www.elsevier.com/locate/yexcr

Experimental Cell Research 297 (2004) 461–471

K+ channel activity and redox status are differentially required for JNK

activation by UV and reactive oxygen species

Jie Gao,a,1 Dan Wu,a,1 Taylor B. Guo,a,b Qin Ruan,c Tie Li,d Zhenyu Lu,a Ming Xu,a,b,e

Wei Dai,a,b,c,* and Luo Lua,b,d,*

aDepartment of Medical Genetics, Shanghai Second Medical University, PR ChinabHealth Science Center, Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, PR China

cDivision of Molecular Carcinogenesis, Department of Medicine, New York Medical College, Valhalla, NY 10595, USAdDivision of Molecular Medicine, Harbor-UCLA Medical Center, School of Medicine, University of California Los Angeles, Torrance, CA 90502, USA

eDepartment of Cell Biology Neurobiology and Anatomy, University of Cincinnati College of Medicine, Cincinnati, OH 45267, USA

Received 2 January 2004, revised version received 19 March 2004

Available online 6 May 2004

Abstract

Upon exposure to ultraviolet (UV) radiation, osmotic changes or the presence of reactive oxygen species (ROS) c-Jun N-terminal kinases

(JNKs) are rapidly activated. Extensive studies have elucidated molecular components that mediate the activation of JNKs. However, it

remains unclear whether activation of JNKs by various stress signals involves different pathways. Here we show that K+ channel activity is

involved in mediating apoptosis induced by UV but not by H2O2 in myelocytic leukemic ML-1 cells. Specifically, JNKs were rapidly

phosphorylated upon treatment of ML-1 cells with UV and H2O2. UV-induced, but not H2O2-induced, JNK-1 phosphorylation was inhibited

by pretreatment with 4-aminopyridine (4-AP), a K+ channel blocker. 4-AP also blocked UV-induced increase in JNK activity as well as p38

phosphorylation. Immunofluorescent microscopy revealed that phosphorylated JNKs were concentrated at centrosomes in ML-1 cells and

that these proteins underwent rapid subcellular translocation upon UV treatment. Consistently, the subcellular translocation of JNKs induced

by UV was largely blocked by 4-AP. Furthermore, UV-induced JNK activation was blocked by NEM, a sulfhydryl alkylating agent also

affecting K+ current. Both UV- and H2O2-induced JNK activities were inhibited by glutathione, suggesting that the redox status does play an

important role in the activation of JNKs. Taken together, our findings suggest that JNK activation by UVand H2O2 is mediated by distinct yet

overlapping pathways and that K+ channel activity and redox status are differentially required for UV- and H2O2-induced activation of JNKs.

D 2004 Elsevier Inc. All rights reserved.

Keywords: JNKs; UV; Reactive oxygen species; K+ channel; Centrosome

Introduction

Ultraviolet (UV) irradiation induces genotoxic stress,

which frequently results in apoptosis. This response appears

to be mediated by various signaling components [1–3]. At

the cell membrane, UV irradiation elicits rapid clustering and

internalization of membrane receptor molecules for epider-

mal growth factor (EGF), tumor necrosis factor (TNF), and

interleukin-1 (IL-1) [4]. Extensive studies show that UV

irradiation strongly activates some MAP kinase family mem-

0014-4827/$ - see front matter D 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.yexcr.2004.03.047

* Corresponding authors. Fax: +1-914-594-4726.

E-mail addresses: wei_dai@nymc.edu (W. Dai),

llu@rei.edu (L. Lu).1 These authors contributed equally to the work reported in this study.

bers such as c-JunN-terminal kinases (JNKs) and p38 [5–7].

In addition, UV activates ATR and Chk1 [8,9], two kinases

mediating the DNA damage checkpoint response. One of the

consequences of elevated kinase activities of JNK, p38, and

Chk1 is the activation of transcription factors including p53,

NF-nB, and/or AP-1 by phosphorylation [1,2,10]. JNKs can

also be activated by other stressors such as reactive oxygen

species (ROS) [9,11]. ROS can be produced by a variety of

cells in response to stimulation of mechanical forces, growth

factors, and cytokines. Many studies have shown that oxi-

dants participate in both proliferative and apoptotic responses

[12,13]. Although the molecular basis for divergent

responses to oxidants remains unclear, it is believed that both

the intensity and the context of stimulation play key roles in

determination of the outcome.

J. Gao et al. / Experimental Cell Research 297 (2004) 461–471462

K+ channel activity is widely distributed in the cell

membrane to stabilize cell membrane electrophysiological

properties. In myeloblastic cells, fibroblastic cells, and neu-

rons, cytokines modulate the activity of K+, resulting in cell

proliferation or apoptosis [14,15]. Single-channel recording

using the cell-attached mode reveals that exposure to UV-C

irradiation (45 mJ/cm2) strongly stimulates K+ channel ac-

tivity within 30 s. A quick loss of intracellular K+ ions causes

membrane fluctuations and cell shrinkage, leading to activa-

tion of known surface receptors and cytoplasmic signaling

components. In the previous work, we have shown that

activation of a voltage-gated K+ channel in ML-1 cells by

UV-C irradiation promotes apoptosis. The K+ channel in

these cells is sensitive to 4-aminopyridine (4-AP, IC50 = 80

AM), a specific voltage-gate K+ channel blocker [16]. The

effects of UV irradiation-activated K+ channels are observed

at both whole-cell and single-channel levels. Blockade of K+

channels with a similar low dose of 4-AP almost completely

prevents UV-induced apoptosis and suppresses UV-stimulat-

ed JNK pathway, indicating that UV-activated K+ channels

do mediate apoptosis in ML-1 cells [17]. It has been shown

that the loss of intracellular K+ activates interleukin 1a-

converting enzyme (ICE) [18–20]. Recent studies indicate

that UV irradiation-induced activation of ICE and JNK-1

occurs after the stimulation of K+ channels and to the loss of

intracellular K+. This mechanism has been implicated in

apoptosis in neuronal cells [19]. In addition, suppression of

K+ channel activity protects corneal epithelial cells from UV-

induced DNA fragmentation and cell death. However, this

suppression of K+ channel activity does not prevent corneal

epithelial apoptosis induced by etoposide because it directly

inhibits topoisomerase II activity in the nucleus rather than

affecting UV-induced cell membrane events.

