See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/8687645

Stress-Dependent Regulation of FOXO Transcription Factors by the SIRT1

Deacetylase

Article in Science · April 2004

DOI: 10.1126/science.1094637 · Source: PubMed

CITATIONS READS
3,141 1,199

17 authors, including:

Anne Brunet Lora B Sweeney

Stanford University Institute of Science and Technology Austria
222 PUBLICATIONS 47,718 CITATIONS 19 PUBLICATIONS 3,821 CITATIONS

SEE PROFILE SEE PROFILE

James Sturgill Katrin F Chua

University of California, San Diego Stanford University
5 PUBLICATIONS 3,840 CITATIONS 55 PUBLICATIONS 16,469 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Haim Y Cohen on 27 May 2014.

The user has requested enhancement of the downloaded file.

 Stress-Dependent Regulation of FOXO
Transcription Factors by the SIRT1 Deacetylase
Anne Brunet, et al.
Science 303, 2011 (2004);
DOI: 10.1126/science.1094637

This copy is for your personal, non-commercial use only.

If you wish to distribute this article to others, you can order high-quality copies for your
colleagues, clients, or customers by clicking here.

Permission to republish or repurpose articles or portions of articles can be obtained by

following the guidelines here.

The following resources related to this article are available online at www.sciencemag.org
(this information is current as of October 28, 2010 ):

Updated information and services, including high-resolution figures, can be found in the online
version of this article at:

Downloaded from www.sciencemag.org on October 28, 2010

http://www.sciencemag.org/cgi/content/full/303/5666/2011

Supporting Online Material can be found at:

http://www.sciencemag.org/cgi/content/full/1094637/DC1
A list of selected additional articles on the Science Web sites related to this article can be
found at:
http://www.sciencemag.org/cgi/content/full/303/5666/2011#related-content
This article cites 26 articles, 11 of which can be accessed for free:
http://www.sciencemag.org/cgi/content/full/303/5666/2011#otherarticles

This article has been cited by 703 article(s) on the ISI Web of Science.

This article has been cited by 92 articles hosted by HighWire Press; see:
http://www.sciencemag.org/cgi/content/full/303/5666/2011#otherarticles

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2004 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.
 REPORTS
the FOXO family member FOXO3 is acety-
Stress-Dependent Regulation of lated within cells and that the acetylation of
FOXO3 is increased in response to hydro-

FOXO Transcription Factors by gen peroxide (H2O2) and slightly enhanced

in response to heat shock (Fig. 1A) (26). In

the SIRT1 Deacetylase

contrast, FOXO3 acetylation was not af-
fected by ultraviolet (UV ) irradiation or
growth factor stimulation (Fig. 1A) (26).
Anne Brunet,1* Lora B. Sweeney,1 J. Fitzhugh Sturgill,1 Katrin F. Thus, FOXO3 acetylation appears to be
Chua,2 Paul L. Greer,1 Yingxi Lin,1 Hien Tran,1 Sarah E. Ross,1 specifically induced in response to oxida-
Raul Mostoslavsky,2 Haim Y. Cohen,3 Linda S. Hu,1 Hwei-Ling tive stress stimuli. Consistent with this ob-
Cheng,2 Mark P. Jedrychowski,4 Steven P. Gygi,4 David A. servation, we found that protein acetylases,
including p300/CBP-associated factor
Sinclair,3 Frederick W. Alt,2 Michael E. Greenberg1† (PCAF), interact with FOXO in a stress-
inducible manner (fig. S1).
The Sir2 deacetylase modulates organismal life-span in various species. How- We tested whether SIRT1 might be re-
ever, the molecular mechanisms by which Sir2 increases longevity are largely cruited to FOXO3 to possibly regulate
unknown. We show that in mammalian cells, the Sir2 homolog SIRT1 appears FOXO3 acetylation. In the absence of stress
to control the cellular response to stress by regulating the FOXO family of

