Interaction between fortilin and transforming growth factor-beta stimulated clone-22 (TSC-22)...

Aotg

*

E

1

FEBS Letters 582 (2008) 1210–1218

Interaction between fortilin and transforming growthfactor-beta stimulated clone-22 (TSC-22) prevents apoptosis via

the destabilization of TSC-22

Jeong Heon Leeb,1, Seung Bae Rhoc,1, Sang-Yoon Parkc, Taehoon Chuna,*

a Division of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul 136-701, Republic of Koreab Department of Obstetrics and Gynecology, Chonbuk National University Medical School, Jeonju 561-712, Republic of Korea

c Research Institute, National Cancer Center, Goyang, Gyeonggi 411-769, Republic of Korea

Received 20 November 2007; revised 10 January 2008; accepted 18 January 2008

Available online 4 March 2008

Edited by Angel Nebreda

Abstract Yeast two-hybrid screening was conducted using a hu-man ovary cDNA library to search for a novel binding proteinusing transforming growth factor-beta stimulated clone-22(TSC-22). The selected protein was fortilin, which has been char-acterized as a nuclear anti-apoptotic protein. Overexpression offortilin in ovarian carcinoma cells reversed TSC-22-mediatedapoptosis, and the inhibition of fortilin expression via small inter-fering RNA (siRNA) resulted in an increase in the apoptosis ofovarian carcinoma cells. Moreover, fortilin overexpression pro-moted the degradation of TSC-22. Thus, an interaction betweenfortilin and TSC-22 prevents apoptosis via the destabilization ofTSC-22 in ovarian carcinoma cells.Structured summary:

MINT-6173230, MINT-6173253:

TSC22 (uniprotkb:Q15714) physically interacts (MI:0218)

with fortilin (uniprotkb:P13693) by co-immunoprecipitation

(MI:0019)

MINT-6173217:

TSC22 (uniprotkb:Q15714) binds (MI:0407) fortilin (uni-

protkb:P13693) by pull-down (MI:0096)

MINT-6173240, MINT-6173270:

TSC22 (uniprotkb:Q15714) physically interacts (MI:0218)

with fortilin (uniprotkb:P13693) by two-hybrid (MI:0018)

� 2008 Federation of European Biochemical Societies. Pub-lished by Elsevier B.V. All rights reserved.

Keywords: Apoptosis; Fortilin; Ovarian carcinoma cell;TSC-22; Yeast two hybrid

1. Introduction

TGF-b stimulated clone 22 (TSC-22) was initially identified

as one of the genes induced by TGF-b1 treatment [1,2]. TSC-22

is evolutionarily conserved to a significant degree from fruit

bbreviations: DAPI, 4,60-diamidino-2-phenylindole; GST, glutathine S-transferase; ONPG, o-nitrophenyl b-DD-galactopyranoside; siforilin, siRNA, small interfering RNA; THG-1, TSC-22 homologousene-1; TSC-22, transforming growth factor-beta stimulated clone-22

Corresponding author. Fax: +82 2 3290 3507.-mail address: [email protected] (T. Chun).

Co-first authors.

0014-5793/$34.00 � 2008 Federation of European Biochemical Soc

doi:10.1016/j.febslet.2008.01.066

--

ieties. Pu

flies to mammals [3,4]. The structure of TSC-22 includes a leu-

cine zipper-like motif, thereby suggesting that TSC-22 func-

tions as a transcriptional factor, but lacks the conventional

DNA binding motif at the N-terminal region [1]. Recently, sev-

eral roles of TSC-22 have been proposed, including the regula-

tion of embryogenesis [3,5], and the pro-apoptotic factor in

several cancer cells [6–9].

TSC-22 exists within the cytoplasm in the steady state and

translocates to the nucleus, and there activates signal transduc-

tion to TSC-22 [10]. After nuclear translocation, TSC-22 may

form a homodimer or heterodimer with TSC-22 homologous

gene-1 (THG-1), and operate as a transcriptional repressor

of several genes [11]. This result suggests that TSC-22 may

mediate apoptosis via a caspase-3 dependent pathway [12].

However, the precise mechanism underlying the regulation of

TSC-22 activity has yet to be precisely elucidated.

In order to characterize the TSC-22-dependent apoptotic

pathway, we first employed a yeast two-hybrid system to

screen a human ovary cDNA library for novel TSC-22 binding

proteins. We identified fortilin, a recently characterized nuclear

anti-apoptotic factor. The results of transient transfection

analyses then demonstrated that fortilin overexpression blocks

TSC-22-mediated apoptotic activity via the induction of TSC-

22 degradation. Therefore, these results support the notion

that a dynamic equilibrium existing between pools of TSC-22

and fortilin regulates an apoptotic homeostasis.

2. Materials and methods

2.1. Yeast two-hybrid screeningFor bait construction with human TSC-22, cDNA encoding for full-

length human TSC-22 was subcloned into the EcoRI and XhoI restric-tion sites of the pGilda cloning vector. The resultant plasmid, pGilda-TSC-22, was then introduced into the EGY48 yeast strain [MATa,his3, trp1, ura3-52, leu2::pLeu2-LexAop6/pSH18-34 (LexAop-lacZ re-porter)] via a modified lithium acetate technique [13]. The ovary cDNAlibrary was then constructed by cloning the cDNA fragments into theEcoRI and XhoI restriction sites of pJG4-5, thus generating B42 fusionproteins (Clontech, Palo Alto, CA, USA). The cDNAs encoding forthe B42 fusion proteins were introduced into competent yeast cells thatalready harbored pGilda-TSC-22, and the transformants were selectedfor tryptophan prototrophy (plasmid marker) on synthetic medium(Ura, His, Trp) containing 2% (w/v) glucose.