Although extensive studies have elucidated the molecular

mechanisms that mediate the activation of JNKs, it remains

unclear whether activation of JNKs by various stress signals

involves different pathways. To determine whether JNK

activation by different genotoxic stressors was mediated

by the same cellular components, we measured activities

and subcellular localization of JNKs in ML-1 cells exposed

to either UV irradiation or H2O2. We observed that JNK

activation by UVand H2O2 is mediated by distinct pathways

and that K+ channel activity is required for UV-induced, but

not H2O2-induced, activation of JNKs. Moreover, redox

status of the cell can affect both UV- and H2O2-induced

activation of JNKs.

Fig. 1. Activation of a 4-AP-sensitive K+ current by UV irradiation. Whole-

cell currents were measured after exposure of ML-1 cells to UV light (60 J/

m2) in the presence or absence of 4-AP (1 mM). The membrane potential

was depolarized from a holding potential of � 60 to 0 mV. Currents were

normalized as I/Ic where Ic and I represent amplitudes of the K+ current

measured at 0 mV before and after UV light and H2O2 stimulation,

respectively. Data were obtained from four independent experiments and

plotted as means with standard error bars. Asterisks represent the significant

difference ( P < 0.01).

Experimental procedures

Materials

Rabbit polyclonal antibodies against JNKs and p38 as

well as to the activated (phosphorylated) versions of these

proteins were obtained from Cell Signaling Technology

(Beverly, MA, USA). Mouse monoclonal antibodies against

human g-tubulin, the protease inhibitor cocktail, and 4-

aminopyridine (4-AP) were purchased from Sigma (MO,

USA). H2O2 was purchased from Yuanda Peroxides Co. Ltd.

(Shanghai, China). ML-1 and U937 cell lines, both myelo-

blastic, were obtained from American Type Cell Collection

(Manassas, VA, USA). The labeling kit for DNA strand

breaking was from Phoenix Flow Systems, Inc. (San Diego,

CA, USA). Okadaic acid and sodium orthovanadate were

purchased from Calbiochem (San Diego, CA, USA) and

[g32P]-ATP was from New England Nuclear (MA, USA).

Cell culture and treatments

ML-1 and U937 cells were cultured in RPMI1640 con-

taining 7.5% heat-inactivated fetal bovine serum (PAA

Laboratories, Linz, Austria) containing antibiotics (100

Ag/ml penicillin and 50 Ag/ml streptomycin sulfate) in a

humidified incubator at 37jC with 5% CO2. For UV

irradiation experiments, cells were exposed to UV-C at an

intensity of 60 J/m2 in a UV cross-linker (UVP Products,

Upland, CA, USA) for 10, 30, 60, 120, and 240 min (unless

otherwise specified). Cells (3 � 105/ml) were also exposed

to H2O2 at 0.3 mM and 1 mM in a serum-depleted medium

for various times. In some experiments, ML-1 cells (3 �105/ml) were treated with 4-AP, a K+-channel blocker, at

different concentrations for 30 min before UV or H2O2

treatment. ML-1 cells were also treated with glutathione

(1 mM), N-acetylmaleimide (0.1 mM), dimethylsulfoxide

(0.2%), or mannitol (2 mM) for 30 min before UV or H2O2

treatment.

ell Research 297 (2004) 461–471 463

Patch clamp studies

For whole-cell K+ current recording, the nystatin-perfo-

rated-patch technique was used. Pipettes with a resistance of

3–4 mV when filled with 150 mM KCl solution were

manufactured with a two-stage puller (PP-83, Narishige).

J. Gao et al. / Experimental C

Fig. 2. K+ channel activity is necessary for UV- but not H2O2-induced apoptosis. (

times in the presence or absence of 4-AP (1 mM). TUNEL staining as described in

cell death. Green staining denotes the cells undergoing apoptosis. (B) Quantificati

H2O2 in the presence or absence of 4-AP. Data (mean F SE) were summarized

The pipette tip was filled with a nystatin-free solution

containing 140 mM KCl, 2 mM MgCl2, 0.5 mM CaCl2, 2

mM ATP, 0.05 mM GTP, 1 mM EGTA, and 10 mM HEPES

(pH 7.2). Pipette was back-filled with the same solution

with 200 Ag/ml of nystatin. The bath solution was composed

of 140 mM NaCl, 2 mM KCl, 1 mM CaCl2, 10 mM HEPES

A) ML-1 cells were exposed to UV (60 J/m2) or H2O2 (0.3 mM) for various

Materials and methods was performed to visualize the extent of programmed

on of percent ML-1 cell death at various times after stimulation with UVor

from three independent experiments.

J. Gao et al. / Experimental Cell Research 297 (2004) 461–471464

(pH 7.4). An Axopatch 200A patch-clamp amplifier was

used to measure whole-cell current. Data were collected and

analyzed with pCLAMP software (Axon Instruments, Inc.).