Downloaded from www.sciencemag.org on October 28, 2010

stimuli, when growth factors are present,
Forkhead transcription factors, a family of proteins that function as sensors of SIRT1 is a nuclear protein (10, 11), whereas
the insulin signaling pathway and as regulators of organismal longevity. SIRT1 FOXO transcription factors are localized in
and the FOXO transcription factor FOXO3 formed a complex in cells in response the cytoplasm (16, 18). However, immuno-
to oxidative stress, and SIRT1 deacetylated FOXO3 in vitro and within cells. staining experiments revealed that, when
SIRT1 had a dual effect on FOXO3 function: SIRT1 increased FOXO3’s ability growth factors are present, oxidative stress
to induce cell cycle arrest and resistance to oxidative stress but inhibited stimuli, including H2O2, menadione, and heat
FOXO3’s ability to induce cell death. Thus, one way in which members of the shock, triggered the translocation of FOXO3
Sir2 family of proteins may increase organismal longevity is by tipping FOXO- from the cytoplasm to the nucleus (Fig. 1B).
dependent responses away from apoptosis and toward stress resistance. In the absence of growth factors, FOXO3 was
already in the nucleus, and under these con-
Although all species have a defined life-span, ate the effects of the Sir2 family on longevity ditions oxidative stress stimuli did not affect
simple modifications of the environment, in- or whether Sir2 has other important sub- FOXO3 subcellular localization (Fig. 1B)
cluding caloric restriction or sublethal levels strates that account for its effect on longevity (27). In contrast to FOXO3, SIRT1 was al-
of stress, can substantially affect organismal is still unclear. ways nuclear, in the absence or presence of
longevity. A pivotal regulator of organismal In C. elegans, the ability of Sir2 to extend stress stimuli (26). Thus, under conditions of
life-span is the Sir2 (silencing information life-span depends on the presence of Daf-16, oxidative stress, FOXO3 and SIRT1 are both
regulator 2) gene (1). Deletion of Sir2 in a member of the FOXO family of Forkhead present within the nucleus.
yeast abolishes the increase in life-span in- transcription factors (5, 14, 15). We hypoth- To determine whether FOXO3 and SIRT1
duced by caloric restriction or sublethal lev- esized that, in higher organisms, FOXO tran- interact in response to oxidative stress, we per-
els of stress, indicating that Sir2 is a mediator scription factors might be direct substrates of formed coimmunoprecipitation experiments in
of signals that promote longevity (2, 3). Ex- SIRT1. In mammals, there are four evolution- cells that expressed a hemagglutinin epitope
pression of extra copies of Sir2 is sufficient to arily conserved FOXO family members (HA)–tagged form of FOXO3 and a Flag-
increase life-span in yeast (4) and in Caeno- (FOXO1, FOXO3, FOXO4, and FOXO6) tagged version of SIRT1 (28). SIRT1 and
rhabditis elegans (5), suggesting that the role that are negatively regulated by the insulin- FOXO3 interacted, and this interaction was in-
of Sir2 in extending organismal life-span is phosphoinositide-3 kinase (PI3K)–Akt sig- creased in response to several types of oxidative
conserved through evolution. naling pathway (16–19). Mammalian FOXO stress stimuli (H2O2, menadione, and heat
Sir2 is a nicotinamide adenine dinucleotide factors control various biological functions, shock), but not in response to UV or gamma
(NAD)– dependent histone deacetylase (6, 7). including cell cycle arrest at the G1-S and irradiation or to growth factors (Fig. 1C).
In mammals, there are seven members of the G2-M checkpoints (20, 21), detoxification of We tested whether the interaction between
Sir2 family, termed sirtuins (SIRTs), among reactive oxygen species (ROS) (22, 23), re- FOXO3 and SIRT1 was detectable when both
which SIRT1 is the closest homolog of yeast pair of damaged DNA (21), and apoptosis proteins were expressed at their endogenous
Sir2 (8). We will here refer to mammalian Sir2 (16, 24). Because the ability to detoxify ROS levels. Immunoprecipitation experiments
as SIRT1. In addition to deacetylating histones, and to repair damage is correlated with in- with an antibody to FOXO3 or a control
SIRT1 also deacetylates other proteins, includ- creased organismal longevity in many species preimmune immunoglobulin G (IgG) re-
ing MyoD and the tumor suppressor p53 (9– (25), these particular functions of FOXO vealed that endogenous FOXO3 interacted
13). However, whether these proteins medi- transcription factors may be relevant to with endogenous SIRT1 in 293T cells and in
FOXO’s ability to control longevity. cerebellar granule neurons and that this inter-
1
Division of Neuroscience, Children’s Hospital, and We sought to identify the conditions un- action increased in response to oxidative
Department of Neurobiology, 2Howard Hughes Med- der which FOXO transcription factors are stress (Fig. 1, D and E) (26). These results
ical Institute, Children’s Hospital, Center for Blood highly acetylated in mammalian cells, be- indicate that the interaction between endoge-
Research (CBR) Institute for Biomedical Research 3De- cause these might be conditions in which nous FOXO3 and SIRT1 is detectable in
partment of Pathology, 4Department of Cell Biology,
Harvard Medical School, Boston, MA 02115, USA.
SIRT1 regulates FOXO transcription factors. primary cells and immortalized cell lines and
We treated cells with various extracellular that the FOXO3-SIRT1 interaction is en-
*Present address: Department of Genetics, Stanford
University, Stanford, CA 94305, USA.
stimuli and performed Western blot experi- hanced in response to oxidative stress.
†To whom correspondence should be addressed. E- ments with antibodies to acetylated lysine We next asked if the presence of FOXO3
mail: [email protected] (Fig. 1A). These experiments revealed that in the nucleus was sufficient to promote the

