| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Cell and Molecular Biology (E.T.M., C.A.P., J.C.,
M.H.-D.) Northwestern University Medical School Chicago,
Illinois 60611
Department of Biochemistry, Molecular Biology
and Cell Biology (A.M., K.E.M.) Northwestern University
Evanston, Illinois 60208 Intramural Research Support Program
(G.S.) SAIC-Frederick National Cancer Institute
Frederick Cancer Research and Development Center Frederick,
Maryland 21702-1201
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Activation of both p42/p44 ERK/MAPK and p38 MAPK has been documented
during ovarian response to hormonal stimulation. ERK/MAPK activation is
elicited by treatment of granulosa cells with epidermal growth factor
(14), FSH (15, 16, 17), or LH (16), and in luteal cells by treatment with
PGF2
(18). The consequent phosphorylation/activation of the
downstream ERK/MAPK target protein kinase p90rsk
was demonstrated in FSH-treated immature rat granulosa cells (15) (J.
Cottom, Y. Park, E. T. Maizels, L. Salvador, J. C. R. Jones,
R. V. Schillace, D. W. Carr, P. Cheung, C. D. Allis, J. L. Jameson, and
M. Hunzicker-Dunn, submitted). Additionally, FSH treatment of
immature granulosa cells elicits p38 MAPK activation (19, 20), with
resultant phosphorylation of the MAPKAPK-2/-3 substrate HSP-27 and
modulation of cell shape (19), and UV light, a stress stimulus,
activates p38 MAPK in bovine luteal cells (21).
In the rat, activation of preovulatory follicular granulosa cell LH
receptors by the proestrus LH surge, or by pharmacological treatment
with the LH receptor agonist, human CG (hCG), causes ovulation and
differentiation of the ovulated follicle into a corpus luteum.
LH/hCG-induced ovulation and luteal formation are accompanied by a
pattern of distinct biochemical changes (reviewed in Ref. 22),
including, for example, down-regulation followed by later reappearance
of the LH receptor and aromatase, the transient induction of the
progesterone receptor, the induction of PGHsynthase-2, and the
down-regulation of inhibin expression. Protein phosphorylation
is recognized to be a key ovarian response to hormonal stimulation.
LH/hCG-induced cAMP formation and consequent protein kinase A
(PKA)-mediated phosphorylation of target proteins are required for both
ovulation and luteal formation. However, once luteal formation is
initiated, the mechanism for the LH/hCG-stimulated cAMP production
becomes desensitized (23), the requirement for continued cAMP-dependent
signaling is lost, and a state of cAMP nonresponsiveness is thought to
characterize the maturing luteal cell (24, 25). Nevertheless,
phosphorylation of luteal proteins, notably CREB, remains evident (25);
therefore, kinases other than PKA would be expected to control
phosphorylation of target proteins during luteal maturation. Based on
our evidence that the p38 MAPK pathway is required for immature
granulosa cell response to FSH (19) and that this pathway is known to
be regulated by a large number of input signals from many G
protein-coupled receptors as well as growth factor receptors in other
cellular models (1), we postulated that the p38 MAPK path would serve
to phosphorylate essential targets during the cAMP-nonresponsive luteal
maturation phase. We undertook to examine components of the p38 MAPK
pathway for evidence of regulation during hCG-induced luteal maturation
with special interest in MAPKAPKs, as those kinases could regulate the
phosphorylation of CREB (11, 12, 13, 26). We found that expression of
MAPKAPKs was inversely regulated. While MAPKAPK-2 expression diminished
with luteal development, MAPKAPK-3 was newly induced during this
developmental transition. Moreover, MAPKAPK-3 induction was accompanied
by upstream kinase activation, and by downstream substrate
phosphorylation; thus, MAPKAPK-3 is uniquely poised to subserve the
role of critical kinase during luteal maturation.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
p38 MAPK, analogous with other MAPK family members (1), is activated by dual phosphorylation on threonine and tyrosine within the TXY motif in the activation loop (28). For p38 MAPK, these phosphorylations are catalyzed by upstream dual-specificity kinases, the MAPK-kinases MKK6 and MKK3 (2, 3). In turn, these MKKs are activated by phosphorylation on homologous serine and threonine residues (29). We tracked the phosphorylation states of these phosphor-ylation-dependent kinases as a measure of their activation states, by means of immunoblotting with phospho-specific antibodies.
First, immunoblots were performed on ovarian lysates prepared
from rats at various times post PMSG and hCG injections to detect
activation of MKK6 and MKK3 by using a phospho-specific antibody that
recognizes activation-specific phosphorylation sites (29),
phosphoserine 207 of MKK6 and the corresponding phosphoserine 189 of
MKK3, respectively (Fig. 1
, upper
panel). Protein levels were determined by immunoblotting with
control antibodies specific for MKK6 and MKK3 (Fig. 1, middle and
lower panels, respectively). Both the 35-kDa MKK3 and the 37-kDa
MKK6 showed a small degree of activation by 1 h in response to
both PMSG (Fig. 1, top panel, lane 2) and hCG (lane 6) in
follicular and periovulatory phases. However, while MKK3 failed to show
sustained activation as luteal maturation progressed (Fig. 1, lanes
9–12), strong sustained activation of the 37-kDa MKK6 accompanied
luteal maturation (Fig. 1, top panel, lanes 8–12).
|
, upper panel).
