The splicing factor ASF/SF2 is associated with
TIA-1-related/ TIA-1-containing ribonucleoproteic
complexes and contributes to post-transcriptional
repression of gene expression
Nathalie Delestienne
1
, Corinne Wauquier
1
, Romuald Soin
1
, Jean-Franc¸ois Dierick
2,
*,
Cyril Gueydan
1,
and Ve
´
ronique Kruys
1,
1 Laboratoire de Biologie Mole
´
culaire du Ge
`
ne, Faculte
´
des Sciences, Universite
´
Libre de Bruxelles, Gosselies, Belgium
2 Biovalle
´
e, Proteomics Unit, Charleroi, Belgium

doi:10.1111/j.1742-4658.2010.07664.x
TIA-1-related (TIAR) protein is a shuttling RNA-binding protein impli-
cated in several steps of RNA metabolism. In the nucleus, TIAR contrib-
utes to alternative splicing events, whereas, in the cytoplasm, it acts as a
translational repressor on specific transcripts such as adenine and uridine-
rich element-containing mRNAs. In addition, TIAR is involved in the
general translational arrest observed in cells exposed to environmental
stress. This activity is encountered by the ability of TIAR to assemble
abortive pre-initiation complexes coalescing into cytoplasmic granules
called stress granules. To elucidate these mechanisms of translational
repression, we characterized TIAR-containing complexes by tandem affinity
purification followed by MS. Amongst the identified proteins, we found the
splicing factor ASF ⁄ SF2, which is also present in TIA-1 protein complexes.
We show that, although mostly confined in the nuclei of normal cells,
ASF ⁄ SF2 migrates into stress granules upon environmental stress. The
migration of ASF ⁄ SF2 into stress granules is strictly determined both by
its shuttling properties and its RNA-binding capacity. Our data also indi-
cate that ASF ⁄ SF2 down-regulates the expression of a reporter mRNA
carrying adenine and uridine-rich elements within its 3¢ UTR. Moreover,
tethering of ASF ⁄ SF2 to a reporter transcript strongly reduces mRNA
translation and stability. These results indicate that ASF ⁄ SF2 and TIA
proteins cooperate in the regulation of mRNA metabolism in normal cells
and in cells having to overcome environmental stress conditions. In addi-
tion, the present study provides new insights into the cytoplasmic function
of ASF ⁄ SF2 and highlights mechanisms by which RNA-binding proteins
regulate the diverse steps of RNA metabolism by subcellular relocalization
upon extracellular stimuli.
Structured digital abstract
l
MINT-7715509: ASF ⁄ SF2 (uniprotkb:Q6PDM2)andTIAR (uniprotkb:P70318) colocalize (MI:0403)

studied regulatory sequences, the adenine and uridine-
rich elements (AREs) located in the 3¢ UTR of
mRNAs are considered to regulate the stability and ⁄ or
traductibility of 8% of all human mRNAs [2]. RNA-
binding proteins comprise other key components of
the post-transcriptional regulation of gene expression.
These proteins are predominantly composed of well-
conserved RNA-binding domains mediating RNA con-
tact, and auxiliary domains involved in protein–protein
interactions and sub-cellular targeting [3,4]. TIA-1-
related (TIAR) protein belongs to the RNA recogni-
tion motif (RRM) family of RNA-binding-proteins. It
is a shuttling protein [5] involved in multiple aspects of
RNA metabolism. In the nucleus, this protein acts as a
regulator of the alternative splicing of diverse pre-
mRNAs such as those encoding Fas, msl-2, FGFR-2
and calcitonin ⁄ CGRP [6–8]. In the cytoplasm, TIAR
has been shown to regulate the translation of various
mRNAs bearing AREs in their 3¢ UTR. For example,
mRNAs encoding human matrix metallinoproteinase-
13 and b
2
-adrenergic receptor are translationaly
repressed by TIAR [9,10]. In addition to the transla-
tional regulation of specific mRNAs, TIAR is involved
in a broader translational repression mechanism that
takes place in cells having to overcome environmental
stress such as UV irradiation, thermic variations or
oxidative shock [11]. Thus, although predominantly
nuclear at steady state, TIAR exerts both nuclear and

Q9CZM2), Rpl14 (uniprotkb:Q9CR57), Rpl13a (u niprotkb :P19253), Rpl13 (uniprotkb:
P47963), Rpl8 (uniprotkb:P62918)andRpl5 (uniprotkb:P47962)bytande m a ffinity purification
(
MI:0676)
l
MINT-7715427: TIA-1 (uniprotkb:P52912) physically interacts (MI:0915) with ASF ⁄ SF2 (uni-
protkb:
Q6PDM2)byanti tag coimmunoprecipitation (MI:0007)
l
MINT-7715264: TIAR (uniprotkb:P70318) physically interacts (MI:0915) with ASF ⁄ SF2 (uni-
protkb:
Q6PDM2)byanti tag coimmunoprecipitation (MI:0007)
l
MINT-7715309: TIAR (uniprotkb:P70318) physically interacts (MI:0915) with Ddx21 (uni-
protkb:
Q8K2L4)byanti tag coimmunoprecipitation (MI:0007)
l
MINT-7715416: TIAR (uniprotkb:P70318) physically interacts (MI:0915) with ASF ⁄ SF2 (uni-
protkb:
Q07955)byanti tag coimmunoprecipitation (MI:0007)
N. Delestienne et al. ASF ⁄ SF2 in TIAR-mediated regulatory pathways
FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS 2497
as well as on the interactions taking place in these par-
ticles.
The present study aimed to identify the proteins
assembling with TIAR and to characterize their role in
TIAR cytoplasmic functions. We used a tandem affin-
ity purification (TAP) approach to isolate TIAR-asso-
ciated proteins and identified them by MS [14]. The
association of these proteins with TIAR was further