Channel activity was determined as a fraction of currents

measured in UVor H2O2 in the presence or absence of 4-AP

(I) and in control conditions (Ic). All experiments were

performed at room temperature (21–23jC). Results were

analyzed by the Student’s t test.

TUNEL assay

ML-1 cells cultured in 100-mm dishes were treated with

4-AP or vehicle for 5 min before exposure to UV-C for 5

min. Cells harvested at 5, 30, 60, and 180 min post UV

exposure were spun onto slides through cytocentrifuge.

After fixing in methanol, cells on slides were rinsed three

times in PBS and subjected to DNA strand-breaking label-

ing via the APO-BrdU procedure according to the protocol

provided by the supplier. BrdU incorporated into the 3VOHgroups via the action of terminal deoxynucleotidyl trans-

Fig. 3. UVand H2O2 can both induce JNK phosphorylation. ML-1 cells were expo

Cell extracts were prepared and equal amounts of cell lysates were blotted for th

migrated faster than pJNKp46 band was also detected by the phospho-JNK anti

antibody), we designated this band as nonspecific (NS). As a loading control, the s

The experiment was repeated for at least three times and identical results were o

ferease was recognized by the anti-BrdU monoclonal anti-

body conjugated with FITC. The specific signals were

detected by fluorescent microscopy. Apoptotic cells (green)

in each treatment group were scored, and statistical signif-

icance of difference in the apoptotic rate between cells

treated with or without 4-AP was determined using the

Student’s t test.

Western blot analysis

ML-1 cells (4 � 106) were lysed in 160 Al of a lysis

buffer (20 mM Tris–HCl, pH 7.5, 137 mM NaCl, 1.5 mM

MgCl2, 2 mM EDTA, 10 mM sodium pyrophosphate, 25

mM h-glycerophosphate, 10% glycerol, 1% Triton X-100,

1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl

fluoride) supplemented with 1 � protease inhibitor cocktail

and 10 AM okadaic acid. After incubation on ice for 10

min, cell lysates were then precleared by centrifugation at

13,000 � g for 25 min. Equal amounts of cell lysates (20

Ag) were fractionated by SDS-PAGE followed by electro-

sed to either UV (A and B) or H2O2 (C and D) for different lengths of time.

e phosphorylation status of JNK-1 and JNK-2. One prominent protein that

body. As its status is unknown (failed to be detected by the general JNK

ame blots were also blotted for all forms of JNKs using a pan-JNK antibody.

btained.

Fig. 4. K+ channel activity is involved in UV- but not H2O2-induced JNK

phosphorylation. ML-1 cells were exposed to UV (A, 60 J/m2) or H2O2 (B,

0.3 mM) for 60 min in the presence of various concentrations of 4-AP. Cell

extracts were prepared and equal amounts of cell lysates were blotted for

the phosphorylation status of JNK-1. As a loading control, the same blots

were also blotted for all forms of JNKs. The experiment was performed at

least three different times and representative results were shown.

Fig. 5. UV induces p38 phosphorylation that can be blocked by 4-AP

treatment. ML-1 cells were exposed to UV (60 J/m2) for the indicated

periods of time in the absence (A) or for 1 h in the presence (B) of different

concentrations of 4-AP. Cell extracts were prepared from the treated cells

and Western blots were probed using antibodies against either phosphor-

ylated or total p38 proteins. The experiment was repeated for three times as

before.

J. Gao et al. / Experimental Cell Research 297 (2004) 461–471 465

transferring to nitrocellulose membrane. After blocking

with 5% nonfat milk in Tris-buffered saline containing

0.1% Tween 20 (TBST), the nitrocellulose membranes

were incubated with primary antibodies in TBST contain-

ing 5% nonfat milk overnight at 4jC. After washing three

times with TBST, the membranes were then incubated with

a goat anti-rabbit immunoglobulin G (IgGs) conjugated

with horse radish peroxidase (1:2000). Specific signals

were detected with enhanced chemiluminescent procedure

using a Supersignal West Pico Kit (Pierce, Rockford, IL,

USA).

Fluorescence microscopy

Localization of JNKs was determined by double immu-

nofluorescence analysis of a centrosomal marker. ML-1

cells spun onto culture slides via a Cytospin were fixed in

methanol for 5 min at room temperature. Fixed cells were

treated with 0.1% Triton X-100 in PBS for 5 min on ice,

and then washed three times with ice-cold PBS. After

blocking with 2.0% bovine serum albumin (BSA) in PBS

for 15 min, cells were incubated for 1 h with rabbit

polyclonal antibodies to JNKs and a mouse monoclonal

antibody to g-tubulin in 2% BSA solution. After the

primary antibody incubation, cells were washed three times

with PBS and then incubated with Rhodamine-Red-X-

conjugated anti-mouse IgGs and/or fluorescein isothiocya-

nate (FITC)-conjugated anti-rabbit IgGs (Jackson Immuno

Research) for 1 h in the dark. Fluorescence microscopy was

performed on a Nikon microscope and images were cap-

tured using a digital camera (Optronics) using Optronics

MagFire, Image-Pro Plus softwares. Statistical significance

of difference in the percentage of cells with diffused (non-

centrosomal) JNKs between cells treated with UV and cells

treated with UV plus 4-AP was determined by the Student’s

t test.