Protein levels were determined by immunoblotting with control antibody
specific for p38 MAPK (Fig. 2A, lower panel). p38 MAPK
showed a small degree of activation in response to PMSG in the
follicular phase (Fig. 2A, top panel, lane 2). p38 MAPK
underwent strong biphasic activation in response to hCG, with initial
acute activation 1 h post hCG (Fig. 2A, lane 6), and a second
sustained chronic activation phase accompanying luteal maturation (Fig. 2A, lanes 8–11).
|
, upper
panel), as well as protein expression of ERK/MAPKs (control
antibody, Fig. 2B, lower panel), were performed on ovarian
lysates prepared at various times post PMSG and post hCG injections.
ERK/MAPK displayed slight activation post PMSG in the follicular phase
(Fig. 2B, lane 3). Strong acute ERK/MAPK activation was detected at
1 h post hCG in a manner similar to p38 MAPK; however, ERK/MAPKs
failed to undergo a second sustained activation phase as luteal
maturation progressed. The lack of a second sustained luteal ERK/MAPK
activation phase suggests that, unlike p38 MAPK, the ERK/MAPKs are not
poised to participate in phosphorylation events during luteal
maturation.
Inverse Modulation of MAPKAPK-2 and MAPKAPK-3 Expression
Accompanying Hormone-Induced Follicular Maturation and
Luteinization
We proceeded to investigate the ovarian pattern of expression of
the two related p38 MAPK target protein kinases, MAPKAPK-2 and
MAPKAPK-3, in vivo during hormone-induced follicle
maturation, ovulation, and luteinization. We studied alterations in
ovarian expression of each MAPKAPK at various time points post PMSG
injection or post hCG injection.
First, immunoblots to detect protein expression of both MAPKAPKs were
performed on ovarian lysates prepared at various times post PMSG and
post hCG injections. As seen in Fig. 3A, MAPKAPK-2 protein, detected as two isoforms at 47 and 54 kDa (30, 31),
was abundant at early time points throughout follicular maturation
(lanes 1–8), and then decreased markedly at later time points post
ovulation, as luteal maturation progressed (lanes 9–12). In contrast
to the decreased expression of MAPKAPK-2 noted above, MAPKAPK-3 protein
expression, detected at 42 kDa, was minimal at early time points during
follicular development (Fig. 3B, lanes 1–7), and was strongly induced
as luteal maturation progressed, increasing from 33 h post hCG
onward through the end of the observation period. (Fig. 3B, lanes
8–11).
|
, top panel), detected in
the earliest time points post hCG (weakly at 0 and more strongly at 1
and 8 h post hCG), diminished in the luteal phase slices (24 h and
33 h post hCG), indicating that loss of MAPKAPK-2 mRNA expression
preceded the loss of MAPKAPK-2 protein expression observed during the
luteal maturation phase (Fig. 3A). In contrast, MAPKAPK-3 mRNA signal
(Fig. 4A, middle panel; Fig. 4B), was minimal in the
earliest time points post hCG (Fig. 4A, 0 and 1 h; also Fig. 4B, 4 h), appeared in periovulatory follicles (8 h post hCG), and was
sustained in forming corpora lutea (24 and 33 h post hCG),
indicating that induction of MAPKAPK-3 mRNA paralleled the subsequent
induction of MAPKAPK-3 protein expression observed during luteal
formation and maturation (Fig. 3B). Hybridization with a cRNA probe
specific for inhibin was performed on the same slices to serve as a
control to verify ovarian response to hCG, since expression of inhibin
, prominent in preovulatory follicles, is recognized to decrease
during gonadotropin-induced ovulation and luteinization (32). As
expected, inhibin mRNA expression was strongly evident in
follicular structures in 0 and 1 h post hCG slices and diminished
thereafter (Fig. 4A, bottom panel).