2.2. Quantitation of interactionThe activity of the interaction between TSC-22 and fortilin was ver-

ified via measurements of the relative levels of b-galactosidase expres-

blished by Elsevier B.V. All rights reserved.

J.H. Lee et al. / FEBS Letters 582 (2008) 1210–1218 1211

sion. The b-galactosidase assay was conducted in accordance with thepreviously described protocols [14], with slight modification. In brief,yeast cells (EGY48 strain) containing each construct were cultured inyeast synthetic media (Ura, His, Trp) with 2% (w/v) glucose until theyachieved mid-growth phase. The cells were then transferred to a yeastmedium (Ura, His, Trp) containing 2% (w/v) galactose and 0.2% di-methyl sulfoxide (Me2SO). After transformation, equivalent numbersof cells were lysed in 0.7 ml of Z buffer (60 mM Na2HPO4, 40 mMNaH2PO4, 10 mM KCl, 1 mM MgSO4, and 50 mM b-mercap-toethanol, pH 7.0) containing 50 ll of 0.1% SDS and 50 ll of chloro-form for 30 s at 30 �C. b-Galactosidase activity was then measured viathe addition of 140 ll of 4 mg/ml r-nitrophenyl b-DD-galactopyranoside(ONPG). The reaction was conducted at 30 �C until a yellow colorformed, and was then quenched via the addition of 0.4 ml of 1 MNa2CO3. The samples were then briefly centrifuged to remove anyremaining cell debris, and the absorbance was measured at wave-lengths of 420 and 550 nm. b-Galactosidase activity was calculatedusing the formula units = [1000 · (A420 � 1.75 · A550)]/(time · vol-ume · A600).

2.3. In vitro pull-down assaycDNA encoding for human TSC-22 was subcloned into the EcoRI

and XhoI restriction sites of the pGEX4T-1 vector (GE Healthcare,Piscataway, NJ, USA) to generate the glutathione S-transferase(GST) fusion protein (GST-TSC-22). Human fortilin cDNA was li-gated into pET28a (Novagen, Madison, WI, USA) using EcoRI andXhoI, thereby generating a histidine fusion protein (His-fortilin).The GST and histidine fusion proteins were purified using either aGST column (GE Healthcare) or an Ni2+ column (Novagen), in accor-dance with the manufacturer�s protocols. The purified GST-TSC-22(2 lg) was then mixed with His-fortilin (2 lg). The mixtures were sub-sequently added to 20 ll of a GST column matrix (glutathione sephar-ose 4B, GE Healthcare) and incubated for 30 min at 25 �C. The slurrywas pelleted via centrifugation and washed. The pellet of the gel matrixwas resuspended in 20 ml of elution buffer (10 mM glutathione, 50 mMTris–HCl, pH 8.0) and incubated for 20 min at 25 �C in order to elutethe bound GST fusion proteins. The eluted proteins were separated viagel electrophoresis, and the proteins were detected by Coomassie stain-ing.

2.4. Co-immunoprecipitationFor co-immunoprecipitation, SKOV-3 ovarian cancer cells were

washed in PBS and resuspended in cell lysis solution (50 mM Tris–HCl, pH 7.2, 150 mM NaCl, 1% Triton X-100, 1 lg/ml leupeptin ml,1 lg/ml pepstatin ml, 2 lg/ml aprotinin, 200 lg/ml PMSF). The lysateswere then incubated with goat anti-fortilin antibody (sc-20426, SantaCruz Biotechnology), goat anti-TSC-22 antibody (sc-27844, SantaCruz Biotechnology) or goat IgG (sc-2028, Santa Cruz Biotechnology),and precipitated with protein A-agarose (GE Healthcare). The precip-itated proteins were resolved via SDS gel electrophoresis, transferredonto Immobilon P membranes (Millipore Corporation, Billerica,MA, USA), and immunoblotted with anti-TSC-22 antibody or anti-fortilin antibody using the ECL system (GE Healthcare). b-Actin anti-body (sc-69879, Santa Cruz Biotechnology) was employed as an inter-nal control.

2.5. Subcloning of deletion mutants of fortilin and TSC-22Three deletion mutants (Met1-Asp52, Asn53-Phe110, Gln111-Ala144) of

TSC-22 were isolated by PCR using the combination of the followingprimers (TSC-22-F1, 5 0-CGGGAATTCATGAAATCCCAATGGT-GT-3 0; TSC-22-R1, 5 0-ATTCTCGAGTCAGTCAATAGCTACCAC-3 0; TSC-22-F2, 5 0-CGGGAATTCAACAAAATCGAGCAAGCT-3 0;TSC-22-R2, 50-ATTCTCGAGTCAAAACTGGGCAAGCTG-30; TSC-22-F3, 5 0-CGGGAATTCCAGGCCCAGCTGCAGACT- 3 0; TSC-22-R3, 5 0-CGGCTCGAGCTATGCGGTTGGTCCTGA-3 0). PCRproducts spanning each fragment were cloned into the EcoRI andXhoI restriction sites of the pGilda. Three deletion mutants (Met1-Gly69, Val70-Ala119, Glu120-Cys172) of fortilin were isolated by poly-merase chain reaction (PCR) using the combination of the followingprimers (fortilin-F1, 5 0-CGGGAATTCATGATTATCTACCGG-GAC-30; fortilin-R1, 5 0-CGGCTCGAGTCAACCAGTGATTACT-GT-3 0; fortilin-F2, 5 0-CGGGAATTCGTCGATATTGTCATGAAC-3 0; fortilin-R2, 5 0-ATTCTCGAGTCATGC AGCCCCTGTCAT-3 0;fortilin-F3, 5 0-CGGGAATTCGAACAAATCAAGCACATC-3 0; for-

tilin-R3, 5 0-CGGCTCGAGTTAACATTTTTCCATTTC-30). PCRproducts spanning each fragment were cloned into the EcoRI andXhoI restriction sites of the pJG4-5. Each constructed plasmid wasintroduced into yeast EGY48 expressing either fortilin or TSC-22 hy-brid protein.

