TAR DNA-binding protein 43 (TDP-43) regulates
stress granule dynamics via differential regulation
of G3BP and TIA-1
Karli K. McDonald1,2, Anaı ¨s Aulas1,2, Laurie Destroismaisons1,2, Sarah Pickles1,2,
Evghenia Beleac1,2, William Camu3, Guy A. Rouleau1,2and Christine Vande Velde1,2,∗
1Centre d’excellence en neuromique de l’Universite ´ de Montre ´al, Centre de recherche du CHUM (CRCHUM) and
2De ´partement de me ´decine, Universite ´ de Montre ´al, 1560 rue Sherbrooke Est, Montre ´al, QC, Canada H2L 4M1
and3Unite ´ de Neurologie Comportementale et De ´ge ´ne ´rative, Institute of Biology, 34967 Montpellier, France
Received October 11, 2010; Revised December 21, 2010; Accepted January 14, 2011
TAR deoxyribonucleic acid-binding protein 43 (TDP-43) is a multifunctional protein with roles in transcription,
pre-messenger ribonucleic acid (mRNA) splicing, mRNA stability and transport. TDP-43 interacts with other
heterogeneous nuclear ribonucleoproteins (hnRNPs), including hnRNP A2, via its C-terminus and several
hnRNP family members are involved in the cellular stress response. This relationship led us to investigate
the role of TDP-43 in cellular stress. Our results demonstrate that TDP-43 and hnRNP A2 are localized to
stress granules (SGs), following oxidative stress, heat shock and exposure to thapsigargin. TDP-43 contrib-
utes to both the assembly and maintenance of SGs in response to oxidative stress and differentially regulates
key SGs components, including TIA-1 and G3BP. The controlled aggregation of TIA-1 is disrupted in the
absence of TDP-43 resulting in slowed SG formation. In addition, TDP-43 regulates the levels of G3BP
mRNA, a SG nucleating factor. The disease-associated mutation TDP-43R361Sis a loss-of-function mutation
with regards to SG formation and confers alterations in levels of G3BP and TIA-1. In contrast, a second
mutation TDP-43D169Gdoes not impact this pathway. Thus, mutations in TDP-43 are mechanistically diver-
gent. Finally, the cellular function of TDP-43 extends beyond splicing and places TDP-43 as a participant
of the central cellular response to stress and an active player in RNA storage.
TAR deoxyribonucleic acid (DNA)-binding protein
(TDP-43) was first described in the transcriptional regulation
sequence motifs in TAR DNA (1). In addition to this initially
described role, TDP-43 is now known to be involved in
several aspects of RNA metabolism, including transcription,
alternative splicing, pre-messenger ribonucleic acid (mRNA)
stability and mRNA transport (2–4). TDP-43 is composed of
414 amino acids and has all of the structural features character-
istic of a heterogeneous nuclear ribonucleoprotein (hnRNP),
including two highly conserved ribonucleic acid (RNA) recog-
nition motifs (RRM1 and RRM2), a nuclear localization signal
and a glycine-rich C-terminal tail (5,6). The glycine-rich C-
terminal region is required for its exon skipping and inhibitory
splicing activities and, as with other hnRNPs, this domain med-
iates protein–protein interactions (7). Indeed, a portion of this
region (residues 321–366) mediates a direct interaction
between TDP-43 and hnRNP A2 (2). Protein–protein
interactions between hnRNPs are suspected to contribute to
RNA–protein complex formation as well as direct RNA–
protein interaction between hnRNPs and mRNAs (8).
In human cells, hnRNPs are concentrated in the nucleus
in physiologically normal conditions. However, a subset
(ex. hnRNP A1, K and Q) continuously shuttle between the
nucleus and cytoplasm (9). hnRNPs are involved in the exten-
sive processing of pre-mRNAs in the nucleus, which are sub-
sequently transported to the cytoplasm. Several studies
indicate that certain hnRNPs are directly involved in the cellu-
lar response to various stress stimuli. For example, the acti-
vation of the p38 stress-signaling pathway in mammalian
∗To whom correspondence should be addressed. Tel: +1 5148908000; Fax: +1 5144127602; Email: email@example.com
# The Author 2011. Published by Oxford University Press. All rights reserved.
For Permissions, please email: firstname.lastname@example.org
Human Molecular Genetics, 2011
HMG Advance Access published January 27, 2011
at CHUM- HOTEL-DIEU on January 28, 2011
cells results in both the phosphorylation and cytoplasmic
accumulation of hnRNP A1 in stress granules (SGs) (10).
Similarly, hnRNP Q redistributes to the cytoplasm and par-
tially co-localizes to SGs and processing bodies (PBs) under
specific stress conditions [e.g. thapsigargin (THAP), heat
shock (HS) and arsenite] (11).
Depending on the type of cellular stress encountered, a
variety of signaling pathways can be activated which ulti-
mately modulate gene expression patterns either transcription-
ally or post-transcriptionally (12). RNA-binding proteins
(RBPs) play a major role in post-transcriptional regulation
during stress yielding global repression of protein translation
(12,13). This is facilitated by the formation of SGs which
are cytoplasmic domains housing translationally arrested
mRNAs (10). SGs are also now considered to be dynamic
triage centers that sort mRNA for storage, decay or
re-initiation during stressful conditions (14,15). The assembly
of SGs can be induced by a variety of stimuli, including HS,
hypoxia, osmotic and oxidative stress, and typically involves
the phosphorylation of the eukaryotic initiation factor [eukary-
otic translation initiation factor 2 alpha (eIF2a)]. This phos-
depletion of the eIF2–GTP–tRNA-met ternary complex,
thus permitting the RBP TIA-1 to bind the 48S complex
instead of the ternary complex. This promotes polysome disas-
sembly and the consequent recruitment of mRNAs to SGs
(10). SGs gradually disperse once the stress is removed (12).