|
The small heat shock protein HSP-27 from rodent sources can be
phosphorylated by MAPKAPKs on two potential phosphorylation sites
corresponding to serine 15 and serine 86 (34). Phosphorylation on one
or both of these sites results in the appearance of HSP-27
phosphoisoforms displaying distinct migration positions on
two-dimensional isoelectric focusing (IEF)/SDS PAGE gels (19, 35). Two-dimensional immunoblots, to detect phosphoisoforms of HSP-27
as a measure of in vivo HSP-27 phosphorylation state, were
performed on ovarian lysates prepared at various times post PMSG and
post hCG injections. As seen in Fig. 5
(top), all two-dimensional blots displayed spots
corresponding to basic unphosphorylated HSP-27 isoforms (designated by
arrowhead a), more acidic monophosphorylated
HSP-27 phosphoisoforms (designated by arrowheads
b and b'), and diphosphorylated HSP 27
phosphoisoforms (designated by arrowhead c), with no
detectable alteration in HSP-27 phosphoisoform content. We calculated
relative HSP-27 phosphorylation levels from observed densities of spots
corresponding to mono- or diphosphorylated HSP-27 phosphoisoforms (Fig. 5, bottom). No modulation of HSP-27 phosphorylation levels
was observed in ovarian lysates obtained during either follicular or
luteal phases of development. Thus, the phosphorylation state of HSP-27
did not correlate with the activation state of the p38 MAPK/MAPKAPK-3
axis during the luteal phase; however, HSP-27 phosphorylation could
be maintained at constant levels by alternative mechanisms,
e.g. through phosphorylation by other recognized HSP-27
kinases (35), or alternatively through the action of HSP-27
phosphatases (36, 37). We have recently observed that the 
isoform
of protein kinase C (PKC-), an efficient HSP-27 kinase (35), is
detected in ovarian extracts in a constitutively activated state during
luteinization (L. M. Salvador, E. Maizels, E. Miyamoto, H.
Yamamoto, and M. Hunzicker-Dunn, in preparation) and would
account for the constant phosphorylation of HSP-27 during this
transition.
|
We evaluated the phosphorylation state of CREB in vivo
during hormone-induced follicle maturation, ovulation, and
luteinization. Immunoblots were performed on ovarian lysates prepared
at various times post PMSG and post hCG injections to detect
phosphorylation of CREB using a phospho-specific antibody that
recognizes CREB phosphorylated on serine 133 (Fig. 6, upper panel). Protein
levels were determined by immunoblotting with control antibody specific
for CREB (Fig. 6, lower panel). Increased phosphorylation of
CREB on serine 133 was detected in response to PMSG treatment (Fig. 6, lanes 2 and 3). Additionally, CREB underwent biphasic phosphorylation
in response to hCG. Phosphorylation of CREB increased acutely at 1
h post hCG (Fig. 6, lane 6), decreased to a nadir at 8 h post hCG
(Fig. 6, lane 7), and then began to rise again by 48 h post hCG
(Fig. 6, lane 9). Densitometric values normalized for protein content,
shown graphically in the lower section of Fig. 6, indicated
that CREB phosphorylation levels had risen to approximately one third
the maximal value by 48 h post hCG, a level compatible with
the extent of induction of MAPKAPK-3 that has occurred at this
time. Strong CREB phosphorylation was maintained as luteal maturation
progressed (Fig. 6, lanes 10–12). The sustained CREB phosphorylation
observed during later time points of luteal maturation is consistent
with CREB functioning as an in vivo phosphorylation target
for the newly induced MAPKAPK-3.
|
-32P]ATP, and the resulting
32P incorporation into CREB was demonstrated by
SDS-PAGE and autoradiography. Assays were performed on MAPKAPK-3 immune
complexes collected from ovarian extracts obtained at various times
post hCG injection in vivo (Fig. 7A). Results showed that there was
minimal [32P]phosphate incorporation into CREB
by immune complexes prepared from early time points (0 and 1 day post
hCG) consistent with the low level of MAPKAPK-3 protein expression at
these times. In contrast, immune complexes collected at later time
points during the luteal maturation phase (2 and 4 days post hCG)
catalyzed increased [32P]phosphate
incorporation into CREB, indicating sustained activation of MAPKAPK-3
accompanying luteal maturation. Control immune complexes, collected
either in the absence of MAPKAPK-3 antibody or in the absence of
ovarian extract, showed no detectable phosphorylation of CREB (Fig. 7A, right panel). These assay results provide direct evidence
that newly induced MAPKAPK-3 undergoes sustained activation in
vivo during luteal maturation, consistent with sustained
activation of upstream kinases MKK6 and p38 MAPK, and the sustained
phosphorylation of CREB. Additionally, these immune complex kinase
assays confirmed previous reports indicating that MAPKAPK-3 can
directly phosphorylate CREB in vitro (12, 26).
|
, left panel), and these cells
have previously been reported to show persistent CREB phosphorylation
in a cAMP-nonresponsive manner (25). Luteinized granulosa cells were
treated with or without p38 MAPK inhibitor SB 203580 for 4 h and
then harvested and analyzed for CREB phosphorylation by immunoblotting.
As seen in Fig. 7B, right panel, CREB serine 133
phosphorylation was prominent in control cells and was markedly reduced
by treatment with SB 203580. These data are consistent with CREB
phosphorylation being a p38 MAPK pathway-dependent event in this
cellular model of luteinization.
A recent immunofluorescence study of cultured luteinized granulosa
cells found phosphorylated CREB largely in the soluble compartment, a
compartment in which transcriptional function of phosphorylated CREB
would be abrogated, rather than the expected nuclear compartment in
which phosphorylated CREB functions as a transcription factor (25).