nent of stress granules and a novel RNA-binding
protein involved in ARE-mediated post-transcriptional
regulation. Therefore, the results obtained in the pres-
ent study support the recent findings showing that
members of the SR proteins family, including
ASF ⁄ SF2 and SRp20, have important roles in the
cytoplasmic control of mRNA metabolism [15–17].
Results
Identification of TIAR-associated proteins
The tandem affinity purification procedure originally
developed in yeast by Rigaut et al. [18] was used to
identify proteins interacting with TIAR in mammalian
cells. Therefore, plasmids encoding the TAP alone or
fused to the carboxy-terminal extremity of TIAR were
generated and the ability of the TIAR-TAP fusion pro-
tein to recapitulate TIAR activities was analyzed
before being used to identify interacting partners. We
thus measured the capacity of the TIAR-TAP protein
with respect to activating the inclusion of TIA-1 alter-
native exon 6A from a transcript derived from a repor-
ter minigene, as previously described for the wild-type
TIAR protein [19]. 293T cells were transiently trans-
fected with the pCI-6-6A-7 minigene in combination
with plasmids encoding TAP alone or TIAR-TAP.
Inclusion of exon 6A in the reporter transcript upon
TIAR-TAP overexpression was subsequently analyzed
by RT-PCR. As shown in Fig. 1A, the expression of
TIAR-TAP but not of TAP alone led to an increased
accumulation of reporter transcript containing exon
6A, thereby indicating that TIAR-TAP recapitulated

and Flag-IP products were analyzed by western blot
analysis with anti-Flag and anti-HA sera. The specific-
ity of the interactions was evaluated by IP of the unre-
lated BOIP-Flag protein [20]. As shown in Fig. 2A,
all the candidates identified except DDX21 were
ASF ⁄ SF2 in TIAR-mediated regulatory pathways N. Delestienne et al.
2498 FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS
specifically immunoprecipitated with TIAR-Flag, both
in the absence and the presence of RNAse A, thereby
indicating that TIAR association with these proteins is
specific and reliant on protein–protein interactions. By
contrast, DDX21 became undetectable in the TIAR-
Flag IP pellet upon RNAse A treatment, suggesting
that its association with TIAR occurs via RNA inter-
mediates.
Subcellular localization of TIAR-associated
proteins
As noted above, TIAR exerts both nuclear and cyto-
plasmic functions. In the cytoplasm, it is an invariant
component of SGs appearing in response to diverse
environmental stresses that induce a general translation
arrest [12]. We investigated the capacity of TIAR inter-
acting candidates to migrate into SGs upon oxidative
stress. COS cells were transiently transfected to express
the HA-tagged partners and were subsequently treated
with arsenite (1 mm for 30 min) to induce SGs. Indi-
rect fluorescence microscopy revealed that TIAR-asso-
ciated proteins predominantly accumulated in the
nucleus in normal conditions. However, upon oxida-
tive stress, only ASF ⁄ SF2 protein migrated in TIAR-

36.5
21.5
14.4
31
48.7
–
TIAR-
TA P
Inputs Eluates
WB anti-TIAR
TIAR-TAP
TIAR-CBP
TIAR
TA P
TA P
TIAR-TAP
6
6A 7
6
7
–++
RT-PCR
WB anti-TIAR
TIAR-TAP
TIAR
C
B
Fig. 1. Functional characterization of TIAR-TAP protein and purification of TIAR-TAP complexes. (A) Upper panel: RT-PCR analysis of exon
6A inclusion in minigene reporter transcript upon overexpression of TIAR-TAP protein. The alternatively spliced RNA species are indicated.
Lower panel: analysis of TIAR-TAP expression in 293T cells by western blot analysis using anti-TIAR sera. (B) Western blot analysis of TIAR-

ASF/SF2
hnRNP M DDX21
p68
BOIP-flag
TIAR-flag
BOIP-flag
HA-p68
WB anti-flag
WB anti-HA
HA-ASF/SF2
RNaseA
RNaseA
w/o w/o with
+
+
–
–+–+
++++++
–
+–+–
w/o w/o with
+
+
–
–+–+
++++++
–
+–+–
w/o w/o with
w/o w/o with

were analyzed by SDS-PAGE and western
blot analysis (WB) with anti-flag or anti-HA
sera. Transfected DNAs are indicated. The
experiments were performed in the absence
or presence of RNAse A in the cell lysate.
(B) Sub-cellular distribution of TIAR partners.
COS cells were transfected with the DNA
constructs encoding the HA-tagged interact-
ing candidates and were treated with
arsenite (1 m
M for 30 min). Cells were fixed
and stained with mouse anti-HA and goat
anti-TIAR sera. Secondary Alexa
594-coupled donkey anti-mouse (red) and
FITC-coupled anti-goat sera (green) were
used to reveal HA-tagged proteins and TIAR,
respectively. Merged figures correspond to
superpositions of signals corresponding to
HA-tagged proteins and TIAR. Nuclei were
stained with 4¢,6-diamidino-2-phenylindole
(blue).
ASF ⁄ SF2 in TIAR-mediated regulatory pathways N. Delestienne et al.
2500 FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS
is also associated with TIA-1 (Fig. 3B), thereby indicat-
ing that ASF ⁄ SF2 can interact with both TIA proteins.
Western blot analysis on whole cell extracts and
purified nuclear and cytoplasmic fractions revealed
that ASF ⁄ SF2 cytoplasmic accumulation upon oxida-
tive stress was a result of the relocalization of the pro-
tein and not an increase of ASF ⁄ SF2 gene expression

