| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 278, Issue 34, 31593-31602, August 22, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶
From the
Department of Biomedical Sciences,
Colorado State University, Fort Collins, Colorado 80523 and the
Department of Biomedical Sciences, Cornell
University, Ithaca, New York 14853
Received for publication, April 23, 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
T3-1 cells expressing
endogenous GnRHR or Chinese hamster ovary cells expressing an epitope-tagged
GnRHR were fractionated through a sucrose gradient. We found the GnRHR and
c-raf kinase constitutively localized to low density fractions independent of
hormone treatment. Partitioning of c-raf kinase into lipid rafts was also
observed in whole mouse pituitary glands. Consistent with GnRH induced
phosphorylation and activation of c-raf kinase, GnRH treatment led to a
decrease in the apparent electrophoretic mobility of c-raf kinase that
partitioned into lipid rafts compared with unstimulated cells. Cholesterol
depletion of T3-1 cells using methyl-
-cyclodextrin disrupted
GnRHR but not c-raf kinase association with rafts and shifted the receptor
into higher density fractions. Cholesterol depletion also significantly
attenuated GnRH but not phorbol ester-mediated activation of extracellular
signal-related kinase (ERK) and c-fos gene induction. Raft
localization and GnRHR signaling to ERK and c-Fos were rescued upon repletion
of membrane cholesterol. Thus, the organization of the GnRHR into low density
membrane microdomains appears critical in mediating GnRH induced intracellular
signaling. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Consistent with the presence of G-proteins in lipid microdomains, several
members of the superfamily of G-protein-coupled receptors
(GPCR)1 have been
shown to partition into lipid rafts and caveolae
(9–13).
Often, the translocation of GPCR into membrane microdomains appears to require
ligand activation; however, -adrenergic receptor subtypes have been
found distributed between both caveolin containing and bulk plasma membrane
fractions (10,
14). Herein we have sought to
determine the membrane localization of the mammalian type I gonadotropin
releasing hormone receptor (GnRHR). An atypical member of the rhodopsin-like
family of GPCR, the GnRHR is located on the cell surface of pituitary
gonadotropes
(14–16).
The binding of the hypothalamic decapeptide GnRH to the pituitary GnRHR not
only stimulates but is obligatory for the synthesis and secretion of
luteinizing hormone (LH) (15,
16). In the absence of GnRH
input to the pituitary gland, LH production and, consequently, gonadal
function in mammals ceases
(17,
18). Thus, the GnRHR
represents the site that mediates the primary stimulatory input to gonadotrope
cells. Upon ligand activation, it is well established that agonist-occupied
GnRHR couples to Gq/11 leading to stimulation of
phospholipase C, formation of inositol 1,4,5-trisphosphate (IP3)
and diacylglycerol (DAG), elevation of intracellular free calcium and
activation of one or more isoforms of protein kinase C (PKC)
(19,
20). These early events
underlie GnRH activation of multiple mitogen-activated protein kinase (MAPK)
signaling cascades including p38 MAPK, c-Jun N-terminal kinase (JNK) and
extracellular signal-regulated kinase (ERK)
(21). ERK activation by GnRH
is dependent on both PKC and calcium influx via L-type voltage-gated calcium
channels and proceeds through a c-raf kinase-dependent mechanism
(21,
22).
Although classified as a member of the rhodopsin class of GPCR, the GnRHR
displays several unique structural characteristics of which perhaps the most
striking is an extremely short C-terminal cytoplasmic domain of only 1–2
amino acids (23,
24). In more prototypical
GPCRs, this domain is quite extensive and contains phosphorylation sites for
G-protein-coupled receptor kinases, second messenger-regulated kinases (such
as protein kinase A) and casein kinases. In some GPCRs, these phosphorylation
events allow for interaction with -arrestins and subsequent receptor
deactivation and internalization
(25,
26). Consistent with the lack
of an extensive C-terminal tail, the GnRHR does not appear to undergo
arrestin-dependent internalization or desensitization
(27–31).
In fact, the GnRHR has been considered as a naturally occurring
internalization resistant mutant
(28). Thus, the GnRHR is both
structurally and functionally unusual.
