
20 (Fig. 1A). The RT-PCR analyses were expected to
generate a 387 bp fragment for STAT5A_DE18 and
a 439 bp fragment for STAT5A_FL. The PCR prod-
uct of STAT5A_FL was detected in multiple tissues,
such as the cerebrum, kidney, and liver (Fig. 1B).
On the other hand, the PCR product of STA-
T5A_DE18 was detected in the brainstem, heart and
lung, and thymus (Fig. 1B), but not in the cerebrum
or cerebellum. Expression of the variant was also
observed in the N2a mouse neuroblastoma cell line
(Fig. 1B).

The genomic structure of STAT5A is highly con-
served in humans and mice. Human STAT5A has 20
exons, and exon 18 consists of 52 bp, similar to the
case for mouse STAT5A. Therefore, we examined
whether the splicing variant was expressed in the
human brainstem. To minimize the amplification of
STAT5B, the analysis was performed by nested
RT-PCR (Fig. 2A). Human brainstem cDNAs were
synthesized from total RNA extracts of the human
pons. The first PCR amplification was performed using
STAT5A-specific primers designed within the 5¢-UTR
and 3¢-UTR. Using the first-round PCR products as
templates, the second PCR amplification was per-
formed. The 411 bp product for STAT5A_DE18 was
detected in the human pons and mouse brainstem by
nested RT-PCR (Fig. 2A, lanes 2 and 3). Furthermore,
we reconfirmed that this variant was not expressed in
the mouse cerebrum (Fig. 2A, lane 1). In the case of
human STAT5A_DE18, deletion of exon 18 caused a
frameshift at Ala688, which led to a premature stop
codon at amino acid 690, resulting in a truncated
C-terminus. Human STAT5A_DE18 was shorter than
the STAT5Ab isoform, encoded by another splicing
Y. Watanabe et al. A novel variant form suppresses full-length STAT5A
FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS 6313
variant of STAT5A, and also lacked the transactiva-
tion domain and tyrosine residue (Fig. 2B).
Immunoblotting and immunocytochemical
analyses of STAT5A_DE18
To examine the expression and the intracellular locali-

zation of STAT5A_ DE18, we produced a polyclonal
antibody against STAT5A_DE18. A rabbit was
immunized with a glutathione S-transferase (GST)–
STAT5A_DE18_C fusion protein, and a polyclonal
antibody against STAT5A_DE18 was affinity-purified
using a thioredoxin–STAT5A_DE18_C-immobilized
column. We performed immunoblotting analyses to
examine the specificity of the antibody. The antibody
specifically detected exogenous Flag–STAT5A_DE18,
and did not cross-react with STAT5A_FL (Fig. 3).
The antibody was subsequently used for immunoblot-
ting and immunocytochemistry of STAT5A_DE18. In
order to examine its intracellular localization, Flag–
STAT5A_DE18 or Flag–STAT5A_FL was transiently
expressed in HeLa cells, and the cells were immuno-
stained with antibodies against STAT5A_DE18 or
STAT5A_FL. Flag–STAT5A_DE18 exhibited a dot-
like localization in HeLa cells (Fig. 4B), whereas Flag–
STAT5A_FL was diffusely localized to the cytoplasm
and nucleus (Fig. 4A). To confirm that this unusual
localization of STAT5A_DE18 was not due to over-
expression of Flag–STAT5A_DE18 in HeLa cells,
we generated HeLa cells stably expressing a
STAT5A_DE18–enhanced green fluorescent protein
(EGFP) fusion protein. The localization of this fusion
protein in the stably transfected cells was similar to
that observed in the transiently transfected cells. Con-
focal laser microscopy revealed that some dot-like
structures colocalized with LysoTracker Red, a lyso-
somal marker (Fig. 4C). However, the dot-like struc-

tures did not colocalize with markers for mitochondria
or the endoplasmic reticulum (data not shown). These
results indicated that STAT5A_DE18 was localized in
the cytoplasm as dot-like structures. Next, the nuclear
transport of STAT5A_DE18 was investigated using an
EGFP fusion protein. An expression vector for
STAT5A_FL–EGFP or STAT5A_DE18–EGFP was
transfected in T47D cells, which endogenously express
the prolactin (PRL) receptor [15,16]. STAT5A_FL–
1kb
ΔΔE18
793 amino acids
STAT5A_FL
Mouse STAT5A gene
Y
694
exon 18
(52 bp)
797 amino acids
STAT5A_ΔΔE18
688
α-helical
coiled-coil
DNA binding
Trans-
activation
SH2
Pro-rich
(693-701)
G3PDH