motifs mediating ASF ⁄ SF2 sub-cellular distribution
and recruitment into SGs (Fig. 5A). These constructs
were transfected into COS cells and the intracellular
distribution of the expressed proteins was analyzed by
fluorescence microscopy (Fig. 5B). Previous studies
[27] reported that the deletion of the carboxy-terminal
arginine-serine (RS)-rich domain markedly increased
the proportion of ASF ⁄ SF2 accumulated in the cyto-
plasm. We observed that the RS1 sub-domain appears
to be the main nuclear import determinant within the
RS domain because the mutant lacking this motif
(DRS1) accumulated in the cytoplasm. By contrast,
ASF ⁄ SF2 nucleo-cytoplasmic distribution is modified
neither by the removal of the RS2 sub-domain, nor by
point mutations disrupting RRM1 (FF-DD mutant)
[28] or RRM2 RNA-binding activity (W134A mutant)
[29]. However, combined inactivation of RRM1 and 2
RNA-binding activities led to a major accumulation of
ASF ⁄ SF2 in the cytoplasmic compartment.
ASF ⁄ SF2 nuclear export determinants were investi-
gated by analyzing the nucleo-cytoplasmic distribution
of wild-type and mutated forms of ASF ⁄ SF2 after
exposure of the transfected cells to actinomycin D and
cycloheximide. This treatment inhibits the transcrip-
tion-dependent nuclear import of ASF ⁄ SF2, allowing
the observation of ASF ⁄ SF2 nuclear export [30]. As
previously observed, a massive relocalization of wild-
type ASF⁄ SF2 to the cytoplasmic compartment was
detected upon actinomycin D exposure. By contrast,
the W134A mutant remained mostly nuclear under the

WB anti-flag
WB anti-flag
HA-ASF/SF2
WB anti-HA
WB anti-ASF/SF2
Fig. 3. (A) Co-immunoprecipitation of endogenous ASF ⁄ SF2 with
TIAR. The experiment was performed as described in Fig. 2A
except that 293T cells were transfected with BOIP-flag or TIAR-
Flag constructs only. ASF ⁄ SF2 detection was performed by wes-
tern blot analysis using anti-ASF ⁄ SF2 serum. (B) Co-immunoprecipi-
tation of ASF ⁄ SF2 with TIA-1 protein. The experimental procedure
was identical to that described in Fig. 2A.
N. Delestienne et al. ASF ⁄ SF2 in TIAR-mediated regulatory pathways
FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS 2501
Arsenite (min)
A
C
D
B
Arsenite (min)
0
0 120
30 60 120
Untreated
ASF/SF2 TIAR Merged
ASF/SF2 TIAR Merged
ASF/SF2 RFP-Dcp1 Merged
Arsenite
Arsenite
Heat shock

lyzed as described in (C). Arrowheads indicate eIF3b-positive foci detectable in HA-ASF ⁄ SF2 overexpressing cells.
ASF ⁄ SF2 in TIAR-mediated regulatory pathways N. Delestienne et al.
2502 FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS
serine (S) residues in the RS domain were replaced by
negatively charged aspartic acid (D) residues. Interest-
ingly, this phosphomimetic mutant was relocalized to
the cytoplasm to the same extent as the wild-type pro-
tein upon cycloheximide ⁄ actinomycin D treatment,
demonstrating the ability of this mutant to efficiently
exit the nucleus (ASF ⁄ SF2 RD; Fig. 5C). We then
analyzed the capacity of ASF ⁄ SF2 mutants to migrate
into SGs upon arsenite exposure and observed that all
of them migrated into SGs, except the double FF-DD
W134A mutant (Fig. 5D). These observations reveal
the importance of ASF ⁄ SF2 RNA-binding ability for
its sub-cellular distribution and the independent capac-
ity of both RRMs to address ASF ⁄ SF2 to SGs. More-
over, the fact that FF-DD and W134A mutants
cannot exit the nucleus (Fig. 5C) but are able to
migrate to SG (Fig. 5D) suggests that ASF ⁄ SF2
migrating in SGs originates from the cytoplasmic frac-
tion or that arsenite induces an alternative nuclear
export pathway relying on other export determinants.
To test this hypothesis, we generated DNA constructs
expressing wild-type or mutated ASF ⁄ SF2 in fusion
with a protein normally confined to the nucleus. This
protein corresponds to the nucleoplasmin core domain
(NPc) fused to the classical nuclear localization signal
(NLS) of hnRNP K. Several studies have shown that,
once carried into the nucleus, this protein does not