2.6. Small interfering RNA (siRNA) constructionThe siRNA oligonucleotide sequence targeting TSC-22 (5 0-

GGCGATGGATCTAGGAGTT-3 0) corresponded to nucleotides27–45 in the human sequence, and the siRNA oligonucleotide sequencetargeting fortilin (5 0-AAGGTACCGAAAGCACAGT-30) corre-sponded to nucleotides 179–197 in the human sequence. siRNA wassynthesized using an siRNA construction kit (Ambion, Austin, TX,USA) and was then transfected by oligofectamine (Invitrogen), inaccordance with the manufacturer�s protocols. After transfection ofsiRNA, the mRNA and protein expression levels of both fortilin andTSC-22 were compared to transfectants with either cDNA constructsof pEGFPC1-fortilin or pEGFPC1-TSC-22 by RT-PCR and immuno-blot as described above.

2.7. Cell viability assaySKOV-3 ovarian cancer cells were plated in six-well plates and trans-

fected with the indicated cDNA and/or siRNA. Three days after trans-fection, the cells were harvested, stained with trypan blue, and countedunder a light microscope. Cell viability assay was also assessed with thepresence of caspase-3 specific inhibitor, z-DEVD-fmk (50 lM; BDP-harmingen, San Diego, CA, USA).

2.8. Apoptosis assaySKOV-3 ovarian cancer cells were plated onto six-chamber slides

and transfected with the indicated cDNA and/or siRNA. For the eval-uation of nuclear morphology, the nuclei were fixed in methanol andstained with 4,6 0-diamidino-2-phenylindole (DAPI, 1 lg/mL, Boehrin-ger Mannheim, Mannheim, Germany) for 15 min and rinsed twicewith PBS, then examined under a fluorescence microscope.

2.9. In vitro caspase-3 activity assayCaspase-3 enzymatic activity in the cell lysates was evaluated using

acetyl-DEVD-7-amino-4-trifluoromethyl coumarin, in accordancewith the manufacturer�s protocols (BDPharmingen). Activity was as-sessed using a Spectramax 340 microplate reader (Molecular Devices,Sunnyvale, USA) in fluorescence mode with an excitation wavelengthof 400 nm and an emission wavelength of 505 nm. Enzyme activity wascalculated and indicated as fluorescence in accordance with the for-mula provided by the manufacturer (BDPharmingen). Caspase-3 enzy-matic activity in the cell lysates was also visualized by immunoblotusing anti-caspase-3 antibody (610324, BDPharmingen).

2.10. Determination of TSC-22 protein stabilitySKOV-3 ovarian cancer cells were transfected with mock (an expres-

sion vector only without insert), fortilin siRNAs, or co-transfectedwith TSC-22 and the indicated pEGFPC1-fortilin cDNA mutant.Three days after transfection, the cells were treated with cycloheximide(Sigma) at a final concentration of 100 lg/ml. The cell lysates were sub-sequently harvested at sequential time points after treatment, resolvedvia SDS gel electrophoresis, transferred onto Immobilon P membranes(Millipore Corporation, Billerica, MA, USA), and immunoblottedwith anti-TSC-22 antibody or anti-GFP antibody (sc-8334, Santa CruzBiotechnology) using the ECL system (GE Healthcare). b-Actin anti-body was utilized as an internal control.

3. Results and discussion

3.1. Identification of fortilin as a TSC-22 binding protein

The human ovary cDNA library fused to the gene for the

transcription activator, pJG4-5, was introduced into yeast cells

harboring pGilda-TSC-22 as bait. Approximately, 3 · 106

independent transformants were pooled and spread again onto

the selection media (Ura�, His�, Trp�, Leu�) containing 2%

(w/v) galactose in order to induce cDNA expression. If a

FgIppffa

1212 J.H. Lee et al. / FEBS Letters 582 (2008) 1210–1218

B42-tagged protein interacts with the TSC-22, the transcrip-

tion of the LEU2 gene is activated, thus allowing the host cells

to grow on a leucine-deficient synthetic medium. Among the

seven colonies obtained on the selection media, a total of five

colonies evidenced galactose dependency. The plasmids were

isolated from the selected yeast cells and introduced into Esch-

erichia coli KC8 in order to isolate the plasmids harboring the

pJG4-5-cDNA inserts. The plasmids were then isolated by the

plasmid marker, trp, in the E. coli host, and the purified plas-

mids were sequenced. A GenBank homology search using the

BLAST program showed that all five of the plasmids encoded

for human fortilin (GenBank accession number: NM_003295).

In order to verify this result, we determined the binding activ-

ity between TSC-22 and fortilin via measurements of the rela-

tive b-galactosidase expression level. As is shown in Fig. 1A,

the b-galactosidase activity between TSC-22 and fortilin was

clearly observed, whereas very little b-galactosidase activity

was observed between TSC-22 and the empty vector (vector

only).

TT

0

50

25

75

100

β-ga

lact

osid

ase

activ

ity (u

nit)

Mw 1 2

31.2 kDa42.6 kDa

85 kDa128 kDa213 kDa

IP

WB : Anti-TSC-22

WB : Anti-β-actin

Whole cell

lysate goat IgG Anti-Fortilin

A

B

C

ig. 1. Interaction analysis between human fortilin and TSC-22. (A) Bindinalactopyranoside (ONPG) assays. b-Galactosidase activity was measured vianteractions between TSC-22 (GST-TSC-22) and fortilin (His-fortilin) were conurified GST-TSC-22 (2 lg) was mixed with His-E-fortilin (2 lg). The proteinroteins were separated via gel electrophoresis, and the proteins were detectedortilin; 4, GST-TSC-22 and His-fortilin, respectively. The results are represenortilin with TSC-22 in human ovarian carcinoma cells, SKOV-3. IP means intibodies. The results are representative of three separate experiments.