FL (439 bp)
ΔΔE18 (387 bp)
1234 56 78910111213
500
400
300
(bp)
A
B
Fig. 1. Gene structure and expression of a novel STAT5A splicing variant. (A) The mouse STAT5A gene contains 20 exons. The translation
initiation codon and stop codon are located in exons 3 and 20, respectively. STAT5A_FL encodes a 793 amino acid protein composed of a
a-helical coiled-coil domain, a DNA-binding domain, an SH2 domain, and a C-terminal transactivation domain. The deletion of exon 18 (52 bp)
in the STAT5A_DE18 mRNA results in a translational frameshift from Ala688. The STAT5A_DE18 protein lacks the Tyr694 phosphorylation
site and transactivation domain, and contains a new reading frame with a proline-rich sequence (amino acids 693–701). (B) The expression
of STAT5A_DE18 in mouse tissues was analyzed by RT-PCR. PCR was performed using primers designed in exons 16 and 20 (A, arrows).
The RT-PCR analyses were expected to generate a 439 bp fragment for STAT5A_FL and a 387 bp fragment for STAT5A_DE18 (upper panel).
RT-PCR amplification of glyceraldehyde-3-phosphate dehydrogenase (G3PDH) is indicated as an internal control (lower panel). Lane 1: cere-
brum. Lane 2: brainstem. Lane 3: cerebellum. Lane 4: heart. Lane 5: lung. Lane 6: kidney. Lane 7: liver. Lane 8: thymus. Lane 9: mammary
gland. Lane 10: adrenal gland. Lane 11: N2a cell line. Lane 12: STAT5A_FL cDNA. Lane 13: STAT5A_DE18 cDNA.
A novel variant form suppresses full-length STAT5A Y. Watanabe et al.
6314 FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS
EGFP was translocated into the nucleus after PRL
treatment (Fig. 5A,B), whereas the dot-like structures
of STAT5A_DE18–EGFP remained in the cytoplasm
despite PRL treatment (Fig. 5C,D).
Aggregate formation by STAT5A_DE18
The characteristic localization of STAT5A_DE18 sug-
gested that this protein may form aggregates. There-
fore, we investigated whether it could form insoluble
aggregates. Flag–STAT5A_DE18 was exogenously

expressed in HeLa or N2a cells, and its solubility in
the detergent Triton X-100 was examined. Transfected
cells were separated into 0.5% Triton X-100-soluble
and 0.5% Triton X-100-insoluble fractions, and the
amounts of Flag–STAT5A_DE18 in the fractions were
quantified by immunoblotting with the antibody
against STAT5A_DE18. Flag–STAT5A_DE18 was
recovered only in the insoluble fractions (Fig. 6A),
confirming that the dot-like structures were aggregates
of STAT5A_DE18. Furthermore, we examined whether
endogenous STAT5A_FL was also present in these
aggregates, as other STAT isoforms, namely STAT3b
and STAT5b, form heterodimers with their full-length
forms [17,18]. To this end, the soluble and insoluble
fractions were analyzed by immunoblotting with the
antibody against STAT5A_FL. In the STAT5A_DE18-
expressing cells, endogenous STAT5A_FL was recov-
ered not only in the soluble fractions but also in the
insoluble fractions (Fig. 6A). These findings were
confirmed by immunocytochemistry. N2a cells trans-
fected with the expression plasmids for STAT5A_
DE18–EGFP and Flag–STAT5A_FL were immuno-
stained with the antibody against STAT5A_FL.
STAT5A_DE18–EGFP formed massive aggregates in
N2a cells (Fig. 6C, arrows). In EGFP-positive cells,
coexpressed STAT5A_FL was also localized to massive
B
A
Fig. 2. Human STAT5A_DE18 variant. (A)
Expression of STAT5A_DE18 in the human

brainstem was confirmed by nested RT-
PCR. In the first PCR amplification, mouse
and human STAT5A were specifically ampli-
fied using primers designed within the
3¢-UTR and 5¢-UTR. The second PCR amplifi-
cation was expected to generate a 411 bp
fragment for STAT5A_DE18 and a 463 bp
fragment for STAT5A_FL. Lane 1: mouse
cerebrum. Lane 2: mouse brainstem.
Lane 3: human pons. Lane 4: mouse
STAT5A_FL plasmid. Lane 5: mouse STA-
T5A_DE18 plasmid. Size markers (M) are
shown on the left. (B) Comparisons of the
mouse and human amino acid sequences of
STAT5A_FL, STAT5A_DE18, and STAT5Ab.
The frameshift regions of STAT5A_DE18 are
underlined. Shaded letters, bold letters and
the open box show the proline-rich region,
tyrosine phosphorylation residues, and SH2
domains, respectively.
p
F
l
a
g
-
C
M
V
-

6
a
p
F
l
a
g
-
S
T
AT
5
A
_
F
L
p
F
l
a
g
-
S
T
A
T
5
A
_
Δ

E
1
8
FL
Δ
E18
83
83
kDa
kDa
Anti-STAT5A_FL
Anti-STAT5A_
Δ
E18
p
F
l
a
g
-
C
M
V
-
6
a
p
F
l
a

g
-
S
T
AT
5
A
_
F
L
p
F
l
a
g
-
S
T
A
T
5
A
_
Δ
E
1
8
Fig. 3. Immunoblotting analysis of STAT5A_DE18. The immunor-
eactivity of the polyclonal antibody against STAT5A_DE18 was con-
firmed by immunoblotting analysis. HeLa cells were transfected

