Stress-induced Reversal of MicroRNA Repression and mRNA P-body Localization in Human Cells

Abstract

In metazoa, microRNAs (miRNAs) imperfectly base-pair with the 3'-untranslated region (3'UTR) of mRNAs and prevent protein accumulation by either repressing translation or inducing mRNA degradation. Examples of specific mRNAs undergoing miRNA-mediated repression are numerous, but whether the repression is a reversible process remains largely unknown. Here, we show that cationic amino acid transporter 1 (CAT-1) mRNA and reporters bearing the CAT-1 3'UTR or its fragments can be relieved from the miRNA miR-122-induced inhibition in human hepatoma cells in response to different stress conditions. The derepression of CAT-1 mRNA is accompanied by its release from cytoplasmic processing bodies (P bodies) and its recruitment to polysomes, indicating that P bodies act as storage sites for mRNAs inhibited by miRNAs. The derepression requires binding of HuR, an AU-rich-element-binding ELAV family protein, to the 3'UTR of CAT-1 mRNA. We propose that proteins interacting with the 3'UTR will generally act as modifiers altering the potential of miRNAs to repress gene expression.

Full text

miRNAs are 20- to 22-nucleotide-long regulatory
RNAs expressed in plants and metazoan animals. Current
estimates indicate that hundreds of different miRNAs are
encoded in individual genomes, with the number of
human miRNAs possibly reaching 1000. Approximately
30% of all human genes are predicted to be subject to
miRNA regulation. Specific functions and target mRNAs
have been assigned to only a few dozen miRNAs, but it is
apparent that miRNAs participate in the regulation of
many different biological processes. Changes in miRNA
expression are observed in human pathologies and some
miRNAs were shown to act as oncogenes or tumor
supressors (for review, see Ambros 2004; Bartel 2004;
Wienholds and Plasterk 2005).
miRNAs regulate gene expression posttranscriptionally,
by base-pairing to target mRNAs. In animals, most inves-
tigated miRNAs form imperfect hybrids with sequences in
the 3′UTR, with the miRNA 5′-proximal “seed” region
(positions 2–8) providing most of the pairing specificity
(for review, see Filipowicz 2005; Tomari and Zamore
2005). The miRNA association results in translational
repression, frequently accompanied by a considerable
degradation of the mRNA by a non-RNAi mechanism (for
review, see Pillai 2005; Valencia-Sanchez et al. 2006).
Argonaute (AGO) proteins are the essential and best-
characterized components of miRNPs. In mammals, only
one of the four AGO proteins, AGO2, is competent to cat-
alyze cleavage of mRNA in the RNA interference (RNAi)-
like mechanism (Liu et al. 2004; Meister et al. 2004). On the
other hand, all four AGO proteins, AGO1–4, appear to
function in miRNA repression (Liu et al. 2004; Meister et
al. 2004; Pillai et al. 2004). The mechanism of translational
inhibition by miRNAs is not well understood. Some natural
or model miRNAs interfere with the initiation of protein
synthesis (Humphreys et al. 2005; Pillai et al. 2005),
although others may affect more downstream steps in trans-
lation (Olsen and Ambros 1999; Petersen et al. 2006). The
AGO proteins and repressed mRNAs are enriched in the
cytoplasmic P bodies (Jakymiw et al. 2005; Liu et al. 2005;
Pillai et al. 2005). The P-body association may represent a
secondary event, which follows the translation inhibition
step. Since mRNA catabolic enzymes also reside in P bod-
ies, the relocation likely results in the reported mRNA
degradation (Pillai 2005; Valencia-Sanchez et al. 2006).
To date, miRNAs have been primarily identified as
negative regulators of expression of cellular mRNAs, and
it remains unknown whether the inhibition of a specific
mRNA can be effectively reversed. Clearly, the ability to
disengage miRNPs from the repressed mRNA, or render
them inactive, would make miRNA regulation much
more dynamic and also more responsive to specific cellu-
lar needs. In this work, we present evidence that CAT-1
mRNA, which encodes the high-affinity cationic amino
acid transporter and which is translationally repressed by
miR-122 in Huh7 hepatoma cells, can be relieved from
the miR-122 repression by subjecting Huh7 cells to dif-
ferent stress conditions. The derepression is accompanied
by the release of CAT-1 mRNA from P bodies and its
recruitment to polysomes, consistent with miR-122
inhibiting translational initiation. We provide evidence
that the stress-induced up-regulation is mediated by bind-
ing of HuR, an AU-rich element (ARE)-binding protein,
to the 3′UTR of CAT-1 mRNA.