www.sciencemag.org SCIENCE VOL 303 26 MARCH 2004 2011

 REPORTS
interaction of FOXO3 with SIRT1. A mutant tein complex, FOXO3 is a substrate of form of FOXO3 in vitro only in the pres-
form of FOXO3 that is constitutively nuclear SIRT1. To determine if SIRT1 directly ence of NAD. The deacetylation of FOXO3
because its three Akt phosphorylation sites deacetylated FOXO3, we first acetylated by rSIRT1 was partially inhibited by BML-
have been mutated to alanines (16) did not FOXO3 in vitro by incubating FOXO3 with 210, an inhibitor of SIRT1, but not by
interact with SIRT1 in the absence of stress PCAF or the p300 acetylase, in the pres- Trichostatin ( TSA), an inhibitor of class I
stimuli (Fig. 1F). Thus, the presence of ence of acetyl– coenzyme A (acetyl-CoA), and class II histone deacetylases (HDACs)
FOXO3 in the nucleus appears not to be and then used the acetylated form of (Fig. 2B). Thus, SIRT1 appears to deacety-
sufficient to promote the interaction between FOXO3 as a substrate for recombinant late FOXO3 in vitro.
SIRT1 and FOXO3. We have found by mass SIRT1 (rSIRT1) (28). rSIRT1 deacetylated To determine if SIRT1 might contribute to
spectrometry that FOXO3 is acetylated at FOXO3 that had been acetylated by PCAF FOXO3 deacetylation within cells, we incu-
five different lysine residues and phosphoryl- or p300 in vitro only in the presence of the bated 293T cells with SIRT1 chemical inhib-
ated at eight serine or threonine residues in cofactor NAD (Fig. 2A). itors (nicotinamide, BML-210, and splitomi-
the presence of stress stimuli (fig. S2). There- We next tested whether SIRT1 deacety- cin) and with TSA. We then immunopreci-
fore, a combination of stress-induced FOXO3 lated FOXO3 that had been acetylated within pitated Flag-FOXO3 from these cells and
acetylation and phosphorylation events may cells in response to oxidative stress. We pu- analyzed the immune complex by Western
promote the interaction between SIRT1 and rified FOXO3 from 293T cells that were blot with antibodies to acetylated lysine (28).
FOXO3. treated in the presence of H2O2 and incu- Treatment with SIRT1 inhibitors alone or
The observations that SIRT1 and FOXO bated the purified acetylated form of with TSA alone had no effect on FOXO3

Downloaded from www.sciencemag.org on October 28, 2010

interact and that FOXO3 is acetylated FOXO3 with rSIRT1 (Fig. 2B). rSIRT1 acetylation (Fig. 2C). By contrast, incubation
raised the possibility that, within this pro- deacetylated the stress-induced acetylated of cells with TSA and nicotinamide together