Kinase assays

Immunocomplex kinase assays were performed essen-

tially as described Ref. [19]. In brief, ML-1 cells treated

with UV or H2O2 for various times were lysed as

described above. An equal amount of cell lysates (0.5

mg) was subjected to immunoprecipitation with antibodies

to JNKs. The resulting precipitates were resuspended in a

kinase buffer as described [5]. The kinase reaction was

initiated by the addition of [g-32P]ATP (2 ACi) and ATF-

2. After incubation for 30 min at 37jC, the reaction

mixtures were analyzed by SDS-PAGE followed by im-

munoblotting for JNKs and by autoradiography for phos-

pho-ATF-2. Each assay was repeated for at least three

times and parallel results were obtained. Changes of

kinase activity were presented as a fraction of band

densities (as arbitrary units) of treated cells over the

control cells on X-ray films.

J. Gao et al. / Experimental Cell Research 297 (2004) 461–471466

Results

We have previously shown that genotoxic stressors such

as UV irradiation and ROS exposure induce apoptosis in

established cell lines [5,12,17]. We have also shown that K+

channels are involved in UV-induced apoptosis [5,21]. To

determine whether K+ channel activity is an essential part of

apoptosis induced by various genotoxic stresses, we ex-

posed ML-1 cells to UV light or H2O2 in the presence or

absence of 4-AP. Effects of UV irradiation or H2O2 on K+

activity were measured using the nystatin-perforated whole-

cell patch clamp. The whole-cell current was recorded that

resulted from depolarization of the membrane potential

Fig. 6. Subcellular localization of JNKs before and after exposure to UV irradiation

tubulin (red). DNA was stained with DAPI (blue). Stained cells were examined

different periods of time in the absence or presence of 4-AP (1 mM) were stained

fluorescence microscopy. (C) ML-1 cells were exposed to UV light for various tim

discrete staining (dot) of centrosomal JNKs was scored using a fluorescent micros

represent the average of three independent experiments.

from a holding potential of �60 to 0 mV. Normalized K+

currents in the presence or absence of 1 mM 4-AP were

determined by a fraction of UV-induced current versus

control current amplitudes (I/Ic) where I and Ic represent

currents measured with and without stimulations, respec-

tively. Upon exposure of ML-1 cells to UV light, the

amplitude of the K+ current was markedly increased. The

UV-evoked K+ current was sensitive to 4-AP as it was

completely blocked by this compound (Fig. 1). However,

H2O2 treatment failed to activate K+ channel as compared

with the untreated control (Fig. 1), suggesting a difference

in K+ requirement in ML-1 cells upon exposure to various

genotoxic stresses.

. (A) ML-1 cells were double-stained with antibodies to JNKs (green) and g-

by fluorescent microscopy. (B) ML-1 cells exposed to UV (60 J/m2) for

with the JNK antibody. The subcellular location of JNKs was examined by

es in the presence or absence of 4-AP. The percent of ML-1 cells without

cope. At least two hundred cells from each treatment were examined. Data

Fig. 7. K+ current is necessary for UV-induced JNK activation.ML-1 (A) and

U937 (B) cells were exposed to UV (60 J/m2) in the presence or absence of 4-

AP (1 mM). Cell extracts were prepared and equal amounts of cell lysates

were immunoprecipitated with the antibody to JNK. The JNK immunopre-

cipitates were used for in vitro kinase assays using recombinant ATF-2 as

substrate as described in Materials and methods. After reaction, the mixture

was fractionated on denaturing polyacrylamide gel followed by autoradiog-

raphy. Immunoprecipitation efficiency was also determined by blotting with

the antibody to JNKs. JNK activities were quantified by densitometric

scanning. Data represent the average of three independent experiments.

J. Gao et al. / Experimental Cell Research 297 (2004) 461–471 467

To determine whether K+ channel activity is essential for

ROS-induced apoptosis, we treated ML-1 cells with H2O2

or UV for various times in the presence or absence of 4-AP.

TUNEL assays revealed that ML-1 cells underwent apopto-

sis in a time-dependent manner (Fig. 2A). By 3 h post UV

exposure, nearly 50% of ML-1 cells were apoptotic (Fig.

2B). In contrast, no significant apoptosis was detected when

UV-irradiated ML-1 cells were pretreated with the K+

blocker 4-AP (Figs. 2A and B). H2O2 treatment also

induced rapid apoptosis, and pretreatment of ML-1 cells

with 4-AP did not obviously affect the rate of cell death

(Figs. 2A and B), suggesting that K+ channel activity is not

involved in apoptotic response initiated by H2O2.

Transcription factors such as AP-1, NF-nB, and p53 are

rapidly activated upon exposure of cells to genotoxic

stresses [22,23]. As JNKs are known to directly regulate

these transcription factors, we measured phosphorylation

status of JNKs in ML-1 cells exposed to UV for various

times. We observed that both JNK-1 (p54) and JNK-2 (p46)

were rapidly phosphorylated upon exposure to UV at a dose

of 60 J/m2, and that a higher dose of UV treatment (300 J/

m2) apparently neither boosted nor shortened the induction

of JNK phosphorylation (Figs. 3A and B). Although the

protein level of JNK-2 was much lower than that of JNK-1,

its phosphorylation level was comparable to that of JNK-1.

Upon H2O2 exposure, phosphorylation of JNKs was in-

creased in a manner similar to that induced by UV (Figs. 3C

and D). Again, the peak phosphorylation of JNKs was

around 1 h post exposure. It was noted that JNK phospho-

specific antibody detected a band that migrated faster than

the JNK-2. As it was not detected by the pan-JNK antibody,

we designated it as a nonspecific signal (NS). Nonetheless,

activation of JNK-1 by both UV and H2O2 was very

consistent in various experiments.