Given its unique structural and functional attributes we sought to assess
the membrane distribution of the mammalian GnRHR. Herein, we find that unlike
other GPCRs, such as the muscarinic acetylcholine receptor and B2 bradykinin
receptor that are uniformly distributed throughout the plasma membrane and
localize to caveolae only in the presence of ligand
(11,
13), the GnRHR constitutively
resides in a low density membrane microdomain in both the gonadotrope-derived
T3-1 cell line that expresses the endogenous GnRHR gene
(32) and Chinese hamster ovary
(CHO) cells. We also find that c-raf kinase, a target of GnRHR signaling, is
constitutively, but not exclusively, localized to low density membrane
fractions. Furthermore, disruption of raft organization by cholesterol
depletion attenuates GnRHR activation of ERK and induction of c-fos
gene expression suggesting that the sublocalization of GnRHR and c-raf kinase
to membrane microdomains is functionally significant.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
q/11, anti-c-raf kinase, anti-b-raf
kinase, anti-p-ERK, and anti-ERK-1 antibodies were purchased from Santa Cruz
Biotechnology, Inc. (Santa Cruz, CA). The antibody against
H+-ATPase was a generous gift from Dr. Klaus Bayenbach (Cornell
University). The anti-flotillin-1 and pan-caveolin antibodies were purchased
from BD Transduction Laboratories (San Jose, CA). The anti-tubulin antibody
and methyl--cyclodextrin (CD) were obtained from Sigma. The anti-HA
antibody was purchased from Roche Applied Science (Indianapolis, IN).
Anti-goat, anti-rabbit, and anti-mouse secondary antibodies were from Pierce,
Santa Cruz Biotechnology, or Bio-Rad. [3H]inositol (14 Ci/mmol) was
obtained from Amersham Biosciences.
D-Ala6-desGly10-GnRH-Pro9-ethylamine
([D-Ala6]GnRH) was purchased from Sigma.
Cell Culture—T3-1 cells were maintained in high
glucose DMEM from Mediatech (Herndon, VA) containing 2 mM
glutamine, 100 units of penicillin/ml, 100 µg streptomycin/ml, 5% fetal
bovine serum, and 5% horse serum. CHO cells were maintained in high glucose
DMEM containing 2 mM glutamine, 100 units of penicillin/ml, 100
µg of streptomycin/ml, 10% fetal bovine serum, and 1x nonessential
amino acids from Invitrogen. All cells were grown in 5% CO2 at 37
°C in a humidified environment.
Construction of Hemagglutinin-tagged GnRH Receptor Attached to Green Fluorescent Protein (GFP)—The construction of the GnRHR-GFP fusion cDNA has been described (33). A 2-step PCR procedure was used to attach the HA sequence (YDYDVPDYA) immediately adjacent to the initiation codon of the GnRHR-GFP fusion protein. Appropriate placement of the HA tag was confirmed by sequencing.
Preparation of Cholesterol-loaded CD (CLCD)—Cholesterol (200 mg) was dissolved in 1 ml of chloroform. In a separate beaker, 1 g of CD was dissolved in 2 ml of methanol. A 0.45-ml aliquot of the cholesterol solution was added to the CD solution, stirred, and then placed under a stream of nitrogen gas until the chloroform and methanol evaporated. The resulting crystals were allowed to dry for 24 h and stored in glass at room temperature. A CLCD working solution was prepared by adding 50 mg of the CLCD crystals to 1 ml of serum-free DMEM, warming to 37 °C, and vortexing briefly.
Detergent-free Preparation of Lipid Rafts—T3-1 cells,
a generous gift from Dr. Pam Mellon, or CHO cells were grown to confluence in
150-mm tissue culture plates. Cells were harvested in phosphate-buffered
saline (PBS) and centrifuged for 3 min at 300 x g. Cells were
resuspended in PBS to a final volume of 1 ml and administered either vehicle
or 100 nM GnRH for 15 min at 37 °C followed by centrifugation
for 3 min at 300 x g. Detergent-free lipid raft preparations
were conducted according to Song et al.
(34). The cell pellet was
resuspended in 2 ml of 500 mM sodium carbonate buffer (pH 11).
Cells were then homogenized using a Wheaton loose fitting glass dounce
homogenizer (10 strokes) followed by sonication (three 20-s bursts) on ice
using a Branson 250 sonicator. Two ml of 90% sucrose prepared in MES buffer
(20% glycerol, 150 mM NaCl, 2 mM EDTA, 25 mM
MES, pH 6.5) was added to the homogenized samples yielding a final
concentration of 45% sucrose in a total volume of 4 ml. A discontinuous
sucrose gradient was then layered on the surface of the 45% fraction (4 ml of
35% sucrose, 4 ml of 5% sucrose in MES containing 250 mM sodium
carbonate). Isopycnic ultracentrifugation was then carried out at 38,000 rpm
using a SW 41 rotor for 16–20 h at 4 °C. Following
ultracentrifugation, 645-µl samples were collected representing a total of
18 fractions. Proteins that migrated to the interface of the 5 and 35%
gradients (approximately fractions 6 and 7) were considered to be
raft-associated (34).