Based on the results of that study, we wished to identify the
subcellular localization of phosphorylated CREB in our in
vivo luteinization model. Ovarian extracts obtained at various
times post hCG injection in vivo were separated by
centrifugation into soluble fractions and nuclear enriched-particulate
fractions, respectively. We additionally prepared soluble and nuclear
enriched-particulate fractions from luteinized granulosa cells to serve
as controls. Immunoblots to visualize phosphorylated CREB were
performed on both soluble and nuclear-enriched particulate fractions
(Fig. 7C). Results showed that luteinized granulosa cells displayed
phosphorylated CREB primarily in the soluble fraction (Fig. 7C, lane
7), confirming the report of Gonzales-Robayna et al. (25).
Notably in contrast, phosphorylated CREB was localized to the
nuclear-enriched particulate fraction at all time points examined using
the in vivo luteinization protocol (Fig. 7C, lanes 2, 4, and
6). Thus, phosphorylated CREB was retained in a functional compartment
in vivo during luteal maturation.
In light of the temporal correlation of upstream kinase
activation, MAPKAPK-3 expression and activation, and sustained CREB
serine 133 phosphorylation during luteal maturation, we wished to
directly evaluate the ability of MAPKAPK-3 to modulate CREB
transcriptional activity in ovarian cells. We employed the sensitive
GAL4-linked CREB transcriptional activity assay system (40, 41). This
assay system measures GAL4-driven transcription activated in response
to phosphorylation of the activating serine residue (corresponding to
serine 133) of the CREB moiety within a fusion protein comprised of
full-length CREB linked to the DNA-binding region of the yeast
transcription factor GAL4. GRMO2 cells, representing a stable mature
granulosa cell line (42), were cotransfected with MAPKAPK-3 cDNAs
together with CREB-GAL4 fusion protein expression vector and GAL4
binding site-luciferase reporter gene construct (Fig. 8). Cotransfection with 50 ng wild-type
MAPKAPK-3 cDNA increased CREB-GAL4-mediated reporter activity, to
levels approximately 3-fold over basal activity seen with empty vector,
a modest but significant increase (P < 0.05, Fig. 8).
Notably, the modest reporter activity achieved by MAPKAPK-3 in the
absence of exogenous p38 MAPK activators likely represents partial
activation of wild-type MAPKAPK-3 supported by inclusion of serum in
GRMO2 culture medium, as serum serves as only a mild activator for
MAPKAPK-3 (8). The strong p38 MAPK activators anisomycin and arsenite
were tested as exogenous agents that might further activate MAPKAPK-3;
however, these treatments were toxic to GRMO2 cells at the
concentrations required to activate p38 MAPK and thus could not be
further evaluated. CREB-GAL4-mediated reporter activity in the presence
of wild-type MAPKAPK-3 corresponded to approximately 60% of that
achieved through activation of PKA in the presence of forskolin, and
approximately 20% of that achieved through transfection of PKA
catalytic subunit cDNA (not shown). Importantly, in contrast to
wild-type MAPKAPK-3 cDNA, 50 ng of the kinase-dead K>M MAPKAPK-3
mutant cDNA (8) failed to support increased CREB-GAL4-mediated reporter
activity, indicating that the kinase activity of MAPKAPK-3 is necessary
for the positive effect of MAPKAPK-3 on CREB transcriptional activity.
Thus MAPKAPK-3 can signal to enhance CREB transcriptional activity in
ovarian cells in a kinase-dependent manner, i.e. through
phosphorylation. Based on the results of these cotransfection
experiments, MAPKAPK-3 induced during luteal maturation in
vivo would be capable of signaling to enhance CREB transcriptional
activity through phosphorylation.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
MAPKAPK-2 (44) and -3 (6, 7) are closely related kinases with 75%
amino acid sequence homology (7), and shared structural features that
include corresponding N-terminal SH3 binding domains, C-terminal
nuclear localization sequences, C-terminal -helices, and analogous
activating phosphorylation sites (7). The mechanism of activation has
been defined for MAPKAPK-2: phosphorylation of activating sites by p38
MAPK (45) induces a conformational change that moves the inhibitory
C-terminal helix-regulatory region away from the kinase catalytic
domain, resulting in activation (46). Conservation of the involved
structures in MAPKAPK-3 (7) suggests that the mechanism of activation
would be comparable. MAPKAPK-2 and -3 share common upstream activators
and overlapping (6, 8, 12), although not identical (7), substrate
specificities. Indeed, MAPKAPK-2 and MAPKAPK-3 have often been treated
as interchangeable (47, 48). However, the striking inverse pattern of
expression of MAPKAPKs during follicle maturation, ovulation, and
subsequent luteal development in the rat suggests that these enzymes
may, in fact, play unique rather than interchangeable roles in the
ovary.