with a control vector (pFlag–CMV-6a), pFlag–STAT5A_FL or pFlag–
STAT5A_DE18 for 48 h. Immunoblotting analyses of the cell
extracts were performed with polyclonal antibody against
STAT5A_FL (left panel) or STAT5A_DE18 (right panel).
Y. Watanabe et al. A novel variant form suppresses full-length STAT5A
FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS 6315
aggregates (Fig. 6B,D, arrows). This coaggregation
was not observed in cells expressing STAT5A_FL
alone (Fig. 6, arrowheads). These results indicate that
expression of STAT5A_DE18 leads to coaggregation
with STAT5A_FL.
STAT5A_DE18 suppresses STAT5A activity
We investigated whether the aggregation affected cell
viability or the transcriptional activity of STAT5A. The
viabilities of STAT5A_DE18-expressing cells were mea-
sured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetra-
zolium bromide (MTT) and dead cell protease-based
cytotoxicity assays. In the MTT assays, overexpression
of STAT5A_DE18 did not decrease the viability of N2a
cells as compared with the vector control, even at 48 h
after transfection (Fig. 7A). Furthermore, the dead cell
protease-based cytotoxicity assay, which is highly sensi-
tive, confirmed that there was no significant difference
in the viability of STAT5A_DE18-expressing cells
(Fig. 7B). Next, the transcriptional activity of STAT5A
was measured in T47D cells, which express
STAT5A_DE18 exogenously. A luciferase reporter gene
joined to the mouse b-casein promoter containing a
c-activated sequence was constructed to monitor the
activity of STAT5A. In vector-transfected T47D cells,

PRL stimulation resulted in a 3.3-fold increase in
reporter gene expression (Fig. 7C). On the other hand,
expression of STAT5A_DE18 reduced the activation of
PRL-stimulated STAT5A to  2.2-fold (Fig. 7C). The
expression of STAT5A_DE18 suppressed the PRL-
induced activity of STAT5A by 33%. Furthermore, we
observed nuclear translocation of STAT5A_FL by
immunocytochemistry using the antibody against
STAT5A_FL. T47D cells were transfected with
Flag–STAT5A_FL alone (Fig. 7D, upper panel) or
both Flag–STAT5A_FL and STAT5A_DE18–EGFP
(Fig. 7D, lower panel), and this was followed by incu-
bation with or without PRL for 24 h. After fixation,
the cells were immunostained with the antibody against
ΔΔE18-EGFP
Flag-FL
ΔΔ
Flag- E18
A
B
C
Fig. 4. Immunocytochemistry of STAT5A_DE18. The HeLa cells
expressing Flag–STAT5A_FL (A) or Flag–STAT5A_DE18 (B) were
analyzed by immunocytochemistry using the antibodies against
STAT5A_FL or STAT5A_DE18 (green). Nuclei were stained with
DAPI (blue). The small panels on the right represent the immunocy-
tochemical images of untransfected HeLa cells. Bar: 20 lm. (C)
HeLa cells were stably transfected with pEGFP–STAT5A_DE18. Liv-
ing cells were stained with LysoTracker Red (red), and observed
using a confocal laser microscope. STAT5A_DE18–EGFP (green) is

detected as dot-like structures and localized to the cytoplasm. A
few dots of STAT5A_DE18–EGFP are colocalized with the lyso-
some marker (arrows). Bar: 20 lm.
A novel variant form suppresses full-length STAT5A Y. Watanabe et al.
6316 FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS
STAT5A_FL and observed by fluorescence micros-
copy. When cells were transfected with Flag–STA-
T5A_FL alone, Flag–STAT5A_FL predominantly
translocated into the nucleus after PRL treatment
(Fig. 7D, upper right panel). In cotransfected cells, the
PRL-induced translocation of Flag–STAT5A_FL into
the nucleus was inhibited by its coaggregation with
STAT5A_DE18–EGFP (Fig. 7D, lower right panel).
These results are consistent with the hypothesis that
aggregation of STAT5A_DE18 suppresses the tran-
scriptional activity of STAT5A.
Discussion
We isolated a novel STAT5A splicing variant from
the mouse brainstem. The STAT5A_DE18 variant
lacked the transactivation domain and a tyrosine
residue. Many STAT isoforms have previously been
reported to be generated by alternative splicing and
proteolytic processing [17,19]. STAT1, STAT3,
STAT4, STAT5A and STAT5B mRNAs are alterna-
tively spliced at the 3¢-end, resulting in the production
of b-isoforms truncated at the transactivation domain.
STAT5A b-isoforms and STAT5B b-isoforms are
generated by insertion of intron 18, and lack only
the transactivation domain [18,20]. These STAT
b-isoforms are phosphorylated on the tyrosine residue

after stimulation by cytokines or hormones, and
translocate into the nucleus, but fail to activate
transcription [18,21]. Unlike the b-isoforms, the
STAT5A_DE18 variant was not phosphorylated,
because it lacked the tyrosine residue. Moreover,
STAT5A_DE18 did not translocate into the nucleus
in T47D cells after PRL treatment, indicating a dis-
tinct property of STAT5A_DE18 as compared with
the b-isoforms. However, STAT5A_DE18 clearly
suppressed the activity of STAT5A, similar to the
case for the b-isoforms. It has been reported that the
phosphorylated b-isoforms form heterodimers with
full-length STATs and decrease their activities [18].
These heterodimers can translocate into the nucleus
and bind to target sequences on DNA, but fail to
activate transcription [18]. On the other hand, we
demonstrated that the STAT5A_DE18-mediated
suppression was caused by coaggregation of STA-
T5A_FL and STAT5A_DE18 in cultured cells,
although the precise mechanism for the coaggregation
remains to be determined. It is known that unphos-
phorylated STAT5A monomers can dimerize via
interactions between their b-barrel (amino acids 332–
470) and four-helix bundle (amino acids 138–331)
domains [22]. STAT5A_DE18 also contains these
domains, suggesting that heterodimers of unphosphor-
ylated STAT5A_FL and STAT5A_DE18 are probably
formed in the soluble condition prior to their coag-
gregation. This coaggregation could be the cause of
the decrease in functional STAT5A, resulting in

