(Received for publication, September 12, 1996, and in revised form, November 7, 1996)
From the Vollum Institute, the Division of
Cardiology, and the § Department of Pathology, Oregon Health
Sciences University, Portland, Oregon 97201
The c-Jun N-terminal protein kinases (JNKs), also called stress-activated protein kinases, are members of the growing family of serine/threonine kinases in the mitogen-activated protein (MAP) kinase superfamily. Like other MAP kinases, JNKs are activated via phosphorylation on adjacent threonine and tyrosine residues and can be inactivated by a unique family of dual specificity phosphatases, called MAP kinase phosphatases (MKPs). MKPs are encoded by immediate early genes and induced in response to environmental stressors and growth factor stimulation. Two prevalent isoforms of MKP, MKP1 and MKP2, are co-expressed in a wide variety of cell types. In this study, we examined the actions of MKP1 and MKP2 on JNK1 and JNK2. JNK1 phosphorylation and activation was inhibited by expression of both MKP1 and MKP2, although MKP1 selectivity toward JNK1 appeared significantly higher than that of MKP2. In contrast, JNK2 activity was inhibited by either phosphatase to similar degrees. Both MKP1 and MKP2 were highly effective at blocking the activation of the physiological target of JNK activation, the transcription factor c-Jun. In PC12 cells, MKP1 and MKP2 are transcriptionally induced following stimulation by nerve growth factor. In these cells, UV light-evoked JNK activation was reduced by pretreatment with nerve growth factor. Therefore, JNKs may be selective targets of MKP action in certain cells.
Mitogen-activated protein (MAP)1 kinases are a unique family of serine/threonine kinases that are activated via reversible phosphorylation and mediate signal transduction for multiple extracellular stimuli. In mammalian cells, at least two MAP kinase pathways have been extensively characterized: the extracellular signal-regulated protein kinase (ERK) cascade, which is responsible for signal transduction involving growth and differentiation, and the stress-activated protein kinase (1) or c-Jun N-terminal protein kinase (JNK) cascade, which mediates the cellular responses to proinflammatory cytokines and genotoxic stress (2). ERKs and JNKs are specifically activated by distinct isoforms of upstream MAP/ERK kinases (MEKs or MKKs) and can be inactivated in vitro (3) and in vivo (3-8) by a novel class of dual specificity phosphatases collectively termed MAP kinase phosphatases (MKPs). The contrasting cellular responses dictated by ERK and JNK activation are regulated in part by the coordinate effects of MEKs and MKPs upon their respective substrates.
MKPs are the products of immediate early genes whose mRNA is rapidly induced following such divergent stimuli as serum, epidermal growth factor and NGF (3, 9), short-wave UV light and DNA-alkylating agents (8). MKP activity is directed toward the homologous tripeptide phosphorylation sites required for ERK and JNK activation, TEY and TPY, respectively (10-14), and MKP actions have thus far been confined exclusively to the MAP kinase family of proline-directed kinases. The prototype dual specificity phosphatase (VH1) was identified first in vaccinia; however, homologs were soon isolated in yeast (Yop51 and MSG5), humans (CL100, PAC1, VH3, and B23), mice (MKP1, 3CH134, and Erp) (15-21), and rats (MKP2 and VH6) (9). Of the current isoforms, PAC1 is present only in hematopoietic cells (19), while MKP1 and MKP2 are expressed more ubiquitously and with overlapping tissue distributions (9). Sequence comparisons between MKP1 and MKP2 demonstrate a highly conserved catalytic core but substantial disparity within the N-terminal domain. Although significant differences in substrate specificity, post-translational regulation, or subcellular localization might be expected, none have yet been identified (6, 9).