Since TDP-43 shares so many features with other hnRNPs,
it is reasonable to suspect that it may also play a role in cellu-
lar stress responses. Indeed, it has recently been reported that
TDP-43 is localized to SGs following oxidative stress or pro-
teasome inhibition (16–18). However, there is currently no
data to determine the functional role for TDP-43 in SGs and
the cellular stress response.
We report here that TDP-43 and its binding partner hnRNP
A2 are components of SGs arising from oxidative stress. Fur-
thermore, TDP-43 down-regulation influences the stoichi-
ometry of other SG protein components, including TIA-1
and G3BP. Moreover, TDP-43 contributes to SG formation
and maintenance. In patient lymphoblastoid cells, at least
one amyotrophic lateral sclerosis (ALS)-causing mutation in
TDP-43 impacts SG formation and TIA-1 and G3BP levels.
Our data suggest that not all TDP-43 mutations have the
same mechanism and clearly define an active role for
TDP-43 in the cellular response to oxidative stress.
Endogenous TDP-43 is localized to SGs
Various types of cellular stress insults are known to affect the
expression and localization of several hnRNPs with some
localizing to SGs (10,14). Since TDP-43 is a bona fide
hnRNP family member, we investigated whether cellular
stress could also affect endogenous TDP-43. To address this,
HeLa cells were exposed to three well-established cellular
stress conditions: sodium arsenite (SA), HS and THAP. SA
treatment is a well-characterized model of oxidative stress,
while THAP induces ER stress via calcium pump dysregula-
tion.Endogenous TDP-43 waslocalizedtodistinct
cytoplasmic puncta upon SA, HS and THAP treatment,
whereas TDP-43 remained largely in the nucleus of untreated
cells (Fig. 1A). Double-labeling demonstrates that these
TDP-43 puncta co-localize with the well-described SG
marker, TIA-1 in SA-, HS- and THAP-treated cells, consistent
with published reports on TDP-43 localization to SGs with
oxidative stress (16,17) (Fig. 1A). Biochemically, the for-
mation of SGs is marked by the enhanced protein insolubility
of SG proteins (and consequent depletion of the soluble pool)
such as is described for TIA-1 (19). Indeed, soluble TDP-43
protein levels were decreased in HS- and SA-treated cells
(37 and 56%, respectively) compared with untreated cells,
and there was a reciprocal increase in TDP-43 in the insoluble
fractions (Fig. 1B). However, THAP treatment did not yield
similar changes in protein solubility.
To determine whether TDP-43 is also a resident of other
RNA granules, such as PBs, we double-labeled SA-stressed
HeLa cells for the PB marker, GW182. TDP-43 was not
robustly co-localized with GW182-labeled PBs (Fig. 1C),
but rather was often located in close proximity to PBs. This
is consistent with published data demonstrating the close jux-
tapositioning of SGs and PBs (20). Thus, TDP-43 is a resident
of SGs but not PBs following acute exposure to certain stress-
ful stimuli, in agreement with recent reports (16–18).
A cell’s encounter with external stress is marked by the
global repression of protein translation, and this is reflected
by the phosphorylation of the eIF2a (12). Since we observed
eIF2a phosphorylation most robustly in SA-treated samples,
we focused on SA as a stress paradigm for subsequent exper-
iments (Fig. 1D).
TDP-43 impacts SG assembly and disassembly
Several cellular proteins are required for SG assembly, and
their reduction is associated with a muted SG response
(19,21–23). SGs are dynamic entities such that oxidative
stress induces TIA-1 to redistribute from the nucleus to the
cytoplasm and aggregate to form SGs (12). This process
occurs over ?30 min (19). Once the stress is removed, SGs
continue to aggregate and increase in size for ?1–2 h
before gradually resolving. Typically, SGs are completely
resolved within 4 h of SA treatment. To determine whether
TDP-43 might impact SG dynamics, we assessed the kinetics
of SG formation and resolution via indirect immunofluores-
cent labeling for TIA-1 in cells depleted of TDP-43 via
small interfering RNA (siRNA) and then exposed to SA.
The number of cells exhibiting SGs was determined at 0, 15,
20 and 30 min during SA (‘stress’) exposure. SG-positive
cells were defined as cells containing at least two TIA-1-
labeled foci with a minimum area of 0.75 mm2. Compared
with control siRNA, siRNA directed to exon 1 of TDP-43
was very effective in reducing endogenous TDP-43 expression
(84%, n ¼ 5; Fig. 2A). We observed that the assembly of
TIA-1-labeled SGs was delayed in cells depleted of TDP-43
compared with cells treated with control siRNA (Fig. 2B).
In particular, we noted a modest but significant reduction in
the number of cells with SG formation at 30 min following
SA exposure (97 versus 81%, P ¼ 0.03) as well as a marked
change in SG size which will be discussed below. Similar
2Human Molecular Genetics, 2011
at CHUM- HOTEL-DIEU on January 28, 2011
findings were observed in neuroblastoma cells (Supplementary
Material, Fig. S1A).
To assess the impact of TDP-43 on SG resolution, we also
assessed SG labeling at various time points following the
removal of SA (‘release’). After 30 min of oxidative stress,
TIA-1-labeled SGs are detectable in 97% of control siRNA
cells, and are small, numerous and clearly defined (Fig. 2C,
tinues to be .94%; however, SGs become larger and fewer in
number due to the fusion of several smaller SGs and acquire a
well defined, compact shape (12). SGs then begin to gradually
resolve and, by 270 min (4 h release), few if any cells contain
SGs (Fig. 2C). Our observations of this process are consistent
with the published kinetics of SG disassembly (12). In
TDP-43-depleted cells, there is a striking difference in SG
dynamics such that TDP-43-depleted cells are slower to form
morphologically distinct SGs compared with control (Fig. 2C,
right panel). This difference is most obvious in the first 30 min
after SA treatment as the number of SG-positive cells continues
to be reduced at 45 and 60 min post-stress, compared with
control siRNA cells (Fig. 2B).