CD treated samples were prepared as above with the exception that
T3-1 cells in monolayer were incubated in 12 ml of serum-free medium
containing 2% CD for1hat37 °C followed by2hof serum-free medium alone.
Control cells were incubated in serum-free medium without CD over the same
time period. For the cholesterol repletion studies, T3-1 cells were
treated with CD as described above. Following CD treatment, cells were washed
with PBS and incubated in 12 ml of serum-free medium containing 1 mg/ml of
CLCD for 2 h. Raft samples were then prepared as described above.
Triton X-100 Preparation of Lipid Rafts—T3-1 or CHO
cells were grown to confluence in monolayer cultures. Cells were harvested and
treated with hormone as in the detergent-free raft method. Following
centrifugation, the cell pellet was resuspended to a final volume of 500 µl
in PBS. 500 µl of 2x Triton X-100 lysis buffer (0.2% Triton X-100
prepared in MES buffer) was then added to yield a final 1x lysis buffer
concentration (Triton X-100 concentration = 0.1%). Cells were then homogenized
either by 3 passes through a 30-gauge needle or douncing. PBS was added to
adjust the final volume to 1 ml. 1 ml of 80% sucrose (in MES buffer) was added
to the samples to yield a 40% sucrose fraction in a final volume of 2 ml. A
discontinuous sucrose gradient was then prepared by layering 2 ml of each
sucrose fraction (80, 60, 40%-containing sample, 30, 20, and 10%) in a 12-ml
ultracentrifuge tube (35)
(Figs. 1D and
2B) or layering 4 ml
of 45, 35, and 5% sucrose in a 12-ml ultracentrifuge tube
(Fig. 6). Samples were
subjected to isopycnic ultracentrifugation in an SW 41 rotor for 20 h at
37,000 rpm. Following ultracentrifugation, 645-µl samples were collected
representing a total of 18 fractions and Western analyses were conducted. In
Fig. 7, whole mouse pituitaries
were obtained following euthanasia, minced, rinsed free of blood in PBS, and
lysed in a 0.1% Triton lysis buffer (in MES) by douncing. Whole pituitary
lysates were adjusted to 45% sucrose and layered within a sucrose gradient (45
(sample), 35, and 5% sucrose) in a total volume of 
3.5 ml to reduce a
potential dilution effect of the larger gradients described for T3-1
and CHO cells. These samples were centrifuged as described above in an SW50.1
rotor. Following centrifugation, gradients were harvested in 10 equal
fractions.
|
|
|
|
Silver Staining—Sucrose gradient fractions from T3-1
cells prepared with carbonate lysis were resolved by SDS-PAGE. The resolved
proteins within the gel were visualized by silver staining using methods as
described by Blum et al.
(36). Identical results were
obtained using gradient fractions from T3-1 lysates generated from 0.1%
Triton X-100 lysis buffer (data not shown).
Western Blots—Samples representing individual fractions were
subjected to SDS-polyacrylamide gel electrophoresis (acrylamide:bis-acrylamide
ratio of 29:1) and electroblotted to nitrocellulose or polyvinylidene
difluoride membranes (PerkinElmer Life Sciences). In
Fig. 5B, SDS-PAGE was
carried out using a 37.5:1 acrylamide:bis-acrylamide ratio to favor the
separation of phosphorylated and non-phosphorylated c-raf kinase. Membranes
were blocked in 5% nonfat dried milk in Tris-buffered saline (TBS) or TBS
containing 0.1% Tween-20 (TBST). Anti-GnRHR antibody (1:500 dilution in 5%
milk) was incubated for 8 h at 4 °C on an orbital shaker. Blots were
washed for 30 min (3 washes x 10 min) with TBS and then incubated with
anti-goat HRP (1:5,000) for 2 h. Anti-Gq/11 antibody was
used at a 1:500 dilution for 2 h at room temperature followed by washing with
TBS and incubation with anti-rabbit HRP (1:5,000) for 2 h. Anti-HA and
anti-flotillin antibodies were used at a 1:1,000 dilution with an incubation
time of 1 h. Blots were washed and then incubated with a 1:10,000 dilution of
anti-mouse HRP for 1 h at room temperature. Anti-c-raf, b-raf, and
H+-ATPase antibodies were used at a dilution of 1:1,000 and
incubations were overnight at 4 °C. Blots were washed in TBST and then
incubated with a 1:5,000 dilution of anti-rabbit HRP for 2 h at room
temperature. Anti-caveolin-1, anti-tubulin, and anti-pan caveolin were used at
a 1:2,000 dilution with a 1:10,000 dilution of the appropriate secondary
antibody for 1 h. All blots were washed for 60 min (6 x 10 min) with TBS
or TBST after secondary antibody and then visualized by chemiluminescence
using either Pierce SuperSignal or PerkinElmer Western Lightening
reagents.