synergized to induce a massive accumulation of the
reporter protein in the cytoplasm, exceeding that of
the reporter protein fused with the full-length
ASF ⁄ SF2 protein, or with the RRM1 alone. Most
likely, this difference is partly a result of the absence
of the RS domain that contributes to nuclear import.
Altogether, our results indicate that ASF ⁄ SF2 nuclear
export relies on the RNA-binding capacity of both
RRMs, whereas the RS domain is dispensable. More-
over, RRM1 but not RRM2 is necessary and sufficient
to promote this cytoplasmic redistribution. However,
in the context of the wild-type ASF⁄ SF2 protein, both
RRMs play equally important roles.
Overexpression of ASF/SF2 down-regulates the
expression of an ARE-containing reporter mRNA
The association of ASF ⁄ SF2 with TIAR and TIA-1
led us to investigate whether ASF ⁄ SF2 might modulate
the expression of ARE-containing genes. Accordingly,
we tested the effect of overexpressing ASF ⁄ SF2 on the
expression of Renilla luciferase (Rluc) reporter genes
carrying (or not) eight AUUU direct repeats in the 3¢
UTR. These reporter genes were placed under the con-
trol of a bidirectional cytomegalovirus promoter medi-
ating the transcription of another reporter gene
encoding firefly luciferase (Fluc) (Fig. 6A). This strat-
egy ensured that the ratio between the control (Fluc)
and the reporter (Rluc) genes was strictly conserved in
all experiments. These reporter constructs were trans-
fected in 293T cells in combination with plasmids
encoding ASF ⁄ SF2, TTP (a mRNA destabilizing

PDADV
GSAQDL
GSAQDL
RDPSYG(RD)
8
NDRDRDYSPRRDRGSPRYSPRHDRDRDRT
PDADV
RRM2 RS
RD
RG
RS1 RS2
Untreated
HA HA
ActD/CHX
ASF/SF2 ΔRS2
ASF/SF2 FFDD
ASF/SF2 W134A
ASF/SF2 FFDD-W134A
ASF/SF2 RD
HA
HA-ASF WT
HA-ASF/SF2
HA-ASF ΔRS1
HA-ASF ΔRS2
HA-ASF FF-DD
HA-ASF W134A
HA-ASF FF-DD W134A
HA-ASF/SF2 WT
HA-ASF/SF2 WT
HA-ASF/SF2 FF-DD

NPc-ASF-NLS
NPc-ASF FFDD-NLS
NPc-ASF FFDD-NLS
NPc-ASF W134A-NLS
NPc-ASF W134A-NLS
NPc-RRM1-NLS
NPc-RRM1-NLS
NPc-RRM1 FFDD-NLS
NPc-RRM1 FFDD-NLS
NPc-RRM2-NLS
NPc-RRM2-NLS
NPc-RRM1 RRM2-NLS
NPc-RRM1 RRM2-NLS
TIAR Merged Flag TIAR Merged
Fig. 5. Subcellular distribution of ASF-SF2 mutants in actinomycin D- or arsenite-treated COS cells. (A) Schematic representation of
ASF ⁄ SF2 mutants. The amino acids bordering the different domains composing ASF ⁄ SF2 as well as the mutated residues are indicated. The
dotted lines indicate the deleted region in the different mutants. (B, D) Subcellular distribution of ASF ⁄ SF2 mutants in untreated COS cells
(B) and in COS cells treated with arsenite (D). Cells were fixed and the localization of the proteins was performed as described in Fig. 2B.
(C) Subcellular distribution of ASF ⁄ SF2 wild-type and mutated forms upon inhibition of transcription. Transfected COS cells were treated for
3 h with cycloheximide (20 lgÆmL
)1
) and actinomycin D (5 lgÆmL
)1
). HA-fused ASF ⁄ SF2 wild-type and mutant proteins were detected with
mouse anti-HA serum and alexa 594-coupled donkey secondary anti-mouse serum. (E) Subcellular distribution of Npc-NLS-Flag alone or in
fusion with ASF ⁄ SF2 domains. The experiment was performed as described in (C), except that Npc-NLS-Flag proteins were detected with
anti-Flag serum.
N. Delestienne et al. ASF ⁄ SF2 in TIAR-mediated regulatory pathways
FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS 2505
ASF/SF2 down-regulates the expression of a

WB anti-flag
Rluc
WB anti-HA
AU8 AU8AU0
AU0 AU8 AU8AU0 AU8AU0
1.00
0.29
0.54
0.60
0.95
TTP flag HA-ASF/SF2
HA-ASF/SF2
HA-ASF/SF2
FF-DD W134A
HA-ASF/SF2
FF-DD W134A
HA-ASF/
SF2ΔRS
HA-ASF/SF2
SF2ΔRS
Normalized ratio AU8/AU0
0
(AUUU)
8
Globin 3′UTR
Renilla luciferase
Fig. 6. Overexpression of ASF ⁄ SF2
down-regulates the expression of a reporter
containing AU-rich elements. (A) Schematic
representation of the bidirectional FLuc ⁄