The interaction between TSC-22 and fortilin was further

evaluated via in vitro pull-down assays and co-immunoprecip-

itation. For the in vitro pull-down analysis, the GST–TSC-22

fusion protein was mixed with His–fortilin, and the mixture

was purified using a GST column matrix. The eluted proteins

were separated via gel electrophoresis and the proteins were

detected by Coomassie staining. As is shown in Fig. 1B, the

His–fortilin fusion protein, which measured approximately

25 kDa, was co-purified with the GST–TSC-22 fusion protein

(lane 4). No protein bands were detected when no input was

added to the GST column matrix (lane 1). A single band of

approximately 45 kDa, equivalent to the GST–TSC-22 fusion

protein, was observed when the GST–TSC-22 fusion protein

was added to the GST column matrix (lane 2). However, no

bands were observed when the His–fortilin fusion protein

was added to the GST column matrix (lane 3).

For co-immunoprecipitation, SKOV-3 ovarian cancer cells

were washed in PBS and resuspended in cell lysis solution.

Immunoprecipitation was subsequently conducted using

SC-22 + Fortilin : 93.24 ± 2.26SC-22 + Vector only : 2.73 ± 0.05

3 4

TSC-22

Fortilin

IP

WB : Anti-Fortilin

WB : Anti-β-actin

Whole cell

lysate goat IgG Anti-TSC-22

g activity of TSC-22 and fortilin measured by r-nitrophenyl b-DD-the addition of ONPG. Data are expressed as the means ± S.E.M. (B)firmed via in vitro pull-down assays. For GST pull-down analysis, themixtures were then subjected to GST affinity purification. The elutedby Coomassie staining. Lanes: 1, no input; 2, GST-TSC-22; 3, His-

tative of three separate experiments. (C) Co-immunoprecipitation ofmmunoprecipitation and WB means immunoblotting with indicated


anti-fortilin antibody or anti-TSC-22 antibody with whole

cell lysates. After immunoprecipitation, the precipitated pro-

teins were immunoblotted with anti-TSC-22 antibody or

anti-fortilin antibody. As is shown in Fig. 1D, TSC-22 was

co-immunoprecipitated with fortilin (upper left panel),

whereas no interactions were observed between isotype con-

trol (goat IgG) and TSC-22 (upper left panel). Also, fortilin

was co-immunoprecipitated with TSC-22 (upper right panel),

whereas no interactions were observed between isotype con-

trol (goat IgG) and fortilin (upper right panel). Immunoblot-

ting using anti-TSC-22 antibody and anti-fortilin antibody

verified that an equal quantity of both fortilin and TSC-22

was present in the whole cell lysates. The whole cell lysates

from both samples harbored the equivalent proteins when

immunoblotted using anti-b-actin antibody (lower panel).

These results clearly demonstrated that TSC-22 and fortilin

interact with each other on the physiological level in

SKOV-3 ovarian cancer cells.

A

B

C

Met1 Cys172

Met1 Gly69

Glu120 Cys172

Val70 Ala119

TSFo

β-galactosiactivity (u

Met1 Ala144

Met1 Asp52

Asn53 Phe110

Gln111 Ala144

FoTS

β-galactosiactivity (u

FortilinTSC-22

1-69Vector only

β-galactosidaseactivity (unit) 1.29 ±0.19 1.3

Fig. 2. Mapping of the critical interaction region between TSC-22 and fortransformed into EGY48 yeast cells to test a protein–protein interaction wiseparate experiments. Data are shown as means ± S.E. (A) Left panel shodeletion mutant and full-length TSC-22 fusion proteins in the yeast two-hydetermined in the yeast two-hybrid system. The values of b-galactosidacorresponding lanes. The values of b-galactosidase activity in negative contrshows the schematic representation of cDNA constructs for each fortilin dhybrid system. Right panel shows the result of protein–protein interaction dactivity (U) measured by ONPG assays are indicated below the correspon(vector only) of each construct were below 1.56 ± 0.25. (C) Interaction betthree TSC-22 deletion (1Met-Asp52, 53Asn-Phe110 and 111Gln-Ala144) fusion

3.2. Mapping of the interaction region between TSC-22 and

fortilin

To identify the fortilin binding region of TSC-22, cDNA

constructs containing three TSC-22 deletion mutants were de-

signed as shown in Fig. 2A. These truncated regions were pre-

dicted to be predominantly a-helices. In the two-hybrid

system, the full-length human fortilin cDNA and either plas-

mid containing a full-length human TSC-22 cDNA (left panel

in Fig. 2A, full) or plasmids containing three truncation mu-

tant forms (left panel in Fig. 2A, Met1-Asp52, Asn53-Phe110,

Gln111-Ala144) of cDNAs were co-transformed into EGY48

yeast cells. Cells containing full-length TSC-22 cDNA and also

one deletion mutant (Asn53-Phe110) grew on the Ura, His, Trp

and Leu deficient plates. Yeast cells transformed with the other

deletion mutants (Met1-Asp52 and Gln111-Ala144) failed to

grow (right panel in Fig. 2A). To confirm this result, we deter-

mined the binding activity of these constructs by measuring the

relative expression level of b-galactosidase. As shown in right

C-22rtilin

FullFull

Full1-69

Full70-119

Full120-172

dasenit) 93.24 2.26 87.35 2.13 2.21 0.32 3.47 0.58

rtilinC-22

FullFull

Full1-52

Full53-110

Full111-144

dasenit) 91.24± ± ± ±2.26 1.39 0.23 79.98 2.47 1.77 0.52

1-691-52

1-6953-110

1-69111-144

± ± ±1 0.35 83.87 6.07 1.73 0.74

± ± ± ±

tilin using the yeast two-hybrid system. The cDNA constructs were co-thin the yeast two-hybrid system. The results are representative of threews the schematic representation of cDNA constructs for each TSC-22brid system. Right panel shows the result of protein–protein interactionse activity (U) measured by ONPG assays are indicated below the

ols (vector only) of each construct were below 1.05 ± 0.05. (B) Left paneleletion mutant and full-length fortilin fusion proteins in the yeast two-etermined in the yeast two-hybrid system. The values of b-galactosidaseding lanes. The values of b-galactosidase activity in negative controls

ween cDNA constructs for a fortilin deletion mutant (1Met-Gly69) andproteins in the yeast two-hybrid system.