SGs in TDP-43-depleted cells have more diffuse TIA-1
labeling (i.e. TIA-1 is not completely localized as distinct
foci) and SGs appear smaller with a less-defined and more
irregular morphology. Thus, SGs in TDP-43-depleted cells
are visibly different compared with the prominent SGs
formed in control siRNA cells following the same stress
stimulus (Fig. 2C). Quantification of SG size reveals that
average SG size is reduced 43% in TDP-43-depleted cells
compared with controls(0.83+0.08versus1.46+
Figure 1. TDP-43 is localized to SGs. (A) HeLa cells cultured on cover slips were treated with 0.5 mM SA (30 min), 1 mM THAP (50 min; TH) and 438C HS
(30 min), or left as untreated (UN) and subsequently immunolabeled for TDP-43 and the SG marker TIA-1. TDP-43 localization to SGs is indicated by line scans
of the merged images showing the overlap between red and green signals. Scale bar, 10 mm. (B) RIPA-soluble and -insoluble extracts were prepared from
stressed cells and then immunoblotted for TDP-43. Actin is included as a loading control. Histograms report the mean+SEM of three independent experiments.
∗P , 0.05. (C) TDP-43 was not markedly co-localized with the PB marker, GW182. (D) Phosphorylation of eIF2a was assessed in the four stress conditions.
Human Molecular Genetics, 20113
at CHUM- HOTEL-DIEU on January 28, 2011
0.12 mm2; P ¼ 0.0002) (Fig. 2D). A similar reduction in SG
size was also observed in human neuroblastoma cells
(SK-N-SH) depleted of TDP-43 (Supplementary Material,
Our data indicate that the lag in SG assembly in siTDP-43
cells is overcome by 90 min, such that the number of cells
with SGs is equivalent in both TDP-43 and control siRNA
cells,andSG sizeand morphology are
However, at 150 min, only 1% of TDP-43 siRNA cells have
obvious SGs, while SGs persist in 31% of control siRNA
cells (P ¼ 0.04) (Fig. 2C). Therefore, depletion of TDP-43
results in a lag in SG assembly such that SG nucleation and
secondary aggregation seem to be delayed. Furthermore,
SGs in TDP-43-depleted cells are smaller and only later
attain normal-appearing SGs, suggesting a role for TDP-43
in TIA-1 aggregation and SG coalescence. Lastly, these SGs
are not sustained and quickly resolve. Thus, endogenous
TDP-43 contributes to the establishment of SGs and is
required to maintain SGs.
To determine whether TDP-43 depletion confers increased
vulnerability to exogenous stress, we evaluated cell death
via Annexin V labeling after SA exposure (Supplementary
Figure 2. TDP-43 regulates SG dynamics. (A) TDP-43 protein levels are reduced by siRNA. Values indicate TDP-43 present as determined by densitometry and
are the average of five independent experiments. Data were normalized to actin. (B) SG formation and resolution were assessed in HeLa cells transfected with
control or TDP-43 siRNA for 72 h and subsequently treated with SA. Cover slips were collected at 0, 15, 20 and 30 min after addition of SA to assess formation
(‘stress’). After 30 min, the media were replaced and cover slips were collected at 45, 60, 90, 150 and 270 min after the addition of SA (‘Release’). SGs were
identified by TIA-1 labeling and cells were scored as positive when they had at least two foci of a minimal size of 0.75 um2. Four fields per condition, represent-
ing at least 100 cells, were imaged and the number of cells containing SGs was counted. The means of three independent experiments+SEM are plotted.∗P ,
0.05. (C) SG morphology in siTDP-43 or siControl HeLa cells treated with SA at ‘release’ time points labeled with TIA-1. Scale bar, 10 mm. (D) SGs are smaller
in cells transfected with TDP-43 siRNA. The average area+SEM is plotted at 30 min post-stress.
4 Human Molecular Genetics, 2011
at CHUM- HOTEL-DIEU on January 28, 2011
Material, Fig. S2). Twenty-four hours after the removal of SA,
TDP-43-depleted cells were found to be more sensitive than
control siRNA cells following acute exposure to oxidative
stress (65+3.5 versus 49+6%; P ¼ 0.05). Thus, TDP-43
positively contributes to cellular recovery following an acute
It has been reported that exogenous expression of some SG
components is sufficient to nucleate SGs (19,22). Since our
earlier data suggest a role for TDP-43 in SG nucleation, we
assessed SG formation in cells transiently transfected with
green fluorescent protein-tagged full-length TDP-43. In our
hands, overexpression of TDP-43 was not sufficient to nucle-
ate SGs but does itself correctly localize to SGs following SA
treatment (data not shown) (16,18).
TDP-43 differentially regulates SG nucleating proteins
TIA-1 and G3BP are both considered primary nucleators of
SGs (24). We hypothesized that TDP-43 may regulate the
levels of these proteins, either at the protein or mRNA level.
Immunoblot analysis of steady-state levels of key SG proteins,
including TIA-1, TIAR and G3BP, revealed that TIA-1 was
up-regulated 130% (P ¼ 0.03) in TDP-43 siRNA cells, while
G3BP was down-regulated 79% (P ¼ 0.05; Fig. 3A). At the
revealed marked alterations in steady-state mRNA levels.
Specifically, we noted a 2.6-fold increase (P ¼ 0.01) in
TIA-1, while G3BP mRNA was reduced 3-fold (P ¼ 0.007;
Fig. 3B). A trend towards upregulation of TIAR was also
noted at both the protein and mRNA level, but it did not
reach statistical significance (P ¼ 0.059). Moreover, the
effect of TDP-43 depletion was selective for TIA-1 and
G3BP since another SG component HuR remained unchanged
(Fig. 3B). Thus, TDP-43 differentially modulates the mRNA
levels of SG-nucleating proteins.