8MS2. (B) Expression of the tagged proteins in the co-transfection experiment was monitored by western blot using anti-Flag serum. (C)
Sub-cellular localization of ASF ⁄ SF2-MS2-CP in normal and arsenite-treated cells. COS cells were transiently transfected with ASF ⁄ SF2-MS2-
CP encoding plasmid and were subsequently treated (or not) with arsenite. Cells were fixed and ASF ⁄ SF2-MS2-CP protein was detected
with anti-Flag serum as described in Fig. 4. (D) Quantification of Rluc mRNA accumulation normalized to Fluc mRNA values. After transfec-
tion, cytoplasmic RNA was isolated and analyzed by northern blot using Fluc and Rluc riboprobes. The radioactive signals were quantified by
the STORM 820 equipment and
IMAGEQUANT software (Molecular Dynamics). The Rluc ⁄ Fluc ratios are indicated as a percentage of the nor-
malized value obtained for Rluc 0MS2. (E) The effect of tethering ASF ⁄ SF2 or TTP to Rluc reporter on the protein expression per cytoplas-
mic mRNA was calculated as the ratios of the normalized Rluc activities compared to the normalized ratio of Rluc cytoplasmic mRNA of one
representative experiment.
ASF ⁄ SF2 in TIAR-mediated regulatory pathways N. Delestienne et al.
2506 FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS
other cellular proteins for the RNA-binding site and
independent of any intermediate docking factor
[26,40]. DNA constructs encoding CP alone or in
fusion with ASF ⁄ SF2 or TTP were co-transfected in
293T cells with plasmids encoding Fluc and Rluc
proteins under the control of the bidirectional cyto-
megalovirus promoter. In these experiments, the b-glo-
bin 3¢ UTR flanking the Rluc coding sequence
Protein
A
BD
CE
WB anti-flag
Localization
Cytoplasmic RNA
Protein/Cytoplasmic RNA
0MS2 8MS2
CP TTP-CP ASF/

100
65.4
65.52
9.55
20.01
*
*
**
**
**
6.3
4.6
39.1
60
40
20
Rluc/Fluc (% of control)
Rluc/Fluc (% of control)
Protein/Cytoplasmic RNA (% of control)
0
120
100
80
60
40
20
0
120
100
80

the tested proteins were expressed at similar levels in
Rluc 8MS2 and Rluc 0MS2 transfected cells (Fig. 7B).
In addition, ASF ⁄ SF2-CP spatial distribution was sim-
ilar to endogenous ASF ⁄ SF2 because it is mainly
nuclear at equilibrium and migrates into SGs upon
stress (Fig. 7C). The effect of ASF ⁄ SF2-CP on cyto-
plasmic Rluc mRNA accumulation was measured and,
to normalize the transfection and recovery efficiencies,
the accumulation of Fluc mRNA was measured in the
same RNA samples (Fig. 7D). CP and TTP-CP were
included as controls. The steady-state level of Rluc
mRNA containing eight MS2 binding sites was slightly
reduced upon expression of CP alone. The tethering of
TTP led to a decrease of 8MS2 Rluc reporter mRNA
to the same extent as that at the protein level
(Fig. 7E), therefore confirming the mRNA-destabiliz-
ing activity of TTP as previously described [39,42].
ASF ⁄ SF2 significantly decreased 8MS2 mRNA accu-
mulation, although not to the same extent as that
observed at the protein level (four- to five-fold less),
thereby suggesting that ASF ⁄ SF2 down-regulates both
mRNA stability and translation (Fig. 7D, E). It is
worth noting that the deletion of the RS domain sig-
nificantly alleviated ASF⁄ SF2 repressive activity as
observed for ARE reporter mRNA (Fig. 7A).
Discussion
In the present study, we identified new candidate pro-
teins associated with TIAR. They include ribosomal
proteins, the transcription factor UBF1 and proteins
involved in RNA metabolism, such as hnRNP M and

TIAR and TIA-1 protein complexes including
ASF ⁄ SF2 might be involved in redundant molecular
processes. However, the physiological existence of
ASF ⁄ SF2-TIA complexes is conditioned by the tissue
distribution of TIA protein [46], which is clearly more
restricted compared to ASF ⁄ SF2 [48].
ASF ⁄ SF2 migration into SGs strongly depends on its
transit in the cytoplasm, as well as on its ability to bind
RNA. Therefore, this process most likely results from
the sequestration of ASF ⁄ SF2-bound transcripts in
such cytoplasmic structures. Because both RRMs can
independently mediate ASF ⁄ SF2 migration into SGs, it
can be speculated that, although acting synergistically
for optimal interaction with RNA, both RRMs mediate
interactions with distinct motifs present in mRNA mol-
ecules addressed to SGs upon stress. ASF ⁄ SF2 is a
bona fide SG component and does not get associated
with other cytoplasmic structures such as processing
bodies. Similar to other SG components, it displays the
capacity of spontaneous SG assembly upon overexpres-
sion. However, down-regulation of ASF ⁄ SF2 expres-
sion does not alter SG assembly upon stress (Fig. S2).
Altogether, these results suggest that ASF ⁄ SF2 and
TIA proteins participate in common mechanisms of
translational repression in response to stress.
The data obtained in the present study further
describe the molecular determinants for ASF ⁄ SF2 sub-
cellular distribution. We demonstrated that the
removal of the RS1 domain, but not of the RS2
ASF ⁄ SF2 in TIAR-mediated regulatory pathways N. Delestienne et al.

