THE JOURNAL OF BIOLOGICAL CHEMISTRY
44675–44682, 2003
© 2003 by The American Society for Biochemistry and Molecular Biology, Inc.
Phosphorylation of Nrf2 at Ser40 by Protein Kinase C in Response to
Antioxidants Leads to the Release of Nrf2 from INrf2, but Is
Not Required for Nrf2 Stabilization/Accumulation in the
Nucleus and Transcriptional Activation of Antioxidant
Response Element-mediated NAD(P)H:Quinone
Oxidoreductase-1 Gene Expression*
Received for publication, July 15, 2003, and in revised form, August 20, 2003
Published, JBC Papers in Press, August 28, 2003, DOI 10.1074/jbc.M307633200
David A. Bloom and Anil K. Jaiswal‡
From the Department of Pharmacology, Baylor College of Medicine, Houston, Texas 77030
The antioxidant response element (ARE) and tran-
scription factor Nrf2 regulate basal expression and an-
tioxidant induction of NAD(P)H:quinone oxidoreduc-
tase-1 (NQO1) and other detoxifying genes. Under
normal conditions, Nrf2 is targeted for proteasomal deg-
radation by INrf2. Oxidative stress causes release of
Nrf2 from INrf2. Nrf2 translocates to the nucleus, binds
to the ARE, and activates gene expression. In this study,
we demonstrate that protein kinase C (PKC) plays a
significant role in the regulation of ARE-mediated NQO1
gene expression and induction in response to t-butylhy-
droquinone. Treatment of HepG2 cells with the PKC
inhibitors staurosporine and calphostin C repressed
ARE-mediated induction of a luciferase reporter as well
as that of the endogenous NQO1 gene. Similar experi-
ments with inhibitors of MEK/ERK, p38, phosphatidyli-
nositol 3-kinase, and tyrosine kinases failed to repress
ARE-mediated gene expression. The PKC inhibitor stau-
rosporine blocked the nuclear translocation of Nrf2,
suggesting that Nrf2 might be the target for PKC regu-
lation. A Prosite search revealed the presence of seven
putative PKC sites in mouse Nrf2. The PKC site at Ser40
is conserved among species and lies in the Neh2 domain,
which interacts with INrf2. We demonstrate that phos-
phorylation of Ser40 is necessary for Nrf2 release from
INrf2, but is not required for Nrf2 stabilization/accumu-
lation in the nucleus and transcriptional activation of
ARE-mediated NQO1 gene expression. A peptide that
competes with endogenous Nrf2 for INrf2 binding was
able to induce ARE activity more effectively than t-bu-
tylhydroquinone, and Nrf2 that accumulated in the nu-
cleus as a result was not phosphorylated.
Antioxidant response element (ARE)1-mediated expression
and coordinated induction of detoxifying enzyme genes, includ-
* This work was supported by National Institutes of Health Grant
GM47466. The costs of publication of this article were defrayed in part
by the payment of page charges. This article must therefore be hereby
marked “advertisement” in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
‡ To whom correspondence should be addressed: Dept. of Pharmacol-
ogy, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030.
Tel.: 713-798-7691; Fax: 713-798-3145; E-mail: ajaiswal@bcm.tmc.edu.
1
The abbreviations used are: ARE, antioxidant response element;
NQO1, NAD(P)H:quinone oxidoreductase-1; JNK1, c-Jun N-terminal
kinase-1; MAPK, mitogen-activated protein kinase; MEK, mitogen-ac-
tivated protein kinase/extracellular signal-regulated kinase kinase;
PKC, protein kinase C; ERK, extracellular signal-regulated kinase;
t-BHQ, t-butylhydroquinone.
This paper is available on line at http://www.jbc.org
This is an Open Access article under the CC BY license.
Vol. 278, No. 45, Issue of November 7, pp.
Printed in U.S.A.
ing NAD(P)H:quinone oxidoreductase-1 (NQO1), glutathione
S-transferase Ya, ␥-glutamylcysteinyl synthetase, and heme
oxygenase-1, contribute significantly to cellular protection
against redox cycling, oxidative stress, and neoplasia (reviewed
in Refs. 1 and 2). Specific nuclear protein complexes bind to the
AREs of various genes (reviewed in Refs. 1 and 2). Among these
proteins, the role of Nrf2 (NF-E2-related factor-2) is the most
clearly established (reviewed in Refs. 1 and 2). Nrf2 is a basic
leucine zipper protein that does not homodimerize, but that
requires heterodimerization with another leucine zipper pro-
tein to be active (reviewed in Refs. 1 and 2). c-Jun and small
Maf proteins have been shown to heterodimerize with Nrf2;
these complexes bind to the ARE and alter ARE-mediated gene
expression (reviewed in Refs. 1 and 2).
A cytosolic inhibitor of Nrf2, Keap1/INrf2, was identified (3,
4). Under normal conditions, it is believed that INrf2 retains
Nrf2 in the cytoplasm (3, 4). When cells are challenged by
chemically induced oxidative stress, Nrf2 is released from
INrf2. Nrf2 then translocates to the nucleus, resulting in the
activation of ARE-mediated gene expression. Recently, reports
have shown that the interaction of Nrf2 with INrf2 targets
Nrf2 for proteasomal degradation (5–7). Nrf2 accumulates in
the nucleus of cells from INrf2-deficient mice, in a manner
similar to challenge with chemicals that induce oxidative stress
(7).
In vitro studies have demonstrated that several kinases,
including JNK1, MAPK, p38, and MEK, are involved in regu-
lating ARE-mediated gene expression (8 –11), yet other studies
have implicated phosphatidylinositol 3-kinase and tyrosine ki-
nases as the kinases responsible (12, 13). However, another
report has disputed the involvement of p38 and MEK kinases
and instead has indicated that protein kinase C (PKC) is the
major kinase involved in antioxidant induction of ARE-medi-
ated gene expression (14). In vitro studies from the same group
showed preferential phosphorylation of a serine residue in the
Neh2 domain of Nrf2 by PKC (15).