Fig. 1. Interaction of SIRT1 and

FOXO3 in response to oxidative
stress. (A) Acetylation of FOXO3 in
cells in response to stress stimuli.
293T cells expressing the Flag-tagged
form of FOXO3 were treated with
H2O2 (500 ␮M, 1 hour), UV radiation
(500 J/m2, 1 hour), or heat shock (HS,
42°C, 1 hour). The acetylation of im-
munoprecipitated FOXO3 was as-
sessed by Western blot with the anti-
bodies to acetylated lysine (Acetyl-K)
and total levels of FOXO3 by Western
blot with an antibody to Flag. (B) Lo-
calization of FOXO3 in response to
stress. CCl39 fibroblasts were trans-
fected with the HA-tagged form of
FOXO3 and were treated with heat
shock (42°C) in the absence or presence of growth factors. Immunoflu- SIRT1 by Western blot with an antibody to SIRT1. Amounts of
orescence experiments were done with an antibody to HA. Data repre- immunoprecipitated FOXO3 were assessed by Western blot with an
sent the mean and error of two independent experiments. C, cytoplasmic antibody to FOXO3. Amounts of endogenous SIRT1 in the cell ex-
FOXO3; N, nuclear FOXO3. (C) Interaction of transfected FOXO3 and tracts were determined with an antibody to SIRT1. (E) Effects of stress
SIRT1. 293T cells were cotransfected with the HA-tagged version of on the interaction between FOXO3 and SIRT1. 293T cells were either
FOXO3 and a Flag-tagged version of human SIRT1. Cells were treated nonstimulated or stimulated by H2O2 (500 ␮M, 1 hour), and FOXO3
with a variety of stimuli, including UV radiation (500 J/m2, 1 hour), H2O2 and SIRT1 interaction was analyzed as in (D). The fact that antibodies
(500 ␮M, 1 hour), menadione (Me, 20 ␮M, 1 hour), heat shock (42°C, 1 to endogenous FOXO3 appear to coimmunoprecipitate less endoge-
hour), gamma irradiation (␥IR, 20 gray, 1 hour), or IGF-1 (100 ng/ml, 1 nous Sir2 in response to oxidative stress than the antibodies to
hour). Flag-tagged SIRT1 was immunoprecipitated (IP), and the immune Flag-FOXO3 (C) may be explained by a difference in the affinity of
complexes were assessed for the presence of FOXO3 by Western blot these two different antibodies for their antigenic determinants. (F)
(WB) with the antibody to HA. The amounts of HA-FOXO3 present in Interaction of mutant FOXO3 with SIRT1. 293T cells were cotrans-
the cell extracts were analyzed with the antibody to HA and amounts of fected with the Flag-tagged form of wild-type SIRT1 and with the
immunoprecipitated SIRT1 with the antibody to Flag. (D) Interaction of HA-tagged form of either wild-type FOXO3 (WT) or a mutant of
endogenous FOXO3 and SIRT1. 293T cells were stimulated by H2O2 (500 FOXO3 in which all three phosphorylation sites have been replaced by
␮M, 1 hour). Endogenous FOXO3 was immunoprecipitated with a mouse alanines (TM). Cells were incubated in the presence (⫹) or absence (–)
polyclonal antibody to FOXO3 (FOXO3) or a control preimmune mouse of H2O2 (500 ␮M, 1 hour). The immune complexes were analyzed
serum (PI). The immune complexes were assessed for the presence of as in (C).

2012 26 MARCH 2004 VOL 303 SCIENCE www.sciencemag.org

 REPORTS
led to an increase in the acetylation of ectopi- assessed the acetylation of FOXO3 in wild- (22, 23), cell cycle arrest (p27KIP1) (20), and
cally expressed or endogenous FOXO3 and type or SIRT1 knockout cells, by immuno- cell death (Fas Ligand and BIM) (16, 24).
of endogenous p53 (Fig. 2, C and D). These precipitating acetylated proteins with two dif- To gain insight into the effect of SIRT1 on
results suggest that both SIRT1 and Class I ferent antibodies that recognize many acety- FOXO3-dependent gene transcription, we used a
and II HDACs contribute to FOXO3 deacety- lated sites in proteins (Fig. 2F and fig. S3) fibroblast cell line that expresses an inducible
lation within cells. and by detecting the presence of FOXO3 in form of FOXO3 consisting of a fusion between
To address more definitively whether en- the immune complex with antibodies to the active form of FOXO3 (a T32A/S253A/
dogenous SIRT1 has a role in deacetylating FOXO3. Amounts of acetylated FOXO3 S315A triple mutant) and the ligand-binding do-
FOXO3 in vivo, we assessed FOXO3 acety- were greater in two independent lines of main of the estrogen receptor (FOXO3-ER) (21).
lation in mouse embryonic fibroblasts SIRT1–/– MEFs (KO1 and KO2) than that in In the absence of ligand, FOXO3-ER is main-
(MEFs) derived from SIRT1 knockout mice wild-type cells (Fig. 2F and fig. S3). These tained in an inactive state in these cells. Addition
or from wild-type littermates by immunopre- data indicate that SIRT1 is one of multiple of 4-hydroxytamoxifen (4-OHT ), an artificial
cipitating endogenous FOXO3 and immuno- FOXO3 deacetylases in mammalian cells. ligand for the estrogen receptor, triggers the rap-
blotting with antibodies to acetylated lysine FOXO transcription factors transactivate a id activation of FOXO3-ER (21). In control cell
(13) (Fig. 2E). Acetylation of endogenous series of target genes that have critical roles in lines that do not express FOXO3-ER, 4-OHT
FOXO3 was enhanced in SIRT1–/– MEFs the cellular response to stress stimuli. FOXO has no effect on the expression of FOXO3 target
compared to wild-type MEFs (Fig. 2E), sug- targets include genes that control repair of dam- genes (21).
gesting that endogenous SIRT1 does influ- aged DNA (GADD45) (21), ROS detoxifica- Addition of 4-OHT to FOXO3-ER–