ASF ⁄ SF2 in ARE-mediated post-transcriptional regu-
lation. The significant but less pronounced effect of
ASF ⁄ SF2 compared to TTP might be a result of the
differential capacity of the two proteins to be
recruited to AREs. It is already well established that
AREs can recruit several different ARE-BPs, depend-
ing on their type of AU-rich motifs, as well as the
abundance of each ARE-BP [53]. Interestingly, the
down-regulating effect of ASF ⁄ SF2 is preserved upon
inactivation of its RNA-binding ability by point muta-
tions in both RRM RNA-binding motifs (FF-DD
W134A mutant). These results suggest that ASF ⁄ SF2
is recruited to the ARE reporter mRNA by intermedi-
ate protein–protein interactions, in contrast to its
migration into SGs, which directly relies on RRM
RNA-binding activities. The removal of the RS
domain completely reversed ASF ⁄ SF2 down-regulat-
ing activity on the ARE reporter mRNA. The mutant
lacking the RS domain might become associated with
the ARE reporter mRNA via interactions with ARE-
BPs but is inactive in down-regulating mRNA transla-
tion and stability. Of note, the mutant lacking the
RS domain still becomes associated with TIAR in
co-immunoprecipitation assays (Fig. S3). However,
down-regulation of the ARE reporter mRNA by
ASF ⁄ SF2 could not be abrogated by the expression of
a RNA-binding defective TIAR mutant (data not
shown), therefore suggesting that the targeting of
ARE-containing mRNAs by ASF ⁄ SF2 can occur via
several different interactions.

mRNA 3¢ UTR.
In conclusion, the present study highlights the
involvement of ASF ⁄ SF2 in post-transcriptional down-
regulating mechanisms in both normal and stressed
cells. It appears that, similar to other RNA-binding
proteins, such as AUF1 [55] and HuR [56], ASF ⁄ SF2
differentially modulates the fate of transcripts with
which it becomes associated.
Materials and methods
Materials
Enzymes were purchased from Invitrogen (Carlsbad, CA,
USA) and Roche (Basel, Switzerland); oligonucleotides
were obtained from Sigma (St Louis, MO, USA); and cell
culture media were obtained from Invitrogen.
N. Delestienne et al. ASF ⁄ SF2 in TIAR-mediated regulatory pathways
FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS 2509
DNA constructs
A pcDNA3.1-TAP construct was obtained by inserting the
sequence coding for TAP into the SfuI ⁄ PmeI sites of
the pcDNA3.1(+) plasmid. The sequence coding for the
C-terminal TAP tag was amplified by PCR using pBS1479
as template DNA. pBS1479 was kindly provided by D.
Lafontaine (Universite
´
Libre de Bruxelles, Belgium) and
has been described previously [57]. pcDNA3.1-TIAR-TAP
construct was generated by inserting murine TIARb coding
region (accession number: AAC52870) between the EcoRI
and BamHI sites of the pcDNA3.1-TAP plasmid. pcDNA-
BOIP-Flag construct was generated by inserting PCR prod-

After approximately 3 weeks of selection, when isolated col-
onies appeared, drug resistant clones were isolated by limit
dilutions of transfection pools.
Purification of TIAR-TAP complexes
NIH 3T3 cells (6 · 10
8
) stably transfected with the
pcDNA3.1-TAP or with the pcDNA3.1-TIAR-TAP con-
struct were lysed in 20 mL of lysis buffer containing 50 mm
Tris (pH 8.0), 150 mm NaCl, 0.1% NP40 and a cocktail of
protease inhibitors (Roche). The lysate was cleared by cen-
trifugation for 30 min at 8000 g. The supernatant (200 mg
of protein extract) was incubated overnight with 150 lL
rabbit IgG Sepharose 6 Fast Flow (GE Healthcare UK Ltd
Amersham, Little Chalfont, UK) agarose beads equili-
brated in IgG binding buffer (IgG BB: 0.15 m NaCl, 0.1%
NP40, 50 mm Tris, pH 8.0) in the presence of RNAse A
(10 lgÆmg
)1
extract). After 15 h of binding at 4 °C, beads
were washed twice with 1 mL of IgG BB and once with
1 mL of Tev cleavage buffer (Tev CB: 25 mm Tris, pH 8.0,
0.15 m NaCl, 0.1% NP40, 0.5 mm EDTA and 1 mm dith-
ithreitol). Three hundred units of Tev protease (Invitrogen)
were added and cleavage of the TAP tag was performed in
1 mL of Tev CB for 3 h at room temperature. Proteins
released from the beads were collected in two fractions of
1 mL, and the eluate was adjusted to 2 mm CaCl
2
before

HEK 293T cells were transiently transfected using
FUGENE-6 reagent (Roche) in accordance with the manu-
facturer’s instructions. pcDNA3.1-BOIP-Flag, pcDNA3.1-
TIAR-Flag or pcDNA3.1-TIA-1-Flag (accession number:
AAC52871) [5] plasmids were co-transfected with pRK5-
NHA constructs encoding the candidate partners in
1 · 10
6
cells. Forty eight hours after transfection, cells were
washed twice in ice-cold NaCl ⁄ P
i
and lysed in IP buffer
(50 mm Tris-HCl, pH 8.0, 100 mm NaCl, 0.2% NP40,
ASF ⁄ SF2 in TIAR-mediated regulatory pathways N. Delestienne et al.
2510 FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS
protease inhibitors) for 30 min at 4 °C. Flagged-proteins
were immunoprecipitated by incubating protein extract
(2 mg) with BSA (1%) blocked-anti-Flag M2 (20 lL) affin-
ity gel (Sigma) for 2 h at 4 °C. The beads were washed
three times in IP buffer and were then incubated for 30 min
at room temperature in IP buffer containing RNAse A
(10 lgÆmg
)1
extract) and washed once in IP buffer. Bound
proteins were eluted in Laemmli gel sample buffer, sepa-
rated on 12.5% SDS-PAGE and transferred to nitrocellu-
lose for western blot analysis. Western blot analysis of HA-
and Flag-tagged proteins was performed as described previ-
ously [60] using anti-HA serum (dilution 1 : 5000) and M2
monoclonal anti-Flag serum (dilution 1 : 2000), respec-