Our laboratory has long-term interest in Nrf2 regulation of
ARE-mediated gene expression with special emphasis on
NQO1 gene expression. Because of the conflicting reports on
the kinases involved, we felt further studies were necessary. In
addition, there is no information available on the kinase(s)
involved in the induction of ARE-mediated human NQO1 gene
expression in response to antioxidants. Furthermore, it is un-
known if phosphorylation of Nrf2 is required only for its release
from INrf2 or is required also for stabilization/accumulation of
Nrf2 in the nucleus and transcriptional activation of ARE-
mediated gene expression.
44675
44676 PKC Regulation of ARE-mediated NQO1 Gene Expression
EXPERIMENTAL PROCEDURES
Construction of Plasmids—The construction of pGL2-NQO1-ARE,
pcDNA-Nrf2, and pcDNA-INrf2 was as described previously (3). The
plasmid pcDNA-Nrf2 was used as the template for all mutations of
Nrf2. The seven putative PKC sites in Nrf2 were mutated using the
QuikChange® II site-directed mutagenesis kit (Stratagene, La Jolla,
CA). Briefly, two complementary PAGE-purified primers that con-
tained the desired mutation were produced (Integrated DNA Technol-
ogies, Inc., Coralville, IA). Fifteen cycles of PCR (94 °C for 1 min, 48 °C
for 0.5 min, and 68 °C for 8 min) were performed using pcDNA-Nrf2 as
the template, the primer pair containing the mutation, and PfuUltra
HF DNA polymerase. The parental plasmid was then digested with
DpnI, and the remaining nicked, circular, mutated plasmid was trans-
formed into maximum efficiency DH5␣ chemically competent cells (In-
vitrogen). The forward primers for each mutation are as follows:
Nrf2⌬PKC1, 5⬘-GTG TTT GAC TTT GCT CAG CGA CAG AAG GAC-3⬘;
Nrf2⌬PKC2, 5⬘-GTG CCC CTG GAG CTG TCA AAC AGA ACG GCC-3⬘;
Nrf2⌬PKC3, 5⬘-CCA ATG TGA AAA TGC AGC AAA AAA AGA AGT
TCC-3⬘; Nrf2⌬PKC4, 5⬘-GTG AAA AGA CAA ACA TGC AGC CCG CTT
AGA GGC-3⬘; Nrf2⌬PKC5, 5⬘-CCT TGT TCC CAA AGC TAA GAA GCC
AGA TAC-3⬘; and Nrf2⌬PKC6, 5⬘-CAA GAA GCC AGA TGC TAA GAA
AAA CTA GG-3⬘. The reverse primers are the exact complements of the
forward primers. The pcDNA-Neh2 and pcDNA-Neh2⌬S plasmids ex-
pressing Neh2 and Neh2⌬S domains, respectively, were created by
PCR, followed by TA cloning into the pcDNA3.1 vector. pcDNA-Nrf2
was used as the template for Neh2, and pcDNA-Nrf2⌬PKC1 was used
as the template for the Neh2⌬S domain. The same primer set was used
to PCR both constructs: 5⬘-CCC TCA GCA TGA TGG ACT TGG-3⬘
(forward) and 5⬘-GGC CGG CTG AAT TTG GGG AGG-3⬘ (reverse).
Kinase Inhibitor/Luciferase Reporter Assay—HepG2 cells were
grown in monolayer cultures by procedures described previously (3).
The cells were cotransfected with a reporter construct containing the
human NQO1 gene ARE driving luciferase expression and a Renilla
luciferase reporter as an internal control. Thirty-six h after transfec-
tion, the cells were pretreated with inhibitors to various kinases
(PD98059 for MEK/ERK, SB23580 for p38, wortmannin for phosphati-
dylinositol 3-kinase, genistein for tyrosine kinase, and calphostin C and
staurosporine for PKC) for 2 h. All the inhibitors were purchased from
Calbiochem and were the highest purity available. The cells were
treated with a combination of kinase inhibitor and 50 ␮M t-butylhydro-
quinone (t-BHQ) for 16 h. The cells were harvested, washed, homoge-
nized, and analyzed for luciferase activity. The dual luciferase reporter
assay from Promega (Madison, WI) and a Packard Topcount NXT
luminometer (PerkinElmer Life Sciences) were used to measure lucif-
erase activity according to the manufacturers’ protocols.
NQO1 Activity—HepG2 cells were pretreated with calphostin C for
2 h, and then fresh medium was added with Me2SO, with 50 ␮M t-BHQ,
or with 50 ␮M t-BHQ plus increasing concentrations of calphostin C.
Two h after treatment, the cells were harvested; cytosolic extracts were
prepared; and NQO1 activity was assayed by a previously described
method (16). The cytosolic extracts from HepG2 cells treated with
t-BHQ and with t-BHQ plus calphostin C were also resolved by SDS-
PAGE and Western-blotted, and the NQO1 protein level was visualized
using antibodies to NQO1 and actin.
Bacterial Expression and Purification of Nrf2—Wild-type Nrf2 and
Nrf2⌬PKC1 through Nrf2⌬PKC6 mutants were subcloned into the
pProExHTb vector from Invitrogen. These clones were transformed into
subcloning efficiency DH5␣ competent cells (Invitrogen). The bacteria
were grown at 37 °C and induced with 0.5 mM isopropyl-␤-D-thiogalac-
topyranoside when they reached A590 ⫽ 0.5. The bacteria were grown
for another 2 h at 37 °C and then harvested. His6-tagged Nrf2 was
purified on a nickel column (QIAGEN Inc.) under denaturing conditions
according to the manufacturer’s protocol. Before Nrf2 was eluted, the
column was washed with diminishing amounts of urea. Nrf2 was then
eluted under native conditions with 500 mM imidazole.