Upstream components of the p38 MAPK cascade displayed discrete patterns of activation in response to gonadotropin treatments in vivo. p38 MAPK underwent initial activation in response to both PMSG and hCG, as well as a sustained chronic activation phase accompanying luteal maturation. Initial responses to PMSG and hCG were accompanied by activation of upstream MKKs MKK3 and MKK6, while the luteal phase response was accompanied by activation of MKK6 exclusively. The initial modest follicular phase activation of p38 MAPK in response to PMSG is consistent with the previous finding that FSH could stimulate p38 MAPK activation in immature rat granulosa cells (19, 20). The initial acute activation of p38 MAPK in response to hCG activation is interesting and is consistent with the possibility that p38 MAPK-dependent signaling may participate in the mediation or modulation of hCG-stimulated events critical to ovulation, such as the induction of PG synthase-2 or progesterone receptor (43, 49, 50). The sustained chronic activation of p38 MAPK during luteal maturation would provide for the observed sustained activation of the newly induced luteal kinase MAPKAPK-3, as monitored by immune complex kinase assays.
What stimuli support extended p38 MAPK activation and thus MAPKAPK-3 activation during luteal maturation in the ovary in vivo? While the gonadotropins FSH and hCG indeed elicit acute activation of p38 MAPK in immature and preovulatory granulosa cells, respectively (Refs. 19, 20 ; L. M. Salvador, E. Maizels, E. Miyamoto, H. Yamamoto, and M. Hunzicker-Dunn, in preparation), and treatment with hCG elicited p38 MAPK activation acutely in the periovulatory period in the current study, hCG would be unlikely to be responsible for prolonged luteal phase activation, as LH receptors would have undergone both desensitization and down-regulation in response to the ovulatory stimulus (23). In addition to gonadotropins, a number of cytokines and growth factors are recognized to participate in control of ovarian function (51, 52). Interleukin-1ß (IL-1ß), a cytokine known to elicit p38 MAPK activation in target cells (53), is transiently induced in the periovulatory period (54, 55, 56) and may participate in the periovulatory activation of p38 MAPK. However, IL-1ß levels have fallen by 48 h post hCG (54), and thus this cytokine would be unlikely to represent the prolonged luteal-phase p38 MAPK stimulus. The growth factors insulin-like growth factor-1 (IGF-I) and fibroblast growth factor (FGF) can each elicit p38 MAPK/MAPKAPK activation leading to CREB phosphorylation in other experimental systems (11, 13), suggesting that these growth factors would be capable of serving as initiating stimuli during luteal maturation. IGF-I is well recognized to synergize with gonadotropins to elicit granulosa cell function and differentiation (52, 57). Moreover, IGF-I receptor levels (58) and basic FGF levels (59) are elevated during early luteal development in the rat. Finally, the proangiogenic growth factor vascular endothelial growth factor (VEGF), a critical factor in corpus luteum formation and growth (60, 61, 62, 63), is appreciated to activate p38 MAPK in target cells (64). Thus, the growth factors IGF-I, FGF, and VEGF comprise a set of likely candidates for luteal-phase stimulus to support the prolonged activation of p38 MAPK and MAPKAPK-3. Further experiments will focus on delineating the contribution of each of these critical growth factors.
CREB was evaluated as a potential ovarian target for newly induced and activated MAPKAPK-3 during luteal maturation. The nuclear transcription factor CREB requires phosphorylation on serine 133 to bind the coactivator CREB-binding protein (CBP) and recruit transcriptional machinery (65). CREB was initially described as a PKA substrate (66); however, CREB is now recognized to serve as a phosphorylation target for a number of distinct kinases in several diverse signaling pathways. Several groups have demonstrated CREB serine 133 phosphorylation catalyzed by kinases downstream of activated p38 MAPK (11, 67, 68, 69). p38 MAPK-dependent CREB phosphorylation was catalyzed by MAPKAPK-2 or a closely related kinase such as MAPKAPK-3 in FGF-treated fibroblasts (11), and by MAPKAPK-3 in IGF-I-treated PC12 cells (13). Moreover, MAPKAPK-3 readily phosphorylates CREB in vitro (12, 26), a finding we have confirmed in the current study through immune complex kinase assays.
CREB underwent clear phosphorylation coinciding with the induction of and activation of MAPKAPK-3. We observed three peaks of CREB phosphorylation in vivo, the first during hormone-induced follicular maturation, the second acute peak in response to hCG in the periovulatory follicle, and the third chronic peak during luteal maturation. The first and second peaks of CREB phosphorylation correspond to previously described cAMP-responsive PKA-mediated events (25, 70, 71). In contrast, the third luteal-phase CREB phosphorylation represents what is thought to be a cAMP-nonresponsive event (25); therefore, this phase of phosphorylation is expected to be catalyzed by a CREB kinase other than PKA. Notably, we detected substantial sustained phosphorylation of CREB accompanying the increased expression and activation of MAPKAPK-3 as well as the activation of upstream kinases MKK6/p38 MAPK as luteal maturation progressed.