Early investigations proposed a role for MKPs as inactivators of ERKs following serum stimulation (3, 22); however, subsequent studies have demonstrated that growth factor-induced transcription of MKP1 and MKP2 did not correspond temporally with ERK inactivation (9, 23), and other mechanisms governing cessation of ERK activity following these stimuli have since been proposed (24-27). More recently, MKPs have been implicated in the regulation of JNK activity (8, 28). Ten isoforms of JNK have been isolated, but analysis of their nucleotide sequences suggests that all are the likely products of alternative processing of the transcripts of only three genes (28). Of these numerous isoforms, JNK1, originally isolated from fetal brain (11) and JNK2, isolated from HeLa (12) and Jurkat (29) cells, have been studied most extensively. JNK3 and its isoforms were only recently characterized (28). MKP actions upon JNKs have been studied previously, but with the goal of simply establishing whether MKPs could inactivate JNK (8, 28) or as an element of investigations toward potential substrate specificities of MKPs between MAP kinase families (6). Direct comparisons of substrate specificities for MKPs among JNK isoforms has not been closely examined. We therefore studied in detail the actions of MKP1 and MKP2 on the kinase activity of the two best characterized isoforms of JNK, JNK1 and JNK2, and compared these effects with the inhibition of JNK-induced gene activation via the transcription factor c-Jun. Additionally, we examined the potential role of MKP induction in regulation of the JNK activity following growth factor activation of the ERK signal transduction cascade.
The anti-ERK2 (C-14) antibody, anti-JNK1 (FL)
antibody, and the anti-MEK kinase (C-22) antibody were purchased from
Santa Cruz Biotechnology. The anti-flag antibody (murine M2) was
purchased from Kodak Scientific Imaging Systems. The anti-PAC1 antibody was provided courtesy of Dr. Steven Pelech (Kineteck, Vancouver, British Columbia, Canada). GST-c-Jun was bacterially expressed using a
plasmid construct provided by Dr. David Parker (Oregon Health Sciences
University), and purified protein was prepared according to published
protocols (30). [-32P]ATP was purchased from DuPont
NEN, and NGF was obtained from Boehringer Mannheim.
COS-7 cells were provided by ATCC and were grown at 5% CO2 in Dulbecco's modified Eagle's medium containing 10% fetal calf serum, and 5% L-glutamine. PC12-GR5 cells were provided by Rae Nishi (Oregon Health Sciences University) and were grown at 5% CO2 in Dulbecco's modified Eagle's medium containing 5% fetal calf serum, 10% horse serum, and 5% L-glutamine. Cells were serum-starved for 16 h in Dulbecco's modified Eagle's medium alone prior to treatment with NGF (100 ng/ml).
PlasmidsFull-length MKP1 cDNA (1.9-kilobase pair
HindIII-BamHI fragment) was provided by Dr.
Nicholas Tonks (Cold Spring Harbor Laboratories). A truncated MKP2
cDNA fragment (2.4-kilobase pair Eco-RV fragment) subcloned into
pcDNA3 (Invitrogen) under the CMV promoter was cloned and
previously characterized by our laboratory (9). Full-length JNK1 and
ERK2 cDNA subcloned into the pcDNA3 vector were provided by Dr.
Richard Maurer (Oregon Health Sciences University). Full-length JNK2,
fused to the flag epitope and subcloned into pCMV5 was provided by Dr.
Roger Davis (Howard Hughes Medical Institute, Worcester, MA). A
constitutively active truncated MEK kinase (MEKK) in pCMV5,
pCMV5-MEKK, was provided by Gary Johnson (National Jewish Center for
Immunology and Respiratory Medicine, Denver, Colorado). The
5xGal4-E1B-luciferase plasmid was a gift of Richard Maurer (Oregon
Health Sciences University) and has been described previously (9). The
Gal4-c-Jun plasmid was a gift of Richard Goodman (Vollum Institute).
80% confluent COS-7 cells were
co-transfected with the combinations of plasmids indicated using the
standard calcium phosphate co-precipitation method (Life Technologies,
Inc.). The parent vector, pcDNA3, was added in all experiments to
equalize the total amount of transfected DNA. Cells were incubated in
precipitate overnight and then washed three to five times (until free
of precipitate) and allowed to recover in serum-containing media for an
additional 24 h prior to either serum starvation and treatment or
harvesting. For experiments involving MEKK and JNK1, COS-7 cells were
co-transfected with 1 µg of the constitutively active truncated MEK
kinase (MEKK), 10 µg of JNK1, and 10 µg of either MKP1 or MKP2.
In other experiments, the quantities of transfected cDNAs are
indicated.