RESULTS
CAT-1 mRNA as a Model to Investigate
Reversibility of miRNA-mediated Repression
To investigate whether mRNA can be relieved from
miRNA-mediated repression, we looked at the mRNA
encoding the high-affinity cationic amino acid trans-
porter, CAT-1, a member of the CAT (or SLC7A1-4)
family of system y
+
transporters. CAT-1, which facili-
Stress-induced Reversal of MicroRNA Repression and mRNA
P-body Localization in Human Cells
S.N. BHATTACHARYYA,* R. HABERMACHER,* U. MARTINE,
†
E.I. CLOSS,
†
AND W. FILIPOWICZ*
*Friedrich Miescher Institute for Biomedical Research, 4002 Basel, Switzerland;
†
Department of Pharmacology,
Johannes Gutenberg University, 67, 55101 Mainz, Germany
In metazoa, microRNAs (miRNAs) imperfectly base-pair with the 3′-untranslated region (3′UTR) of mRNAs and prevent
protein accumulation by either repressing translation or inducing mRNA degradation. Examples of specific mRNAs under-
going miRNA-mediated repression are numerous, but whether the repression is a reversible process remains largely
unknown. Here, we show that cationic amino acid transporter 1 (CAT-1) mRNA and reporters bearing the CAT-1 3′UTR
or its fragments can be relieved from the miRNA miR-122-induced inhibition in human hepatoma cells in response to
different stress conditions. The derepression of CAT-1 mRNA is accompanied by its release from cytoplasmic processing
bodies (P bodies) and its recruitment to polysomes, indicating that P bodies act as storage sites for mRNAs inhibited by
miRNAs. The derepression requires binding of HuR, an AU-rich-element-binding ELAV family protein, to the 3′UTR of
CAT-1 mRNA. We propose that proteins interacting with the 3′UTR will generally act as modifiers altering the potential
of miRNAs to repress gene expression.
Cold Spring Harbor Symposia on Quantitative Biology, Volume LXXI. © 2006 Cold Spring Harbor Laboratory Press 978-087969817-1 513
513-522_Bhattacharyya_Symp71.qxd 2/8/07 1:42 PM Page 513

tates uptake of arginine and lysine in mammalian cells, is
expressed ubiquitously, but its levels vary significantly in
different cells and tissues and its expression is known to
undergo extensive regulation at both transcriptional and
posttranscriptional levels (for review, see Hatzoglou et al.
2004). CAT-1 regulation has been studied intensively in
rat C6 glioma cells, where transcription of the gene and
stability and translation of the mRNA are up-regulated in
response to different types of cellular stress (Yaman et al.
2002; Haztoglou et al. 2004). In human hepatoma Huh7
cells, CAT-1 mRNA activity is regulated by a liver-spe-
cific miRNA, miR-122. The human CAT-1 3′UTR con-
tains several potential target sites for miR-122 (see Fig.
2A). Experiments performed with endogenous CAT-1
mRNA (Bhattacharyya et al. 2006) and reporters contain-
ing its 3′UTR (or fragments thereof) indicated that miR-
122 has a repressive effect on translation and, in the case
in some chimeric reporter mRNAs, may also cause lim-
ited destabilization of the target RNA (Chang et al. 2004;
Bhattacharyya et al. 2006). Maintenance of low CAT-1
activity in liver cells is important to avoid hydrolysis of
the plasma arginine by arginase, a highly expressed
enzyme in hepatocytes, which catalyzes the last step of
the urea cycle (hydrolysis of arginine to ornithine and
urea). However, under certain conditions, e.g., when urea
cycle enzymes are down-regulated or during liver regen-
eration, CAT-1 expression is induced, most likely to sus-
tain hepatocellular protein synthesis.
Amino Acid Deprivation Induces Translational
Up-regulation of CAT-1 mRNA in Huh7 Cells
In cultured rat C6 glioma cells, CAT-1 protein expres-
sion increases in response to amino acid deprivation and
other forms of cellular stress. Both transcriptional and
posttranscriptional regulation were reported to contribute
to the increase (Hatzoglou et al. 2004). We investigated
the effect of amino acid starvation on expression of CAT-
1 protein and mRNA in hepatoma Huh7 cells. The CAT-
1 protein level increased markedly after 1 hour of
starvation and then remained unchanged during several
additional hours of amino acid depletion (Fig. 1A). In con-
514 BHATTACHARYYA ET AL.
Figure 1. Starvation of Huh7 cells induces expression of CAT-1 protein independent of transcription. (A) Stress-induced expression of
CAT-1 in Huh7 cells is a posttranscriptional event. Huh7 cells were cultured for 0–4 hours in amino-acid-depleted medium in the absence
or presence of indicated inhibitors. Proteins were treated with PGNase F and analyzed by western blotting using indicated antibodies.
Two lanes on the right contained protein extracts of DLD-1 cells, abundant in CAT-1. Positions of glycosylated and deglycosylated
CAT-1 are indicated by an arrowhead and arrow. (B) Analysis of RNA isolated from Huh7 cells starved of amino acids in the presence
of indicated inhibitors. (Upper panel) Real-time PCR quantification of CAT-1 mRNA. Normalized values are means (+/–
SD) from three
independent experiments, with GAPDH mRNA serving as an internal control. (Two lower panels) Northern analysis of miR-122 and
staining of the gel with ethidium bromide. Details of methodological procedures used for experiments described in this and other figures
are described in Bhattacharyya et al. (2006). (Reprinted, with permission, from Bhattacharyya et al. 2006 [© Elsevier].)
513-522_Bhattacharyya_Symp71.qxd 2/8/07 1:42 PM Page 514

trast, the substantial increase in CAT-1 mRNA level,
measured by either real-time polymerase chain reaction
(PCR) (Fig. 1B) or northern blotting (not shown)
(Bhattacharyya et al. 2006), was only detectable after 3–4
hours of starvation. In HepG2 hepatoma cells, which do
not express miRNA miR-122, no appreciable change in
either CAT-1 protein or mRNA level was observed during
4 hours of starvation (data not shown). The early induction
of the CAT-1 protein in Huh7 cells was independent of
RNA polymerase II transcription, since treatment with
inhibitors of RNA polymerase II, either actinomycin D
(ActD) or α-amanitin (α-Am), had no effect (Fig. 1A).