Fcae


panel in Fig. 2A, b-galactosidase assay results confirmed that

either of these mutants (Met1-Asp52 and Gln111-Ala144) cannot

bind to fortilin. However, we cannot exclude that the fortilin

binding site of TSC-22 is located near the junction of these mu-

tants since b-galactosidase activity in one deletion mutant

(Asn53-Phe110) is lower than full-length TSC-22 (right panel

in Fig. 2A).

Subsequently, cDNA constructs containing three fortilin

truncation mutants were designed to localize the TSC-22 bind-

ing region of fortilin (left panel in Fig. 2B, full). These trun-

cated regions were predicted to be predominantly a-helices.

In the two-hybrid system, full-length human TSC-22 cDNA

and either full-length human fortilin cDNA (left panel in

Fig. 2B, full) or three truncation mutants (left panel in

Fig. 2B, Met1-Gly69, Val70-Ala119, Glu120-Cys172) were co-

transformed into EGY48 yeast cells. Cells containing full-

length fortilin cDNA and also the one deletion mutant

(Met1-Gly69) grew on the Ura, His, Trp and Leu deficient

plates. Yeast cells transformed with the other deletion mutants

(Val70-Ala119, Glu120-Cys172) failed to grow (right panel in

Fig. 2B). We also quantitated the binding activity of these con-

structs by measuring relative expression level of b-galactosi-

dase. Results on b-galactosidase assay also indicated that the

critical fortilin region for binding TSC-22 resided within

Met1-Gly69 (right panel in Fig. 2B).

To confirm these results, one deletion mutant of fortilin

(Met1-Gly69) and either vector cDNA or three truncation mu-

tants of TSC-22 (Met1-Asp52, Asn53-Phe110, Gln111-Ala144)

were co-transformed into EGY48 yeast cells (Fig. 2C). Consis-

tent with results from Fig. 2A and B, cells containing fortilin

deletion mutant (Met1-Gly69) with one TSC-22 deletion mu-

GFP-TSC-22GFP-TSC-22

+ siRNA

Tsc-22 mRNA

Gapdh mRNA

Anti-TSC-22

Anti-β-actin

GFP-TSC-22GFP-TSC-22

+ siRNA G

ig. 3. The construction of siRNA system that inhibits TSC-22 and fortilin.DNA alone or co-transfected each GFP fused cDNA with each siRNA. Afternd analyzed the mRNA or protein expression levels by RT-PCRs and immunoxperiments.

tant (Asn53-Phe110) only grew on the Ura, His, Trp and Leu

deficient plates whereas the other deletion mutants (Val70-

Ala119, Glu120-Cys172) failed to grow. Subsequent results on

b-galactosidase assay were also agreed with these results

(Fig. 2C).

3.3. Overexpression of TSC-22 can induce apoptosis in SKOV-3

ovarian cancer cells via caspase-3 dependent pathway

To evaluate the functional consequences of the interaction

taking place between TSC-22 and fortilin, we conducted a cou-

ple of experiments involving mammalian cell apoptosis. First,

we attempted to construct siRNA system that inhibits TSC-22

and fortilin since both proteins endogenously express in

SKOV-3 ovarian cancer cells. To test whether siRNA transfec-

tion efficiently blocks endogenous TSC-22 or fortilin, we mea-

sured both mRNA and protein expression levels of siRNA and

overexpression co-transfectants as well as overexpression

transfectants. As shown in Fig. 3, the mRNA and protein

expression levels of both TSC-22 and fortilin in siRNA and

overexpression co-transfectants (lane 2 in Fig. 3) were signifi-

cantly decreased comparing with overexpression transfectants

(lane 1 in Fig. 3).

Next, we determine whether TSC-22 overexpression induces

apoptosis in SKOV-3 cells. As is shown in Fig. 4A, the viabil-

ity of the SKOV-3 cells is gradually reduced in a transfected

TSC-22 cDNA dose-dependent manner, and nearly 50% of

cells died when 1 lg of TSC-22 cDNAs was transfected. This

result is also supported by immunoblot analysis using cas-

pase-3 antibody. Only activated form of caspase-3 was ob-

served in TSC-22 transfectant, not in both mock and siTSC-

22 (siRNA of TSC-22) transfectants (Fig. 4B). To confirm

GFP-FortilinGFP-Fortilin

+ siRNA

Fortilin mRNA

Gapdh mRNA

Anti-Fortilin

Anti-β-actin

FP-FortilinGFP-Fortilin

+ siRNA

SKOV-3 ovarian cancer cells were transfected with each GFP fusedtransfection, the cells were visualized under fluorescence microscope,blots. Size bar, 20 lm. The results are representative of three separate

A

B

β-actin

Mock siTSC-22 TSC-22

Pro-caspase-3

Cleaved caspase-3

C

TSC-22 cDNA (μg)

0

20

40

60

80

100

Cel

l via

bilit

y (p

erce

ntag

e)

0.2 0.4 0.6 0.8 1.0

Anti-β-actin

Anti-TSC-22

0.2 0.4 0.6 0.8 1.0 TSC-22 cDNA (μg)

0Mock

+z-DEVD-fmk

TSC-22+

z-DEVD-fmk

Mock TSC-22

Cel

l via

bilit

y (%

)