Transfected cells were washed
twice in phosphate-buffered saline (PBS) and then lysed in buffer
containing 10% sucrose, 1% Nonidet P-40, 20 mM Tris-HCl
(pH 8.0), 137 mM NaCl, 10% glycerol, 2 mM
EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 µg/ml
leupeptin, 1 mM sodium orthovanadate, and 10 mM
sodium fluoride. Lysates were centrifuged at low speed to remove
nuclei, and the supernatant was assayed by the method of Bradford (31)
to determine total protein concentration. Samples containing 100 µg
of total protein were then immunoprecipitated in a total of 500 µl of
lysis buffer with an agarose-conjugated antibody to ERK2 (C-14) (Santa
Cruz Biotechnology) overnight at 4 °C. Immunoprecipitated ERK2 was then washed three times in wash buffer containing 50 mM
Tris-HCl (pH 7.3), 0.2% Nonidet P-40, 0.5 M NaCl, 1 mM phenylmethylsulfonyl fluoride, and 5% sucrose (w/v).
Kinase activity was determined by incubation with 10 µg of myelin
basic protein and 5-10 µCi of [-32P]ATP in 50 µl
of assay buffer containing 80 mM Hepes (pH 7.4), 80 mM MgCl2, 0.1 mM ATP, 2 mM sodium orthovanadate, and 20 mM sodium
fluoride for 30 min at 30 °C. Reactions were terminated via the
addition of 65 µl of Laemmli sample buffer and analyzed by
SDS-polyacrylamide gel electrophoresis. Quantitation was performed following scanning the gel with a PhosphorImager 445SI (Molecular Dynamics, Inc.).
Cells were transfected
as indicated above. Treated and untreated cells were then washed twice
in PBS or HPS (20 mM Hepes-KOH (pH 7), 150 mM
NaCl) and harvested in a stress-activated protein kinase lysis buffer
containing 20 mM Hepes-KOH (pH 7.4), 2 mM EDTA,
50 mM -glycerophosphate, 10% glycerol, 1% Triton
X-100, 1 mM dithiothreitol, 1 mM sodium
orthovanadate, 0.4 mM phenylmethylsulfonyl fluoride, 0.5 µg/ml aprotinin, and 0.5 µg/ml leupeptin. Lysates were centrifuged
at low speed to separate nuclei, and the supernatant was assayed for
total protein concentration via the method of Bradford (31). Samples
containing 100 µg of total protein were immunoprecipitated overnight
at 4 °C with an agarose-conjugated anti-JNK1 (FL) antibody (Santa
Cruz Biotechnology), or, for JNK2 kinase assays, anti-M2 antibody
(Eastman Kodak Co.) and Protein G-Sepharose (Pierce). The
immunoprecipitated JNK was then washed twice in LiCl buffer (500 mM LiCl, 100 mM Tris-HCl (pH 7.6), 1 mM dithiothreitol, and 0.1% Triton X-100) and three times
in assay buffer (20 mM MOPS (pH 7.2), 10 mM
MgCl2, 2 mM EGTA, 1 mM
dithiothreitol, and 0.1% Triton X-100) and assayed for kinase activity
via incubation with 3 µg of GST-c-Jun and 1 µCi of
[
-32P]ATP in 50 µl of assay buffer for 30 min at
30 °C. Reactions were terminated by the addition of 65 µl of
Laemmli sample buffer, and samples were analyzed by SDS-polyacrylamide
gel electrophoresis. Quantitation was determined by scanning the gel
using a PhosphorImager.
Stimulation of JNK activity
with ultraviolet light was achieved using the available bactericidal
lamps employed in tissue culture hoods. UV energy was measured using an
ultraviolet light meter (Questar Microsystems, Inc., Woodinville, WA;
model LT1000). UV output was determined to be from 0.05 to 0.20 J
s1 m
2. Cells were treated for a duration of
60 s or a total delivered UV dose of 3-12 J m
2.