However, the protein accumulation was inhibited by
cycloheximide (CHX), an inhibitor of translational
elongation (Fig. 1A). Amino acid starvation and treatment
of cells with ActD or α-Am had no effect on the level of
miR-122 or Ago2, essential components of miRNPs (Fig.
1A,B). However, inclusion of either inhibitor prevented
the late (4-hour time point) accumulation of CAT-1
mRNA, demonstrating that they effectively inhibit RNA
polymerase II transcription in Huh7 cells.
It was important to exclude the possibility that CAT-1
protein turns over rapidly in cells grown in the medium rich
in amino acids, and its apparent induction by starvation is
due to an increase in protein stability. To this end, we
have performed western analysis of lysates prepared from
cells grown in either the presence or absence of CXH.
The analysis indicated that instead of stabilizing the
CAT-1 protein, the starvation increased its turnover
(Bhattacharyya et al. 2006). The accelerated decay of
CAT-1 in stressed cells provides a plausible explanation
for the observation that steady-state levels of the protein
do not continue to increase at times beyond 1 hour post-
starvation. Taken together, the data indicate that starva-
tion of Huh7 cells results in a de novo synthesis of the
CAT-1 protein from the preexisting mRNA pool.
Response to Amino Acid Starvation and Other
Types of Stress Is Mediated by the
CAT-1 3′UTR and Involves miR-122
To test whether the translational induction described
above is mediated by the CAT-1 mRNA 3′UTR, we
measured the effect of starvation on the activity of dif-
ferent RL-cat reporters (Fig. 2A) in Huh7 cells. In RL-
catA, the RL-coding region is fused to the 2.5-kb 3′UTR
found in a short form of CAT-1 mRNA. In RL-catB, the
3′-proximal 1-kb region (referred to as region D) of the
CAT-1 3′UTR is deleted. The remaining part contains
three predicted miR-122-binding sites identified in the
REVERSAL OF
MIRNA-MEDIATED REPRESSION 515
Figure 2. Effect of different types of cellular stress on activity of RL reporters in Huh7 and HepG2 cells. (A) Schemes of reporters
bearing different segments of the CAT-1 3′UTR fused to the RL-coding region. Positions of potential miR-122-binding sites (1–3)
and the approximately 1-kb region D and its AU-rich subregion ARD are indicated. (B) Expression of the RL-catA reporter is specif-
ically up-regulated in stressed Huh7 cells (upper panel) but not HepG2 cells (lower panel). Cells transfected with indicated reporters
were starved for 2 hours (Starved) or grown in the presence of either thapsigargin (TG; 2 hr) or arsenite (Ars; 30 min). ActD or CHX
was added at the time of the shift to stress conditions. Values are normalized to activities in nonstressed (Fed) cells, which were set
to 1. (Reprinted, with permission, from Bhattacharyya et al. 2006 [© Elsevier].)
513-522_Bhattacharyya_Symp71.qxd 2/8/07 1:42 PM Page 515

2.5-kb CAT-1 3′UTR. In RL-catC, the CAT-1 3′UTR is
shortened further to eliminate the region containing the
miR-122 sites. Reporter activity, normalized to activity
of the coexpressed firefly luciferase (FL), was tested in
transfected Huh7 and HepG2 cells (Fig. 2B). In Huh7
cells, an approximately fourfold induction of RL activity
was observed upon starvation of cells transfected with
the reporter containing miR-122 sites, RL-catA, but not
with that devoid of miRNA sites, RL-catC. Starvation
increased expression of RL-catB by approximately 30%
(Fig. 2B). As in the case of endogenous CAT-1 protein,
the stress-induced expression of RL from RL-catA was
inhibited by addition of CHX but not ActD (Fig. 2B).
Likewise, other stress conditions such as the ER stress
(induced by thapsigargin) or oxidative stress (induced by
arsenite) stimulated expression of RL-catA approxi-
mately 2.5-fold, whereas the effect on other reporters
was either minimal or absent (Fig. 2B). Amino acid
starvation and treatment with thapsigargin or arsenite
had no effect on the level of RL-catA mRNA, indicating
that the effect was posttranscriptional (Bhattacharyya
et al. 2006).
To test whether the stress-mediated activation of RL-
catA indeed involves miR-122, we investigated whether
responses observed in Huh7 cells can be reproduced in
HepG2 cells, which do not express miR-122. Exposure of
HepG2 cells to different forms of stress had no effect on
expression of any RL-cat reporter (Fig. 2B, lower panel).
However, when HepG2 cells were cotransfected with
miR-122, up-regulation of RL-catA was clearly evident
in starved cells, with activity of RL-catB and RL-catC
remaining unchanged (Bhattacharyya et al. 2006).
HuR Protein Interacts with CAT-1 3′UTR and
Is Essential for the Derepression
The stress-inducible RL-catA reporter differs from
the reporter not undergoing activation, RL-catB, by the
presence of an additional segment (region D) of the
CAT-1 3′UTR (Fig. 2A). We found that region D is
essential for the stress-induced relief of the miR-122-
mediated repression in Huh7 cells. Moreover, this
region can confer stress inducibility on a reporter inhib-
ited by let-7 RNA, an abundant miRNA expressed in
HeLa cells. The let-7-specific RL-3xBulge reporter
(Pillai et al. 2005) was used to further dissect CAT-1
region D. It was found that the central part of region D,
containing sequences rich in A+U and U residues and
referred to hereafter as region ARD (see Fig. 2A), is
essential for mediating the stress-induced reversal of
miRNA repression (Bhattacharyya et al. 2006).