|
ERK and c-Fos Activation Assays—Monolayers of T3-1
cells in 6-well tissue culture plates were subjected to 1 of 3 treatments.
First, control cells were washed with PBS and then incubated with 1 ml of
serum-free medium for 3 h. Second, for cholesterol depletion, cells were
washed with PBS and incubated with 1 ml of serum-free medium containing 2% CD
for 1 h. Cells were then washed with PBS and incubated for2hin serum-free
medium. Third, for cholesterol repletion, cells were washed with PBS and
administered 1 ml of 2% CD in serum-free medium for 1 h. Cells were then
washed with PBS and incubated for 2 h in serum-free medium containing 1 mg/ml
CLCD. Following the 3-h treatment protocols, cells were washed with PBS and
serum-free medium was replaced containing either 0 or 100 nM GnRH.
After a 30- or 60-min incubation, cells were washed in ice cold PBS and lysed
in RIPA buffer containing 20 mM Tris (pH 8.0), 137 mM
NaCl, 10% glycerol, 1% Nonidet P-40, 0.1% SDS, 0.5% deoxycholate, and 0.2
mM phenylmethylsulfonyl fluoride. 6x sample buffer (300
mM Tris-HCl, pH 6.8, 60% glycerol, 30 mM dithiothreitol,
6% SDS) was added to yield a final concentration of 1x. Aliquots (15
µl) of each lysate were heated to 95 °C for 5 min and subjected to
SDS-PAGE and Western analysis. Nitrocellulose membranes were incubated for 2 h
with either a phospho-ERK or c-Fos antibody (both at 1:1,000 dilutions)
followed by a 2-h incubation with a 1:2,000 dilution of HRP conjugated
secondary antibody. Phospho-ERK blots were then stripped at room temperature
with 100 mM 2-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl
(pH 6.7) heated to 50 °C for 30 min. After stripping, membranes were
washed twice for 15 min with TBS and blocked with 5% milk for 1 h. Blots were
then re-probed with a 1:10,000 dilution of an anti-ERK antibody that
recognizes ERK-1 and ERK-2 independent of phosphorylation state. Following
washing in TBS, blots were incubated with a 1:2,000 dilution of anti-rabbit
HRP, and immunoreactive bands were visualized by chemiluminescence.
Cholesterol Assays—T3-1 cells were either incubated
in serum-free medium alone for 3 h, cholesterol-depleted or
cholesterol-depleted followed by incubation with CLCD as described under
"ERK and c-Fos Activation Assays" above. At the end of the 3-h
treatment period, all cells were washed and harvested in PBS. Cells were
pelleted by centrifugation and resuspended to a final volume of 500 µl in
PBS. An equal volume of 2x lysis buffer containing 0.2% Triton X-100 in
MES buffer cooled to 4 °C was then added to solubilize the plasma
membrane. Cholesterol content was determined using Infinity cholesterol
reagent (Sigma). Briefly, lysates were diluted 1:5 with cholesterol reagent
and samples were incubated for 10 min at 37 °C followed by
spectrophotometric analysis (absorption at 500 nm). Cholesterol concentration
was determined using a 6-point standard curve generated with known
concentrations of cholesterol. To standardize cholesterol content, protein
concentrations were determined using a BCA protein assay (Pierce). Data are
expressed as mg of cholesterol per mg of protein.