To determine whether JNK activation by UV and H2O2

was affected by K+ channel activity, ML-1 cells were

pretreated with various concentrations of 4-AP before ex-

posure to UV or H2O2. Whereas phosphorylation of JNK1

induced by UV was significantly inhibited by 4-AP at a

concentration higher than 0.5 mM, its phosphorylation

induced by H2O2 was largely unaffected by 4-AP even at

a 2 mM concentration (Figs. 4A and B). We focused on the

activation of JNK1 in this series of experiments as its

phosphorylation was more consistent in our assays. The

phosphorylation of p38 was also greatly induced by UV

irradiation and the peak phosphorylation was around 2

h after UV treatment (Fig. 5A). Likewise, 4-AP significantly

inhibited UV-induced phosphorylation of p38 (Fig. 5B)

although the level of inhibition was not nearly as complete

as that of JNKs.

JNKs are primarily localized in the cytosol. A significant

fraction of activated JNKs is thought to be localized in the

nucleus where they target substrates such as c-Jun. A recent

study showed that JNKs are present in the cell in both soluble

and insoluble forms [24]. Whereas soluble JNKs are present

in the cytoplasm, insoluble JNKs are localized to the

centrosomes [24]. We examined the subcellular localization

of JNKs in ML-1 cells via fluorescent microscopy. A low

level of JNK staining was detectable throughout the cyto-

plasm of ML-1 cells and JNK signals detected as a bright dot

were present in each cell (Fig. 6A, arrow). Immunostaining

with g-tubulin, a centrosome-specific protein, confirmed that

JNKs were concentrated at the centrosomes (Fig. 6A). We

next examined the subcellular localization of JNKs in ML-1

J. Gao et al. / Experimental Cell Research 297 (2004) 461–471468

cells exposed to UV for various times. Five minutes post UV

exposure, JNK staining that started to disperse no longer

confined to the centrosomes (Fig. 6B). The dissipation of

JNKs throughout the cell persisted after UV treatment for 30

min or longer (Figs. 6B and C). On the other hand, when

cells were pretreated with 4-AP, centrosomal JNKs were not

significantly dispersed and JNKs remained at the centro-

somes or the centrosomal region in most ML-1 cells (Figs.

6B and C). In addition, little nuclear JNK signals were

detected 30 min after UV treatment (Fig. 6B).

The enhanced phosphorylation status and subcellular

translocation of JNKs strongly suggest that K+ channel is

required for the activation of JNKs. We directly measured

JNK activities in ML-1 cells exposed to UV in the presence

or absence of 4-AP. Immunocomplex kinase assays revealed

that UV treatment strongly activated JNKs and that pre-

treatment with 4-AP completely abolished the activation

(Fig. 7A), again indicating the requirement of K+ channel in

Fig. 8. Redox status plays an important role in the activation of JNKs upon UV irra

(60 J/m2, C and D) in the presence or absence of glutathione (GSH), N-acetylm

amounts of cell lysates were immunoprecipitated with antibody to JNK. The JNK

vitro substrate as described in Materials and methods. GSH and NEM but not DM

activities were quantified by densitometric scanning. Data represent the average o

JNK activation by UV. The requirement of the K+ channel

activity in JNK activation was also confirmed in U937 cells

(Fig. 7B). We then asked whether the redox status or

sulfhydryl group was involved in regulating JNK activation

by H2O2 and UV. ML-1 cells were pretreated with reducing

agents before exposure to H2O2 as well as to UV. JNK

activity was increased upon H2O2 exposure that peaked

around 1 h in ML-1 cells, and treatment with glutathione

(GSH), a potent reducing agent, significantly abolished

H2O2-induced activation of JNKs (Figs. 8A and B). Pre-

treatment of the cells with DMSO and mannitol, the latter

being a hydroxyl radical scavenger, did not affect JNK

activity (Fig. 8B). To determine the effect of reducing agents

on JNK activation by UV, ML-1 cells were pretreated with

GSH and N-acetylmaleimide (NEM), the latter being a

strong sulfhydryl alkylating agent that also affects K+

current in vivo, before UV irradiation. Immunocomplex

kinase assays revealed that NEM completely inhibited

diation. ML-1 cells were exposed to either H2O2 (0.3 mM, A and B) or UV

aleimide (NEM), dimethyl sulfoxide (DMSO), or Mannitol (Man). Equal

immunoprecipitates were analyzed for kinase activities using ATP-2 as in

SO and Man selectively block H2O2- and UV-induced JNK activation. JNK

f three independent experiments.

J. Gao et al. / Experimental Cell Research 297 (2004) 461–471 469

UV-induced JNK activities in vitro (Fig. 8C). Consistently,

the reducing agent GSH also blocked the kinase activity

induced by UV (Fig. 8D), indicating that redox status does

play an important role in the activation of JNKs upon H2O2

or UV irradiation.

Discussion

In the present study, we determined whether JNK acti-

vation by different genotoxic stressors was mediated by the

same cellular components by measuring activities and

subcellular localization of JNKs in ML-1 cells exposed to

either UV irradiation or H2O2. Our results suggest that JNK

activation by UV and H2O2 is mediated by distinct yet

overlapping pathways and that K+ channel activity and

redox status are differentially required for UV- and H2O2-

induced activation of JNKs.

We found that exposure of ML-1 cells to UV and H2O2

rapidly induces phosphorylation and activation of JNKs.