HuR is an AU-rich element (ARE)-binding protein, a
member of the ELAV family of proteins, implicated in
different aspects of posttranscriptional regulation (for
review, see Brennan and Steitz 2001; Katsanou et al.
2005; Lal et al. 2005). In response to various types of
cellular stress, HuR is mobilized from the nucleus to the
cytosol, where it may modulate translation or increase
the stability of different target mRNAs, including CAT-
1 mRNA in rat glioma cells (Yaman et al. 2002). We
found that HuR relocates from the nucleus to the cytosol
upon amino acid starvation also in Huh7 cells (Fig. 3A)
and that the RNA-mediated depletion of HuR, using two
different small interfering RNAs (siRNAs) (Fig. 3B),
eliminates the RL-catA response in comparison to con-
trol cells or cells treated with control siRNA (Fig. 3C).
We tested, by native gel analysis, if the
32
P-labeled ARD
fragment, which is essential for mediating the stress-
induced derepression, can interact with a recombinant
GST-HuR fusion protein. As shown in Figure 3D, puri-
fied GST-HuR but not GST, formed a complex with the
ARD fragment, and the complex was competed by an
excess of unlabeled fragment ARD but not the ΔARD
portion of region D. Additional experiments indicated
that CAT-1 mRNA can be specifically immunoselected
from the soluble cytoplasmic fraction of starved Huh7
cells by the anti-HuR antibody. We have also demon-
strated that HuR binding per se does not cause activation
of the RL reporter devoid of miR-122 sites and not
repressed by the miRNA (Bhattacharyya et al. 2006).
Taken together, the above experiments indicate that HuR
has a role in the stress-induced activation of CAT-1 mRNA
and RL reporters undergoing repression mediated by
either miR-122 or let-7 RNA.
Stress Induces HuR-dependent Relocation of
CAT-1 mRNA from P bodies
mRNA reporters repressed by miRNAs were found to
localize in P bodies (Liu et al. 2005; Pillai et al. 2005).
We analyzed the intracellular localization of the
endogenous CAT-1 mRNA and RL-cat reporters in cells
grown under different conditions. In situ hybridization
revealed that in nonstarved Huh7 cells, CAT-1 mRNA
is concentrated in P bodies, as demonstrated by its colo-
calization with the P-body marker, GFP-Dcp1a (Fig.
4A). P-body enrichment of the mRNA was abolished
when cells were transfected with the 2′-O-methyl
oligonucleotide complementary to miR-122 but not
with control anti-miR-15 oligonucleotide (Fig. 4B),
indicating that the localization was dependent on miR-
122. Most importantly, in Huh7 cells grown for 2 hours
under amino acid deficiency, a condition that markedly
increases CAT-1 protein without an effect on the mRNA
level (see Fig. 1), CAT-1 mRNA was no longer
detectable in P bodies (Fig. 4A). Starvation did not pro-
duce an appreciable decrease in the miR-122 signal in P
bodies (Bhattacharyya et al. 2006), arguing for an effect
specific for the CAT-1 mRNA and possibly only a lim-
ited number of other mRNAs among the many regulated
by miR-122 in liver cells (Krutzfeldt et al. 2005). As
expected, CAT-1 mRNA did not colocalize to P bodies
in HepG2 cells. However, in HepG2 cells transfected
with miR-122, CAT-1 mRNA was enriched in P bodies,
further supporting the idea that the repression and P-
body localization of CAT-1 mRNA are controlled by
miR-122 (Bhattacharyya et al. 2006).
To find out whether the relocation of CAT-1 mRNA
from P bodies, like the activation of its expression, also
requires HuR, we studied CAT-1 distribution in Huh7
cells in which HuR protein had been knocked down by
RNAi. The knockdown had no effect on the P-body
516 BHATTACHARYYA ET AL.
513-522_Bhattacharyya_Symp71.qxd 2/8/07 1:42 PM Page 516

REVERSAL OF MIRNA-MEDIATED REPRESSION 517
Figure 3. HuR relocates from the nucleus to the cytoplasm in stressed Huh7 cells, interacts with the CAT-1 3′UTR, and is required for the
stress-induced derepression of the RL-catA reporter. (A) Starvation-induced relocation of HuR. Huh7 cells, either nonstarved (Fed) or
starved for 2 hours (Starved) were fixed, and the localization of HuR (red) was determined using mouse anti-HuR monoclonal antibodies.
(B) Two different siRNAs effectively deplete HuR. Western blot was performed 72 hours after siRNA transfection. (C) Knockdown of
HuR prevents starvation-induced derepression of RL-catA. Huh7 cells were cotransfected with RL-cat reporters and indicated siRNAs;
72 hours after transfection, cells were transferred for an additional 2 hours to a medium either with or without amino acids. The values are
means from three transfections +/– standard deviation. (D) Recombinant GST-HuR protein interacts with the CAT-1 ARD fragment.
32
P-labeled ARD RNA was incubated with either purified GST-HuR or GST alone in the absence or presence of indicated cold competi-
tors, and complexes were analyzed on a native gel. (Reprinted, with permission, from Bhattacharyya et al. 2006 [© Elsevier].)
enrichment of CAT-1 mRNA in control cells. However,
it prevented mobilization of the mRNA from these struc-
tures upon amino acid starvation. In cells transfected with
nonspecific siRNA, the relocation did take place, simi-
larly as in nontreated Huh7 cells (Fig. 4C).