Transmission Electron Microscopy (TEM)—CHO and T3-1
cells were cultured to 60–70% confluence then scraped from dishes
in ice-cold Hepes-buffered saline (HBS; pH 7.4) and washed once in HBS. The
cell pellets were fixed in 2.5% glutaraldehyde in 0.1 M sodium
cacodylate buffer, pH 7.4 for 30 min at room temperature followed by 1.5hat4
°C. The cells were then washed three times 10 min each in 0.1 M
sodium cacodylate buffer (pH 7.4). The cells were postfixed in 2% osmium
tetroxide for 1 h at room temperature, followed by washes in sodium cacodylate
buffer as described above. Cells were then dehydrated through a standard
graded ethanol series followed by an acetone rinse. Gradual infiltration of
epon araldite resin followed, after which the cells were placed into Beem
capsules and cured in a 60 °C oven. The resin blocks were cut on a
Reichert OmU2 ultramicrotome and 70-nm sections were stained with uranyl
acetate and lead citrate. The grids were viewed in a Tecnai 12 Biotwin
transmission electron microscope (FEI Corp). For quantitation of each cell
type, 100 cells were visualized and the number of putative caveolae was
obtained. Digital images were captured with a Gatan Model 791 Multiscan
Camera.
125I-[D-Ala6]GnRH Binding
Assay—[D-Ala6] GnRH was radioiodinated using
a glucose-oxidase procedure and purified by chromatography in QAE-Sephadex as
described by Wagner et al.
(37). Following a 1-h
incubation in either serum-free medium alone or serum-free medium containing
2% CD, T3-1 cells were washed with PBS and incubated for 2 h in
serum-free medium alone. Cells were harvested in ice-cold PBS. Following
centrifugation, cell pellets were resuspended in assay buffer (10
mM Tris-HCl, 0.1% bovine serum albumin, .01 mM
CaCl2) to a final concentration of 1 x 107
cells/50 µl. Triplicate 12 x 75 mm assay tubes were prepared
containing 50-µl aliquots of cell suspension and 5 x 104
cpm of 125I-[D-Ala6]GnRH (61.4 pM)
in 50-µl assay buffer in the presence or absence of 50 µl of
non-radioactive [D-Ala6]GnRH (340 nM). The
total volume for each tube was adjusted to 250 µl by addition of ice-cold
assay buffer. Following a 2-h incubation at 4 °C, 3 ml of ice-cold assay
buffer was added, and samples were immediately centrifuged for 15 min at
16,000 x g. The supernatants were decanted and radioactivity in
the cell pellets was quantitated using an Apex automatic gamma counter
(Micromedic systems, Horsham, PA). Specific binding was determined by
subtracting the cpm in samples containing
125I-[D-Ala6]GnRH in the presence of
unlabeled [D-Ala6]GnRH from the cpm in samples
containing only 125I-[D-Ala6]GnRH samples. To
standardize binding for protein concentration, a BCA protein assay (Pierce)
was performed. The binding activity presented in
Fig. 3A
(inset) was adjusted for protein concentration.
|
[3H]Inositol Assays—Phospholipase C activity was
assessed by quantifying cellular accumulation of phosphorylated inositol using
previously described methods
(38). Briefly, T3-1
cells were plated overnight in 24-well culture plates. The DMEM culture medium
was washed from the cells with serum-free M199 culture medium (Mediatech;
Herndon, VA) supplemented to 0.37% (w/v) sodium bicarbonate, 20 mM
HEPES buffer, 100 units of penicillin/ml, and 100 µg streptomycin/ml. After
washing, cells were incubated at 37 °C for 5 h in 0.3 ml of serum-free
M199 containing 2 µCi of myo-[2-3H]inositol. The labeled cells
were then washed with serum-free DMEM containing 5 mM LiCl and
incubated for an additional hour at 37 °C in either 1 ml of serum-free
DMEM with 5 mM LiCl or 1 ml of serum-free DMEM containing 5
mM LiCl and 2% CD. After 1 h of cholesterol depletion, this medium
was removed and then control and cholesterol-depleted cells remained untreated
or were challenged with 100 nM GnRH in 1 ml serum-free DMEM
containing 5 mM LiCl. These treatment conditions were maintained at
37 °C for 1 h, after which the medium was removed, and cells were
immediately lysed by addition of 0.05 ml of RIPA buffer (described above) and
1 ml of water heated to 95 °C. The cells were then frozen overnight and
thawed at room temperature. Cell lysates from each individual well were
collected and loaded separately onto Dowex 1-X8, 200–400 mesh,
formate-form columns with an approximate bed volume of 0.4 ml. Free,
unphosphorylated inositol was eluted from the lysate by the addition of 10 bed
volumes of water. After collection of the eluent containing the free inositol,
total inositol phosphates were collected by the addition of 10 bed volumes of
1 M ammonium formate in 0.1 M formic acid. The amounts
of radioactivity in both the free and phosphorylated inositol eluents were
quantitated using a Beckman LS-5000CE liquid scintillation counter. Data are
presented as phosphorylated inositol expressed as a percentage of the total
[3H]inositol.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
T3-1 cells was conducted to assess the efficacy of protein separation
using sucrose gradients (Fig.