Moreover, UV-induced phosphorylation, activation and sub-

cellular translocation of JNKs are inhibited by 4-AP, indi-

cating that K+ channel activity is essential for the JNK

functional activation in vivo. Indeed, UV irradiation acti-

vates a 4-AP-sensitive K current. Interestingly, H2O2 also

induces phosphorylation and activation of JNKs in a kinet-

ics similar to that induced by UV. However, JNK activation

by H2O2 does not activate K currents and is not affected by

4-AP, suggesting that there exist distinct pathways of JNK

activation that can be distinguished by K+ channel depen-

dence. Although the exact nature of these molecular path-

ways remains unclear, UV light directly irradiates the

surface membrane of the cell, which may trigger the

activation of K+ channels in the membrane, an essential

early process in the signaling cascade that leads to JNK

activation. The increased K+ channel activity may result in a

change in intracellular K+ concentration and a change in cell

volume, leading to activation of stress response proteins

such as JNKs and p38. On the other hand, H2O2 is thought

to be able to penetrate through the cell surface membrane,

bypassing the requirement for K+ channel and directly

activating downstream molecules in the JNK activation

cascade. As we have shown, UV irradiation induces an

activation of p38, which is consistent with previous studies

in other cell types [25,26]. Relative to JNK/SAPK activa-

tion, activation of p38 in ML-1 cells in response to UV

irradiation was much less sensitive to suppression of K+

channels by 4-AP (>1.0 mM, Fig. 5B). It remains unknown

as to what is the effect of UV-induced p38 activation on the

survival of ML-1 cells. At present, we cannot exclude that

there may be potential side effects of 4-AP directly on JNK

activities or indirectly on its signaling pathway other than

blocking K+ channel.

Numerous studies reveal that K+ channel activity is

affected by apoptotic inducers including reactive oxygen

species [27], Fas ligand and TNF [28], and anticancer drugs

[29]. There is growing evidence showing that K+ channel

activities are probably involved in programmed cell death.

Neurons undergoing apoptosis exhibited an upregulation of

outward K+ currents. This enhancement of outward K+

current, induced by serum deprivation and staurosporine,

can be prevented by the K+ channel blockers 4-AP and

TEA. It has also been observed that the K+ channel opener

cromakalim induces neuronal apoptosis [30]. In addition,

the K+ channel blocker 4-AP inhibits the shrinkage of

human eosinophils undergoing apoptosis induced by cyto-

kine withdrawal [31], and a combination of two K+ channel

blockers, TEA and 4-AP, inhibited IL-1b release from LPS-

stimulated monocytes [32]. Thus, blockade of UV irradia-

tion-induced activation of K+ channels by 4-AP in ML-1

cells provides further evidence that inhibition of possible K+

efflux, the consequent membrane hyperpolarization, and

decrease in cell volume prevent activating a particular

signaling system leading to cell apoptosis.

UV irradiation also causes oxidative stress [33]. Pretreat-

ment of ML-1 cells with GSH and NEM prevents the

activation of JNKs by UV, suggesting that UV may trigger

an oxidative stress on cell surface membrane, an event before

the activation of K+ channel. It is also possible that K+

channel activity can be directly affected by the oxidative

stress. For example, disulfide bond formation, largely influ-

enced by the redox status, could be an essential part that

regulates the K+ channel molecules. It has been shown that

the antioxidant agent GSH plays a central part in the cellular

defense against UV radiation and that an increase in GSH

concentration may provide photoprotection in many cell

types [34]. We have demonstrated that GSH blocks UV-

induced JNK activation in ML-1 cells in a concentration

range of 5–15 mM (data not shown). A moderate concen-

tration of GSH prevents UV- and H2O2-induced JNK acti-

vation in ML-1 cells, the target(s) of which appears to lie at

the very upstream of the signaling cascade. Moreover, NEM,

a protective alkylating agent of sulfhydryl group only [35],

prevents UV-induced JNK activation without affecting the

effect of H2O2-induced JNK activation. It suggests that NEM

may interact with sulfhydryl group in the channel protein to

inhibit UV-induced changes of the channel conformation.

The mechanism of NEM effect is quite different from the

effect of 4-AP. 4-AP is a specific blocker of voltage-depen-

dent K+ channels and it binds to the pore region of the

channel to inhibit K+ ion fluxes [36]. Free radical scavengers

DMSO and mannitol are also used to study whether UV

induced K+ channel activation and downstream JNK activa-

tion. The membrane permeable and impermeant scavengers,

respectively, did not affect UV-induced response in ML-1

cells, supporting the notion that a conformational change of

the channel protein in response to UV irradiation may be

required to activate the channel.

A direct consequence of enhanced K+ channel activity is

a change in the cell volume due to water loss, which

inevitably affects the cytoskeletons of the cell. Centrosome,

the major microtubule organization center of the cell, may

J. Gao et al. / Experimental Cell Research 297 (2004) 461–471470

function to sense cellular (e.g., microtubule) stresses. Reg-

ulatory molecules that reside at the centrosome can be

activated because the microtubule stress initiated by the cell

volume changes. This notion is supported by several lines of

existing evidence: (i) JNKs are localized to the centrosomes,

and UV irradiation causes their rapid subcellular transloca-

tion; (ii) UV-induced subcellular translocation is largely

blocked by 4-AP treatment; (iii) nocodazole, an agent that

depolymerizes microtubules, also activates MEKK1 [37], an

upstream activator of JNKs, and p38 [38]. An alternative

explanation for JNK subcellular localization at the centro-

somes is that the centrosome serves as an intracellular

transportation depot. Signaling molecules move rapidly

along the tracks of microtubules and are modified (e.g.,

phosphorylation) at the centrosome so that the signals on the

cell surface membrane can be directed to various organelles

including the nucleus. The fact that many signaling protein

kinases, such as MEK1, ERKs, and Plks [39] (unpublished

data), and checkpoint proteins, including p53 and BRCAs

[40], are found at the centrosomes supports this possibility

although it remains to be elucidated as to the exact role of

JNKs as centrosomal proteins.