Treated and untreated cells were washed twice in PBS and lysed in stress-activated protein kinase buffer as described above. Lysates were centrifuged at low speed to remove nuclei, and total protein concentration was determined by the Bradford method (31). Samples containing 25 µg of total protein were resolved on a 12% SDS-polyacrylamide gel and transferred onto an Immobilon P membrane using a semidry transfer apparatus (Bio-Rad). Membranes were blocked using blocking buffer containing 10% (w/v) dry milk (Carnation) in PBS with 0.15% Tween. Membranes were probed with anti-JNK1 antibody (diluted at 1:10,000) (Santa Cruz Biotechnology, Inc.), anti-MEKK antibody (diluted at 1:1000), anti-PAC1 antibody (diluted at 1:1000), or anti-phosphotyrosine antibody (diluted at 1:1000). Other samples containing 100 µg of total protein were immunoprecipitated overnight with an anti-JNK1 antibody, and the immunoprecipitate was washed in LiCl and assay buffers as described for the JNK immune complex assay above. These samples were then resolved on a 12% SDS-polyacrylamide gel and transferred to Immobilon P membrane and probed with anti-phosphotyrosine antibody (diluted at 1:1000). A horseradish peroxidase-conjugated secondary antibody was utilized to allow detection of the appropriate bands using enhanced chemiluminescence (Amersham Corp.).
Jun/Gal AssayAs described previously, COS-7 cells were co-transfected using the calcium phosphate method with the various combinations of the following plasmids: 1 µg of pCMV5-MEKK, 5-20 µg of pcDNA3-MKP1, 5-20 µg of pcDNA3-MKP2, and 10 µg of pcDNA3-JNK1. To determine the transcriptional activation of c-Jun, we utilized a mammalian two-hybrid system. Cells were also transfected with 3 µg of 5xGal4-E1B-luciferase with or without 3 µg of Gal4-c-Jun, as indicated. As with all studies, the parent vector pcDNA3 was added to each set of transfections to equalize the total amount of transfected DNA. Twelve hours after transfection, cells were washed free of precipitate and allowed to recover in serum-containing medium for an additional 36 h. Cells were then harvested for the luciferase assay. Briefly, cells were washed twice in PBS, harvested in 1 ml of PBS, following low speed centrifugation, and lysed by freeze-thawing three times in 100 mM K2PO4, pH 7.8. The lysate was centrifuged at high speed, and the supernatant was assayed for luciferase activity using a luminometer (AutoLumat LB953). Protein assays were performed on the remaining supernatant by the method of Bradford (31), and all raw data were normalized to light units/µg of total protein.
COS-7
cells were initially chosen for study because of their ability to
replicate and overexpress plasmids containing the SV-40 origin of
replication, thus allowing direct monitoring of all transfected and
expressed proteins. To confirm the efficacy of COS-7 cells as a model
for MKP activity in vivo, we examined the ability of MKP1
and MKP2 to inactivate ERK2 following stimulation with a truncated
Raf-1, BxB-Raf, that is a constitutive activator of ERKs (32). COS-7
cells were transfected with a cDNA encoding ERK2 and co-transfected
with vector alone, 1 µg of BxB-Raf, or BxB-Raf plus 10 µg of MKP1
or MKP2. Cells were harvested 48 h after transfection, and ERK
activity was determined via an immune complex assay. Transfection of
vector alone or ERK2 alone produced no detectable ERK activity (Fig.
1). Transfection of BxB-Raf with ERK2 produced a large
increase in ERK activity as assayed by
[-32P]ATP-labeled myelin basic protein, while
co-transfection of 10 µg of either MKP1 or MKP2 cDNA reduced
assayed ERK2 activation nearly to base line. These results confirm that
MKPs can be expressed in COS-7 cells to inactivate a co-transfected
substrate.
MEKK-induced JNK1 Activity Is Blocked by Co-transfection of MKP1 and MKP2
Using COS-7 cells, we then examined the capability of
either MKP to inactivate JNK1. Constitutively active MEKK (MEKK) is a potent stimulus for JNK activation (33, 34). No JNK1 kinase activity
was evident in lysates prepared from cells transfected with vector or
JNK1 alone (Fig. 2A). Co-transfection with
MEKK produced robust JNK1 activation, but this activation was almost completely eliminated following transfection with the cDNA for MKP1
and was significantly reduced in the presence of MKP2 (Fig. 2A). The activation of endogenous JNKs by
MEKK was not
detectable in this assay and is presumably secondary to the small
percentage of the cells transfected with
MEKK. These results
demonstrate the specificity of this assay for transfected JNK1.