The phosphorylation of the eukaryotic initiation factor
eIF2a at serine 51 is an important early initiating step in SG
assembly in response to oxidative stress (10). To determine
whether eIF2a signaling was intact in TDP-43-depleted
cells, we immunoblotted control and TDP-43 siRNA cell
lysates in the presence and absence of SA with an antibody
specific for serine-51 phospho-eIF2a. Following SA treat-
ment, wedid notobserve
phospho-eIF2a levels in the presence or absence of TDP-43,
placing TDP-43 downstream of this step of the stress response
(Fig. 3C). Moreover, in untreated conditions, the depletion of
TDP-43 itself did not induce phosphorylation of eIF2a
suggesting that the removal of TDP-43 does not outright
trigger an intracellular stress response.
TDP-43 impacts G3BP SGs
Given the significant down-regulation of G3BP in TDP-43
siRNA cells, we investigated the ability of cells to form
G3BP-labeled SGs immediately following SA treatment.
Control siRNA cells treated with SA contain numerous large
SGs labeled with both G3BP and TIA-1. In contrast, when
TDP-43 is reduced, there is a generally more diffuse labeling
of G3BP and the number of cells with SGs is reduced (Fig. 4).
In cells which do form SGs, it appears that the number of
G3BP-labeled SGs is markedly reduced and these SGs are
much reduced in size compared with their control counter-
parts. TIA-1 and G3BP co-localization appears to be main-
tained (Fig. 4).
Endogenous hnRNP A2 is a resident of SGs
hnRNP A2 directly interacts with TDP-43, albeit the func-
tional significance of this interaction remains to be defined
(2). Therefore, we hypothesized that hnRNP A2 may also be
a component of SGs and/or influenced by TDP-43. Following
SA exposure, endogenous hnRNP A2 redistributed so that it
was co-localized with TIA-1 (Fig. 5A). Moreover, triple label-
ing revealed that hnRNP A2 co-localized with TDP-43 in SGs
(Fig. 5B) indicating that endogenous TDP-43 and hnRNP A2
are residents of the same SGs in conditions of oxidative stress.
We also examined whether TDP-43 is required for hnRNP A2
localization to SGs. Thus, we examined the localization of
hnRNP A2 to SGs in cells treated with TDP-43 siRNA and
subjected to SA treatment. As before, fewer cells with SGs
were observed and SG size was decreased following TDP-43
depletion (Fig. 5C). hnRNP A2 labeling in the absence of
TDP-43 showed few distinct puncta which did colocalize to
TIA-1-marked SGs. However, the cells appeared to have an
Figure 3. TDP-43 regulates G3BP and TIA-1. (A) Western blot analysis of
soluble fractions of control and TDP-43 siRNA cells indicates decreased
levels of G3BP and increased accumulation of TIA-1 and to a lesser extent
TIAR. Data were normalized to actin via densitometry. Data from two to
three independent experiments are expressed as the mean fold change+
SEM relative to siControl cells,∗P , 0.05. (B) qPCR analysis of G3BP,
TIA-1 and TIAR. HuR remained unchanged. Data normalized to b-actin
and fold change are plotted.∗P , 0.01. (C) TDP-43 siRNA is not sufficient
to induce eIF2a phosphorylation, and this event is not disrupted by TDP-43
Human Molecular Genetics, 20115
at CHUM- HOTEL-DIEU on January 28, 2011
increased patchy distribution within the cytosol compared with
control siRNA cells. Thus, TDP-43 influences hnRNP A2
localization to SGs (Fig. 5C).
SG assembly is disrupted by TDP-43R361Smutation
In a cellular overexpression model, mutant fused in sarcoma/
translocated in liposarcoma (FUS/TLS) has recently been
described to be sufficient to induce SGs (25–27). Thus, we
hypothesized that mutations in TDP-43 may alter SG for-
mation in response to oxidative stress. In order to avoid arti-
facts potentially introduced by transient overexpression of
TDP-43, we took advantage of patient lymphoblasts expres-
sing physiological levels of wild-type and two different
TDP-43 mutations, TDP-43D169G(located in RRM1) and
(located in C-terminus). We verified that
steady-state levels of the TDP-43 protein were not reduced
in the different cells (Fig. 6A). In fact, a slight increase in
mutant TDP-43 protein levels was observed such that
1.2-fold (P ¼ 0.04) and 1.3-fold (P ¼ 0.01) relative to controls
cells expressing wild-type TDP-43. Lymphoblasts were
treated with SA, and SG formation was scored as described
earlier using TIA-1 as a marker. In control cells expressing a
non-pathogenic silent polymorphism (A66A), distinct SGs are
robustly present in 10% of the population following oxidative
stress (Fig. 6B). In contrast, there was a 2-fold reduction in the
number of cells forming SGs in cells expressing TDP-43R361S
mutation (P , 0.03; Fig. 6C). Interestingly, SG formation in
cells expressing TDP-43D169Gwas comparable with control.
To determine whether mutations in TDP-43 conferred an
effect on the accumulated levels of SG components, we exam-
ined G3BP and TIA-1 protein levels by immunoblot. While no
proteins accumulated to
significant differences in these proteins were detected in cells
expressing TDP-43D169G, the G3BP protein was decreased
26% (P ¼ 0.04) and TIA-1 was increased 20% (P ¼ 0.02) in
TDP-43R361Scells (Fig. 6D). These data are similar to what is
observed with siTDP-43 and thus suggest that TDP-43R361Sis
a loss-of-function mutation with respect to SG dynamics.
Our studies indicate that TDP-43 contributes to the cellular
response to acute stress. Specifically, endogenous TDP-43 is
recruited into SGs which are considered to be key elements
in the protective response to cellular stress. Moreover, our
data demonstrate that TDP-43 participates in regulating SGs
such that depletion of TDP-43 delays SG nucleation and sec-
ondary aggregation via the differential deregulation of key
nucleating factors TIA-1 and G3BP at the mRNA level. Fur-
thermore, the number and size of TIA-1- and G3BP-positive
SGs are reduced in cells depleted of TDP-43 and subsequently
treated with oxidative stress. TDP-43 therefore contributes
positively to SG assembly and their maintenance, as well as
cellular survival following acute oxidative stress. Our use of
TDP-43 siRNA demonstrates that formed SGs are unable to
persist and that they resolve quickly in cells depleted of
TDP-43. It is well accepted that following the removal of a
stress, SGs disassemble, and the majority of released
mRNAs are recruited back to the translation machinery (28).