Acknowledgments

We thank the members in the joint laboratory of Drs. Lu,

Dai, and Xu for helpful discussions. We are also grateful to

Drs. Guoqiang Chen and Ying Jin for the use of equipment,

and Xueyu Chen for assistance. This work was supported in

part by a start-up fund from the SSMU, a core support from

the Chinese Academy of Sciences, a One Hundred Talent

Grant from the Chinese Academy of Sciences (M.X.), a

National Key Program (973) for Basic Research of China

grant (NO2002CB512805), and a grant from E-Institute of

Shanghai Municipal Education Commission.

References

[1] M. Buscher, H.J. Rahmsdorf, M. Litfin, M. Karin, P. Herrlich, Acti-

vation of the c-fos gene by UV and phorbol ester: different signal

transduction pathways converge to the same enhancer element, On-

cogene 3 (1988) 301–311.

[2] Y. Devary, R.A. Gottlieb, L.F. Lau, M. Karin, Rapid and preferential

activation of the c-Jun gene during the mammalian UV response,

Mol. Cell. Biol. 11 (1991) 2804–2811.

[3] P. Herrlich, H. Ponta, H.J. Rahmsdorf, DNA damage-induced gene

expression: signal transduction and relation to growth factor signal-

ing, Rev. Physiol., Biochem. Pharmacol. 119 (1992) 187–223.

[4] C. Rosette, M. Karin, Ultraviolet light and osmotic stress: activation

of the JNK cascade through multiple growth factor and cytokine

receptors, Science 274 (1996) 1194–1197.

[5] T. Li, W. Dai, L. Lu, Ultraviolet-induced junD activation and apopto-

sis in myeloblastic leukemia ML-1 cells, J. Biol. Chem. 277 (2002)

32668–32676.

[6] C. Hagemann, J.L. Blank, The ups and downs of MEK kinase inter-

actions, Cell Signal. 13 (2001) 863–875.

[7] G. Fritz, B. Kaina, rhoB encoding a UV-inducible Ras-related small

GTP-binding protein is regulated by GTPases of the Rho family and

independent of JNK, ERK, and p38 MAP kinase, J. Biol. Chem. 272

(1997) 30637–30644.

[8] Z. Guo, A. Kumagai, S.X. Wang, W.G. Dunphy, Requirement for Atr

in phosphorylation of Chk1 and cell cycle regulation in response to

DNA replication blocks and UV-damaged DNA in Xenopus egg

extracts, Genes Dev. 14 (2000) 2745–2756.

[9] K. Tobiume, A. Matsuzawa, T. Takahashi, H. Nishitoh, K. Morita, K.

Takeda, O. Minowa, K. Miyazono, T. Noda, H. Ichijo, ASK1 is

required for sustained activations of JNK/p38 MAP kinases and ap-

optosis, EMBO Rep. 2 (2001) 222–228.

[10] S.Y. Fuchs, V. Adler, M.R. Pincus, Z. Ronai, MEKK1/JNK signaling

stabilizes and activates p53, Proc. Natl. Acad. Sci. U. S. A. 95 (1998)

10541–10546.

[11] S. Le, T.J. Connors, A.C. Maroney, c-Jun N-terminal kinase specifi-

cally phosphorylates p66ShcA at serine 36 in response to ultraviolet

irradiation, J. Biol. Chem. 276 (2001) 48332–48336.

[12] S. Xie, Q. Wang, L. Luo, Q. Ruan, T. Liu, M. Jhanwar-Uniyal, Z.

Darzynkiewicz, F. Traganos, W. Dai, Proteasome-dependent down-

regulation of p21(Waf1/Cip1) induced by reactive oxygen species, J.

Interferon Cytokine Res. 22 (2002) 957–963.

[13] A.K. Garg, B.B. Aggarwal, Reactive oxygen intermediates in TNF

signaling, Mol. Immunol. 39 (2002) 509–517.

[14] C.D. Bortner, F.M. Hughes Jr., J.A. Cidlowski, A primary role for K+

and Na+ efflux in the activation of apoptosis, J. Biol. Chem. 272

(1997) 32436–32442.

[15] L. Wang, P. Zhou, R.W. Craig, L. Lu, Protection from cell death by

mcl-1 is mediated by membrane hyperpolarization induced by K(+)

channel activation, J. Membr. Biol. 172 (1999) 113–120.

[16] L. Lu, T. Yang, D. Markakis, W.B. Guggino, R.W. Craig, Alterations

in a voltage-gated K+ current during the differentiation of ML-1

human myeloblastic leukemia cells, J. Membr. Biol. 132 (1993)

267–274.

[17] S. Xie, Q. Wang, H. Wu, J. Cogswell, L. Lu, M. Jhanwar-Uniyal, W.

Dai, Reactive oxygen species-induced phosphorylation of p53 on

serine 20 is mediated in part by polo-like kinase-3, J. Biol. Chem.

276 (2001) 36194–36199.

[18] J.B. Schulz, M. Weller, T. Klockgether, Potassium deprivation-in-

duced apoptosis of cerebellar granule neurons: a sequential require-

ment for new mRNA and protein synthesis, ICE-like protease activity,

and reactive oxygen species, J. Neurosci. 16 (1996) 4696–4706.