Transient Transfection of MKP1 and MKP2 Inhibit MEKK-induced JNK1 Tyrosine Phosphorylation
To establish whether the reduction in
JNK1 kinase activity by MKP1 and MKP2 corresponded to reductions in
JNK1 tyrosine phosphorylation, cell lysates were prepared and examined
for JNK1 phosphotyrosine content by immunoprecipitation with an
agarose-conjugated anti-phosphotyrosine antibody followed by Western
blotting with anti-JNK1 antibody (Fig. 2B). No detectable
phosphorylated JNK was seen in cells transfected with vector alone or
with vector containing JNK1 in the absence of the activated MEKK.
Co-transfection of 1 µg of MEKK produced marked phosphorylation of
the substrate JNK1. More importantly, the addition of 10 µg of MKP1
resulted in a significant reduction in JNK1 phosphorylation to levels
seen with transfection of JNK1 alone, while co-transfection of MKP2
with MEKK reduced JNK1 phosphotyrosine content by more than 50%. Thus,
both MKP1 and MKP2 dephosphorylate and inactivate MEKK-stimulated
JNK1.
Because the results of the
above Western blots and immune complex assays may have been contingent
not only upon MKP's actions upon JNK1 but upon the degree of plasmid
replication and translation of the transfected cDNAs encoding JNK1,
MKP1 or MKP2, and MEKK, we examined protein expression of all
transfected components via Western blotting. Lysates from COS-7 cells
exhibited high levels of protein expressed from transfected cDNAs
as determined by anti-MEKK (A), anti-JNK1 (B),
and antiserum to PAC1 (C) that recognizes both MKP1 and MKP2
(Fig. 3). Furthermore, despite co-transfection with
multiple plasmids, levels of protein expression for each transfected
component were nearly identical. These results also demonstrate that,
under the conditions of this assay, the JNK1 antisera recognized only
transfected JNKs and did not recognize endogenous JNK1 or other JNKs
reported to be recognized by this antibody (12). Also, these data show
that variations in JNK1 activity and tyrosine phosphorylation were
dependent solely upon the presence or absence of co-transfected MKP1 or
MKP2.
JNK1 Inactivation by MKP1 and MKP2 Is Dose-dependent
Although both MKP1 and MKP2
demonstrate phosphatase activity toward JNK1, only MKP1 consistently
produced complete elimination of all detectable JNK1 phosphorylation
and activity. Therefore, we examined the dose dependence of both
phosphatases on JNK1 activity in COS-7 cells (Fig. 4).
In the presence of 1 µg of MEKK, co-transfection of only 5 µg of
MKP1 cDNA produced marked diminution of JNK1 activity as determined
by immune complex assay, whereas a similar amount of MKP2 cDNA
resulted in a reduction of only approximately 50% of measured JNK1
activity. Thirty µg of transfected MKP2 cDNA were required to
produce the equivalent reduction in JNK1 activity produced by 5 µg of
transfected MKP1 cDNA. This suggests that, in COS-7 cells
overexpressing both JNK and MKPs, MKP1 may be more selective for JNK1
than MKP2.
Transient Transfection of MKP1 or MKP2 Inhibits UV Light-stimulated JNK1 Activity in COS-7 Cells
To examine whether JNK1 was
inhibited by MKPs in the presence of stimuli other than MEKK, we
examined JNK activity following exposure to UV light. Treatment of
COS-7 cells with UV light (3-12 J/m2) produced JNK
activation with a time course consistent with that reported by other
investigators for different cell types (11). Prior to examining the
action of transfected MKPs on JNK activity following UV stimulation, we
determined whether activation of transfected JNK1 could be detected
above activated endogenous JNKs. Since the expression of transfected
MKPs should be limited to the cells that are co-transfected with JNK1,
the activation of endogenous JNKs by UV light will persist, even
following complete inhibition of transfected JNKs by co-transfected
MKPs. Total JNK1 activity as determined by immune complex assay was
more than 150% of that in cells transfected with vector alone (Fig.
5). Moreover, co-transfection of JNK1 cDNA and
subsequent overexpression of JNK1 did not alter the time course of
activation. Transfection of MKPs in the absence of co-transfected JNK1
did not result in decreased endogenous JNK activity following UV
stimulation,2 suggesting that MKP
expression was limited to a fraction of the cells. Co-transfection of
MKP1 or MKP2 assayed at both 30 and 60 min following UV light exposure
resulted in significant decreases in JNK activity to levels approaching
that seen in the absence of transfected JNK1. The addition of MKP1
reduced UV-stimulated JNK1 activity to those levels seen in COS-7 cells
transfected with vector alone, while MKP2 was less effective in
diminishing JNK1 activation, especially at 30 min following UV
treatment. Thus, both MKP1 and MKP2 function to inhibit JNK activation
in response to multiple upstream stimuli. Under both conditions, MKP1
exhibited increased activity toward JNK1, as compared with MKP2.