suppressed transcription of its downstream genes.
STAT5Ab is expressed in early myeloid lineages
[23], whereas the STAT5A_DE18 variant was expressed
in the mouse brainstem, thymus, and lung. Moreover,
exon 5 of STAT5A is alternatively spliced by heteroge-
neous ribonucleoprotein L-like at the transition from
naı
¨
ve to activated T-cells [24]. STAT5A transcripts
variously undergo tissue-specific or cell type-specific
alternative splicing, suggesting that these variants are
involved in specific functions. The physiological role of
STAT5A_DE18 may involve the regulation of STAT5
function in the brainstem, considering that STA-
T5A_DE18 suppresses the activity of STAT5. The
functions of STAT5 in the central nervous system have
recently been reported. Intracerebroventricular admin-
istration of granulocyte–macrophage colony-stimulat-
ing factor and leptin activated neuronal STAT5 and
reduced food intake [10,25]. The activation of STAT5
following leptin administration was observed not only
in the hypothalamus but also in areas of the brain-
stem, such as the raphe obscurus, raphe pallidus,
dorsal motor nucleus of the vagus, and solitary tract
*
*
*
*
EGFP
-PRL

+PRL
FL-EGFP
E18-EGFP
*
*
*
*
*
*
AB
CD
EF
Fig. 5. Nuclear transport analysis of STAT5A_DE18. (A–D) STA-
T5A_FL–EGFP (A, B), STAT5A_DE18–EGFP (C, D), or EGFP (E, F)
were transiently expressed in T47D cells. Cells were cultured in
serum-free medium with (+) or without ()) PRL (10 ngÆmL
)1
) for
5 h, and observed under a fluorescence microscope. Asterisks
indicate nuclei. Bar: 20 lm.
Y. Watanabe et al. A novel variant form suppresses full-length STAT5A
FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS 6317
nucleus [25]. Furthermore, neuron-specific STAT5A
and STAT5B knockout mice develop severe obesity
with hyperphagia, impaired thermal regulation in
response to cold, hyperleptinemia, and insulin resis-
tance [10]. These STAT5-mediated functions in the
central nervous system may be controlled by STA-
T5A_DE18. From the results of RT-PCR and immu-
noblotting analyses, however, the expression level of

STAT5A_DE18 seemed to be very low in the normal
brain. Furthermore, this variant protein tended to
form aggregates even in stably transfected cells,
suggesting that high-level expression of this variant
might lead to pathological conditions. The formations
of aggregates and inclusion bodies in the brain are
pathognomonic features of many neurodegenerative
diseases, such as Alzheimer’s disease, Parkinson’s dis-
ease, and Huntington’s disease [26,27]. Recent studies
have revealed that mutant huntingtin aggregates inter-
act with several transcription factors, such as CREB-
binding protein, TATA-binding protein, and NF-Y,
resulting in reduced expression of their target genes
[28,29]. This inhibition of functional transcription fac-
tors may be associated with the normal functions of
huntingtin and ⁄ or involved in the pathology of Hun-
tington’s disease. We observed that the transcriptional
activation of STAT5A_FL was suppressed by STA-
T5A_DE18 aggregates. Moreover, it has been reported
that defective mutations of STAT5B are involved in
A
BCD
EFG
Fig. 6. Aggregate formation of STAT5A_DE18 and STAT5A_FL. (A) pFlag–STAT5A_DE18 or control vector was transfected into N2a (left pan-
els) or HeLa (right panels) cells. The Triton X-100-soluble (S) and Triton X-100-insoluble (P) fractions were analyzed by immunoblotting with
antibody against STAT5A_DE18 (upper panels) or antibody against STAT5A_FL (lower panels). N2a cells were transfected with Flag–STA-
T5A_FL and STAT5A_DE18–EGFP (B–D) or Flag–STAT5A_FL alone (E–G). Flag–STAT5A_FL was detected with a primary antibody against
STAT5A_FL and Alexa Fluor 546-conjugated secondary antibody (B, E). The localization of STAT5A_DE18–EGFP was observed under a fluo-
rescence microscope (C, F). The merged image containing DAPI staining (blue) is shown in (D) and (G). Colocalization of the signals appears
yellow. Bar: 20 lm.