In Vitro Kinase Assay—Equal amounts of purified Nrf2 or Nrf2⌬PKC
mutants were used as the substrate for purified PKC consisting pri-
marily of the ␣- and ␤-isozymes (Promega). The purified proteins were
incubated with the PKC enzyme and [␥-32P]ATP in PKC buffer (20 mM
Hepes (pH 7.4), 1 mM dithiothreitol, 10 mM MgCl2, 1.7 mM CaCl2, and
0.1 mg/ml phosphatidylserine) for 1 h at 30 °C. The proteins were then
resolved by SDS-PAGE and visualized by autoradiography.
Immunoprecipitation and Phospho-specific Antibodies—Hepa-1 cells
were grown in 150-mm dishes. The cells were treated with Me2SO or 50
␮M t-BHQ for 2 h, washed twice with phosphate-buffered saline,
scraped into 10 ml of phosphate-buffered saline, and pelleted by cen-
FIG. 1. Effect of PKC inhibitors on t-BHQ induction of ARE-
mediated gene expression. HepG2 cells were seeded in 6-well dishes.
Twenty-four h later, they were cotransfected with a reporter construct
expressing the luciferase gene under the control of the NQO1 gene ARE
and a control reporter expressing Renilla luciferase. Thirty-six h after
transfection, the cells were pretreated with staurosporine, calphostin C,
or Me2SO (DMSO; control) for 2 h and then treated with Me2SO, with
t-BHQ, or with t-BHQ plus inhibitor for 16 h. The cells were lysed, and
the lysates were analyzed by dual luciferase assay. The results repre-
sent the mean ⫾ S.E. of three independent experiments, with each
experiment done in triplicate.
trifugation. Cytosolic and nuclear extracts were prepared as follows.
The cell pellet was suspended in 10 volumes of isotonic buffer (0.15 M
NaCl, 1.5 mM MgCl2, 10 mM Tris-HCl (pH 7.4), 0.5% (v/v) Nonidet P-40,
1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 50 ␮M NaF, 1
mM Na3VO4, and 1 ␮M okadaic acid) and homogenized in a Dounce
homogenizer 10 times with a loose pestle. The nuclei were pelleted at
3500 rpm at 4 °C. The cytosolic extract was removed, and the pelleted
nuclei were lysed using radioimmune precipitation assay buffer (50 mM
Tris (pH 8.0), 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1%
SDS, 1 mM phenylmethylsulfonyl fluoride, 50 ␮M NaF, 1 mM Na3VO4,
and 1 ␮M okadaic acid). Both extracts were cleared by centrifugation at
12,000 rpm for 10 min. Nrf2 was immunoprecipitated from the nuclear
extracts using anti-Nrf2 antibody (H-300, Santa Cruz Biotechnology,
Inc., Santa Cruz, CA). Equal amounts of precipitate were run on an
SDS-polyacrylamide gel, and Western blotting was performed with
antibody specific to Nrf2 or phosphoserine (Sigma), phosphothreonine
(Cell Signaling, Salem, MA), or phosphotyrosine (Cell Signaling). The
Neh2 and Neh2⌬S domains were pulled down using Ni2⫹-nitrilotriace-
tic acid-Sepharose beads (QIAGEN Inc.) according to the manufactur-
er’s protocol.
Similar assays were also performed with Hepa-1 cells transfected
with the pcDNA-Neh2-V5 and pcDNA-Neh2⌬S-V5 plasmids in separate
experiments. The cytosolic fractions from the transfected cells were
prepared by the procedures described above. Cytosolic proteins (1 mg)
were immunoprecipitated with anti-V5 antibodies. Equal amounts of
immunoprecipitated proteins from the transfected cells were separated
by 12% SDS-PAGE, Western-blotted, and probed with anti-V5 and
anti-phosphoserine antibodies.
Nrf2 Localization and Immunofluorescence—Hepa-1 cells were
grown on Lab-Tek II chamber slides. The cells were pretreated with 10
nM staurosporine for 1 h and then treated with 50 ␮M t-BHQ alone or
with 50 ␮M t-BHQ plus 10 nM staurosporine for 2 h. The cells were fixed
using formalin and permeabilized with cold acetone. Immunofluores-
cence was then performed using anti-Nrf2 antibody H-300 and fluores-
cein isothiocyanate-labeled secondary antibody (Rockland, Philadel-
phia, PA). The cells were stained with Hoechst stain to visualize the
nuclei. A Nikon Eclipse TE2000-U fluorescent microscope fitted with a
Photometrics CoolSnap CF camera and the appropriate filters was used
to capture the fluorescent images.
RESULTS
To determine which kinase pathway(s) is involved in ARE-
mediated induction, HepG2 cells were transfected with a lucif-
erase reporter plasmid driven by the NQO1 gene ARE. Thirty-
six h after transfection, the cells were treated with 50 ␮M
t-BHQ for 16 h, resulting in ⬃3.5-fold induction over basal
activity. This induction was inhibited by cotreatment with the
PKC inhibitor staurosporine (Fig. 1). Similar results were also
observed with a second PKC inhibitor, calphostin C (Fig. 1).