MKP1 and MKP2 Inhibit JNK-dependent Gene Expression
The physiological target of JNK1 is the transcription
factor c-Jun, which is transactivated following phosphorylation of
serine residues in the NH2-terminal domain (35). c-Jun is a
component of the AP-1 transcriptional complex and, when activated,
induces the transcription of numerous immediate early genes including c-fos and c-jun itself (36). Multiple JNKs have
been identified at present (28), showing characteristic binding and
phosphorylation of c-Jun substrates (29). To measure transactivation of
c-Jun, we used a mammalian two-hybrid system involving a chimeric
Gal4/c-Jun cDNA construct to direct expression of the 5xGal4-E1B
luciferase plasmid. In this assay, activation of luciferase activity is
dependent on the phosphorylation and activation of the Gal4/c-Jun
chimera (7, 35). Transfection of MEKK increased luciferase activity, confirming that MEKK can activate c-Jun in COS-7 cells. Furthermore, this activation is specific; constitutively active MEK-1 (37, 38) did
not stimulate luciferase activity in this assay.2
Co-transfection of either MKP1 or MKP2 cDNAs blocked this
activation in a dose-dependent manner (Fig.
6A), demonstrating that the inactivation of
JNKs by MKPs has physiological consequences within the cell. Interestingly, results of the luciferase assay differed significantly from kinase assays that measured JNK1 activity directly. In contrast to
their actions on transfected JNK1, MKP2 and MKP1 were equally proficient blocking c-Jun activation. Although this difference may
reflect uncoupling of JNK and c-Jun activation, it is also possible
that, unlike the kinase activity of transfected JNK1, c-Jun activation
was only partially dependent on transfected JNKs. We therefore examined
the contribution of endogenous JNKs on c-Jun activation by repeating
the previous experiment in the absence of co-transfected JNK1 (Fig.
6B). c-Jun activation by endogenous JNKs was readily
detectable, although total luciferase activity was decreased in the
absence of transfected JNK1. MKP1 and MKP2 inhibited endogenous JNK
activity in a dose-dependent manner. In these experiments,
MKP2 was a more potent inhibitor of c-Jun activation than was MKP1.
This is in contrast to the actions of MKP2 on the kinase activity of
transfected JNK1. Thus, despite the higher activity toward JNK1, MKP1
is a less effective inhibitor of endogenous JNKs, and the reverse
appears true for MKP2.
JNK2 Activity Is Inhibited by MKPs
JNK2 is a second abundant
isoform of the class of c-Jun N-terminal kinases, and like its
predecessor JNK1, it is activated following cellular stress by similar
dual specific kinases (1, 11). JNK2 has been demonstrated to have a
higher affinity for and activity toward the transcription factor c-Jun,
both in vitro and in vivo (12), suggesting that
it may be the more important of the two isoforms for physiological
regulation of c-Jun activity. To examine the regulation of JNK2 by
MKPs, COS-7 cells were transiently transfected with vector alone, JNK2
alone, or MEKK plus JNK2 along with increasing amounts of MKP1 or
MKP2. Results are shown in Fig. 7. In panel
A, kinase activity was dependent on transfected JNK2,
demonstrating that the JNK1 antisera used recognized JNK2, as reported
(12). The basal activity of transfected JNK2 was minimal but was
markedly increased upon co-transfection of
MEKK. The addition of
either MKP1 or MKP2 significantly blunted JNK2 activation by
MEKK,
and this inhibition, like that of JNK1, showed dose dependence. In
contrast to its inhibition of JNK1, MKP2 was more effective at
inhibiting the kinase activity of JNK2, especially at lower doses. In
panel B, whole cell lysates were immunoprecipitated with a
monoclonal M2 antibody directed at a flag epitope contained within the
JNK2 cDNA. This antibody allowed us to examine transfected JNK2
without the potential contribution of endogenous JNKs (12). When JNK2
was assayed using this antibody, both MKP1 and MKP2 showed similar
inhibition of this kinase activity. Thus, while MKP1 appears better
than MKP2 at inhibiting JNK1, both phosphatases are similar in their
inhibition of transfected JNK2.