Thus, the contribution of TDP-43 to the assembly and disas-
sembly of SGs offers an important mechanism by which
TDP-43 may regulate gene expression in response to stress
as well as cellular recovery/survival.
Interestingly, TDP-43 has been reported to interact with
TIA-1 and TIAR, both core nucleating components of SGs
Figure 4. G3BP SG formation is impaired by TDP-43 siRNA. Formation of G3BP SGs was assessed in HeLa cells transfected with control or TDP-43 siRNA for
72 h and subsequently treated with SA. Cover slips were collected immediately and labeled for TIA-1 and G3BP. Data are representative of two independent
6 Human Molecular Genetics, 2011
at CHUM- HOTEL-DIEU on January 28, 2011
(16,17). Our study shows that TDP-43 directly modulates the
expression of TIA-1 and G3BP (and less so TIAR), providing
a potential mechanism for the impact of TDP-43 on SG
dynamics. Our immunofluorescence data demonstrate that
TDP-43 can modulate TIA-1 aggregation, thus we speculate
that the association between TDP-43 and TIA-1 facilitates
the homotypic interactions of TIA-1 which are required for
SG assembly and maintenance. The down-regulation of
G3BP mRNA levels by TDP-43 also yields fewer SGs.
Thus, we propose that TDP-43 regulates SG formation via
two independent mechanisms. However, we cannot rule out
that the impact of TDP-43 on G3BP is the mechanism
which affects the controlled aggregation of TIA-1.
hnRNP A2 interacts with TDP-43 (2). In the presence of
acute oxidative stress, endogenous hnRNP A2 was also
recruited to SGs. This is the first description of hnRNP A2
localization to SGs and our data indicate that it may be par-
tially dependent on TDP-43. Specifically, the down-regulation
of TDP-43 yielded fewer cells with SGs and smaller hnRNP
A2-labeled SGs. Whether this is due to the direct action of
TDP-43 on hnRNP A2 or a consequence of slowed SG assem-
bly due to TDP-43 regulation of G3BP transcripts and/or
TIA-1 aggregation remains to be clarified.
The localization of TDP-43 to SGs, the regulation of SG
proteins by TDP-43 and the delay in SG assembly and main-
tenance in the absence of TDP-43 suggest a novel function for
Figure 5. Endogenous hnRNP A2 is localized to SGs. HeLa cells cultured on cover slips were treated with SA or left as untreated (UN) and subsequently immu-
nolabeled for (A) hnRNP A2 and TIA-1 or (B) hnRNP A2, TDP-43 and TIA-1. Scale bar, 10 mm. hnRNP A2 co-localization to SGs is quantified by line scans of
the merged images showing the overlap between red and green (and blue) signals. Scale bar, 10 mm. (C) Localization of hnRNP A2 to SGs (marked with TIA-1)
was assessed in HeLa cells transfected with control or TDP-43 siRNA for 72 h and subsequently treated with SA. Cover slips were collected immediately and
labeled for TIA-1 and hnRNP A2. Scale bar, 10 mm. Data are representative of three independent experiments.
Human Molecular Genetics, 20117
at CHUM- HOTEL-DIEU on January 28, 2011
TDP-43 in acute stress. It will be interesting to evaluate the
role of TDP-43 in models of chronic stress. Our observation
that cells expressing the mutation TDP-43R361Shave deregu-
lated G3BP and TIA-1 levels and are hampered in their
ability to form SGs is interesting and suggests a potentially
disease-relevant mechanism. The observations in these cells
are reminiscent of that observed when TDP-43 is depleted,
thus suggesting that TDP-43R361S
mutation with regard to SG dynamics. In a transient overex-
pression culture model, it has previously been reported that
TDP-43R361Sreduces mRNA expression of the histone deace-
tylase HDAC6 (29). Separately, the down-regulation of
TDP-43 results in reduced transcription of HDAC6 (29)
(data not shown). HDAC6 has been published as an important
determinant in SG assembly (30), and TDP-43 and FUS/TLS
have recently been reported to cooperatively regulate
HDAC6 mRNA (31). In addition to its role in SG dynamics,
HDAC6 is also involved in the removal of misfolded proteins
and aggresome formation (32). Large cytoplasmic aggregates
are a feature of ALS, thus deregulation of HDAC6 by
TDP-43 mutations is a very interesting target worthy of
is a loss-of-function
In the context of ALS, mutations in TDP-43 could compro-
mise the cellular stress response such that one could envision
successive cycles of a weakened stress response leading to a
maladaptive state. Successive encounters with oxidative
stress would eventually overcome the cell’s ability to
manage the stress, and ultimately result in cellular demise
(Fig. 7). Furthermore, it remains possible that large pathologi-
cal aggregates arise due to disrupted SG dynamics (and/or pro-
longed exposure to stress) as has been proposed in Parkinson’s
disease (33). This hypothesis has gained support from the
recently described colocalization between SG makers and
neuronal cytoplasmic inclusions in cases of familial ALS
bearing a mutation in the RBP FUS/TLS (27). In addition,
the multiple reports of TDP-43 and FUS/TLS localization to
SGs in response to acute exogenous stress (16–18,25–27)
emphasize that the links between disturbed cellular SG
dynamics and neurodegenerative disease may be relevant to
pathogenesis. Importantly, we report here SG formation is
not disturbed by the TDP-43D169Gmutation, suggesting that
it is mechanistically independent of the reported C-terminal
mutants. The nature of this mutation, predicted to be mRNA
binding, remains to be demonstrated.