i
, permeabilized for 5 min with NaCl ⁄ P
i
, 0.5%
Triton-X 100 at 4 °C and washed again as described above.
Blocking was performed with NaCl ⁄ P
i
containing 10%
BSA for 30 min. After blocking, the coverslips were incu-
bated for 1 h with antibodies diluted in NaCl ⁄ P
i
, 0.1%
Tween 20. Antibodies were used at a dilution of 1 : 50 for
anti-ASF ⁄ SF2 (Zymed Laboratories; Invitrogen), goat anti-
TIAR C18 and anti-eiF3b (Santa Cruz Biotechnology,
Santa Cruz, CA, USA) sera; at a dilution of 1 : 5000 for
mouse anti-FLAG M2 serum; and at a dilution of 1 : 30
000 for anti-HA serum (Sigma). The coverslips were subse-
quently washed three times for 10 min with NaCl ⁄ P
i
, 0.1%
Tween 20 and incubated with the secondary antibody in
NaCl ⁄ P
i
, 0.1% Tween 20. Alexa594-conjugated donkey
anti-goat serum was used at a dilution of 1 : 25 000. In
double staining experiments, a donkey anti-goat serum
conjugated with fluorescein isothiocyanate (FITC) was used
(dilution 1 : 1000). After 1 h, coverslips were washed three
times as above, rapidly rinsed with desionized water

were pelleted, two-thirds of the cytoplasmic extract was
recovered and mixed with an equal volume of SDS buffer
(0.2 m Tris, pH 7.5, 0.3 m NaCl, 25 mm EDTA, 2% SDS)
and twice extracted with phenol ⁄ chloroform, once extracted
with chloroform, and precipitated with ethanol and sodium
acetate [62]. The quality of the RNA samples was verified
by agarose gel electrophoresis before loading on a 1.5%
agarose gel and conducting northern blot analysis. Renilla
and firefly luciferase antisense RNA probes were generated
by in vitro transcription using linearized DNA templates in
the presence of 80 lCi [a-
32
P]UTP (800 CiÆmmol
)1
) and
20 lm UTP. Quantitative analysis of northern blots was
performed using a phosphorimager (STORM 820; Molec-
ular Dynamics, Sunnyvale, CA, USA) and imagequant
software (Molecular Dynamics).
Acknowledgements
We thank Sylvain Lestrade for providing excellent
technical assistance in the proteomic analysis, Domi-
nique Weil for the Dcp1-RFP construct, Luc Paillard
N. Delestienne et al. ASF ⁄ SF2 in TIAR-mediated regulatory pathways
FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS 2511
for the bidirectional reporter genes and Fabienne
Konczak for helping us test the splicing activity of
TIAR-TAP protein. This work was funded by the
DGTRE (Re
´

´
nez C, Granneman
S, Se
´
raphin B, Anderson P & Valca
´
rcel J (2000) The
apoptosis-promoting factor TIA-1 is a regulator of
alternative pre-mRNA splicing. Mol Cell 6, 1089–1098.
7 Del Gatto F, Plet A, Gesnel MC, Fort C & Breathnach
R (1997) Multiple interdependent sequence elements
control splicing of a fibroblast growth factor receptor 2
alternative exon. Mol Cell Biol 17, 5106–5116.
8 Zhu H, Hasman RA, Young KM, Kedersha NL & Lou
H (2003) U1 snRNP-dependent function of TIAR in
the regulation of alternative RNA processing of the
human calcitonin ⁄ CGRP pre-mRNA. Mol Cell Biol 23,
5959–5971.
9 Yu Q, Cok SJ, Zeng C & Morrison AR (2003) Transla-
tional repression of human matrix metalloproteinases-
13 by an alternatively spliced form of T-cell-restricted
intracellular antigen-related protein (TIAR). J Biol
Chem 278, 1579–1584.
10 Kandasamy K, Joseph K, Subramaniam K, Raymond
JR & Tholanikunnel BG (2005) Translational control of
beta2-adrenergic receptor mRNA by T-cell-restricted
intracellular antigen-related protein. J Biol Chem 280 ,
1931–1943.
11 Kedersha N & Anderson P (2002) Stress granules: sites
of mRNA triage that regulate mRNA stability and

nucleo-cytoplasmic SR protein functions in viral
IRES-mediated translation initiation. EMBO J 26,
459–467.
18 Rigaut G, Shevchenko A, Rutz B, Wilm M, Mann M
&Se
´
raphin B (1999) A generic protein purification
method for protein complex characterization and prote-
ome exploration. Nat Biotechnol 17, 1030–1032.
19 Le Guiner C, Lejeune F, Galiana D, Kister L,
Breathnach R, Ste
´
venin J & Del Gatto-Konczak F
(2001) TIA-1 and TIAR activate splicing of alternative
exons with weak 5¢ splice sites followed by a U-rich
stretch on their own pre-mRNAs. J Biol Chem 276,
40638–40646.
20 Van Wayenbergh R, Taelman V, Pichon B, Fischer A,
Kricha S, Gessler M, Christophe D & Bellefroid EJ
(2003) Identification of BOIP, a novel cDNA highly
expressed during spermatogenesis that encodes a protein
interacting with the orange domain of the hairy-related
transcription factor HRT1 ⁄ Hey1 in Xenopus and
mouse. Dev Dyn 228, 716–725.
21 Anderson P & Kedersha N (2006) RNA granules. J Cell
Biol 172, 803–808.
22 Tourrie
`
re H, Chebli K, Zekri L, Courselaud B,
Blanchard JM, Bertrand E & Tazi J (2003) The