NGF Blocks JNK Activation following UV Light
Both MKP1 and
MKP2 are transcriptionally induced following growth factor stimulation
in PC12 cells (9). To examine whether the induction of endogenous MKPs
is associated with a concomitant reduction in JNK activity, we measured
UV light-induced JNK activity in PC12 cells with and without prior NGF
stimulation. PC12 cells were serum-starved overnight and then treated
with UV radiation (3-12 J m2) alone or after
pretreatment with NGF 100 ng/ml 2 h prior to exposure. This time
course was chosen, since our laboratory has previously demonstrated
significant increases in MKP mRNA during the first 1-2 h following
NGF treatment (9). Protein levels of MKPs followed this increase as
well.2 As illustrated in Fig. 8, NGF
pretreatment significantly blunted UV light-induced activation of JNK
as compared with cells pretreated with serum-free medium alone. The
ability of induced MKPs to reduce JNK activity may be limited by the UV
treatment itself, which may induce a rapid fall in MKP protein levels
(22, 39). These results demonstrate that NGF pretreatment reduces the
maximal level of JNK activation following UV stimulation with kinetics that are consistent with the induction of MKPs.
MAP kinase phosphatases comprise a class of dual specificity phosphatases that are expressed in a wide variety of tissues and are transcriptionally induced in response to diverse extracellular signals. Some, like PAC1, show restricted tissue distribution, while others, such as MKP1 and MKP2, are broadly expressed, often within the same cells (9). Although a recently identified a novel MKP isoform, MKP3 (40, 41) or Pyst1 (26), has been shown to possess specificity toward ERKs alone, identification of definitive in vivo substrate selectivities between MKP1 or MKP2 has has been elusive. The only demonstrated divergence thus far has been the inhibition of the JNK-related stress-activated protein kinase p38 by MKP1 but not MKP2 in COS-7, NIH-3T3, and HeLa cells (6), although this finding appears to be cell type-specific, since MKP2 has been shown to be a potent inhibitor of p38 in PC12 cells (42). The expansion of the mammalian family of MAP kinases by the discovery of JNK1 and JNK2 and other JNKs (11, 28, 29) raised the possibility that various isoforms of these related kinases might be differentially regulated by MKP1 and MKP2. We subsequently examined the actions of MKP1 and MKP2 on the two well characterized isoforms of JNK, JNK1 and JNK2 (6).
COS-7 cells were chosen for study because of their ability to provide
high levels of expression of transfected kinase to produce a robust
signal that could be easily detected above background activity (43).
The validity of this method was first examined by confirming the known
activity of MKPs on ERKs using BxB-Raf as the upstream activator (9,
32, 44). Using similar techniques, we demonstrated that JNK can be
inactivated in vivo by both MKP1 and MKP2 using MEKK as
the stimulator of JNK activation (Fig. 2). Importantly, JNK activity
was dependent on transfected JNK; both basal and stimulated levels of
endogenous JNKs in untransfected cells were barely detectable compared
with that in transfected cells. Therefore, assayed JNK kinase activity
in cells co-transfected with MKPs is an accurate reflection of the
specific activity of MKPs toward transfected substrate. Overall, in
COS-7 cells, MKP1 exhibited similar activities for both isoforms of
JNK, while MKP2 activity appeared to favor JNK2. This does represent a
minor difference from other recent data suggesting that MKP1 is a more
potent inhibitor of JNK2, and it has been suggested that overexpression
of the products of transfected plasmids in COS-7 cells hinders
interpretation of substrate specificities (6). We believe, however,
that in our experiments, under conditions where protein expression is monitored, findings using this expression system can be quantitative. Tyrosine phosphorylation of JNK1 was also examined (Fig. 2), and inactivation of JNK1 was unsurprisingly found to be associated with
tyrosine dephosphorylation, which paralleled decreased JNK1 activity
tightly. Although not specifically assayed, it is likely that threonine
dephosphorylation followed similar kinetics (45-48).