20

40

60

80

100

Fig. 4. TSC-22 mediated apoptosis via caspase-3 dependent pathwayin human ovarian carcinoma cells, SKOV-3. (A) SKOV-3 ovariancancer cells were transfected with mock (an expression vector onlywithout insert) or TSC-22 cDNA. Three days after transfection, thecells were stained with trypan blue and quantitated to analyze cellviability. Data are expressed as the means ± S.E.M. (B) SKOV-3ovarian cancer cells were transfected with mock (an expression vectoronly without insert), siRNA of TSC-22 (siTSC-22) or TSC-22 cDNA.The cells were harvested and caspase-3 was detected by immunoblotusing anti-caspase-3 antibody. The results are representative of threeseparate experiments. (C) A caspase-3 specific inhibitor, z-DEVD-fmk,was treated in mock or TSC-22 transfectants. The cells were stainedwith trypan blue and quantitated to analyze cell viability. Data areexpressed as the means ± S.E.M.


these results, we treated z-DEVD-fmk, a specific inhibitor of

caspase-3, in TSC-22 transfectant. As a result, z-DEVD-fmk

treatment definitively restores viability of TSC-22 transfectant

(Fig. 4C). Thus, 1 lg of TSC-22 cDNAs was utilized as the

treatment concentration in all of the following experiments.

3.4. Fortilin prevents TSC-22 mediated apoptosis via the

destabilization of TSC-22

Fortilin is a highly conserved hydrophilic nuclear protein,

and is expressed ubiquitously in a variety of cells [15]. The re-

sults of a recent study showed that fortilin overexpression can

prevent stress-induced mammalian cell apoptosis [15,16]. Con-

versely, the inhibition of fortilin in malignant human carci-

noma cells induced an increase in apoptosis [15,16]. Thus,

fortilin may be relevant to cell survival as a negative apoptotic

regulator. However, the exact mechanism underlying the anti-

apoptotic effects of fortilin remains to be clearly elucidated.

Therefore, we attempted to determine whether TSC-22 and

fortilin evidence an additive effect or antagonize each other

upon apoptosis.

As shown in Fig. 5A, nearly 50% of the cells perished after

transfection with TSC-22 cDNA alone as compared to the

mock transfectant, whereas no change in cell viability was ob-

served when the cells were transfected with fortilin cDNA

alone (Fig. 5A). Consistent with previous observation

[15,16], the fortilin siRNA transfectant shows decreased cell

viability, similar level as that of TSC-22 transfectant

(Fig. 5A). Interestingly, cell viability largely recovered upon

co-transfection with TSC-22 and fortilin cDNAs (Fig. 5A).

This suggests that fortilin may antagonize the TSC-22-medi-

ated apoptosis of SKOV-3 cells. In order to verify this result,

we inhibited fortilin expression by siRNA when we transfected

the cells with TSC-22 cDNA. As is shown in Fig. 5A, the via-

bility of TSC-22 and the sifortilin (siRNA of fortilin) co-trans-

fectant was greatly reduced as compared to the TSC-22 and

fortilin co-transfectant.

To verify that the reduction in cell numbers was reflective of

apoptosis, we employed DAPI staining to confirm the ob-

served reduced proliferation in SKOV-3 cells overexpressing

mock (expression vector only), fortilin, sifortilin, TSC-22, or

siTSC-22, or co-transfected with TSC-22 and fortilin cDNAs,

co-transfected with TSC-22 and sifortilin, or co-transfected

with TSC-22 and sifortilin, or triple-transfected with fortilin,

sifortilin and TSC-22. As had been expected, after transfection

with TSC-22 cDNA alone compared to mock transfectant,

nearly 60% of the cells evidenced fragmented nuclei, whereas

no change in cell viability was observed when the cells were

transfected with mock and fortilin cDNA alone. Cells transfec-

ted with a combination of TSC-22 and fortilin evidenced a sig-

nificantly reduced number of cells harboring fragmented

DNA, but cells transfected with fortilin-siRNA did not

(Fig. 5B). We subsequently assessed the caspase-3 activity in

those transfectants. As is shown in Fig. 5C, no significant

changes in caspase-3 activity were observed in the cells trans-

fected with fortilin cDNA alone, as compared to the mock-

transfected cells. A significant upregulation of caspase-3 activ-

ity was detected in the cells transfected with TSC-22 cDNA

alone compared to the mock-transfected cells (Fig. 5C). How-

ever, caspase-3 activity was reduced by almost half in the co-

transfectant harboring both TSC-22 and fortilin cDNAs as

compared to the transfectant harboring TSC-22 cDNA alone

Fortilin+

TSC-22

siFortilin+

TSC-22

siFortilin+

siTSC-22

Fortilin+

siFortilin+

TSC-22

Fortilin siFortilinMock TSC-22 siTSC-220

Cel

l via

bilit

y (%

)

20

40

60

80

100

Mock

Fortilin

Fortilin+TSC-22

siTSC-22

Fortilin+siFortilin+TSC-22

siFortilin+TSC-22

siFortilin TSC-22

siFortilin+siTSC-22

A

B

C

Cas

pase

-3 a

ctiv

ity (f

luor

esce

nce)

0

2000

4000

6000

8000

10000

Fortilin+

TSC-22

siFortilin+

TSC-22

siFortilin+

siTSC-22

Fortilin+

siFortilin+

TSC-22

Fortilin siFortilinMock TSC-22 siTSC-22

Fig. 5. Fortilin inhibits TSC-22 mediated apoptosis in human ovarian carcinoma cells, SKOV-3. Cell viability assay (A), DAPI staining (B) andcaspase-3 activity assay (C) were performed on SKOV-3 ovarian cancer cells transfected with mock (an expression vector only without insert),fortilin, sifortilin, TSC-22 or siTSC-22, or co-transfected with TSC-22 and fortilin cDNAs, co-transfected with TSC-22 and sifortilin, or co-transfected with TSC-22 and sifortilin, or triple-transfected with fortilin, sifortilin and TSC-22. (A) Cell viability assay was quantified via tryptophanblue staining. Data are expressed as the means ± S.E.M. (B) Cells were stained with DAPI to visualize DNA fragmentation for the apoptosis assay.Arrows indicate the observed DNA fragmentations. Size bar, 20 lm. The results are representative of three separate experiments. (C) Caspase-3activity was measured using a microplate reader in fluorescence mode with an excitation wavelength of 400 nm and an emission wavelength of505 nm. Enzyme activity was calculated and indicated as fluorescence in accordance with the formula provided by the manufacturer. Data areexpressed as the means ± S.E.M.