Downloaded from www.sciencemag.org on October 28, 2010

ence FOXO3 acetylation in vivo. We also tion (Mn superoxide dismutase and catalase) expressing cells moderately increased the ex-

Fig. 2. Deacetylation of
FOXO3 by SIRT1 in vitro and in
cells. (A) The Flag-tagged form
of FOXO3 was purified from
293T cells and incubated with
PCAF or the histone acetyl
transferase (HAT) domain of
p300, in the presence of
acetyl-CoA. Acetylated FOXO3
was incubated in the absence
or presence of rSIRT1 and NAD,
then detected by Western blot
with acetyl-K. Total amounts
of FOXO3 were assessed with
an antibody to Flag. (B) The
Flag-tagged form of FOXO3
was purified from 293T cells
stimulated by H2O2 (500 ␮M,
1 hour). Purified FOXO3 was
incubated in the presence or
absence of rSIRT1, NAD, and
the indicated inhibitors. Ace-
tylated FOXO3 and total
amounts of FOXO were measured as in (A). (C) 293T cells expressing FOXO3 was assessed by immunoblotting with antibodies to acetylated
Flag-tagged FOXO3 were incubated for 6 hours in the absence or lysine, amounts of FOXO3 with an antibody to FOXO3, and endogenous
presence of SIRT1 inhibitors [nicotinamide (N), BML-210 (B), and splito- SIRT1 with an antibody to SIRT1. The acetylation of endogenous p53 was
micin (S)] and in the absence or presence of TSA. Flag-tagged FOXO3 was detected with an antibody to acetylated p53. Molecular size markers (kD)
immunoprecipitated, and the acetylation of FOXO3 was assessed by are indicated. (F) MEFs derived from wild-type or two independent litter-
Western blot with the antibodies to acetylated lysine. Total levels of mate SIRT1–/– embryos (KO1 and KO2) were incubated in the presence of
FOXO3 were assessed with the antibody to Flag. The acetylation of the TSA for 2 hours. Endogenous acetylated proteins were immunoprecipitated
endogenous p53 in the same cellular extracts were analyzed by Western with antibodies to acetylated lysines (Cell Signaling and Technology). The
blot with an antibody directed against the acetylated lysine 382 of p53 presence of FOXO3 in the immune complex and FOXO3 levels in the
(30). (D) Endogenous FOXO acetylation analyzed as in (C). Endogenous extract were assessed by Western blot with antibodies to FOXO3. The
SIRT1 was detected with an antibody to SIRT1. (E) MEFs derived from presence of endogenous SIRT1 was detected by Western blot with an
wild-type or SIRT1–/– (KO1) embryos were incubated in the presence of antibody to SIRT1. The control panel represents the IgG that corresponds
TSA for 2 hours. 293T cells were stimulated with nicotinamide and TSA to the anti-acetylated antibodies and controls for the addition of the
for 6 hours (293 lane). Acetylation of immunoprecipitated endogenous same quantity of anti-acetylated antibodies in each sample.

www.sciencemag.org SCIENCE VOL 303 26 MARCH 2004 2013

 REPORTS
pression of the FOXO3 targets GADD45 and (28) showed that LY treatment of MEFs led to sary for FOXO-mediated cell cycle arrest, we
p27 (Fig. 3, A and B). The increased expression the increased expression of the stress resistance compared effects of FOXO3 on cell cycle
of GADD45 and p27 was inhibited in the pres- gene GADD45 (Fig. 3C). This response to LY arrest in MEFs derived from either SIRT1–/–
ence of nicotinamide and TSA, a combination was reduced in two independent lines of mice or their wild-type littermates (Fig. 4B).
of inhibitors for SIRT1 and Class I and II SIRT1–/– fibroblasts (Fig. 3C). Thus, endoge- We infected SIRT1–/– or wild-type MEFs
HDACs that led to the greatest acetylation of nous SIRT1 appears to contribute to increased with a retrovirus that encodes an active form
endogenous FOXO3 (Fig. 3, A and B). How- expression of the stress response gene GADD45 of FOXO3 and assessed cell cycle progres-
ever, FOXO3-induced expression of the cell- in response to inhibition of the PI3K pathway. In sion. The constitutively active form of
death gene BIM was not inhibited by treatment contrast, amounts of BIM mRNA were variable, FOXO3 still induced cell cycle arrest in wild-
of cells with nicotinamide and TSA, indicating and on average, LY treatment increased BIM type and in SIRT1–/– MEFs. However, the
that not all FOXO3 target genes are affected in expression to a similar extent in wild-type and effect of the constitutively active form of
the same way by deacetylase inhibitors. Con- SIRT1–/– cells (Fig. 3D). These results provide FOXO3 on cell cycle arrest was diminished
sistent with the possibility that deacetylases further evidence that SIRT1 does not regulate all in SIRT1–/– MEFs (Fig. 4B). In addition, LY,
may differentially affect various FOXO target FOXO target genes in the same manner. whose presence leads to endogenous FOXO
genes, we found that SIRT1 repressed FOXO- The finding that the presence of SIRT1 activation, was also less effective in inducing
dependent expression of a luciferase reporter enhanced expression of a FOXO target in- cell cycle arrest in SIRT1–/– MEFs compared
gene that is driven by the promoter of the gene volved in stress resistance (GADD45) but to wild-type MEFs (fig. S6). These results
encoding the death cytokine Fas ligand (fig. appeared to somewhat diminish expression of indicate that endogenous SIRT1 may poten-