 PKC Regulation of ARE-mediated NQO1 Gene Expression 44677

FIG. 2. Effect of inhibitors of MEK/ERK, p38, phosphatidylinositol 3-kinase, and tyrosine kinases on t-BHQ induction of ARE-
mediated gene expression. HepG2 cells were grown in monolayers and cotransfected with ARE-luciferase and Renilla luciferase reporter
plasmids. Thirty-six h after transfection, the cells were pretreated with inhibitors for 2 h and then treated with Me2SO, with t-BHQ, or with t-BHQ
plus inhibitors for 16 h. The cells were harvested and lysed, and the lysates were analyzed for luciferase activity. The results are presented as the
mean ⫾ S.E. of three independent experiments, with each experiment done in triplicate. DMSO, Me2SO; PI3-K, phosphatidylinositol 3-kinase.

Inhibitors of MEK/ERK (PD98059), p38 (SB23580), phosphati-

dylinositol 3-kinase (wortmannin), and tyrosine kinases (genis-
tein) failed to block t-BHQ induction of the ARE (Fig. 2). The
small amount of inhibition observed with phosphatidylinositol
3-kinase and p38 inhibitors was not significant (p ⬎ 0.1). In-
terestingly, the inhibitors of MEK/ERK and tyrosine kinases
stimulated NQO1 ARE-mediated gene expression.
Calphostin C was also able to inhibit the induction of NQO1
activity in vivo (Fig. 3). Untransfected HepG2 cells were
treated with calphostin C or with calphostin C plus t-BHQ. The
lysates were then assayed for NQO1 activity, and the amount
of NQO1 was determined by Western blot analysis. Both NQO1
activity and the amount of NQO1 protein decreased in a dose-
dependent manner in response to treatment with the PKC
inhibitor.
To determine whether PKC inhibitors affect the localization
of Nrf2, immunofluorescence was used to follow the distribu-
tion of endogenous Nrf2 before and after treatment. Hepa-1
cells were grown on chamber slides; treated with Me2SO, with
t-BHQ, with staurosporine, or with t-BHQ and staurosporine;
fixed; and probed with anti-Nrf2 antibody (Fig. 4A). In control
cells, Nrf2 was found localized in the cytoplasm and nucleus,
although the fluorescence was weak. After treatment with t-
BHQ, Nrf2 accumulated in the nucleus, resulting in a much
stronger signal. The PKC inhibitor staurosporine was able to
FIG. 3. In vivo effect of calphostin C on NQO1 activity and
block the t-BHQ-induced nuclear accumulation of Nrf2. These protein. HepG2 cells were treated with Me2SO, with t-BHQ, or with
results were also confirmed by Western blot analysis. Cytosolic t-BHQ plus increasing concentrations of calphostin C for 16 h. Cytosolic
and nuclear extracts were prepared from treated and untreated extracts were prepared and analyzed for NQO1 activity (upper panel).
Hepa-1 cells. The extracts were resolved by SDS-PAGE and The cytosolic extracts from HepG2 cells treated with t-BHQ and with
t-BHQ plus calphostin C were also analyzed for NQO1 protein by
probed with anti-Nrf2 antibody (Fig. 4B). Nrf2 was hardly SDS-PAGE and Western blotting (lower panel). Western blots were
visible in any of the cytosolic extracts or in the nuclear extracts probed with antibodies to NQO1 and actin. Anti-actin antibody was
from control cells or from cells treated with staurosporine plus used to show equal loading of proteins in the lanes.
44678 PKC Regulation of ARE-mediated NQO1 Gene Expression
FIG. 4. Effect of staurosporine on t-BHQ-induced nuclear accumulation of Nrf2. A, Hepa-1 cells were grown and treated on chamber
slides. They were fixed and probed with anti-Nrf2 antibody, followed by fluorescein isothiocyanate (FITC)-labeled secondary antibody. The cells
were also stained with Hoechst stain to visualize the nuclei. B, Hepa-1 cells were treated, and cytosolic and nuclear extracts were prepared. The
extracts were resolved by SDS-PAGE and probed with anti-Nrf2 antibody. The membrane was then stripped and probed with anti-actin antibody
to show equal loading. DMSO, Me2SO.
t-BHQ. However, there was an abundance of Nrf2 in the nu-
cleus of t-BHQ-treated cells. These results coincide with the
proposed model that Nrf2 is rapidly degraded in the cytosol as
a result of the Nrf2/INrf2 interaction and is stabilized/accumu-
lated in nuclei after chemically induced oxidative stress (5–7).
Next, we studied the ability of PKC to phosphorylate Nrf2 in
vitro. Histidine-tagged Nrf2 was overexpressed in a bacterial
system and purified using nickel-coated Sepharose beads. This
Nrf2 was used as the substrate in an in vitro kinase assay
utilizing purified PKC as the enzyme (Fig. 5A). Nrf2 was phos-
phorylated by PKC in vitro. Staurosporine was able to inhibit
the phosphorylation, whereas wortmannin, a phosphatidyli-
nositol 3-kinase inhibitor, was not.
To determine whether Nrf2 is phosphorylated in vivo, we
utilized antibodies specific to phosphoserine, phosphothreo-
nine, and phosphotyrosine. Hepa-1 cells were treated with
Me2SO or t-BHQ. Nrf2 was immunoprecipitated from the cy-
tosolic and nuclear extracts, and Western analysis was per-
formed (Fig. 5B). This revealed that Nrf2 that accumulated in
the nucleus in response to t-BHQ treatment was phosphoryl-
ated at a serine residue(s). Phosphorylation of Nrf2 in the
untreated nuclear extracts could not be detected. It is possible
that it is not phosphorylated or that the amount of protein was
not sufficient to detect the phosphorylation. Likewise, we could
not detect phosphorylation with anti-phosphothreonine or anti-
phosphotyrosine antibody. The question arises whether PKC
inhibitors can block the phosphorylation. However, there was
an insufficient amount of Nrf2 in the staurosporine-treated
nuclear extracts to determine phosphorylation.