[19] I. Walev, K. Reske, M. Palmer, A. Valeva, S. Bhakdi, Potassium-

inhibited processing of IL-1 beta in human monocytes, EMBO J.

14 (1995) 1607–1614.

[20] I. Walev, J. Klein, M. Husmann, A. Valeva, S. Strauch, H. Wirtz, O.

Weichel, S. Bhakdi, Potassium regulates IL-1 beta processing via

calcium-independent phospholipase A2, J. Immunol. 164 (2000)

5120–5124.

[21] B. Xu, B.A. Wilson, L. Lu, Induction of human myeloblastic ML-1

cell G1 arrest by suppression of K+ channel activity, Am. J. Physiol.

271 (1996) C2037–C2044.

[22] M.A. Read, M.Z. Whitley, S. Gupta, J.W. Pierce, J. Best, R.J. Davis,

T. Collins, Tumor necrosis factor alpha-induced E-selectin expression

is activated by the nuclear factor-kappaB and c-JUN N-terminal ki-

nase/p38 mitogen-activated protein kinase pathways, J. Biol. Chem.

272 (1997) 2753–2761.

[23] B.B. Zhou, S.J. Elledge, The DNA damage response: putting check-

points in perspective, Nature 408 (2000) 433–439.

[24] R.A. MacCorkle-Chosnek, A. VanHooser, D.W. Goodrich, B.R.

Brinkley, T.H. Tan, Cell cycle regulation of c-Jun N-terminal kinase

activity at the centrosomes, Biochem. Biophys. Res. Commun. 289

(2001) 173–180.

[25] M. Karin, Z. Liu, E. Zandi, AP-1 function and regulation, Curr. Opin.

Cell Biol. 9 (1997) 240–246.

[26] M.L. Lu, M. Sato, B. Cao, J.P. Richie, UV irradiation-induced apop-

tosis leads to activation of a 36-kDa myelin basic protein kinase in

HL-60 cells, Proc. Natl. Acad. Sci. U. S. A. 93 (1996) 8977–8982.

J. Gao et al. / Experimental Cell Research 297 (2004) 461–471 471

[27] F. Duprat, E. Guillemare, G. Romey, M. Fink, F. Lesage, M. Honore,

E. Honore, Susceptibility of cloned K+ channels to reactive oxygen

species, Proc. Natl.Acad. Sci. U. S. A. 92 (1995) 11796–11800.

[28] B. Soliven, S. Szuchet, D.J. Nelson, Tumor necrosis factor inhibits

K+ current expression in cultured oligodendrocytes, J. Membr. Biol.

124 (1991) 127–137.

[29] B. Rouzaire-Dubois, J.M. Dubois, Tamoxifen blocks both prolifera-

tion and voltage-dependent K+ channels of neuroblastoma cells, Cell

Signal. 2 (1990) 387–393.

[30] S.P. Yu, C.H. Yeh, S.L. Sensi, B.J. Gwag, L.M. Canzoniero, Z.S.

Farhangrazi, H.S. Ying, M. Tian, L.L. Dugan, D.W. Choi, Mediation

of neuronal apoptosis by enhancement of outward potassium current,

Science 278 (1997) 114–117.

[31] F. Beauvais, L. Michel, L. Dubertret, Human eosinophils in culture

undergo a striking and rapid shrinkage during apoptosis. Role of K+

channels, J. Leukocyte Biol. 57 (1995) 851–855.

[32] I. Walev, K. Reske, M. Palmer, A. Valeva, S. Bhakdi, Potassium-

inhibited processing of IL-1 beta in human monocytes, EMBO J.

14 (1995) 1607–1614.

[33] D. Kulms, T. Schwarz, Molecular mechanisms involved in UV-in-

duced apoptotic cell death, Skin Pharmacol. Appl. Skin Physiol. 15

(2002) 342–347.

[34] D.P. Steenvoorden, D.M. Hasselbaink, G.M. Beijersbergen van Hene-

gouwen, Protection against UV-induced reactive intermediates in hu-

man cells and mouse skin by glutathione precursors: a comparison of

N-acetylcysteine and glutathione ethylester, Photochem. Photobiol. 67

(1998) 651–656.

[35] M.K. al Shawi, I.L. Urbatsch, A.E. Senior, Covalent inhibitors of P-

glycoprotein ATPase activity, J. Biol. Chem. 269 (1994) 8986–8992.

[36] L. Lu, T. Yang, D. Markakis, W.B. Guggino, R.W. Craig, Alterations

in a voltage-gated K+ current during the differentiation of ML-1

human myeloblastic leukemia cells, J. Membr. Biol. 132 (1993)

267–274.

[37] Y. Xia, Z. Wu, B. Su, B. Murray, M. Karin, JNKK1 organizes a MAP

kinase module through specific and sequential interactions with up-

stream and downstream components mediated by its amino-terminal

extension, Genes Dev. 12 (1998) 3369–3381.

[38] K. Takenaka, T. Moriguchi, E. Nishida, Activation of the protein

kinase p38 in the spindle assembly checkpoint and mitotic arrest,

Science 280 (1998) 599–602.

[39] W. Dai, X. Huang, Q. Ruan, Polo-like kinases in cell cycle checkpoint

control, Front Biosci. 8 (2003) D1128–D1133.

[40] C.X. Deng, Roles of BRCA1 in centrosome duplication, Oncogene 21

(2002) 6222–6227.