Downloaded from www.sciencemag.org on October 28, 2010

S4). Accordingly, in preliminary experiments proapoptotic FOXO targets (Fas ligand and tiate FOXO3’s ability to induce cell cycle
using Affymetrix gene arrays, we found that BIM) led us to consider the possibility that arrest, possibly allowing more time for cells
SIRT1 expression in FOXO3-ER– expressing SIRT1 could modulate the balance between to detoxify ROS and to repair damaged DNA.
fibroblasts promoted increased expression of stress resistance and cell death within cells. FOXO transcription factors can also pro-
one set of genes and repressed expression of We generated FOXO3-ER stable cell lines mote apoptosis (16, 24). We therefore tested
another (29). that expressed either an empty vector or wild- the effects of SIRT1 on FOXO-induced cell
To assess further the role of endogenous type mouse SIRT1 (fig. S5). The activation of death in primary cultures of cerebellar gran-
SIRT1 in FOXO3-dependent gene expression, FOXO3-ER induced cell cycle arrest at the ule neurons (Fig. 4C). Expression of FOXO3
we analyzed the expression of FOXO3 target G1-S phase transition (28) (Fig. 4A). In cells in neurons led to a modest increase in cell
genes in wild-type or SIRT1–/– MEFs that were expressing extra copies of SIRT1, the ability death, and this effect decreased when exoge-
treated with LY 294002 (LY ), a chemical com- of FOXO3 to induce cell cycle arrest at the nous SIRT1 was also expressed (Fig. 4C).
pound that inhibits PI3K, thereby leading to the G1-S transition was enhanced compared to We also tested whether SIRT1 affects
activation of endogenous FOXO3 (16). Quanti- that in cells expressing the empty vector. To FOXO3-mediated cell death in nonneuronal
tative real-time polymerase chain reaction (PCR) determine if endogenous SIRT1 was neces- cell types. In fibroblasts, activation of

Fig. 3. Differential ef-

fects of SIRT1 on
FOXO3 target genes.
(A) Stable cell lines of
Rat1 fibroblasts ex-
pressing the inducible
version of FOXO3
(FOXO3-ER) were sti-
mulated for 8 hours
with 4-OHT, in the
presence or absence
of nicotinamide and
TSA (N⫹T ). Amounts
of BIM, p27, or GAPDH
mRNA were detect-
ed by reverse tran-
scription–PCR. This ex-
periment is represen-
tative of two indepen-
dent experiments. (B)
Cells were treated as
in (A). The expression
of FOXO3 target gene
products was assessed
by Western blot with antibodies to BIM, p27, or GADD45. The total amount of protein expression was assessed
by comparison to a background band recognized by a control antibody. This experiment is representative of
three independent experiments. (C) Reduced FOXO3-dependent expression of GADD45 in SIRT1–/– cells. One
line of wild-type MEFs and two independent lines of SIRT1–/– MEFs (KO1 and KO2) were treated with LY (15
␮M) for 5 hours. The amounts of GADD45 mRNA were quantified with real-time PCR. Results were normalized
to expression levels of hypoxanthine-guanine phosphoribosyltransferase (HPRT). The data here correspond to
the mean and error of five independent experiments for wild-type MEFs and KO1 and three independent
experiments for KO2. Statistical significance was determined by analysis of variance (ANOVA). These results
have also been repeated in another pair of wild-type MEFs and SIRT1–/– MEFs. (D) Analysis of BIM expression
as described in (C).