A Prosite search revealed seven putative PKC sites in Nrf2.
Of these, four are serine residues.2 PKC site 1, Ser40, is located
in the N-terminal Neh2 domain of Nrf2, the domain that in-
teracts with INrf2 (2, 3). This PKC site is conserved across the
species (human, rat, mouse, and chicken).2 Six Nrf2 mutants
2
D. A. Bloom and A. K. Jaiswal, unpublished data.
were made and are referred to as Nrf2⌬PKC1 through
Nrf2⌬PKC6 (Fig. 6A). All seven PKC sites were mutated from
serine/threonine to alanine. The two threonines located next to
each other at positions 417 and 418 were mutated in the same
construct, Nrf2⌬PKC3 (Fig. 6A).
Mutating the putative PKC sites individually did not have a
significant effect on the activity of any of the mutants (Fig. 6B,
data shown only for Nrf2⌬PKC1). However, mutating the PKC
site in the Neh2 domain resulted in INrf2 inhibiting
Nrf2⌬PKC1 more efficiently than wild-type Nrf2 (Fig. 6C). The
inhibition of the other mutants by INrf2 was not affected (data
not shown). In vitro kinase results revealed that Nrf2⌬PKC1
was not phosphorylated by PKC (Fig. 6D).
To further study the effect of this mutation in vivo, con-
structs were created that would express the Neh2 domain or
the Neh2 domain with the PKC1 serine mutation, Neh2⌬S
(Fig. 7A). Both domains were tagged with V5 and His6. HepG2
cells were transfected with increasing amounts of these con-
structs as well as the ARE-luciferase reporter. As expected, the
wild-type Neh2 domain was able to compete with endogenous
Nrf2 for INrf2. This resulted in a concentration-dependent
increase in ARE-luciferase activity (Fig. 7B). At the highest
dose, the Neh2 domain induced ARE activity to the same extent
as t-BHQ in cells that were not transfected with the Neh2
domain. However, the Neh2⌬S domain was much more effi-
cient in inducing ARE activity. At the highest dose, it induced
ARE activity almost twice as much as t-BHQ alone or the
wild-type Neh2 domain.
To demonstrate binding of the Neh2 and Neh2⌬S domains to
INrf2 and that the domains can cause the nuclear accumula-
tion of Nrf2, Hepa-1 cells were transfected with an empty
vector or with vector encoding the Neh2 or Neh2⌬S domain.
The cells were treated with Me2SO or t-BHQ, and cytosolic and
nuclear extracts were made. Western analysis revealed that
the Neh2 and Neh2⌬S domains were able to cause the nuclear
accumulation of Nrf2 (Fig. 7C, upper panel). Treatment with
 PKC Regulation of ARE-mediated NQO1 Gene Expression 44679

FIG. 5. In vitro and in vivo phospho-

rylation of Nrf2 by PKC. A, Nrf2 was
expressed in an isopropyl-␤-D-thiogalacto-
pyranoside-inducible bacterial system.
Purified Nrf2 was phosphorylated in vitro
using a mixture of PKC isoforms. B, en-
dogenous Nrf2 was immunoprecipitated
from cytosolic (C) and nuclear (N) extracts
prepared from Hepa-1 cells treated with
Me2SO or t-BHQ. The immunoprecipi-
tated proteins were resolved by SDS-
PAGE and probed with anti-Nrf2, anti-
phosphoserine, anti-phosphothreonine, or
anti-phosphotyrosine antibody.

FIG. 6. PKC phosphorylation of wild-type and mutant Nrf2. A, shown is a graphical representation of Nrf2 showing the locations of several
important domains and the PKC phosphorylation sites. Nrf2⌬PKC3 contains mutations of two overlapping PKC sites in Nrf2. b-Zip, basic leucine
zipper. B, HepG2 cells were transfected with the NQO1 ARE-luciferase reporter along with wild-type Nrf2 or one of the Nrf2⌬PKC mutants.
Renilla luciferase was included in each transfection as transfection efficiency control. Thirty-six h later, the cells were treated with Me2SO (DMSO)
or t-BHQ for 16 h. The cells were harvested, and luciferase activity was determined. C, HepG2 cells were cotransfected with the NQO1
ARE-luciferase reporter plasmid, Nrf2, or Nrf2⌬PKC1 with or without INrf2. The cells were harvested and analyzed for luciferase activity. D,
bacterially expressed and purified Nrf2 or Nrf2⌬PKC1 was used as the substrate in an in vitro kinase assay following the procedures described
under “Experimental Procedures.” The proteins were analyzed by SDS-PAGE and autoradiographed (upper panel) or stained with Coomassie Blue
(lower panel).

t-BHQ or expression of the Neh2 domain caused nearly an
equal amount of Nrf2 accumulation in the nucleus. However,
expression of the Neh2⌬S domain caused a greater amount of
Nrf2 to accumulate in the nucleus, consistent with the ARE
activity data. The Neh2 and Neh2⌬S domains were equally
expressed (Fig. 7C, middle panel).