We also determined the intracellular localization of
different RL reporters repressed by either miR-122 or
let-7 miRNA. The repressed reporters localized to P bod-
ies, and their relocalization from these structures in
response to amino acid depletion was dependent on the
CAT-1 region D present downstream from the miRNA-
binding sites. Reporters devoid of region D remained
enriched in P bodies in both nonstarved and starved cells
(Bhattacharyya et al. 2006).
In starved Huh7 cells, CAT-1 mRNA becomes
extractable from cells permeabilized with digitonine,
very likely as a consequence of the relocation of mRNA
from P bodies to the cytosol (Pillai et al. 2005;
Bhattacharyya et al. 2006). As shown in Figure 5A, the
amount of CAT-1 mRNA present in the cytosol prepared
from permeabilized Huh7 cells increased already after 20
minutes following the shift to the amino-acid-depleted
medium. Importantly, the shift of the CAT-1 mRNA to
the cytosol upon stressing the cells occurred with a kinet-
ics similar to the appearance of HuR in this fraction and
also paralleled the accumulation of CAT-1 protein (Fig.
5). Hence, the stress-induced derepression of the CAT-1
mRNA translation appears to be a rapid event.
Together, the results demonstrate that CAT-1 mRNA
and RL reporters are concentrated in P bodies when
repressed by the miRNA but are rapidly mobilized from
these structures under conditions, including cellular
stress, that preclude miRNA repression. The experiments
further support a role for HuR and the CAT-1 mRNA
region D in mediating the response induced by the amino
acid stress. Moreover, they indicate that such a response
may be a more general phenomenon, applying to differ-
ent cell types and miRNA–mRNA combinations.
513-522_Bhattacharyya_Symp71.qxd 2/8/07 1:42 PM Page 517

Stress-induced Relocation of CAT-1 mRNA from
P Bodies Is Accompanied by Its Entry to Polysomes
We found previously that repression of protein syn-
thesis by let-7 RNA in HeLa cells is accompanied by a
less effective entry of target reporters to polysomes,
indicative of the translation initiation block (Pillai et al.
2005). Gradient analysis of Huh7 cell extracts indicated
that amino acid starvation results in an increase in the
fraction of CAT-1 mRNA associated with polysomes. In
contrast, β-tubulin mRNA moved toward the top of the
gradient in response to starvation, consistent with a gen-
eral inhibitory effect of stress on translation (Fig. 6A,B).
Treatment of Huh7 cells with anti-miR-122 but not con-
trol anti-let-7a 2′-O-methyl oligonucleotide resulted in
the CAT-1 mRNA shift to polysomes similar to that
induced by starvation (Fig. 6C). Of note, a fraction of
the “repressed” CAT-1 mRNA sedimented faster than
the polysome-associated CAT-1 mRNA present in
stressed or anti-miR-122-treated cells. Possibly, this
material represents P-body aggregates containing CAT-
1 mRNA.
Taken together, these data indicate that relocation of
CAT-1 mRNA from P bodies and relief of the miRNA-
mediated repression is accompanied by recruitment of
CAT-1 mRNA to polysomes, consistent with miR-122
inhibiting translational initiation.
CONCLUSIONS
Our work demonstrates that CAT-1 mRNA, and
reporters bearing the CAT-1 3′UTR or its fragments,
can be relieved from miR-122-mediated repression in
Huh7 cells subjected to amino acid deprivation or the
ER or oxidative stress. Observations that the response to
the amino acid starvation stress can be recapitulated in
other cell lines by either an ectopic supply of miR-122
or the use of chimeric reporters targeted by another
miRNA argue for a general importance of this type of
regulation. We also demonstrate that repressed CAT-1
and reporter mRNAs accumulate in P bodies in an miR-
122-dependent process and that the derepression is
accompanied by the release of the mRNAs from these
structures. The demonstrated stress-induced mobiliza-
tion of mRNAs from P bodies provides evidence that
metazoan P bodies represent sites not only of mRNA
turnover, but also of storage of translationally
repressed mRNAs. So far, such evidence was available
for baker’s yeast, an organism lacking miRNA regula-
tion (Brengues et al. 2005). Time-course experiments
measuring the appearance of CAT-1 mRNA in a soluble
cytosolic fraction and the formation of CAT-1 protein
suggest that reactivation of mRNAs sequestered in P
bodies is a very rapid process. Hence, it is tempting to
speculate that P bodies may act as general storage sites
for mRNAs, which need to be quickly mobilized into
polysomes under specific cellular conditions.