Figure 6. The disease-causing mutation TDP-43R361Simpacts SG formation. (A) Immunoblot of steady-state TDP-43 protein levels in RIPA cell lysates from
control human lymphoblasts or patients expressing the disease-causing mutations TDP-43D169Gand TDP-43R361S. Data are representative of four experiments.
Histogram indicates quantification via densitometry.∗P ¼ 0.04;∗∗P ¼ 0.01. (B) Control and mutant patient cells were treated with SA or left untreated (UN) and
labeled with TIA-1. SGs are indicated with arrows. (C) The number of SG-positive cells was counted from a minimum of three fields representing at least 100
cells. The mean+SEM of four independent experiments is plotted.∗P , 0.03. Scale bar, 10 mm. (D) G3BP, TIA-1 and TIAR protein levels in control and
patient cell lysates. Actin is a loading control. Data are representative of three to six independent experiments. Histogram indicates quantification via densito-
metry.∗P ¼ 0.04;∗∗P ¼ 0.02.
8Human Molecular Genetics, 2011
at CHUM- HOTEL-DIEU on January 28, 2011
We describe a mechanistic role for endogenous TDP-43 in
the cellular stress response which may be a disease relevant
mechanism. Defining the relevance of this aspect of TDP-43
biology in motor neurons and ALS is now an ensuing challenge.
Finally, our data emphasize that the pathogenic affects of
TDP-43 mutations may be mechanistically divergent.
MATERIALS AND METHODS
Plasmids, cell culture and transfection
modified Eagle medium and Iscove’s modified Dulbecco’s
medium, respectively, supplemented with 10% fetal bovine
serum and 1% streptomycin/penicillin/glutamate. For transfec-
tion of siRNA, 125 pmol of custom siRNAs were transfected
with Lipofectamine 2000 (Invitrogen), according to the manu-
facturer’s instructions. Control and TDP-43 siRNA sequences
(Invitrogen). Cells were transfected at 30–50% confluency.
Transfection media were replaced with regular culture media
without antibiotics after 5 h. Cells were collected after 72 h of
Cells were treated with various stresses, including 0.5 mM SA
(30 min, 378C; Sigma), HS (30 min, 438C) and 1 mM THAP
(50 min, 378C; Sigma). For stress recovery experiments, cells
were stressed and then media were replaced, and cells were per-
Cell death was measured using the Annexin V-PE Apoptosis
Detection Kit I (BD Biosciences), according to the manufac-
turer’s instructions. Total Annexin V-positive cells are plotted.
Immunofluorescence and antibodies
Cells grown on cover slips were fixed in 1% formaldehyde in
phosphatebuffered saline (PBS)and subsequently
permeabilized with 0.1% Triton X-100 in PBS for 15 min.
Cover slips were blocked with 0.1% bovine serum albumin
(BSA) in PBS for 15 min and incubated with antibodies to
TDP-43 (1:300; Proteintech), TIA-1 (1:100; Santa Cruz),
hnRNP A2 (1:100; Abnova), G3BP (1:400; BD Biosciences)
and GW182 (1:50; M. Fritzler) diluted in blocking buffer for
1 h at room temperature. Cover slips were subsequently
washed once with 0.1% Triton X-100 in PBS and twice with
0.1% BSA in PBS. Labeling was visualized with the fluores-
cently conjugated secondary antibodies donkey anti-mouse
Texas Red (Jackson Immunochemicals), donkey anti-rabbit
FITC (Jackson Immunochemicals), goat anti-human Alexa
555 (Invitrogen) and donkey anti-goat Alexa 633 (Invitrogen).
Cover slips were washed as before, and mounted with
ProLong Antifade reagent (Invitrogen). Lymphoblasts were
affixed to Superfrost charged slides via cytospin and then sub-
sequently fixed and labeled, as previously described (34).
Images were collected on a Leica SP5 confocal microscope.
Quantification of SG size
SGs were identified by TIA-1 staining and cells were scored as
positive when they had at least two foci of a minimal size of
0.75 mm2. The area of 10 SGs (ranging from 0.75 to 5 mm2),
randomly selected in at least 10 cells per condition, was manu-
ally measured with ImageJ. The average SG size of at least
100 SGs in 10 cells is presented.
Cell lysates and immunoblot analysis
Cells were collected in ice-cold PBS, lysed in RIPA buffer
(150 mM NaCl, 50 mM Tris pH 7.4, 1% Triton X-100, 0.1%
SDS, 1% sodium deoxycholate) and centrifuged at 16 000g.
Supernatants were collected and quantified with the BCA
Protein Assay Kit (Thermo Scientific). For fractionation of
soluble and insoluble components, cells lysed in RIPA
buffer were passed through a 25 G syringe six times and cen-
trifuged at maximum speed. Supernatants (soluble) were
recovered and pellets (insoluble) were resuspended directly
in 1× Laemmli sample buffer. Equal volumes of each fraction
were separated by SDS–PAGE. The following antibodies
were used in immunoblotting: rabbit anti-TDP-43 (1:5000;
Proteintech), goat anti-TIA-1 (1:500; Santa Cruz), goat anti-
TIAR (1:200; Santa Cruz), mouse anti-G3BP (1:600; BD Bio-
sciences), rabbit anti-eIF2a (1:1000; Cell Signaling), rabbit
anti-Phospho eIF2a (1:1000; Cell Signaling) and mouse anti-
actin (1:400 000; MP Biomedicals). Blots were visualized with
peroxidase-conjugated secondary antibodies and ECL Western
Blotting Substrate (Thermo Scientific). Densitometry was per-
formed with ImageJ.