1A). This analysis revealed enrichment of distinct
populations of proteins associated with either low density (fractions 5 and 6)
or high density fractions (fractions >10) suggestive of effective
separation of raft-associated proteins. To assess membrane sublocalization of
the GnRHR, T3-1 cells were administered either vehicle or 100
nM GnRH for 15 min. Cells were then sonicated in a detergent-free
sodium carbonate buffer and fractions separated in a discontinuous sucrose
gradient. Fractions were isolated and subjected to SDS-PAGE followed by
Western blotting using an anti-GnRHR, anti-tubulin, or anti-flotillin
antibody. Immunodetectable GnRHR localized to low density fractions within the
gradient independent of GnRH treatment
(Fig. 1A, lanes
6 and 7). Lanes 6 and 7 represent the interface of the 5 and 35%
sucrose fractions where raft-associated proteins would migrate
(34). The absence of caveolin
expression in T3-1 cells (Fig.
1C) precluded the utility of this protein as a marker
protein for low density membrane microdomains, thus, we used an antibody
directed against flotillin, another raft-associated protein
(39,
40). Using the same detergent
free raft isolation method in T3-1 cells, flotillin appropriately
localized to low density fractions within the sucrose gradient
(Fig. 1B). Not
considered a raft-associated protein
(41), tubulin did not
partition into membrane rafts in these conditions. To confirm GnRHR
localization to membrane microdomains, an independent approach based on
resistance to detergent solubilization was utilized to prepare samples prior
to isopycnic centrifugation. This technique takes advantage of the relative
insolubility of lipid rafts in non-ionic detergents at 4 °C. As with the
detergent-free raft preparations, the GnRHR localized to the low density
fractions regardless of hormone treatment. Tubulin was again localized to high
density fractions (Fig.
1D).
To assess whether GnRHR localization to low density membrane microdomains
was specific to a gonadotrope-derived, non-caveolin expressing cell line, we
next studied an HA-tagged GnRHR-GFP fusion protein stably expressed in CHO
cells (33). Using the
detergent-free raft preparation, we found that the GnRHR migrated to low
density fractions 6 and 7 (the 5/35% interface). As in T3-1 cells, the
migration of the GnRHR in the sucrose gradient was unaffected by hormone
treatment (Fig. 2A).
Similar results were evident when cells were prepared using non-ionic
detergent (Fig. 2B).
Specifically, the GnRHR localized to the low density fractions, and the
migration of the GnRHR in the sucrose gradient was unaffected by treatment
with agonist (GnRH), super-agonist ([D-Ala6]GnRH), or
antagonist (antide). With both the detergent and non-detergent preparations,
tubulin failed to migrate out of the high density fractions of the gradient.
Finally, unlike T3-1 cells, CHO cells express the caveolar raft marker
protein caveolin-1 that appropriately localized to low density fractions.
Thus, GnRHR localization to membrane microdomains or lipid rafts appears to be
independent of cell type or the presence of caveolin-1.
Transmission Electron Microscopy Reveals Topological Differences in the
Plasma Membrane of T3-1 and CHO Cells— The absence
of caveolin expression in T3-1 cells suggests that the GnRHR localizes
to non-caveolar rafts in homologous cells, consistent with expression of
flotillin-1. We wanted to examine the potential topological differences that
could possibly exist in the plasma membrane of T3-1 and CHO cells. The
topological differences should be expressed as the presence of flask-shaped
invaginations with the plasma membrane of CHO cells but not T3-1 cells.
For quantitation, 100 cells were visualized, and the number of flask-shaped
invaginations in the plasma membrane of each cell type was determined
(Fig. 3A). Consistent
with non-detectable levels of caveolin-1, T3-1 cells contain 20-fold
fewer morphologically distinct membrane invaginations relative to CHO cells
(Fig. 3B). These
ultrastructural studies support the notion that GnRHR most likely partitions
to noncaveolar lipid rafts.