Short-wave UV light and -irradiation are reliable stimulators of
stress activated pathways and can activate JNKs in multiple cell types
(8, 11). In Fig. 5, UV irradiation of COS-7 cells produced a
predictable rise in endogenous JNK activity with a time course
consistent with that previously reported for other cell types (8) and
was not altered by transfection of JNK1. Importantly, the augmentation
of JNK activity by the expression of transfected JNK1 permitted the
specific evaluation of the effects of transfected MKPs on JNK1
activity, despite the co-activation of endogenous JNKs. As in the
experiments using
MEKK, MKP1 was more effective than MKP2 in
inhibiting JNK1 following its activation. Although MKP1 and MKP2
demonstrate varying activities toward JNKs when overexpressed in COS-7
cells, we cannot rule out the possibility that endogenous MKPs may have
different selectivities toward JNKs in other cells or when expressed to
lesser degrees.
To examine whether MKP activity was reflected in JNK-dependent gene transcription, we used a mammalian two-hybrid system to measure c-Jun activation (Fig. 6). Contrary to observed effects on kinase activity, here MKP2 was highly efficient at reducing activation of c-Jun both in the presence and absence of co-transfected JNK1. Because c-Jun-coupled luciferase activity represents the effects of JNK (and c-Jun) activity integrated over an extended period of time, it is likely that the apparent discrepancies between kinase and luciferase assays reflect the activity of endogenous JNKs, for which MKP2 may be a more specific inhibitor. Furthermore, although transfection of JNK1 may provide a convenient substrate for the actions of MKPs, the coupling of transfected JNKs to the nuclear target c-Jun in COS-7 cells had not been previously examined (6). Although both JNK1 and JNK2 phosphorylate the NH2-terminal activation domain of c-Jun, activation of c-Jun by JNK2 is as much as 10-fold greater than that of JNK1, correlating with the increased binding of c-Jun to JNK2 (12, 29). To determine whether selective inhibition of JNK2 by MKP2 contributed to the results of the luciferase assays, we examined the actions of each phosphatase on JNK2 kinase activity (Fig. 7), revealing that this isoform was a suitable substrate for either MKP1 or MKP2. We suggest that the measurement of JNK kinase activities may only partially signify MKP's physiologic actions, which may be more appropriately assessed by measurement of the activation of c-Jun.
MKPs are immediate early genes that are rapidly induced by serum (3) and by NGF (9); however, the physiological role of newly synthesized MKPs has not been clearly established. In PC12 cells, NGF produces significant induction of MKP1 mRNA but does not result in diminished ERK activity, suggesting other possible substrates (24). Although JNKs also can be activated via Ras-dependent pathways (35), NGF is a poor activator of JNKs.2,3 To explore whether NGF can reduce JNK activity with kinetics consistent with induction of MKP protein, we examined JNK activity in PC12 cells exposed to UV irradiation. Pretreatment with NGF blunted JNK activation by UV light, with a maximal effect of NGF seen when pretreatment was timed 90-120 min prior to the onset of UV irradiation,2 consistent with a requirement for new protein synthesis. Other studies have also demonstrated potentiation of UV-induced increase in JNK activity after treatment with cyclohexamide, again suggesting that inactivation of JNKs requires new protein synthesis (8, 23). Since MKPs are labile proteins (3), the maximal effect of newly synthesized MKPs on UV-stimulated JNK activity may be blunted by the inhibition of MKP synthesis following UV treatment (39). Despite this limitation, induction of MKPs by NGF correlated with a significant reduction in JNK stimulation, while ERK activity has been shown to remain unaffected during the same time course (9, 24).
The physiological significance of MKP inactivation of JNKs following NGF treatment may be in the regulation of programmed cell death, or apoptosis (49). JNK activation is required for apoptosis in PC12 cells triggered by NGF withdrawal or MEKK activation (50). Recent studies have proposed that the choice between neurotrophic and apoptotic responses in PC12 cells and other cells is controlled by a balance of activity between ERKs and JNKs (50, 51). Since transcription of MKPs is partially dependent on ERK activity (46),3 MKP inhibition of JNKs may provide a mechanism coupling ERK activation to JNK inactivation following NGF stimulation of PC12 cells.
We thank Caroline Rim for providing invaluable technical support, Dr. Anita Misra-Press for many helpful suggestions and insights, Drs. John Denu, Mark Vossler, and Anne Hirsch for critical reading of the manuscript, and Sheri Medford for excellent secretarial assistance.