RNA was extracted with RNAeasy kit (Qiagen) and reverse
transcribed with QuantiTect (Qiagen). Resulting cDNA was
processed for qPCR with SybrGreen (Biorad) according to the
manufacturer’s instructions using the following primer
sets: b-actin exon 5 F: 5′-CGTTGGCATCCACGAAACTA-3′;
b-actin exon 6 R: 5′-AGTACTTGCGCTCAGGAGGA-3′;
TIA-1 exon 12 F: 5′-CATGGAACCAGCAAGGATTT-3′;
Figure 7. Model of TDP-43 in regulation of SGs. Reduced TDP-43 protein
levels or TDP-43 mutations yield a reduction in G3BP levels and disrupt
TIA-1 aggregation. These events yield slowed and diminished SG formation
and poor maintenance. This may increase cellular susceptibility to acute
stress stimuli and contribute to cellular death. This could set up a feed-forward
amplification loop resulting in a maladaptive state in motor neurons, thereby
contributing to an increased vulnerability over time.
Human Molecular Genetics, 20119
at CHUM- HOTEL-DIEU on January 28, 2011
TIA-1 exon 13 R: 5′-CACTCCCTGTAGCCTCAAGC-3′;
TIAR exon 11 F: 5′-GCCAATGGAGCCAAGTGTAT-3′;
TIAR intron 12 R: 5′-CATATGCGGCTTGGTTAGGA-3′;
G3BP exon 11 F: 5′-TAATCGCCTTCGGGGACCTG-3′;
G3BP exon 11 R: 5′-AAGCCCCCTTCCCACTCCAA-3′;
HuR exon 4/5 junction:5′-CGCAGAGATTCAGGTTC
TCC-3′; and HuR exon 5 R: 5′-CCAAACCCTTTGCACT
Data were analyzed by Student’s t-test or one-way ANOVA,
where appropriate. Error bars represent standard error of the
Supplementary Material is available at HMG online.
We thank M. Fritzler (University of Calgary) for the GW182
antibody, L. Chatel-Chaix (Universite ´ de Montre ´al) and D.R.
Foltz (University of Virginia) for experimental advice,
J. Laganiere for help with microscopy, N. Arbour, P. Cossette
and A. Prat for access to equipment and J.A. Parker, P. Dion
and N. Grandvaux for critical reading of the manuscript.
Conflict of Interest statement. None declared.
du Quebec (FRSQ), Canadian Institutes of Health Research
(CIHR) Neuromuscular Research
dian Foundation for Innovation (CFI) (all to C.V.V.). K.K.M. is
a recipient of FRSQ and CIHR Masters studentships. S.P. is a
recipient of the ALS Society of Canada Tim Noe ¨l Studentship.
1. Ou, S.H., Wu, F., Harrich, D., Garcia-Martinez, L.F. and Gaynor, R.B.
(1995) Cloning and characterization of a novel cellular protein, TDP-43,
that binds to human immunodeficiency virus type 1 TAR DNA sequence
motifs. J. Virol., 69, 3584–3596.
2. D’Ambrogio, A., Buratti, E., Stuani, C., Guarnaccia, C., Romano, M.,
Ayala, Y.M. and Baralle, F.E. (2009) Functional mapping of the
interaction between TDP-43 and hnRNP A2 in vivo. Nucleic Acids Res.,
3. Buratti, E., Dork, T., Zuccato, E., Pagani, F., Romano, M. and Baralle,
F.E. (2001) Nuclear factor TDP-43 and SR proteins promote in vitro and
in vivo CFTR exon 9 skipping. EMBO J., 20, 1774–1784.
4. Mercado, P.A., Ayala, Y.M., Romano, M., Buratti, E. and Baralle, F.E.
(2005) Depletion of TDP 43 overrides the need for exonic and intronic
splicing enhancers in the human apoA-II gene. Nucleic Acids Res., 33,
5. Buratti, E., Brindisi, A., Giombi, M., Tisminetzky, S., Ayala, Y.M. and
Baralle, F.E. (2005) TDP-43 binds heterogeneous nuclear
ribonucleoprotein A/B through its C-terminal tail: an important region for
the inhibition of cystic fibrosis transmembrane conductance regulator
exon 9 splicing. J. Biol. Chem., 280, 37572–37584.
6. Lagier-Tourenne, C. and Cleveland, D.W. (2009) Rethinking ALS: the
FUS about TDP-43. Cell, 136, 1001–1004.
7. Ayala, Y.M., Zago, P., D’Ambrogio, A., Xu, Y.F., Petrucelli, L., Buratti,
E. and Baralle, F.E. (2008) Structural determinants of the cellular
localization and shuttling of TDP-43. J. Cell Sci., 121, 3778–3785.
8. Kim, J.H., Hahm, B., Kim, Y.K., Choi, M. and Jang, S.K. (2000)
Protein-protein interaction among hnRNPs shuttling between nucleus and
cytoplasm. J. Mol. Biol., 298, 395–405.
9. Pinol-Roma, S. and Dreyfuss, G. (1993) hnRNP proteins: localization and
transport between the nucleus and the cytoplasm. Trends Cell Biol., 3,
10. Guil, S., Long, J.C. and Caceres, J.F. (2006) hnRNP A1 relocalization to
the stress granules reflects a role in the stress response. Mol. Cell Biol., 26,
11. Quaresma, A.J., Bressan, G.C., Gava, L.M., Lanza, D.C., Ramos, C.H.
and Kobarg, J. (2009) Human hnRNP Q re-localizes to cytoplasmic
granules upon PMA, thapsigargin, arsenite and heat-shock treatments.
Exp. Cell Res., 315, 968–980.
12. Kedersha, N. and Anderson, P. (2002) Stress granules: sites of mRNA
triage that regulate mRNA stability and translatability. Biochem. Soc.
Trans., 30, 963–969.
13. Abdelmohsen, K., Kuwano, Y., Kim, H.H. and Gorospe, M. (2008)
Posttranscriptional gene regulation by RNA-binding proteins during
oxidative stress: implications for cellular senescence. Biol. Chem., 389,
14. Henao-Mejia, J. and He, J.J. (2009) Sam68 relocalization into stress
granules in response to oxidative stress through complexing with TIA-1.
Exp. Cell Res., 315, 3381–3395.