The Neh2 and Neh2⌬S domains tagged with histidines were
precipitated from the cytosolic extract using nickel-coated
Sepharose beads. Western analysis using anti-INrf2 antibody
revealed that both the Neh2 and Neh2⌬S domains bound to
INrf2 in uninduced cells (Fig. 7D). However, after t-BHQ treat-
ment, the wild-type Neh2 domain no longer associated with
44680 PKC Regulation of ARE-mediated NQO1 Gene Expression

FIG. 7. In vivo effect of t-BHQ and the Neh2 and mutant Neh2 domains on ARE-mediated gene expression, stabilization/nuclear
accumulation, and phosphorylation of Nrf2. A, Nrf2 is shown with functional protein domains including Neh2. The Neh2 and mutant Neh2
domains tagged with V5 are also shown. Ser40 in the Neh2 domain was mutated to alanine to generate the mutant Neh2 domain. The construction
of recombinant plasmids expressing Neh2 and mutant Neh2 domains was described under “Experimental Procedures.” Constructs expressing the
wild-type Neh2 domain or the Neh2⌬S domain tagged with V5 and His6 were created. b-Zip, basic leucine zipper. B, the plasmids expressing the
Neh2 or Neh2⌬S domain were transfected into HepG2 cells along with the ARE-luciferase reporter. Thirty-six h after transfection, the cells were
treated with Me2SO (DMSO) or t-BHQ. Forty-eight h after transfection, the cells were harvested, and the extracts were analyzed by dual luciferase
assay. C, Hepa-1 cells were either treated with t-BHQ or transfected with either the Neh2 or Neh2⌬S domain. Cytosolic and nuclear extracts were
prepared, and the lysates were resolved by SDS-PAGE. The proteins were transferred to nylon membranes. The membranes were then probed with
anti-Nrf2 antibody, stripped, probed with anti-V5 antibody, stripped again, and probed with anti-actin antibody. D, the cytosolic extracts were
made from untreated or t-BHQ-treated Hepa-1 cells that had been transfected with either the Neh2 or Neh2⌬S domain of Nrf2 tagged with V5 and
histidines. The Neh2 and mutant Neh2 domains were affinity-purified on nickel-coated Sepharose beads. After washing, the proteins were eluted
and resolved by SDS-PAGE and transferred to nylon membranes. The membranes were probed with anti-INrf2 antibody, stripped, and probed with
anti-V5 antibody. E, the Hepa-1 cells were either treated with t-BHQ or transfected with the Neh2 or Neh2⌬S domain. The t-BHQ-treated and
transfected Hepa-1 cells were homogenized; nuclei were isolated; and nuclear extracts were prepared. Nrf2 was immunoprecipitated from these
nuclear extracts with anti-Nrf2 antibody. The eluted proteins were analyzed by SDS-PAGE and Western blotting. Western blots were probed with
either anti-Nrf2 or anti-phosphoserine antibody. Equal amounts of the immunoprecipitated proteins were loaded in the lanes. F, the Hepa-1 cells
were transfected with plasmid pcDNA-Neh2-V5 or pcDNA-Neh2⌬S-V5, and cytosolic fractions were prepared by standard procedures. Cytosolic
proteins (1 mg) from the transfected cells were immunoprecipitated with anti-V5 antibody. Equal amounts of immunoprecipitated proteins from
each transfection were separated by 12% SDS-PAGE, Western-blotted, and probed with anti-V5 and anti-phosphoserine antibodies.

INrf2. The interaction of the Neh2⌬S domain with INrf2 was
not affected by t-BHQ treatment.
Next, Nrf2 was immunoprecipitated from the nuclear ex-
tracts from the previous experiment (Fig. 7E). Western analy-
sis demonstrated that only Nrf2 from cells induced with t-BHQ
was phosphorylated at serine residues. Nrf2 that accumulated
in the nucleus of cells transfected with either the Neh2 or
Neh2⌬S domain was not phosphorylated. It may be noteworthy
that each lane contains an equal amount of immunoprecipi-
tated Nrf2 in Fig. 7E. It is this reason that the increase in Nrf2
in the presence of t-BHQ in Fig. 7C is not visible in Fig. 7E.
Further studies with Hepa-1 cells transfected with Neh2-V5
and Neh2⌬S-V5 and anti-phosphoserine antibody demon-
strated that Neh2-V5 and not Neh2⌬S-V5 was phosphorylated
(Fig. 7F).
Taken together, these results indicate that the phosphoryl-
ation of Nrf2 at Ser40 by PKC is required for Nrf2 to evade
INrf2-mediated degradation. Phosphorylation is not required
for Nrf2 to accumulate in the nucleus. Furthermore, the un-
phosphorylated form of Nrf2 is able to activate transcription of
the ARE.
DISCUSSION
NQO1 is a phase II detoxifying enzyme that competes with
the one-electron reducing enzyme cytochrome P450 reductase
and catalyzes two-electron reductive metabolism and detoxifi-
cation of quinones (1). NQO1-null mice have been generated,
and studies have shown that NQO1 plays an important role in
regulating the intracellular redox status of cells and protects
against myelogenous hyperplasia and quinone-induced redox
cycling, oxidative stress, and cytotoxicity (17–19). Studies have
also shown that NQO1 protects the skin against benzo-
[a]pyrene- and 9,10-dimethyl-1,2-benzanthracene-induced car-
cinogenesis (20, 21).
Antioxidants and oxidative stress induce a battery of more
than two dozen defensive genes, including NQO1, by an ARE-
 PKC Regulation of ARE-mediated NQO1 Gene Expression 44681

FIG. 8. Role of PKC in transcriptional activation of ARE-mediated gene expression. Two alternative hypotheses are shown. Both require
PKC-mediated phosphorylation of Nrf2, leading to either release of Nrf2 from INrf2 (Hypothesis I) or escape of Nrf2 from INrf2 (Hypothesis II).
The Nrf2 heterodimeric partner is shown as Unknown. This is because Jun and small Maf proteins have been shown to form heterodimers with
Nrf2. However, none of these proteins have been clearly established as the activating heterodimeric partner of Nrf2.