Although miRNAs may affect gene expression in dif-
ferent ways (for review, see Pillai 2005; Valencia-
Sanchez et al. 2006), recent findings indicated that let-7
RNA and some model miRNAs in mammals inhibit
518 BHATTACHARYYA ET AL.
Figure 4. CAT-1 mRNA accumulation in P bodies is miR-122-
dependent, and its stress-induced relocation from P-bodies
requires HuR. Details of cell treatment are indicated at the left of
each row. P bodies were visualized by measuring GFP-Dcp1a
fluorescence (green), and CAT-1 by in situ hybridization with
Cy3-labeled probes (red). DAPI (blue) stained the nucleus. Cells
were starved for 2 hours prior to fixation. Exposure time of the
red channel for starved Huh7 cells (A), the anti-miR-122 row (B),
and starved siControl cells (C) was ten times longer than that for
other images. (Insets) Enlargements of indicated regions. For
quantification and discussion of overlap between CAT-1 mRNA
and GFP-Dcp1a foci, see Bhattacharyya et al. (2006). (A,B)
CAT-1 mRNA is mobilized from P bodies upon starvation of
Huh7 cells (A) or upon transfection with anti-miR-122 but not
anti-miR-15 oligonucleotide (B). Bar, 5 μm. (C) HuR is required
for the stress-induced mobilization of CAT-1 mRNA from
P bodies. Huh7 cells were cotransfected with pGFP-Dcp1a and
either anti-HuR or control siRNA. After 48 hours, a fraction of
the cells was starved for 2 hours. (Reprinted, with permission,
from Bhattacharyya et al. 2006 [© Elsevier].)
513-522_Bhattacharyya_Symp71.qxd 2/8/07 1:42 PM Page 518

REVERSAL OF MIRNA-MEDIATED REPRESSION 519
Figure 5. The stress-induced relocation of HuR
protein and CAT-1 mRNA to the cytosol and the
accumulation of CAT-1 protein are rapid events.
(A) Accumulation of CAT-1 mRNA and HuR
protein in the cytosol upon stressing Huh7 cells
occurs with a similar kinetics. Huh7 cells were
starved for amino acids for increasing time, and
levels of CAT-1 mRNA (upper panels) and HuR
protein (western; lower panels) were measured in
the cytosolic and total extracts of starved cells.
Cytosolic extracts were prepared by permeabiliz-
ing cells with digitonine (Bhattacharyya et al.
2006). (B) Kinetics of the CAT-1 protein accu-
mulation in Huh7 cells subjected to amino acid
deprivation stress, determined by western.
Figure 6. Stress- or anti-miR-122-induced relocation of CAT-1 mRNA from P bodies is accompanied by its recruitment to polysomes.
(A) Distribution of CAT-1 mRNA in extracts from cells fed with amino acids (upper panels) or starved for amino acids (lower pan-
els). RNA extracted from individual fractions was analyzed by northern blots with probes specific for CAT-1 and β-tubulin mRNAs.
Two lanes at the right represent input RNA isolated from fed and starved cells. (B) Quantification of distribution of mRNAs analyzed
in panel A, expressed as a percentage of total radioactivity present in each lane. (C) Distribution of CAT-1 mRNA in extracts from
cells treated with either anti-let-7a or anti-miR-122 (middle panels). A
260
profile of the anti-let-7a extract was similar to that shown
at the top, and distribution of β-tubulin mRNA in both gradients was similar to that shown in panel A, fed cells (data not shown).
Quantification of CAT-1 mRNA is at the bottom. (Reprinted, with permission, from Bhattacharyya et al. 2006 [© Elsevier].)
translation of reporter mRNAs at the initiation step
(Humphreys et al. 2005; Pillai et al. 2005) and that
repressed mRNAs localize to P bodies for either storage
or degradation (Liu et al. 2005; Pillai 2005; Pillai et al.
2005; Valencia-Sanchez et al. 2006). The data presented
in this work indicate that this scenario also applies to the
miRNA-mediated regulation of an endogenous mRNA.
The stress- or anti-miR-122-induced relocalization of
CAT-1 mRNA from P bodies was accompanied by its
increased association with polysomes, consistent with the
miRNA inhibition acting at the initiation step of transla-
tion. Importantly, activation of CAT-1 and reporter
mRNAs by exposing the cells to stress or transfecting
them with anti-miR-122 oligonucleotide had no apprecia-
ble effect on the mRNA level, indicating that miR-122 in
Huh7 cells controls CAT-1 mRNA mainly at the transla-
tional level and not the stability level.
Our results strongly argue for a role of HuR in the
stress-induced activation of mRNAs undergoing miR-
122-mediated repression. The HuR knockdown pre-
vented both the translational activation of repressed
mRNAs and their mobilization from P bodies.
513-522_Bhattacharyya_Symp71.qxd 2/8/07 1:42 PM Page 519

Moreover, a recombinant HuR interacted with the CAT-
1 3′UTR fragment implicated in mediating the stimula-
tory effect of stress (see Fig. 3D), whereas the
endogenous HuR associated with all reporter mRNAs
undergoing derepression but not with their inactive vari-
ants bearing mutations in the predicted HuR-binding
sites (Bhattacharyya et al. 2006). HuR is a ubiquitously
expressed member of the ELAV family of proteins,
which also comprises three neuronal proteins. In
response to different types of cellular stress, HuR is
mobilized from the nucleus to the cytosol, where it may
modulate translation and/or stability of different
mRNAs (for review, see Brennan and Steitz 2001;
Katsanou et al. 2005; Lal et al. 2005). Our data suggest
that at least some of the known effects of HuR, both
translational and stability-related, may be due to the
interference of HuR with the function of miRNAs,
which would result in enhanced translation or stability
of mRNA. HuR shuttles between the nucleus and cyto-
plasm, and HuR has been suggested to bind some ARE-
containing mRNAs in the nucleus and chaperone them
to the cytoplasm (Gallouzi and Steitz 2001; Lal et al.
2005). The in situ experiments demonstrating CAT-1
mRNA and RL reporter localization in P bodies in
unstressed hepatoma and HeLa cells make it very
unlikely that redistribution of mRNA between the
nucleus and the cytoplasm contributes to the effects
described in our work.