GnRH Activation of ERK and c-Fos Expression Is Lost with Cholesterol
Depletion—The microenvironment necessary for raft formation is
sensitive to cholesterol depletion
(42). Thus, to assess the
functional relevance of GnRHR localization in lipid rafts we sought to disrupt
raft organization utilizing the cholesterol-sequestering agent,
methyl--cyclodextrin (CD)
(5). T3-1 cells were
incubated for 1 h in serum-free medium with or without 2.0% CD. 1 h exposure
to 2.0% CD effectively reduced cholesterol content by 55%
(Fig. 4A). To address
the functional consequence of cholesterol depletion and, presumably, raft
disruption we next examined the effects of CD treatment on GnRH activation of
ERK. Consistent with our previous studies
(22,
43), 30 min exposure to GnRH
increased the dual phosphorylated forms of ERK (ERK1 and ERK2)
(Fig. 4B). In
contrast, phosphorylated ERK was not detectable in cells that had been exposed
to 2.0% CD. GnRH activation of c-fos gene expression proceeds through
an ERK-dependent mechanism
(44,
45). Consistent with this
notion, the effects of cholesterol depletion were also evident as a loss of
GnRH activation of c-fos expression
(Fig. 4C). Finally, it
seemed possible that the loss of GnRH activation of ERK and c-fos
gene expression in CD-treated cells might simply reflect a significant
attenuation in the number of cell surface GnRH receptors available for
binding. This would not, however, appear to be the case as cholesterol
depleted T3-1 cells retained the ability to bind
125I-[D-Ala6]GnRH at levels 72% that of
control cells (Fig.
4A, inset).
|
c-raf Kinase Localizes to Low Density Membrane Microdomains in
T3-1 Cells—GnRH receptor activation of the ERK cascade
and c-fos gene expression is thought to proceed through a c-raf
kinase-dependent mechanism
(22). Thus, we were intrigued
with the possibility that the loss of ERK and c-Fos activation associated with
cholesterol depletion may reflect disruption of membrane microdomains
containing both the GnRHR and c-raf kinase. Consistent with this possibility,
we find that c-raf kinase is also localized to low density microdomains in
T3-1 cells prepared using the non-detergent based approach
(Fig. 5A).
Partitioning of c-raf kinase to the low density sucrose fractions did not
appear to be affected by hormone treatment. Thus, c-raf kinase was
constitutively but not exclusively present in lipid rafts. In contrast to
c-raf kinase, b-raf-kinase, a closely related member of the raf kinase family,
was only evident in high density fractions. Thus, partitioning into lipid
rafts appeared to be specific to the c-raf isoform of raf kinase. Also
supportive of the specificity of fractionation, an integral membrane protein
H+-ATPase migrated to high density sucrose fractions. Finally, to
test if raft-associated c-raf kinase is a target for GnRH-mediated
phosphorylation, fractions 6 and 7 were isolated from the sucrose gradient and
subjected to electrophoresis in a low cross-linking gel. Consistent with
GnRH-induced phosphorylation, the electrophoretic mobility of raft-associated
c-raf kinase was reduced with GnRH treatment
(Fig. 5B). Previous
studies have demonstrated that the retarded electrophoretic mobility of c-raf
kinase correlates with GnRH-induced catalytic activation of this enzyme
(22).
GnRH Binding Enhances the Association of c-raf Kinase with Lipid Microdomains—We next sought to confirm raft localization of c-raf kinase using the detergent-based preparation. As with the non-detergent approach, immunoblots of lysates prepared using a 0.1% Triton X-100 lysis buffer revealed constitutive but not exclusive localization of c-raf kinase low density fractions independent of hormone treatment (Fig. 6A). It is interesting to note, however, that increasing the concentration of Triton X-100 from 0.1 to 1.0% resulted in a distinctly different pattern of segregation such that c-raf kinase was evident in the low density fractions only under conditions of GnRH binding (Fig. 6B). Thus, in the unbound state, the association of c-raf kinase in lipid rafts was more susceptible to disruption by non-ionic detergent suggesting that GnRH treatment may act to stabilize or enhance the association of c-raf kinase in lipid microdomains.