(Fig. 5C). Conversely, caspase-3 activity in the co-transfectant

containing both TSC-22 cDNA and sifortilin was restored

markedly, to almost the same levels observed in the cells trans-

fected with TSC-22 cDNA alone. These results clearly showed

Fortilin

TSC-22

β-actin

CHX (h) 0 1 3 5 0 1 3 5

Mock siFortilin

0TSC-22

+Fortilin

(120-172)

TSC-22 TSC-22+

Fortilin(whole)

Mock TSC-22+

Fortilin(1-69)

TSC-22+

Fortilin(70-119)

Cel

l via

bilit

y (%

)

20

40

60

80

100

β-actin

TSC-22

Fortilin

1-69 70-119 120-172

Fortilin

A

B

C

Fig. 6. Stability of TSC-22 is reduced by fortilin. (A) SKOV-3 ovariancancer cells were transfected with mock (an expression vector onlywithout insert) or fortilin siRNAs. After transfection, the cells weretreated with cycloheximide (100 lg/ml). At each time point, the wholecell lysates were prepared and analyzed via immunoblot analysis usingantibodies against TSC-22, fortilin or b-actin to measure the amountsof TSC-22 protein remaining at the respective times. The �h� refers tohours after cycloheximide treatment. The results are representative ofthree experiments. (B) SKOV-3 ovarian cancer cells were co-transfec-ted TSC-22 cDNA with the each pEGFPC1-fortilin cDNA mutant(Met1-Gly69, Val70-Ala119, Glu120-Cys172). The cells were treated withcycloheximide (100 lg/ml). At 5 h after cycloheximide treatment, thewhole cell lysates were prepared and analyzed via immunoblot analysisusing antibodies against TSC-22, GFP or b-actin to measure theamounts of TSC-22 and fortilin. The results are representative of threeexperiments. (C) SKOV-3 ovarian cancer cells were transfected withmock (an expression vector only without insert) or TSC-22 cDNA, orco-transfected TSC-22 cDNA with the each pEGFPC1-fortilin cDNAmutant (Met1-Gly69, Val70-Ala119, Glu120-Cys172). Three days aftertransfection, the cells were stained with trypan blue and quantitated toanalyze cell viability. Data are expressed as the means ± S.E.M.


that fortilin inhibits TSC-22 mediated apoptosis in SKOV-3

human ovarian carcinoma cells.

To define the mechanism by which fortilin inhibits TSC-22

mediated apoptosis, we assessed the stability of the TSC-22

protein in the presence or absence of fortilin in SKOV-3 cells.

In brief, mock transfectant and sifortilin transfectant were

treated with cycloheximide. Then, the cell lysates were har-

vested at 1 h, 3 h and 5 h after treatment, and immunoblotted

with anti-TSC-22 antibody or anti-fortilin antibody. As is

shown in Fig. 6, the level of TSC-22 protein evidenced a signif-

icant reduction in a time-dependent manner in the presence of

fortilin. Conversely, the level of TSC-22 protein evidenced no

significant change for over 5 h in the absence of fortilin

(Fig. 6A). To confirm this result, we co-transfected TSC-22

cDNA with the each pEGFPC1-fortilin cDNA mutant

(Met1-Gly69, Val70-Ala119, Glu120-Cys172) and performed

TSC-22 stabilization assay and cell viability test using these

co-transfectants. For TSC-22 stabilization assay, we harvested

cell lysates at 5 h after cycloheximide treatment, and immuno-

blotted with anti-TSC-22 antibody or anti-GFP antibody. As

shown in Fig. 6B, the level of TSC-22 protein was significant

decreased in co-transfectant containing TSC-22 cDNA and

one deletion fortilin mutant (Met1-Gly69). Also, cell viability

in co-transfectant containing TSC-22 cDNA and one deletion

fortilin mutant (Met1-Gly69) was recovered, compared to

transfectant containing TSC-22 cDNA alone (Fig. 6C). How-

ever, neither the level of TSC-22 protein nor cell viability

was changed in co-transfectants containing TSC-22 cDNA

and the other deletion mutant (Val70-Ala119, Glu120-Cys172).

These results clearly showed that the interaction between for-

tilin and TSC-22 accelerates the destabilization of TSC-22 in

SKOV-3 human ovarian carcinoma cells.

In this study, we have described the specific interaction

occurring between TSC-22 and fortilin, a highly conserved

anti-apoptotic protein [15,16]. The interaction between TSC-

22 and fortilin appears to be physiologically relevant, as this

interaction occurs in both yeast and mammalian cells. Our re-

sults also show that fortilin binds to and destabilizes TSC-22.