15. Anderson, P. and Kedersha, N. (2008) Stress granules: the Tao of RNA
triage. Trends Biochem. Sci., 33, 141–150.
16. Colombrita, C., Zennaro, E., Fallini, C., Weber, M., Sommacal, A.,
Buratti, E., Silani, V. and Ratti, A. (2009) TDP-43 is recruited to
stress granules in conditions of oxidative insult. J. Neurochem., 111,
17. Liu-Yesucevitz, L., Bilgutay, A., Zhang, Y.J., Vanderwyde, T., Citro, A.,
Mehta, T., Zaarur, N., McKee, A., Bowser, R., Sherman, M. et al. (2010)
Tar DNA binding protein-43 (TDP-43) associates with stress granules:
analysis of cultured cells and pathological brain tissue. PLoS ONE, 5,
18. Freibaum, B.D., Chitta, R.K., High, A.A. and Taylor, J.P. (2010)
Global analysis of TDP-43 interacting proteins reveals strong association
with RNA splicing and translation machinery. J. Proteome Res., 9,
19. Gilks, N., Kedersha, N., Ayodele, M., Shen, L., Stoecklin, G., Dember,
L.M. and Anderson, P. (2004) Stress granule assembly is mediated by
prion-like aggregation of TIA-1. Mol. Biol. Cell, 15, 5383–5398.
20. Kedersha, N., Stoecklin, G., Ayodele, M., Yacono, P., Lykke-Andersen, J.,
Fritzler, M.J., Scheuner, D., Kaufman, R.J., Golan, D.E. and Anderson, P.
mRNP remodeling. J. Cell Biol., 169, 871–884.
21. McEwen, E., Kedersha, N., Song, B., Scheuner, D., Gilks, N., Han, A.,
Chen, J.J., Anderson, P. and Kaufman, R.J. (2005) Heme-regulated
inhibitor kinase-mediated phosphorylation of eukaryotic translation
initiation factor 2 inhibits translation, induces stress granule formation,
and mediates survival upon arsenite exposure. J. Biol. Chem., 280,
22. Tourriere, H., Chebli, K., Zekri, L., Courselaud, B., Blanchard, J.M.,
Bertrand, E. and Tazi, J. (2003) The RasGAP-associated endoribonuclease
G3BP assembles stress granules. J. Cell Biol., 160, 823–831.
23. Thomas, M.G., Martinez Tosar, L.J., Desbats, M.A., Leishman, C.C. and
Boccaccio, G.L. (2009) Mammalian Staufen 1 is recruited to stress
granules and impairs their assembly. J. Cell Sci., 122, 563–573.
24. Kedersha, N. and Anderson, P. (2007) Mammalian stress granules and
processing bodies. Methods Enzymol., 431, 61–81.
25. Bosco, D.A., Lemay, N., Ko, H.K., Zhou, H., Burke, C., Kwiatkowski,
T.J. Jr, Sapp, P., McKenna-Yasek, D., Brown, R.H. Jr. and Hayward, L.J.
(2010) Mutant FUS proteins that cause amyotrophic lateral sclerosis
incorporate into stress granules. Hum. Mol. Genet., 19, 4160–4175.
26. Gal, J., Zhang, J., Kwinter, D.M., Zhai, J., Jia, H., Jia, J. and Zhu, H.
(2010) Nuclear localization sequence of FUS and induction of stress
granules by ALS mutants. Neurobiol. Aging. doi:10.1016/
10Human Molecular Genetics, 2011
at CHUM- HOTEL-DIEU on January 28, 2011
27. Dormann, D., Rodde, R., Edbauer, D., Bentmann, E., Fischer, I., Hruscha, Download full-text
A., Than, M.E., Mackenzie, I.R., Capell, A., Schmid, B. et al. (2010)
ALS-associated fused in sarcoma (FUS) mutations disrupt
Transportin-mediated nuclear import. EMBO J., 29, 2841–2857.
28. Mazroui, R., Di Marco, S., Kaufman, R.J. and Gallouzi, I.E. (2007)
Inhibition of the ubiquitin-proteasome system induces stress granule
formation. Mol. Biol. Cell, 18, 2603–2618.
29. Fiesel, F.C., Voigt, A., Weber, S.S., Van den Haute, C., Waldenmaier,
A., Gorner, K., Walter, M., Anderson, M.L., Kern, J.V., Rasse, T.M.
et al. (2010) Knockdown of transactive response DNA-binding
protein (TDP-43) downregulates histone deacetylase 6. EMBO J., 29,
30. Kwon, S., Zhang, Y. and Matthias, P. (2007) The deacetylase HDAC6 is a
novel critical component of stress granules involved in the stress response.
Genes Dev., 21, 3381–3394.
31. Kim, S.H., Shanware, N., Bowler, M.J. and Tibbetts, R.S. (2010)
ALS-associated proteins TDP-43 and FUS/TLS function in a common
biochemical complex to coregulate HDAC6 mRNA. J. Biol. Chem., 285,
32. Boyault, C., Zhang, Y., Fritah, S., Caron, C., Gilquin, B., Kwon, S.H.,
Garrido, C., Yao, T.P., Vourc’h, C., Matthias, P. and Khochbin, S. (2007)
HDAC6 controls major cell response pathways to cytotoxic accumulation
of protein aggregates. Genes Dev., 21, 2172–2181.
33. Olanow, C.W., Perl, D.P., DeMartino, G.N. and McNaught, K.S. (2004)
Lewy-body formation is an aggresome-related process: a hypothesis.
Lancet Neurol., 3, 496–503.
34. Didiot, M.C., Subramanian, M., Flatter, E., Mandel, J.L. and Moine, H.
(2009) Cells lacking the fragile X mental retardation protein (FMRP) have
normal RISC activity but exhibit altered stress granule assembly. Mol.
Biol. Cell, 20, 428–437.
Human Molecular Genetics, 201111
at CHUM- HOTEL-DIEU on January 28, 2011