The demonstration that mRNAs repressed by miRNAs
can depart P bodies to return to active translation indi-
cates that P bodies are dynamic structures, exchanging
their content rapidly with that of the cytosol (Andrei et
al. 2005; Brengues et al. 2005). It is possible that HuR,
following its relocation to the cytoplasm in stressed
cells, shifts the P-body-to-cytosol equilibrium of
repressed mRNAs by binding to AREs in the 3′UTR.
Whether this is accompanied by the dissociation of
miRNPs from the mRNA or just prevents miRNPs from
acting as effectors in the repression remains to be estab-
lished (Fig. 7). It will also be interesting to study other
details of HuR involvement in relieving the miRNP
repression. The findings that several identified protein
ligands of HuR are protein phosphatase PP2A inhibitors
(Brennan et al. 2000) and that HuR can undergo methy-
lation (Li et al. 2002) or synergize with other RNA-bind-
ing proteins (Katsanou et al. 2005) indicate that HuR is a
part of an elaborated network involved in posttranscrip-
tional regulation of gene expression.
Other examples of the reversible action of miRNAs
have recently been identified in neuronal cells. In neu-
rons, many mRNAs are transported along the dendrites
as repressed mRNPs to become translated at the final
destination, dendritic spines, upon synaptic activation.
Such local translation is important for spine develop-
ment, learning, and memory (Sutton and Schuman
2005). miRNA miR-134 is implicated in translational
regulation of Limk1, a protein kinase important for
spine development, in cultured rat neurons. Limk1
mRNA appears to be relieved from the miR-134-medi-
ated repression in dendritic spines in response to extra-
cellular stimuli, in a process involving mTOR
(mammalian target of rapamycin) (Schratt et al. 2006).
In Drosophila, stimulation of olfactory neurons, which
leads to long-term memory formation, is associated with
proteolysis of Armitage, a protein essential for the
assembly of the RNA-induced silencing complex
(RISC)/miRNP complexes. As the result of Armitage
degradation, mRNAs that are normally repressed by
miRNAs, including the one encoding calcium/calmod-
ulin-dependent protein kinase II (CamKII), become
520 BHATTACHARYYA ET AL.
Figure 7. Model of the stress- and HuR-mediated relief of CAT-1 mRNA repression by miR-122.
513-522_Bhattacharyya_Symp71.qxd 2/8/07 1:42 PM Page 520

effectively translated in the synapse (Ashraf et al. 2006).
The regulation in Drosophila differs from the examples
studied in mammalian cells, in that the Armitage deple-
tion most probably indiscriminately prevents the forma-
tion of repressed mRNPs, rather than causing
reactivation of specific mRNAs controlled by miRNAs.
In addition to HuR, three other ELAV proteins,
HuA, HuB, and HuD, are expressed in neurons, and a
role of HuD in stability and translation of some neu-
ronal mRNAs has already been documented (Perrone-
Bizzozero and Bolognani 2002). It will be interesting
to find out whether, similarly to HuR in hepatoma
cells, the other ELAV proteins modulate miRNA-
mediated regulation in neurons. Likewise, it will be
important to determine whether other classes of RNA-
binding proteins interacting with the 3′UTR will act as
modifiers altering the potential of miRNAs to repress
gene expression.
ACKNOWLEDGMENTS
We thank C. Clayton, T. Hobman, Y. Nagamine, C.
Ender, R. Pillai, C. Artus, and E. Bertrand for providing
plasmids and/or antibodies. S.N.B. is a recipient of a long-
term HFSP fellowship. The Friedrich Miescher Institute is
supported by the Novartis Research Foundation.
REFERENCES
Andrei M.A., Ingelfinger D., Heintzmann R., Achsel T.,
Rivera-Pomar R., and Luhrmann R. 2005. A role for eIF4E
and eIF4E-transporter in targeting mRNPs to mammalian
processing bodies. RNA 11: 717.
Ambros V. 2004. The functions of animal microRNAs. Nature
431: 350.
Ashraf S.I., McLoon A.L., Sclarsic S.M., and Kunes S. 2006.
Synaptic protein synthesis associated with memory is regu-
lated by the RISC pathway in Drosophila. Cell 124: 191.
Bartel D.P. 2004. MicroRNAs: Genomics, biogenesis, mecha-
nism, and function. Cell 116: 281.
Bhattacharyya S.N., Habermacher R., Martine U., Closs E.I.,
and Filipowicz W. 2006. Relief of microRNA-mediated
translational repression in human cells subjected to stress.
Cell 125: 1111.
Brengues M., Teixeira D., and Parker R. 2005. Movement of
eukaryotic mRNAs between polysomes and cytoplasmic
processing bodies. Science 310: 486.
Brennan C.M. and Steitz J.A. 2001. HuR and mRNA stability.
Cell. Mol. Life Sci. 58: 266.
Brennan C.M., Gallouzi I.E., and Steitz J.A. 2000. Protein
ligands to HuR modulate its interaction with target mRNAs
in vivo. J. Cell Biol. 151: 11.
Chang J., Nicolas E., Marks D., Sander C., Lerro A., Buendia
M.A., Xu C., Mason W.S., Moloshok T., Bort R., et al.
2004. miR-122, a mammalian liver-specific microRNA, is
processed from hcr mRNA and may downregulate the high
affinity cationic amino acid transporter CAT-1. RNA Biol.
1: 106.