To further delineate the role of the p38 MAPK pathway in CREB
phosphorylation during luteinization, we performed additional studies
with the p38 MAPK inhibitor SB 203580 (38). We found that SB 203580
strongly inhibited CREB phosphorylation in luteinized granulosa cells.
These studies are consistent with a requirement for p38 MAPK- mediated
signaling events in CREB phosphorylation in this luteal model, in good
agreement with in vivo luteal maturation phase profiles.
However, it has been suggested recently that p38 inhibitor SB 203580
can impact other signaling pathways in addition to p38 MAPK.
Specifically, SB 203580 can impact signaling through ERK (72) and
phosphatidylinositol-3 kinase/Akt pathways (73, 74) either through
direct interaction with other kinases (72, 73) or through inhibition of
cross-talk between p38 MAPK and other kinases (74). We can detect both
ERK and Akt activation in the follicular and periovulatory periods
in vivo (Fig. 2B and E. T. Maizels, L. M. Salvador, J. E.
Cottom, and M. Hunzicker-Dunn, manuscript in preparation).
However, we detect neither ERK nor Akt activation during the luteal
maturation phase (48 h to 6 days post hCG); thus neither of these
kinase pathway can be implicated in CREB phosphorylation during the
luteal maturation phase in vivo (Fig. 2B and E. T. Maizels,
L. M. Salvador, J. E. Cottom, and M. Hunzicker-Dunn, manuscript in
preparation). In light of the lack of participation of either ERK or
Akt in luteal-phase downstream signaling events in vivo, the
role of the p38 MAPK path in luteal-phase CREB phosphorylation is
unambiguous.
We investigated phosphorylation of an additional MAPKAPK substrate, the
small heat shock protein HSP-27. In contrast to CREB, we found no
modulation in vivo of the phosphorylation state of HSP-27, a
finding that would be explained by the constitutive activation of an
alternate HSP-27 kinase, PKC- (35), accompanying ovulation and
luteinization (Salvador et al., in preparation).
Findings implicating newly induced MAPKAPK-3 as CREB kinase in vivo and in vitro were reinforced by transfection studies, which yielded direct evidence that MAPKAPK-3 could signal to activate CREB transcriptional activity in a phosphorylation-dependent manner in ovarian cells.
Previous reports indicate that luteal CREB expression and/or function can vary depending on the choice of experimental model of luteinization employed. For example CREB expression is maintained in luteinizing granulosa cells in the rat (25, 70), although CREB protein expression is completely lost upon luteinization in the primate (75), indicating that there is species specificity in luteal CREB expression. Additionally, a recent study detected phosphorylated CREB primarily localized to the soluble compartment in luteinized granulosa cells in primary cell culture (25), a compartment in which phosphorylated CREB would be unable to fulfill its function as a transcription factor. We therefore evaluated the subcellular localization of phosphorylated CREB in our in vivo luteinization model. We were able to confirm that phosphorylated CREB is indeed a soluble protein in cultured luteinized granulosa cells. In contrast, phosphorylated CREB was primarily localized in the nuclear-enriched particulate fraction in the in vivo ovarian samples at all time points, including the 4-day post-hCG time point coinciding with MAPKAPK-3 expression as well as upstream p38 MAPK/MKK6 activation. Thus, phosphorylated CREB is retained in a subcellular compartment compatible with its function as a transcription factor during luteal maturation in vivo.
CREB phosphorylation has been previously implicated in the transcriptional regulation of several important ovarian target genes (43), making it an interesting potential target for MAPKAPKs in the ovary. It is well established that CREB participates in transcriptional activation of the CYP19 aromatase gene (76, 77). Aromatase catalyzes the conversion of androgen precursor to estrogen. Aromatase, initially induced by FSH (or PMSG) in granulosa cells of maturing follicles, is lost in response to the ovulatory LH surge (78, 79) but reappears in the corpus luteum by early- to mid-pregnancy to allow production of estrogen by the maturing corpus luteum (80, 81). Notably, in an analogous in vivo hCG-stimulated rat luteinization model, aromatase mRNA was induced by 3-day post hCG (82), correlating well with the peak of phosphorylated CREB observed during luteal maturation in our study. The strong temporal correlation indicates aromatase as a potential transcriptional target for MAPKAPK-3-catalyzed CREB phosphorylation accompanying luteal maturation.