A novel variant form suppresses full-length STAT5A Y. Watanabe et al.
6318 FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS
the syndrome of growth hormone insensitivity [30,31].
Considering the specific expression of STAT5A_DE18
in the brainstem and its suppression of STAT5 activ-
ity, accumulation of STAT5A_DE18 may be pathogen-
ically involved in certain neurological disorders. This
possibility will be investigated in future studies.
AB
CD
Fig. 7. Viability and reporter analysis. (A, B) N2a or HeLa cells were transfected with pFlag–STAT5A_FL, pFlag–STAT5A_DE18 or control
vector (pFlag–CMV-6a) in serum-containing medium. After transfection, the cell viability was assessed by MTT (A) and cytotoxicity (B) assays.
As a positive control, the viability of rotenone-treated cells was also measured. (C) T47D cells were transfected with pFlag–STAT5A_FL,
pFlag–STAT5A_DE18 or control vector (pFlag–CMV-6a), as well as with pCasein-luc and pSV40-Rluc. The cells were treated with PRL
(10 ngÆmL
)1
), and this was followed by measurement of the luciferase activities. These results are shown as the means ± standard deviation
of three experiments. *P < 0.05 versus control cells without PRL stimulation. Statistical analyses were performed using ANOVA with Tukey’s
HSD post hoc test. (D) T47D cells were transfected with pFlag–STAT5A_FL alone (upper panels) or together with pEGFP–STAT5A_DE18
(lower panels). The cells were then incubated with (+) or without ()) PRL (10 ngÆmL
)1
) for 24 h, and this was followed by immunocytochemi-
cal analysis using a primary antibody against STAT5A_FL and Alexa Fluor 546-conjugated secondary antibody. Nuclei are indicated by
asterisks. DE18–EGFP–positive and DE18–EGFP-negative cells are indicated by arrows and arrowheads, respectively. Bar: 20 lm.
Y. Watanabe et al. A novel variant form suppresses full-length STAT5A
FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS 6319
Experimental procedures
Cell culture and transfection
N2a and HeLa cells were routinely maintained in Ham’s
F12 medium containing 10% fetal bovine serum. The

human breast cancer T47D cell line was cultured in RPMI-
1640 medium containing 10% fetal bovine serum [32]. To
generate a stable cell line, transfected HeLa cells were
passaged into medium containing G418 (400 lgÆmL
)1
) for
10 days. A stable cell line expressing STAT5A_DE18–
EGFP was isolated from the pool of cells by limiting
dilution cloning, and maintained in the same medium.
Plasmids
Mouse STAT5A_FL and STAT5A_DE18 cDNAs were
cloned into the pGEM-T easy plasmid (Promega, Madison,
WI, USA) by PCR, using the primers 5¢-CCGTCAGGA
GCCGTCAGAAG-3¢ and 5¢-CCTGGCGCAAGAACTGA
CAC-3¢. For these amplifications, cDNA libraries from the
mouse brainstem and breast were prepared as previously
described [33]. Briefly, total RNA was extracted from
C57BL ⁄ 6N mouse tissues using TRIzol (Invitrogen, Carls-
bad, CA, USA), and cleaned up with an RNeasy Micro Kit
(Qiagen, Hilden, Germany). cDNAs were synthesized with
a ThermoScript RT-PCR System (Invitrogen), using an oli-
go-dT primer. Flag-tagged STA5A_FL and STAT5A_DE18
constructs were generated by PCR using the primers 5¢ -GA
ATTCTATGGCGGGCTGGATTCAG-3¢ and 5 ¢-GTCG
ACCTACAACTGACGTGGGC-3¢. The PCR fragments
were cloned into pGEM-T easy, and this was followed by
subcloning into the EcoRI–SalI site of pFlag–CMV6a
(Sigma-Aldrich, St Louis, MO, USA). STAT5A_DE18–
EGFP was constructed by PCR using the primers 5¢-GA
ATTCGCCACCATGGCGGGCTGGATTC-3¢ and 5¢-CC

CGGGCCAACTGACGTGGGCTCC-3¢, and the resulting
PCR product containing a Kozak sequence was cloned into
pGEM-T easy. The EcoRI–SmaI fragment of this plasmid
was subcloned into the EcoRI–SmaI site of pEGFP-N1
(TaKaRa Bio, Otsu, Japan). For construction of a reporter
plasmid (pCasein-luc), the mouse b-casein promoter was
inserted into the firefly luciferase reporter gene by PCR
using the primers 5¢-CTTCATAACTGAGGTTAAAGC
C-3¢ and 5¢-GTCCTATCAGACTCTGTGAC-3¢.
PCR analysis
To analyze the expression of mouse STAT5A_DE18,we
designed the specific primers 5¢-CTGCGCTTCAGT
GACTCGGA-3¢ and 5¢-CGTGCCTGGCAACATCCAT
G-3¢, located within exons 16 and 20, respectively. Further-
more, we confirmed the expression of mouse and human
STAT5A_DE18 by nested RT-PCR analysis. As a first step,
STAT5A containing the 5¢-UTR and 3¢-UTR was specifi-
cally amplified from cDNA libraries of the human pons
(TaKaRa Bio) and mouse brainstem, using the primers 5¢
-CTGCTCTCCGCTCCTTCCTG-3¢⁄5¢-CAGAGAGTCTG
GAGTCCACG-3¢ and 5¢-CCGTCAGGAGCCGTCAGAA
G-3¢⁄5¢-GACGTGGGCTCCTCACACTG-3¢, respectively.
The PCR amplification was performed with 35 cycles of
95 °C for 15 s and 60 °C for 180 s. Using the resulting
PCR products as templates, we examined the existence of
exon 18. The primers 5¢-GACCTGCTCATCAACAAGCC
-3¢ and 5¢-CATCCATGGTCTCATCCAGG-3¢ were used
for a second round of PCR amplification. The second
round of PCR amplification was performed with 35 cycles
of 95 °C for 15 s and 60 °C for 45 s. All PCR amplifica-