dependent mechanism (reviewed in Ref. 1). This coordinated
induction of detoxifying genes is of critical importance for an-
tioxidant action and chemoprevention (reviewed in Ref. 1). The
transcription factor Nrf2 is known to bind to the NQO1 ARE as
well as to the AREs of several other detoxifying enzyme genes
and to activate their transcription in response to antioxidants
and xenobiotics (reviewed in Ref. 1). Under normal conditions,
Nrf2 is believed to be retained in the cytoplasm by a cytosolic
inhibitor, INrf2 (3, 4). Recently, it has also been shown that
INrf2 targets Nrf2 for proteasomal degradation (5–7). Treat-
ment of cells with antioxidants disrupts the Nrf2/INrf2 inter-
action (3, 4). When Nrf2 is no longer associated with INrf2, it is
stabilized and accumulates in the nucleus. It forms het-
erodimers with other leucine zipper proteins, leading to the
coordinated activation of NQO1 and other detoxifying enzyme
genes, including glutathione S-transferase Ya, ␥-glutamylcys-
teinyl synthetase, and heme oxygenase-1 (22–25). c-Jun and
small Maf proteins have been shown to be heterodimeric part-
ners of Nrf2, leading to the activation of detoxifying genes (26,
27). However, the role of small Maf proteins is controversial.
Other studies have shown that Nrf2-small Maf heterodimers
repress ARE-mediated gene expression (23, 25). t-BHQ treat-
ment does not affect the levels of Nrf2 mRNA (2). Therefore, we
suspected that t-BHQ treatment leads to the modification of
Nrf2 and that these modifications are critical for ARE-medi-
ated induction of detoxifying enzyme genes.
It is interesting that inhibitors of MEK/ERK and tyrosine
kinases activate ARE-mediated gene expression. It is possible
that phosphorylation of Nrf2 by members of these kinase fam-
ilies facilitates the binding of Nrf2 to INrf2 and therefore
degradation of Nrf2. Tyrosine kinases are involved in regulat-
ing cell growth, so it would not be surprising if this pathway is
linked to defensive measures. Inhibition of cell growth could
serve as a warning to the cell that the environmental condi-
tions are unfavorable. The cell would then activate defensive
genes, through the ARE, to limit cellular damage.
A recent report had demonstrated that Nrf2 is phosphoryl-
ated in vitro by PKC and that this affects the interaction of
Nrf2 with INrf2 in vitro (14). However, the involvement of PKC
had been contradicted by other reports (8 –13). We confirm that
PKC activity is critical for the induction of ARE-mediated gene
expression by t-BHQ. Furthermore, we show for the first time
in vivo evidence that phosphorylation of Ser40 is necessary for
Nrf2 to dissociate from INrf2. Mutation of this residue in the
Neh2 domain resulted in the Neh2⌬S domain/INrf2 interaction
being unresponsive to t-BHQ treatment.
Importantly, inducing Nrf2 nuclear accumulation by overex-
pressing the Neh2 domain resulted in an increase in ARE
activity similar to that caused by t-BHQ induction. The mutant
Neh2⌬S domain induced ARE expression and Nrf2 accumula-
tion much more efficiently than the wild-type Neh2 domain or
t-BHQ. These data point to the fact that phosphorylation of
Nrf2 is not required for it to activate transcription. We dem-
onstrated this by showing that Nrf2 that accumulated in the
nucleus in response to t-BHQ treatment was phosphorylated at
a serine residue(s), but that Nrf2 that accumulated in the
nucleus as a result of Neh2 overexpression was not. Clearly, it
is the association of Nrf2 with INrf2 that is regulated by PKC,
not Nrf2 stability or activity.
The role of PKC phosphorylation of Nrf2 in ARE-mediated
gene expression is shown in Fig. 8. Two alternative hypotheses
are depicted. Briefly, in Hypothesis I, Nrf2 is retained in the
cytoplasm by INrf2. Antioxidants induce the expression of or
activate PKC. PKC then phosphorylates Nrf2 that is bound to
INrf2. This phosphorylation leads to the release of Nrf2 from
INrf2 and the stabilization/nuclear accumulation of Nrf2.
Phosphorylated Nrf2 binds to the ARE and activates ARE-
mediated gene expression. This hypothesis is based on cur-
rently accepted results on Nrf2 interaction with INrf2, antiox-
idant-induced release of Nrf2 from INrf2, and nuclear
localization of Nrf2 (3, 4). However, an alternative hypothesis
that can not be ruled out is possible. In Hypothesis II, INrf2 is
44682 PKC Regulation of ARE-mediated NQO1 Gene Expression
an inhibitor of unphosphorylated Nrf2. INrf2 binds to unphos-
phorylated Nrf2 and stimulates its degradation (5–7). Antioxi-
dants/oxidants/other stresses induce/activate PKC, which
phosphorylates free Nrf2. Phosphorylated Nrf2 escapes INrf2-
mediated degradation and accumulates in the nucleus, where it
activates ARE-mediated gene expression. Hypothesis II is
based on several observations. First, INrf2 targets Nrf2 for
proteasomal degradation (5–7). Second, PKC phosphorylates
Nrf2 in the Neh2 domain, which binds to INrf2 (this study and
Ref. 4). The binding of INrf2 to the Neh2 domain might inter-
fere with the availability of Ser40 for phosphorylation by PKC.
Third, Nrf2 does not require phosphorylation to accumulate in
the nucleus and to activate ARE-mediated gene expression
(this study). Future experiments should select one hypothesis
and rule out the other.
The data suggest that any mechanism that modifies Nrf2,
allowing it to disrupt the Nrf2-INrf2 complex and/or to escape
Nrf2 binding to INrf2, will result in the up-regulation of ARE-
mediated gene expression. Recently, four sulfhydryl groups in
INrf2 that are sensitive to the redox state of the cell were
reported (29). Alteration of these sulfhydryl groups leads to the
loss of the association of Nrf2 with INrf2, leading to the acti-
vation of ARE-mediated gene expression (28). It is possible that
this mechanism is redundant to the phosphorylation of Nrf2 by
PKC or that the two mechanisms work in concert. It is also
possible that these two mechanisms are activated in a stimu-
lus- and/or cell type-specific manner. Likewise, there could be
other cell type-specific factors that disrupt the Nrf2-INrf2 com-
plex or modify Nrf2, preventing its interaction with INrf2 in
response to various stimuli.