TSC-22 and fortilin are both localized predominantly within

the nucleus. During apoptosis, activated TSC-22 moves from

the cytosol to the nucleus and functions as a transcriptional

repressor of several genes [10]. Fortilin is a known nuclear-

residing protein, and myeloid cell leukemia 1 protein is known

to stabilize fortilin via direct interaction [17]. Thus, it is likely

that fortilin destabilizes the activity of TSC-22 within the nu-

cleus. Conversely, the weakening activity of fortilin induced

by treatment with sifortilin reversed the apoptotic activity of

TSC-22. Therefore, the interaction between TSC-22 and forti-

lin appears crucial for the regulation of TSC-22-mediated

apoptosis. The results of this study provide valuable informa-

tion as to the mechanisms by which fortilin exerts its anti-

apoptotic protein properties. Further studies in this regard will

focus on the specific mechanism by which fortilin destabilizes

TSC-22 in the nucleus, and will also concern the possible influ-

ence of fortilin on other pro-apoptotic proteins.

Acknowledgements: This work was supported by a grant from the Kor-ea Health 21 R&D Project, Ministry of Health & Welfare, Republic ofKorea (Project Number: A040004) and by a grant from the NationalCancer Center, Korea (NCC-0510571-3).

References

[1] Shibanuma, M., Kuroki, T. and Nose, K. (1992) Isolation of agene encoding a putative leucine zipper structure that is inducedby transforming growth factor beta 1 and other growth factors. J.Biol. Chem. 267, 10219–10224.

[2] Uchida, D., Omotehara, F., Nakashiro, K., Tateishi, Y., Hino, S.,Begum, N.M., Fujimori, T. and Kawamata, H. (2003) Posttran-scriptional regulation of TSC-22 gene by TGF-beta 1. Biochem.Biophys. Res. Commun. 305, 846–854.


[3] Kania, A., Salzberg, A., Bhat, M., D�Evelyn, D., He, Y., Kiss, I.and Bellen, H.J. (1995) P-element mutations affecting embryonicperipheral nervous system development in Drosophila melanogas-ter. Genetics 139, 1663–1678.

[4] Jay, P., Ji, J.W., Marsollier, C., Taviaux, S., Berge-Lefranc, J.L.and Berta, P. (1996) Cloning of the human homologue of theTGF-b-stimulated clone 22 gene. Biochem. Biophys. Res. Com-mun. 222, 821–826.

[5] Hashiguchi, A., Okabayashi, K. and Asashima, M. (2004) Role ofTSC-22 during early embryogenesis in Xenopus laevis. Dev.Growth Differ. 46, 535–544.

[6] Nakashiro, K., Kawamata, H., Hino, S., Uchida, D., Miwa, Y.,Hamano, H., Omotehara, F., Yoshida, H. and Sato, M. (1998)Down-regulation of TSC-22 (transforming growth factor b-stimulated clone 22) markedly enhances the growth of a humansalivary gland cancer cell line in vitro and in vivo. Cancer Res. 58,549–555.

[7] Ohta, S., Yanagihara, K. and Nagata, K. (1997) Mechanism ofapoptotic cell death of human gastric carcinoma cells mediated bytransforming growth factor b. Biochem. J. 324, 777–782.

[8] Uchida, D., Kawamata, H., Omotehara, F., Miwa, Y., Hino, S.,Begum, N.M., Yoshida, H. and Sato, M. (2000) Over-expressionof TSC-22 (TGF-b stimulated clone-22) markedly enhances 5-Fluorouracil-induced apoptosis in a human salivary gland cancercell line. Lab. Invest. 80, 955–963.

[9] Lu, Y., Kitaura, J., Oki, T., Komeno, Y., Ozaki, K., Kiyono, M.,Kumagai, H., Nakajima, H., Nosaka, T., Aburatani, H. andKitamura, T. (2007) Identification of TSC-22 as a potential tumorsuppressor that is upregulated by Flt3-D835V but not Flt3-ITD.Leukemia 21, 2246–2257.

[10] Hino, S., Kawamata, H., Uchida, D., Omotehara, F., Miwa, Y.,Begum, N.M., Yoshida, H., Fujimori, T. and Sato, M. (2000)Nuclear translocation of TSC-22 (TGF-beta-stimulated clone-22)concomitant with apoptosis: TSC-22 as a putative transcriptionalregulator. Biochem. Biophys. Res. Commun. 278, 659–664.

[11] Kester, H.A., Blanchetot, C., den Hertog, J., van der Saag, P.T.and van der Burg, B. (1999) Transforming growth factor-beta-stimulated clone-22 is a member of a family of leucine zipperproteins that can homo- and heterodimerize and has transcrip-tional repressor activity. J. Biol. Chem. 274, 27439–27447.

[12] Ohta, S., Yanagihara, K. and Nagata, K. (1997) Mechanism ofapoptotic cell death of human gastric carcinoma cells mediated bytransforming growth factor beta. Biochem. J. 324, 777–782.

[13] Ito, H., Fukada, Y., Murata, K. and Kimura, A. (1983)Transformation of intact yeast cells treated with alkali cations.J. Bacteriol. 153, 163–168.

[14] Guarente, L. (1983) Yeast promoters and lacZ fusions designed tostudy expression of cloned genes in yeast. Meth. Enzymol. 101,181–191.

[15] Li, F., Zhang, D. and Fujise, K. (2001) Characterization offortilin, a novel antiapoptotic factor. J. Biol. Chem. 276, 47542–47549.

[16] Graidist, P., Phongdara, A. and Fujise, K. (2004) Antiapoptoticprotein parteners fortilin and MCL1 independently protect cellsfrom 5-fluorouracil-induced cytotoxicity. J. Biol. Chem. 279,40868–40875.

[17] Zhang, D., Li, F., Weidner, D., Mnjoyan, Z.H. and Fujise, K.(2002) Physical and functional interaction between myeloid cellleukemia 1 protein (MCL1) and fortilin. J. Biol. Chem. 277,37430–37438.

Interaction between fortilin and transforming growth factor-beta stimulated clone-22 (TSC-22)...

Documents

Transcript of Interaction between fortilin and transforming growth factor-beta stimulated clone-22 (TSC-22)...