tions were performed using Z-Taq DNA polymerase
(TaKaRa Bio).
Production of an antibody against STAT5A_DE18
To raise mouse STAT5A_DE18-specific antisera, we used
the C-terminal region of STAT5A_DE18 (amino acids 688–
797), which was not found in STAT5A_FL. A cDNA
encoding this unique region, STAT5A_DE18_C, was
inserted into the EcoRI–XhoI site of pGEX-6P-1 (GE
Healthcare, Little Chalfont, UK) or the EcoRI–SalI site of
pThioHisA (Invitrogen), using the primers 5¢-GGAT
CCGGTTCGTCAATGCATCC-3¢ and 5¢-CTCGAGCTAC
AACTGACGTGGGCTCCTCAC-3¢. Expression of the
GST–STAT5A_DE18_C and thioredoxin–STAT5A_DE18_C
fusion proteins was induced in Escherichia coli by treatment
with 1 mm isopropyl b-d-1-thiogalactopyranoside for 4 h at
37 °C. Inclusion bodies containing these proteins were
recovered by centrifugation (20 000 g for 20 min at 4 °C),
and washed with NaCl ⁄ P
i
containing 0.5% Triton X-100
and 1 m m phenylmethanesulfonyl fluoride. The purified
inclusion bodies were separated by SDS ⁄ PAGE, using a
12.5% gel, and the GST–STAT5A_DE18_C in the poly-
acrylamide gel was emulsified with Freund’s Complete
Adjuvant (Difco Laboratories, Detroit, MI, USA). To pro-
duce an antiserum, a Kbl:NZW rabbit was immunized with
this emulsion (Kitayama Labes, Ina, Japan). At 8 weeks
after the immunization, the antiserum was recovered and
subjected to a titration assay. For affinity purification of
the antibody against STAT5A_DE18, inclusion bodies of

thioredoxin–STAT5A_DE18_C were solubilized in 50 mm
phosphate buffer (pH 8.5) containing 8 m urea and 1 mm
phenylmethanesulfonyl fluoride, and rapidly refolded by
10-fold dilution in 50 mm phosphate buffer (pH 8.5). The
soluble thioredoxin–STAT5A_DE18_C was purified with
HiTrap Q HP (GE Healthcare) and immobilized on HiTrap
NHS-activated HP (GE Healthcare). The antiserum was
loaded onto this affinity column, and this was followed by
washing with 1 m NaCl and 1% Triton X-100. The
antibody against STAT5A_DE18 was eluted with 100 mm
A novel variant form suppresses full-length STAT5A Y. Watanabe et al.
6320 FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS
glycine-HCl (pH 2.8), and this was followed by rapid neu-
tralization. The specificity of the purified antibody was con-
firmed by immunoblotting analysis.
Immunoblotting
Cells were washed with NaCl ⁄ P
i
and extracted with ice-cold
NaCl ⁄ P
i
containing 0.5% Triton X-100 and a protease
inhibitor cocktail (Nacalai Tesque, Kyoto, Japan). After
sonication, each sample was fractionated by centrifugation
(20 000 g, 15 min, 4 °C), and the supernatant was recov-
ered as a soluble fraction. The precipitate was washed with
ice-cold NaCl ⁄ P
i
and recovered as an insoluble fraction.
These fractions were extracted in Laemmli buffer and sepa-

rated by SDS ⁄ PAGE, using a 12.5% gel; this was followed
by electroblotting onto poly(vinylidene difluoride) mem-
branes (Millipore, Billerica, MA, USA). After blocking, the
membranes were incubated with a polyclonal antibody
against STAT5A_FL (1 : 1000 dilution; Santa Cruz Bio-
technology, Santa Cruz, CA, USA) or the polyclonal anti-
body against STAT5A_DE18 (1 : 500 dilution) for 12 h at
25 °C. The membranes were then incubated with alkaline
phosphatase-conjugated anti-rabbit IgG (1 : 5000 dilution;
Millipore) for 1 h. Immunopositive signals were detected
with the nitroblue tetrazolium chloride and 5-bromo-4-
chloro-3¢-indolylphosphatase p-toluidine salt reagents.
Immunocytochemistry
N2a and HeLa cells (1 · 10
5
) were plated on round cover
glasses (13 mm in diameter), and transfected with pFlag–
STAT5A_FL and pSTAT5A_DE18–EGFP, using Lipofec-
tamine LTX (Invitrogen). Cells were cultured for 2 days
after transfection, and fixed with 4% paraformaldehyde in
0.1 m phosphate buffer. After being washed with NaCl ⁄ P
i
,
they were incubated with the antibody against STAT5A_
FL or antibody against STAT5A_DE18 (1 : 100 dilution) in
NaCl ⁄ P
i
containing 0.1% Triton X-100 overnight at 4 °C.
After three washes with NaCl ⁄ P
i

, the cells were incubated
with fluorescein isothiocyanate-conjugated or Alexa Fluor
546-conjugated anti-rabbit IgG (1 : 500 dilution; Invitro-
gen) for 4 h at room temperature. After washing and stain-
ing with 4¢,6-diamidino-2-phenylindole (DAPI) (Dojindo,
Kumamoto, Japan), the cells on the cover glasses were
mounted on glass slides in an aqueous mounting medium
(GEL ⁄ MOUNT; Biomeda, Foster City, CA, USA) and
examined under a fluorescence microscope (Olympus,
Tokyo, Japan). HeLa cells stably expressing STA-
T5A_DE18–EGFP were cultured in glass-bottomed culture
dishes (Iwaki, Tokyo, Japan), and organelles were stained
with MitoTracker Orange CM-H
2
TMRos, ER-Tracker
Red, and LysoTracker Red DND-99 (Invitrogen). Confocal
laser microscopy was performed on live cells using a multi-
track analysis (LSM 510 META; Carl Zeiss, Oberkochen,
Germany). EGFP was excited using a 488 nm argon laser,
and emission was recorded through a BP 500–530 nm filter.
Red-emitting dyes were excited with a 543 nm helium–neon
laser, and emission was recorded through an LP 560 nm
filter.
MTT and cytotoxicity assays
For MTT and cytotoxicity assays, 1 · 10
4
cells were cul-
tured in 96-well plates and transfected with control vector,
pFlag–STAT5A_FL or pFlag–STAT5A_DE18, using Lipo-
fectamine LTX. MTT assays were performed using a Cell