Filipowicz W. 2005. RNAi: The nuts and bolts of the RISC
machine. Cell 122: 17.
Gallouzi I.E. and Steitz J.A. 2001. Delineation of mRNA
export pathways by the use of cell-permeable peptides.
Science 294: 1895.
Hatzoglou M., Fernandez J., Yaman I., and Closs E. 2004.
Regulation of cationic amino acid transport: the story of the
CAT-1 transporter. Annu. Rev. Nutr. 24: 377.
Humphreys D.T., Westman B.J., Martin D.I., and Preiss T.
2005. MicroRNAs control translation initiation by inhibit-
ing eukaryotic initiation factor 4E/cap and poly(A) tail func-
tion. Proc. Natl. Acad. Sci. 102: 16961.
Jakymiw A., Lian S., Eystathioy T., Li S., Satoh M., Hamel
J.C., Fritzler M.J., and Chan E.K. 2005. Disruption of GW
bodies impairs mammalian RNA interference. Nat. Cell
Biol. 7: 1267.
Katsanou V., Papadaki O., Milatos S., Blackshear P.J.,
Anderson P., Kollias G., and Kontoyiannis D.L. 2005. HuR
as a negative posttranscriptional modulator in inflammation.
Mol. Cell 16: 777.
Krutzfeldt J., Rajewsky N., Braich R., Rajeev K.G., Tuschl T.,
Manoharan M., and Stoffel M. 2005. Silencing of
microRNAs in vivo with “antagomirs.” Nature 438: 685.
Lal A., Kawai T., Yang X., Mazan-Mamczarz K., and Gorospe
M. 2005. Antiapoptotic function of RNA-binding protein
HuR effected through prothymosin alpha. EMBO J. 24:
1852.
Li H., Park S., Kilburn B., Jelinek M.A., Henschen-Edman A.,
Aswad D., Stallcup M.R., and Laird-Offringa I.A. 2002.
Lipopolysaccharide-induced methylation of HuR, an
mRNA-stabilizing protein, by CARM1. J. Biol. Chem. 277:
44623.
Liu J., Valencia-Sanchez M.A., Hannon G.J., and Parker R.
2005. MicroRNA-dependent localization of targeted
mRNAs to mammalian P-bodies. Nat. Cell Biol. 7: 719.
Liu J., Carmell M.A., Rivas F.V., Marsden C.G., Thomson
J.M., Song J.J., Hammond S.M., Joshua-Tor L., and Hannon
G.J. 2004. Argonaute2 is the catalytic engine of mammalian
RNAi. Science 305: 1437.
Meister G., Landthaler M., Patkaniowska A., Dorsett Y., Teng
G., and Tuschl T. 2004. Human Argonaute2 mediates RNA
cleavage targeted by miRNAs and siRNAs. Mol. Cell 15:
185.
Olsen P.H. and Ambros V. 1999. The lin-4 regulatory RNA
controls developmental timing in Caenorhabditis elegans
by blocking LIN-14 protein synthesis after the initiation of
translation. Dev. Biol. 216: 671.
Perrone-Bizzozero N. and Bolognani F. 2002. Role of HuD
and other RNA-binding proteins in neural development and
plasticity. J. Neurosci. Res. 68: 121.
Petersen C.P., Bordeleau M.E., Pelletier J., and Sharp P.A.
2006. Short RNAs repress translation after initiation in
mammalian cells. Mol. Cell 21: 533.
Pillai R.S. 2005. MicroRNA function: Multiple mechanisms
for a tiny RNA? RNA 11: 1753.
Pillai R.S., Artus C.G., and Filipowicz W. 2004. Tethering of
human Ago proteins to mRNA mimics the miRNA-medi-
ated repression of protein synthesis. RNA 10: 1518.
Pillai R.S., Bhattacharyya S.N., Artus C.G., Zoller T., Cougot
N., Basyuk E., Bertrand E., and Filipowicz W. 2005.
Inhibition of translational initiation by let-7 miRNA in
human cells. Science 309: 1573.
Schratt G.M., Tuebing F., Nigh E.A., Kane C.G., Sabatini
M.E., Kiebler M., and Greenberg M.E. 2006. A brain-
specific miRNA regulates dendritic spine development.
Nature 439: 283.
Sutton M.A. and Schuman E.M. 2005. Local translational con-
trol in dendrites and its role in long-term synaptic plasticity.
J. Neurobiol. 64: 116.
Tomari Y. and Zamore P.D. 2005. Perspective: Machines for
RNAi. Genes Dev. 19: 517.
Valencia-Sanchez M.A., Liu J., Hannon G.J., and Parker R.
2006 Control of translation and mRNA degradation by
miRNAs and siRNAs. Genes Dev. 20: 515.
Wienholds E. and Plasterk R.H. 2005. MiRNA function in ani-
mal development. FEBS Lett. 579: 5911.
Yaman I., Fernandez J., Sarkar B., Schneider R.J., Snider
M.D., Nagy L.E., and Hatzoglou M. 2002. Nutritional con-
trol of mRNA stability is mediated by a conserved AU-rich
element that binds the cytoplasmic shuttling protein HuR. J.
Biol. Chem. 277: 41539.
REVERSAL OF MIRNA-MEDIATED REPRESSION 521
513-522_Bhattacharyya_Symp71.qxd 2/8/07 1:42 PM Page 521

513-522_Bhattacharyya_Symp71.qxd 2/8/07 1:42 PM Page 522