Acknowledgments—We are grateful to Drs. Jefferson Y. Chan and
Yuet W. Kan (both from University of California, San Francisco) for
providing cDNA encoding Nrf2. We thank Dr. S. Dhakshinamoorthy for
helpful suggestions.
REFERENCES
1. Dhakshinamoorthy, S., Long, D. J., II, and Jaiswal, A. K. (2000) Curr. Top.
Cell. Regul. 36, 201–216
2. Jaiswal, A. K. (2000) Free Radic. Biol. Med. 29, 254 –262
3. Dhakshinamoorthy, S., and Jaiswal, A. K. (2001) Oncogene 20, 3906 –3917
4. Itoh, K., Wakabayashi, N., Katoh, Y., Ishii, T., Igarashi, K., Engel, J. D., and
Yamamoto, M. (1999) Genes Dev. 13, 76 – 86
5. Itoh, K., Wakabayashi, N., Katoh, Y., Tetsuro, I., O’Connor, T., and Yamamoto,
M. (2003) Genes Cells 8, 379 –391
6. Nguyen, T., Sherratt, P. J., Huang, H. C., Yang, C. S., and Pickett, C. B. (2003)
J. Biol. Chem. 278, 4536 – 4541
7. McMahon, M., Itoh, K., Yamamoto, M., and Hayes, J. D. (2003) J. Biol. Chem.
278, 21569 –21600
8. Yu, R., Jiao, J. J., Duh, J. L., Tan, T.-H., and Tony Kong, A.-N. (1996) Cancer
Res. 56, 2954 –2959
9. Yu, R., Mandlekar, S., Weber, M. J., Der, C. J., Wu, J., and Tony Kong, A.-N.
(1999) J. Biol. Chem. 274, 27545–27552
10. Yu, R., Mandlekar, S., Lei, W., Fahl, W. E., Tan, T.-H., and Tony Kong, A.-N.
(2000) J. Biol. Chem. 275, 2322–2327
11. Yu, R., Chen, C., Mo, Y. Y., Hebbar, V., Owuor, E. D., Tan, T.-H., and Tony
Kong, A.-N. (2000) J. Biol. Chem. 275, 39907–39913
12. Kang, K. W., Ryu, H., and Kim, S. G. (2000) Mol. Pharmacol. 58, 1017–1025
13. Dieter, M. Z., Freshwater, S. L., Solis, W. A., Nebert, D. W., and Dalton, T. P.
(2001) Biochem. Pharmacol. 61, 215–225
14. Huang, H. C., Nguyen, T., and Pickett, C. B. (2000) Proc. Natl. Acad. Sci.
U. S. A. 97, 12475–12480
15. Huang, H. C., Nguyen, T., and Pickett, C. (2002) J. Biol. Chem. 277,
42769 – 42774
16. Shaw, P. M., Reiss, A., Adesnik, M., Nebert, D. W., Schembri, J., and Jaiswal,
A. K. (1991) Eur. J. Biochem. 195, 171–176
17. Radjendirane, V., Joseph, P., Lee, Y.-H., Kimura, S., Klein-Sazanto, A. J. P.,
Gonzalez, F. J., and Jaiswal, A. K. (1998) J. Biol. Chem. 273, 7382–7389
18. Gaikwad, A., Long, D. J., II, Stringer, J. L., and Jaiswal, A. K. (2001) J. Biol.
Chem. 276, 22559 –22564
19. Long, D. J., II, Gaikwad, A., Multani, A., Pathak, S., Montgomery, C. A.,
Gonzalez, F. J., and Jaiswal, A. K. (2002) Cancer Res. 62, 3030 –3036
20. Long, D. J., II, Waikel, R. L., Wang, X., Perlaky, L., Roop, D. R., and Jaiswal,
A. K. (2000) Cancer Res. 60, 5913–5915
21. Long, D. J., II, Waikel, R. L., Wang, X. J., Roop, D. R., and Jaiswal, A. K. (2001)
J. Natl. Cancer Inst. 93, 1166 –1170
22. Venugopal, R., and Jaiswal, A. K. (1996) Proc. Natl. Acad. Sci. U. S. A. 93,
14960 –14965
23. Wild, A. C., Moinova, H. R., and Mulcahy, R. T. (1999) J. Biol. Chem. 274,
33627–33636
24. Gong, P., Hu, B., Stewart, D., Ellerbe, M., Figueroa, Y. G., Blank, V., Beckman,
B. S., and Alam, J. (2001) J. Biol. Chem. 276, 27018 –27025
25. Nguyen, T., Huang, H. C., and Pickett, C. B. (2000) J. Biol. Chem. 275,
15466 –15473
26. Itoh, K., Chiba, T., Takahashi, S., Ishii, T., Igarashi, K., Katoh, K., Oyake, T.,
Hayashi, N., Satoh, K., Hatayama, I., Yamamoto, M., and Nabeshima, Y.
(1997) Biochem. Biophys. Res. Commun. 236, 313–322
27. Venugopal, R., and Jaiswal, A. K. (1998) Oncogene 17, 3145–3156
28. Dinkova-Kostova, A. T., Holtzclaw, W. D., Cole, R. N., Itoh, K., Wakabayashi,
N., Katoh, Y., Yamamoto, M., and Talalay, P. (2002) Proc. Natl. Acad. Sci.
U. S. A. 99, 11908 –11913