Counting Kit-8 (Dojindo) according to the manufacturer’s
recommendations. The absorbances at 429 or 600 nm were
measured using a Multiskan Plus (Thermo Fisher Scientific,
Waltham, MA, USA). Cytotoxicity assays were performed
using a CytoTox-Glo Cytotoxicity Assay Kit (Promega).
Luminescence was measured using a GENios (Tecan,
Ma
¨
nnedorf, Switzerland). All experiments were repeated
three times.
Reporter assay
T47D cells were transfected with 0.5 lg of pFlag–STA-
T5A_FL, pFlag–STAT5A_DE18, or vector. Simultaneously,
0.5 lg of a firefly luciferase reporter plasmid with the
mouse b-casein promoter (pCasein-luc) and 0.25 lgofa
Renilla luciferase control plasmid (pSV40-Rluc) were
cotransfected using the Lipofectamine LTX and PLUS
reagents (Invitrogen). At 24 h after transfection, the cells
were treated with 10 ngÆmL
)1
recombinant human PRL
(Cedarlane Laboratories, Burlington, Canada) in serum-free
medium for 24 h. The cells were lysed with Passive Lysis
Buffer (Promega), and the luciferase activities were
measured using a Dual-Luciferase Reporter Assay System
(Promega) and a MicroLumat LB96P Luminometer
(Berthold Technologies, Bad Wildbad, Germany). All
experiments were repeated three times.
Acknowledgements
We greatly appreciate the gift of the T47D cell line

from Dr J. Kitawaki (Department of Obstetrics and
Gynecology, Kyoto Prefectural University of Medi-
cine). This work was supported in part by a grant
from the Ministry of Education, Culture, Sports,
Science and Technology, Japan (No. 18500268 to M.
Tanaka) and a grant from the Research Institute for
Neurological Diseases and Geriatrics, Kyoto Prefec-
tural University of Medicine (to M. Tanaka).
References
1 Darnell JE Jr, Kerr IM & Stark GR (1994) Jak-STAT
pathways and transcriptional activation in response to
Y. Watanabe et al. A novel variant form suppresses full-length STAT5A
FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS 6321
IFNs and other extracellular signaling proteins. Science
264, 1415–1421.
2 Levy DE & Darnell JE Jr (2002) Stats: transcriptional
control and biological impact. Nat Rev Mol Cell Biol 3,
651–662.
3 Ihle JN (1996) STATs: signal transducers and activators
of transcription. Cell 84, 331–334.
4 Hoey T & Schindler U (1998) STAT structure and func-
tion in signaling. Curr Opin Genet Dev 8, 582–587.
5 John S, Vinkemeier U, Soldaini E, Darnell JE Jr &
Leonard WJ (1999) The significance of tetramerization
in promoter recruitment by Stat5. Mol Cell Biol 19,
1910–1918.
6 Brockman JL, Schroeder MD & Schuler LA (2002)
PRL activates the cyclin D1 promoter via the Jak2 ⁄ Stat
pathway. Mol Endocrinol 16, 774–784.
7 O’Shea JJ (1997) Jaks, STATs, cytokine signal trans-

duction, and immunoregulation: are we there yet?
Immunity 7, 1–11.
8 Markham K, Schuurmans C & Weiss S (2007) STA-
T5A ⁄ B activity is required in the developing forebrain
and spinal cord. Mol Cell Neurosci 35, 272–282.
9 Mori H, Hanada R, Hanada T, Aki D, Mashima R,
Nishinakamura H, Torisu T, Chien KR, Yasukawa H
& Yoshimura A (2004) Socs3 deficiency in the brain
elevates leptin sensitivity and confers resistance to
diet-induced obesity. Nat Med 10, 739–743.
10 Lee JY, Muenzberg H, Gavrilova O, Reed JA, Berry-
man D, Villanueva EC, Louis GW, Leinninger GM,
Bertuzzi S, Seeley RJ et al. (2008) Loss of cytokine–
STAT5 signaling in the CNS and pituitary gland alters
energy balance and leads to obesity. PLoS ONE 3,
e1639.
11 Tanaka M, Iijima N, Miyamoto Y, Fukusumi S, Itoh
Y, Ozawa H & Ibata Y (2005) Neurons expressing
relaxin 3 ⁄ INSL 7 in the nucleus incertus respond to
stress. Eur J Neurosci 21, 1659–1670.
12 McGowan BM, Stanley SA, Smith KL, White NE,
Connolly MM, Thompson EL, Gardiner JV, Murphy
KG, Ghatei MA & Bloom SR (2005) Central relaxin-3
administration causes hyperphagia in male Wistar rats.
Endocrinology 146, 3295–3300.
13 Miyamoto Y, Watanabe Y & Tanaka M (2008) Devel-
opmental expression and serotonergic regulation of
relaxin 3 ⁄ INSL7 in the nucleus incertus of rat brain.
Regul Pept 145, 54–59.
14 Tanaka M, Watanabe Y & Yoshimoto K (2009)