In summary, we have described the developmental pattern of regulation of components of the p38 MAPK cascade as ovarian follicles undergo PMSG-induced maturation followed by hCG-induced ovulation, luteal formation, and luteal maturation. The closely related kinases MAPKAPK-2 and MAPKAPK-3 underwent inverse changes in expression level, with loss of MAPKAPK-2 mRNA and protein expression, and induction of MAPKAPK-3 mRNA and protein expression accompanying these developmental transitions. During the luteal maturation phase, MAPKAPK-3 induction was accompanied by sustained activation of upstream activating kinases p38 MAPK and MKK6, and by sustained phosphorylation of its substrate CREB. MAPKAPK-3, activated during luteal maturation in vivo, readily catalyzed CREB phosphorylation in immune complex kinase assays, and phosphorylation of CREB was shown to depend on an intact p38 MAPK signaling pathway in a cellular model of luteinization. Wild-type, but not kinase-dead, MAPKAPK-3 enhanced CREB transcriptional activity in cotransfection studies, demonstrating directly MAPKAPK-3’s ability to signal to activate CREB. Thus MAPKAPK-3 is indeed uniquely poised to support luteal maturation through the phosphorylation and activation of the nuclear transcription factor CREB. Further studies will attempt to uncover the stimulus of MKK6/p38 MAPK activation, as well as define transcriptional targets of MAPKAPK-3/CREB during luteal maturation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Animals
Sprague Dawley rats (Charles River Laboratories, Inc. Portage, MI) were housed at Northwestern University animal
care facilities, maintained in accordance with the "Guidelines for
the Care and Use of Laboratory Animals" by protocols approved by the
Northwestern University ACUC committee. With the exception of
luteinized granulosa cell culture experiments performed as described
below, immature female rats (26–27 days old) were injected
subcutaneously with 25 IU PMSG. Indicated animals were further injected
subcutaneously with 25 IU of hCG 48 h following PMSG injection.
Ovaries were harvested at the indicated times post PMSG injection or
post hCG injection and either immediately frozen at -70 C for
subsequent in situ hybridization analysis or subjected to
tissue extract preparations as described below.
Tissue Extract Preparation
Whole ovarian extracts were prepared by homogenization in lysis
buffer (15) containing 10 mM potassium phosphate, pH 7.0, 1
mM EDTA, 5 mM EGTA, 10 mM
MgCl2, 50 mM ß-glycerophosphate, 1
mM Na orthovanadate, 2 mM dithiothreitol, 1
mM phenylmethylsulfonyl fluoride, 0.5% NP-40, and 0.1%
sodium deoxycholate. A clarified lysate, containing both soluble
proteins and detergent-solubilized membrane proteins, was obtained by
centrifuging the homogenate at 12,000 x g for 10 min
at 4 C. Alternatively, subcellular fractionation was performed by
homogenization of ovaries in protease- and
phosphatase-inhibitor-enriched homogenization buffer (PPI buffer) as
described previously (85), followed by centrifugation at 105,000
x g for 70 min. The separated soluble and nuclear-enriched
particulate fractions were prepared for SDS-PAGE by suspension in equal
volumes of SDS-containing sample buffer followed by heat denaturation
(100 C, 5 min).
Protein concentrations were measured by the method of Lowry et al. (86) using crystalline BSA as a standard.
Luteinized Granulosa Cell Culture
Primary culture of luteinized granulosa cells was performed as
previously described (85). Briefly, immature rats were injected sc with
0.15 IU hCG for 2 days. On the third day, rats were injected sc with 10
IU hCG, and ovaries were removed 7 h post injection. Granulosa
cells from large preovulatory follicles were cultured for 9 days in the
presence of 1% FBS as described (85). For inhibitor studies, 9-day
cultured cells were removed from serum for 14 h, and then
subjected to 4-h treatments with vehicle or 10 µM SB
203580, and cell lysates were prepared in the presence of lysis buffer
as described in Tissue Extract Preparation. Alternatively,
9-day cultured cells were harvested and homogenized in the presence of
PPI buffer and subjected to subcellular fractionation as described
above.
Protein Separation
Separation of ovarian lysate proteins was by SDS-PAGE using 10%
or 12% separating gels (87). For two-dimensional gel electrophoresis,
ovarian lysate proteins were separated by isoelectric focusing using
mixed ampholines (4 parts pH range of 5–8 with 1 part pH range 3–10),
and then by SDS-PAGE (87). For immunoblots, proteins were
electrophoretically transferred to Hybond Nitrocellulose C-extra,
incubated with primary antibody overnight at 4 C, and detected by
enhanced chemiluminescence (Amersham Pharmacia Biotech).
Densitometry was analyzed with Molecular Analyst software
(Bio-Rad Laboratories, Inc.).
In Situ Hybridization
Twenty-micrometer sections of frozen ovaries were prepared using
a Reichert 820 cryostat (AO/Reichert, Buffalo, NY) and mounted
onto gelatin-coated glass slides for in situ hybridization
as described previously (88). Hybridization probes used were
[35S]UTP-labeled riboprobes derived from the
full-length rat inhibin -subunit cDNA (89, 90), a 243-bp long
fragment of rat MAPKAPK-2 cDNA corresponding to amino acids 119–199 of
mouse MAPKAPK-2 (44), and a 216-bp long fragment of rat MAPKAPK-3
isolated by RT-PCR and corresponding to amino acids 267–338 of human
MAPKAPK-3 (7). Hybridization was continued for 12–18 h at 47 C in a
humidified chamber. Sense riboprobes were used as controls.