2014 26 MARCH 2004 VOL 303 SCIENCE www.sciencemag.org

 REPORTS
FOXO3 alone is not sufficient to induce cell dependent apoptosis in the presence of stress bility, we find that FOXO3 and p53 interact with
death. However, in the presence of stress stimuli, stimuli. By deacetylating FOXO transcription one another specifically under conditions of ox-
such as H2O2 or the DNA damage–inducing factors, SIRT1 might tip FOXO-dependent re- idative stress (fig. S7). A better understanding of
agent etoposide, activation of FOXO3 did poten- sponses away from cell death and toward stress the network of regulation between SIRT1,
tiate cell death (as measured by cleavage of resistance. Given that increased cellular stress FOXO transcription factors, and p53 may help
caspase 3) (Fig. 4D) (26). Expression of extra resistance is correlated with extended organismal reveal important mechanisms that control stress
copies of wild-type SIRT1 inhibited cell death longevity, such an action might explain the ef- resistance and organismal longevity.
induced by FOXO3 in the presence of stress fects of the Sir2 family of proteins on organismal
stimuli (Fig. 4D). Consistent with a role for longevity. Nevertheless, predicting the effect of References and Notes
1. L. Guarente, Genes Dev. 14, 1021 (2000).
SIRT1 in inhibiting stress-induced cell death in FOXO deacetylation on mammalian longevity is 2. S. J. Lin, P. A. Defossez, L. Guarente, Science 289,
fibroblasts, we found that SIRT1–/– MEFs were difficult, because FOXO transcription factors 2126 (2000).
more sensitive to H2O2-induced cell death than have distinct functions in different tissues. 3. R. M. Anderson, K. J. Bitterman, J. G. Wood, O. Med-
vedik, D. A. Sinclair, Nature 423, 181 (2003).
were wild-type MEFs (fig. S6). SIRT1’s effects on FOXO3 are reminiscent 4. M. Kaeberlein, M. McVey, L. Guarente, Genes Dev. 13,
We have shown that in mammalian cells, in of SIRT1’s effects on the tumor suppressor p53 2570 (1999).
response to oxidative stress, the deacetylase (9, 10). Under conditions of cellular stress, 5. H. A. Tissenbaum, L. Guarente, Nature 410, 227 (2001).
6. S. Imai, C. M. Armstrong, M. Kaeberlein, L. Guarente,
SIRT1 forms a protein complex with the Fork- SIRT1 deacetylation of p53 leads to an inhibition Nature 403, 795 (2000).
head transcription factor FOXO3 that contrib- of apoptosis (9, 10, 13). Given that SIRT1 also 7. J. S. Smith et al., Proc. Natl. Acad. Sci. U.S.A. 97, 6658
utes to deacetylation of FOXO3. SIRT1 differ- reduces FOXO3-induced apoptosis in the pres- (2000).
8. R. A. Frye, Biochem. Biophys. Res. Commun. 273, 793