Regulation of relaxin 3 gene expression via cAMP–
PKA in a neuroblastoma cell line. J Neurosci Res 87,
820–829.
15 Kline JB, Roehrs H & Clevenger CV (1999) Functional
characterization of the intermediate isoform of the
human prolactin receptor. J Biol Chem 274, 35461–
35468.
16 Fang F, Antico G, Zheng J & Clevenger CV (2008)
Quantification of PRL ⁄ Stat5 signaling with a novel
pGL4-CISH reporter. BMC Biotechnol 8
, 11.
17 Lim CP & Cao X (2006) Structure, function, and regu-
lation of STAT proteins. Mol Biosyst 2, 536–550.
18 Wang D, Stravopodis D, Teglund S, Kitazawa J &
Ihle JN (1996) Naturally occurring dominant negative
variants of Stat5. Mol Cell Biol 16, 6141–6148.
19 Meyer J, Ju
¨
cker M, Ostertag W & Stocking C (1998)
Carboxyl-truncated STAT5b is generated by a nucleus-
associated serine protease in early hematopoietic
progenitors. Blood 91, 1901–1908.
20 Yamashita H, Iwase H, Toyama T & Fujii Y (2003)
Naturally occurring dominant-negative Stat5 suppresses
transcriptional activity of estrogen receptors and
induces apoptosis in T47D breast cancer cells. Oncogene
22, 1638–1652.
21 Caldenhoven E, van Dijk TB, Solari R, Armstrong J,
Raaijmakers JA, Lammers JW, Koenderman L & de
Groot RP (1996) STAT3b, a splice variant of transcrip-

tion factor STAT3, is a dominant negative regulator of
transcription. J Biol Chem 271, 13221–13227.
22 Neculai D, Neculai AM, Verrier S, Straub K, Klumpp
K, Pfitzner E & Becker S (2005) Structure of the
unphosphorylated STAT5a dimer. J Biol Chem 280,
40782–40787.
23 Azam M, Erdjument-Bromage H, Kreider BL, Xia M,
Quelle F, Basu R, Saris C, Tempst P, Ihle JN &
Schindler C (1995) lnterleukin-3 signals through
multiple isoforms of Stat5. EMBO J 14, 1402–1411.
24 Oberdoerffer S, Moita LF, Neems D, Freitas RP,
Hacohen N & Rao A (2008) Regulation of CD45
alternative splicing by heterogeneous ribonucleoprotein,
hnRNPLL. Science 321, 686–691.
25 Mu
¨
tze J, Roth J, Gerstberger R & Hu
¨
bschle T (2007)
Nuclear translocation of the transcription factor STAT5
in the rat brain after systemic leptin administration.
Neurosci Lett 417, 286–291.
26 Lansbury PT & Lashuel HA (2006) A century-old
debate on protein aggregation and neurodegeneration
enters the clinic. Nature 443, 774–779.
27 Kopito RR (2000) Aggresomes, inclusion bodies and
protein aggregation. Trends Cell Biol 10, 524–530.
28 Li SH & Li XJ (2004) Huntingtin–protein interactions
and the pathogenesis of Huntington’s disease. Trends
Genet 20, 146–154.

29 Yamanaka T, Miyazaki H, Oyama F, Kurosawa M,
Washizu C, Doi H & Nukina N (2008) Mutant
Huntingtin reduces HSP70 expression through the
sequestration of NF-Y transcription factor. EMBO
J 27, 827–839.
30 Kofoed EM, Hwa V, Little B, Woods KA, Buckway
CK, Tsubaki J, Pratt KL, Bezrodnik L, Jasper H,
Tepper A et al. (2003) Growth hormone insensitivity
A novel variant form suppresses full-length STAT5A Y. Watanabe et al.
6322 FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS
associated with a STAT5b mutation. New Engl J Med
349, 1139–1147.
31 Hwa V, Little B, Adiyaman P, Kofoed EM, Pratt KL,
Ocal G, Berberoglu M & Rosenfeld RG (2005) Severe
growth hormone insensitivity resulting from total
absence of signal transducer and activator of transcrip-
tion 5b. J Clin Endocrinol Metab 90, 4260–4266.
32 Xu B, Kitawaki J, Koshiba H, Ishihara H, Kiyomizu
M, Teramoto M, Kitaoka Y & Honjo H (2007)
Differential effects of progestogens, by type and regi-
men, on estrogen-metabolizing enzymes in human
breast cancer cells. Maturitas 56, 142–152.
33 Watanabe Y, Okui A, Mitsui S, Kawarabuki K,
Yamaguchi T, Uemura H & Yamaguchi N (2004)
Molecular cloning and tissue-specific expression
analysis of mouse spinesin, a type II transmembrane
serine protease 5. Biochem Biophys Res Commun 324,
333–340.
Y. Watanabe et al. A novel variant form suppresses full-length STAT5A
FEBS Journal 276 (2009) 6312–6323 ª 2009 The Authors Journal compilation ª 2009 FEBS 6323