´
ceres JF & Krainer AR (1993) Functional analysis of
pre-mRNA splicing factor SF2 ⁄ ASF structural
domains. EMBO J 12, 4715–4726.
29 Tintaru AM, Hautbergue GM, Hounslow AM, Hung
ML, Lian LY, Craven CJ & Wilson SA (2007) Struc-
tural and functional analysis of RNA and TAP binding
to SF2 ⁄ ASF. EMBO Rep 8, 756–762.
30 Ca
´
ceres JF, Screaton GR & Krainer AR (1998) A specific
subset of SR proteins shuttles continuously between the
nucleus and the cytoplasm. Genes Dev 12, 55–66.
31 Gui JF, Tronche
`
re H, Chandler SD & Fu XD (1994)
Purification and characterization of a kinase specific for
the serine- and arginine-rich pre-mRNA splicing factors.
Proc Natl Acad Sci USA 91, 10824–10828.
32 Colwill K, Feng LL, Yeakley JM, Gish GD, Ca
´
ceres
JF, Pawson T & Fu XD (1996) SRPK1 and Clk ⁄ Sty
protein kinases show distinct substrate specificities for
serine ⁄ arginine-rich splicing factors. J Biol Chem 271,
24569–24575.
33 Huang Y, Yario TA & Steitz JA (2004) A molecular
link between SR protein dephosphorylation and
mRNA export. Proc Natl Acad Sci USA 101,
9666–9670.

Blackshear PJ (2006) Novel mRNA targets for triste-
traprolin (TTP) identified by global analysis of stabi-
lized transcripts in TTP-deficient fibroblasts. Mol Cell
Biol 26, 9196–9208.
43 Shevchenko A, Schaft D, Roguev A, Pijnappel P, Stew-
art AF & Shevchenko A (2002) Deciphering protein
complexes and protein interaction networks by tandem
affinity purification and mass spectrometry. Mol Cell
Proteomics 1, 204–212.
44 Trinkle-Mulcahy L, Boulon S, Wah Lam Y, Urcia R,
Boisvert F-M, Vandermoere F, Morrice NA, Swift S,
Rothbauer U, Leonhardt H et al. (2008) Identifying
specific protein interaction partners using quantitative
mass spectrometry and bead proteomes. J Cell Biol 183,
223–239.
45 Rousseau S, Morrice N, Peggie M, Campbell DG,
Gaestel M & Cohen P (2002) Inhibition of SAP-
K2a ⁄ p38 prevents hnRNP A0 phosphorylation by
MAPKAP-K2 and its interaction with cytokine
mRNAs. EMBO J 21, 6505–6514.
46 Beck AR, Medley QG, O’Brien S, Anderson P & Streuli
M (1996) Structure, tissue distribution and genomic
organization of the murine RRM-type RNA binding
proteins TIA-1 and TIAR. Nucleic Acids Res 24,
3829–3835.
47 Piecyk M, Wax S, Beck ARP, Kedersha N, Gupta M,
Maritim B, Chen S, Gueydan C, Kruys V, Streuli M
et al. (2000) TIA-1 is a translational silencer that
selectively regulates the expression of TNF-a. EMBO J
19, 4154–4163.

ceres JF (2008) The
splicing factor SF2 ⁄ ASF regulates translation initiation
by enhancing phosphorylation of 4E-BP1. Mol Cell 30,
179–189.
55 Liao B, Hu Y & Brewer G (2007) Competitive binding
of AUF1 and TIAR to MYC mRNA controls its trans-
lation. Nat Struct Mol Biol 14, 511–518.
56 Katsanou V, Papadaki O, Milatos S, Blackshear PJ,
Anderson P, Kollias G & Kontoyiannis DL (2005)
HuR as a negative posttranscriptional modulator in
inflammation. Mol Cell 19, 777–789.
57 Puig O, Rutz B, Luukkonen BG, Kandels-Lewis S,
Bragado-Nilsson E & Se
´
raphin B (1998) New constructs
and strategies for efficient PCR-based gene manipula-
tions in yeast. Yeast 14, 1139–1146.
58 Lykke-Andersen J, Shu MD & Steitz JA (2000) Human
Upf proteins target an mRNA for nonsense-mediated
decay when bound downstream of a termination codon.
Cell 103, 1121–1131.
59 Gobom J, Nordhoff E, Mirgorodskaya E, Ekman R &
Roepstorff P (1999) Sample purification and prepara-
tion technique based on nano-scale reversed phase
columns for the sensitive analysis of complex peptide
mixtures by matrix-assisted laser desorption ⁄ ionization
mass spectrometry. J Mass Spectrom 34, 105–116.
60 Rothe
´
F, Gueydan C, Bellefroid E, Huez G & Kruys

2514 FEBS Journal 277 (2010) 2496–2514 ª 2010 The Authors Journal compilation ª 2010 FEBS