Downloaded from www.sciencemag.org on October 28, 2010

entially affected FOXO3 function, potentiating ence of stress stimuli, it is possible that FOXO3 (2000).
FOXO3’s effect on cell cycle arrest and DNA and p53 somehow function together to mediate 9. J. Luo et al., Cell 107, 137 (2001).
repair target genes but attenuating FOXO3- the effects of SIRT1. Consistent with this possi- 10. H. Vaziri et al., Cell 107, 149 (2001).
11. E. Langley et al., EMBO J. 21, 2383 (2002).
12. M. Fulco et al., Mol. Cell 12, 51 (2003).
13. H. L. Cheng et al., Proc. Natl. Acad. Sci. U.S.A. 100,
10794 (2003).
14. K. Lin, J. B. Dorman, A. Rodan, C. Kenyon, Science 278,
1319 (1997).
15. S. Ogg et al., Nature 389, 994 (1997).
16. A. Brunet et al., Cell 96, 857 (1999).
17. G. J. Kops et al., Nature 398, 630 (1999).
18. W. H. I. Biggs, J. Meisenhelder, T. Hunter, W. K. Cavenee,
K. C. Arden, Proc. Natl. Acad. Sci. U.S.A. 96, 7421 (1999).
19. F. M. Jacobs et al., J. Biol. Chem. 278, 35959 (2003).
20. R. H. Medema, G. J. Kops, J. L. Bos, B. M. Burgering,
Nature 404, 782 (2000).
21. H. Tran et al., Science 296, 530 (2002).
22. G. J. Kops et al., Nature 419, 316 (2002).
23. S. Nemoto, T. Finkel, Science 295, 2450 (2002).
24. P. F. Dijkers et al., J. Cell Biol. 156, 531 (2002).
25. T. B. L. Kirkwood, S. N. Austad, Nature 408, 233 (2000).
26. A. Brunet, M. E. Greenberg, unpublished data.
27. In these experiments, the PI3K inhibitor LY was in-
cluded to suppress autocrine insulin-like growth fac-
Fig. 4. Differential effects of SIRT1 on FOXO- tor 1 (IGF-1) signaling, thereby uncoupling Akt from
mediated biological responses. (A) Rat1 fibro- upstream signaling pathways. The use of LY, or dif-
blasts expressing FOXO3-ER, together with ei- ferences between cell lines, may explain the differ-
ence between our findings and those of Nemoto and
ther empty vector (–) or mouse SIRT1 (⫹) Finkel (23) regarding stress-dependent FOXO3 local-
were incubated in the presence or absence of ization in the absence of growth factors.
500 nM 4-OHT for 24 hours. Cell cycle pro- 28. Materials and methods are available as supporting
gression was assessed by bromodeoxyuridine material on Science Online.
(BrdU) incorporation. More than 1000 cells 29. Y. Lin, M. E. Greenberg, unpublished data.
were scored per data point. The data represent 30. K. T. Howitz et al., Nature 425, 191 (2003).
the mean and error bars of four independent 31. We thank J. Sage, S. Paradis, and E. Griffith for helpful
experiments. Results are expressed as the comments on the manuscript; R. A. Weinberg, H. Vaziri,
and S. Imai for their gifts of SIRT1 constructs; and J. Sage
number of cells that have undergone cell cycle and T. Jacks for p53⫹/⫹ and p53–/– MEFs. Supported by a
arrest divided by the number of cells undergo- Senior Scholars Award from the Ellison Foundation, NIH
ing cell cycle arrest in the control (the FOXO3- grant no. PO1 NS35138-17, and Mental Retardation Re-
ER cell line, in the absence of 4-OHT). Statistical significance was determined by ANOVA. (B) Wild-type search Center grant no. NIHP30-HD18655 (M.E.G.). M.E.G.
MEFs (WT) or SIRT1–/– MEFs (KO) were infected with a control retrovirus (–) or with a retrovirus that acknowledges the generous contribution of the F. M. Kirby
encodes the constitutively active form of FOXO3 (⫹). Cell cycle progression was assessed by BrdU Foundation to the Division of Neuroscience. A.B. is sup-
incorporation. The results are expressed as fold cell-cycle arrest induced by FOXO3 encoded retrovirus ported by a fellowship from the Radcliffe Institute for
compared to that induced by the control retrovirus in wild-type MEFs versus SIRT1–/– MEFs. Data Advanced Studies. F.W.A. is a Howard Hughes Medical
Institute investigator and is supported by an Ellison Med-
represent the mean and error bars of three independent experiments conducted in duplicate. Statistical ical Foundation Senior Scholar Award. K.C. is supported by
significance was determined by ANOVA. (C) Cerebellar granule neurons were cotransfected with empty a Pfizer postdoctoral fellowship in Rheumatology/Immu-
vector control (–) or the constitutively active mutant of FOXO3 (⫹) together with an empty vector (⫺) nology, and R.M. by a postdoctoral fellowship from the
or SIRT1 wild-type (⫹) and a vector encoding ␤-galactosidase. Twenty-four hours after transfection, Human Frontiers Science Program.
cells were incubated in the presence of IGF-1 for an additional 24 hours. Transfected cells were stained Supporting Online Material
with the antibody to ␤-galactosidase, and DNA was stained with the Hoechst compound. Experiments www.sciencemag.org/cgi/content/full/1094637/DC1
were performed in duplicate, and 200 cells per coverslip were counted in a blinded fashion. Data Materials and Methods
correspond to the mean and error of three independent experiments. Statistical significance was Figs. S1 to S7
determined by ANOVA. (D) Rat1 fibroblasts expressing FOXO3-ER, together with an empty vector (–) References and Notes
or mouse SIRT1 (⫹) were incubated in the presence or absence of 1 ␮M etoposide (ETO), in the 12 December 2003; accepted 11 February 2004
presence or absence of 500 nM 4-OHT for 16 hours. Apoptosis was detected by Western blot analysis Published online 19 February 2004;
with an antibody to the cleaved form of caspase 3. This experiment is representative of three 10.1126/science.1094637
independent experiments. Include this information when citing this paper.

www.sciencemag.org SCIENCE VOL 303 26 MARCH 2004 2015

View publication stats