Subsequently, the slides were washed to a final stringency of 0.1x SSC
at 65 C after a 1 h treatment with 20 µg/ml RNAse at 37 C.
Slides were then processed for film and emulsion autoradiography
(NTB-2, Eastman Kodak Co., Rochester, NY). Exposure time
on film was 3 days and on emulsion was 2 weeks. After development of
the slides, they were stained with hematoxylin to visualize the nuclei.
The sections were then examined and photographed using a microscope
(Nikon Optiphot, Nippon Kogaku (USA) Inc., Garden City,
NY) or the film autoradiograms were scanned using a Microtek flatbed
scanner.
Immune Complex Kinase Assays
Ovarian lysates (500 µg protein) were subjected to
immunoprecipitation (IP), as described previously (8), in the presence
of 10 µl of MAPKAPK-3-specific antibody and 30 µl agarose-linked
protein A+G in 500 µl incubation volume for 2 h at 4 C.
Additional control IPs were done in the absence of antibody (with 4 day
post hCG lysate), or with antibody but in the absence of lysate.
Complexes were collected by centrifugation, washed three times with
RIPA (10 mM Tris-HCl, pH 7.2, 150 mM NaCl,
1.0% deoxycholate, 1.0% Triton X-100, 0.1% SDS, 1 mM Na
orthovanadate, 40 µg/ml phenylmethylsulfonylfluoride) and then once
with TE (10 mM Tris, pH 7.5, 0.1 mM EGTA).
Complexes were resuspended in 50 µl TE, and assayed for kinase
activity for 7 min at 30 C in 115 µl reaction volume in the presence
of 42 mM -glycerolphosphate, pH 7.0, 8.4 mM
MgCl2, 0.8 mM dithiothreitol, 42
mM ATP, 4.8 µCi [-32P]ATP, and
4.8 µg purified recombinant CREB. Reactions were terminated by
addition of 50 µl of SDS stop solution and heat denaturation (100 C,
5 min).
GRMO2 Cell Culture, Transfection, and Luciferase Assays
Cationic liposomes, prepared as described (91), were used for
transient transfection (92) of GRMO2 cells (42) (provided by N.V.
Innogenetics, Ghent, Belgium) that were cultured as described (42, 93)
in HDTIS (DMEM-F12, 1:1, 10 µg/ml insulin, 5 nM sodium
selenite, 5 µg/ml transferrin, and 100 mg/liter sodium pyruvate)
supplemented with 2% FBS in a humidified incubator at 37 C and 5%
CO2. DNA for transfection was preincubated at
room temperature with lipofection reagent for 20–30 min in OptiMEM and
then added to cells washed with PBS. GRMO2 cells, grown in 12-well
culture dishes, were transfected (per well) with 500 ng of a GAL4
binding site-luciferase reporter plasmid DNA and 50 ng of CREB-GAL4
fusion protein expression construct, and 50 ng of MAPKAPK-3 constructs
(empty MAPKAPK-3 vector, wild-type MAPKAPK-3, or kinase-dead K>M
mutant MAPKAPK-3), as indicated. After 6 h of transfection, the
DNA-lipid mixture was replaced with fresh HDTIS containing 2% FBS.
Cells were incubated for 12 h, washed with PBS, and then subjected
to lysis by gentle agitation on ice in the presence of cell-lysis
buffer (25 mM HEPES pH 7.8, 15 mM
MgSO4, 1 mM dithiothreitol, 0.1%
Triton X-100). Luciferase assays were performed essentially as
described previously (94). One hundred microliters of the cell lysates
were added to 400 µl of assay buffer (25 mM HEPES, pH
7.8, 15 mM MgSO4, 5 mM
ATP, 1 µg/ml BSA), and then 100 µl of 1 mM luciferin
were added and emitted luminescence was measured using a 2010
luminometer (Analytical Luminescence, San Diego, CA) for 10 sec.
Protein content of cell lysates was determined by the Bradford method
(95). Results were analyzed using Student’s t test
(P < 0.05) (96).
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
This research was supported by the following grants: NIH PO1 HD-21921 (M.H.D., K.E.M.), P30 Center grant HD-28048 (K.E.M.) Training Programs in Reproductive Biology T32 HD-07068 (A.M.) and Endocrinology T32 DK-07169 (C.A.P.), and National Cancer Institute contract no. NO1-CO-56000 (G.S.).
1 Current address: Fred Hutchinson Cancer Research Center, Seattle,
Washington 98109. 
2 Current address: Cancer Research Institute, University of
California, San Francisco California 94115.
3 Current address: Wyeth-Ayerst Laboratories, Inc.
Women’s Health Research Institute, Radnor Pennsylvania 19087.
Received for publication March 3, 2000. Revision received February 6, 2001. Accepted for publication February 9, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, and
H2O2. J Biol Chem 271:26981–26988 isoform of protein kinase C. Biochem J 332:703–712