c-raf Kinase Is Present in Low Density Membrane Microdomains in Mouse
Pituitary Cells—To address whether c-raf kinase localization to
lipid rafts was a unique property of the T3-1 cell line, detergent-free
raft preparations were prepared from CHO, JEG3 (choriocarcinoma), and NIH 3T3
cells and subjected to sucrose density gradient centrifugation. As with
T3-1 cells, a raft-associated population of c-raf kinase was detectable
in all 3 cell lines (data not shown). Thus, c-raf kinase is present in low
density membrane microdomains independent of cell type and expression of
caveolin-1. Most importantly, c-raf kinase was also evident in low density
fractions prepared from whole pituitary lysates from adult female mice
(Fig. 7). Pituitary expression
of caveolin-1 allowed us to use this protein as a marker for low density
microdomains. Clearly, we do not suggest that these results reflect a unique
contribution of GnRHR expressing cells (gonadotropes) as there are multiple
cell lineages present in the adult pituitary. These data do, however,
demonstrate that raft localization of c-raf kinase is evident in
non-transformed, GnRHR expressing tissues and is not simply an aberrant
feature of the T3-1 cell line. Unfortunately, we were unable to detect
the GnRHR in mouse pituitaries due to the low number of gonadotropes and
insufficient protein concentrations.
Disruption of Raft Localization of GnRHR by CD Treatment of
T3-1 Cells Is Reconstituted by Cholesterol Repletion—In
Fig. 4, we demonstrate that
GnRH activation of ERK and c-Fos expression is lost in cholesterol-depleted
T3-1 cells. Presumably, the loss of cholesterol is associated with raft
disruption. If correct, then the effects of CD should be revealed as a loss of
both GnRHR and potentially c-raf kinase in low density sucrose fractions.
Furthermore, if the effects of CD are specific to cholesterol depletion then
reconstitution of cholesterol content should accordingly reconstitute raft
microdomains. To directly test these possibilities, T3-1 cells were
incubated in the presence or absence of 2.0% CD for 1 h. Cholesterol repletion
was accomplished by incubation of cholesterol-depleted cells with 1 mg/ml CLCD
for 2 h. Cellular lysates were then prepared in sodium carbonate buffer and
subjected to sucrose density gradient centrifugation. As in the previous study
(Fig. 4), 1 h exposure to CD
resulted in an approximate 55% reduction in cholesterol in CD-treated cells as
compared with control cells (Fig.
8A). Subsequent incubation of cholesterol-depleted cells
with CLCD was effective in restoring cholesterol content to 85% of
control levels (Fig.
8A). Consistent with raft disruption, cholesterol
depletion was associated with a loss of GnRHR from the low density fractions 6
and 7 (Fig. 8B).
Importantly, however, cholesterol repletion was effective in reconstituting
the localization of GnRHR to low density fractions in the sucrose gradient
(Fig. 8B). In
contrast, partitioning of c-raf kinase into lipid rafts was not remarkably
affected by cholesterol depletion (Fig.
8C) suggesting that fundamental differences exist in the
mechanism(s) underlying the partitioning of c-raf kinase and GnRHR into low
density compartments.
|
Cholesterol Repletion Reconstitutes GnRH Activation of ERK and c-Fos
Expression in CD-treated T3-1 Cells—Based on the
data in the previous section, incubation of cholesterol depleted T3-1
cells with CLCD effectively restored cholesterol content and raft localization
of GnRHR. Next, we sought to determine if this is sufficient to reconstitute
GnRH signaling to ERK and c-Fos. Accordingly, GnRH activation of ERK and c-Fos
expression was assessed utilizing the same cholesterol depletion/repletion
paradigm. As in Fig. 4,
GnRH-induced phosphorylation of ERK and c-Fos expression was lost upon
incubation of T3-1 cells for 1 h in 2.0% CD
(Fig. 9). Partial
reconstitution of both parameters was, however, evident after incubation of
cholesterol-depleted cells with CLCD (Fig.
9).
|
Cholesterol Depletion Uncouples GnRHR- but Not Phorbol Ester-mediated
Activation of ERK—To begin to localize the lesion in GnRH signaling
to ERK we next asked if ERK activation in response to phorbol ester (PMA)
treatment was retained in cholesterol-depleted cells. The use of PMA in these
studies would, presumably, directly activate PKC isozymes thus effectively
bypassing the GnRHR, Gq/11, and phospholipase C. Consistent
with earlier studies
(43–46),
both GnRH and PMA induced ERK phosphorylation in control cells
(Fig. 10A). As
expected, cholesterol depletion resulted in a loss of GnRH-induced ERK
phosphorylation. Importantly, however, CD treatment did not visibly compromise
ERK activation in response to PMA. Thus, cholesterol depletion and the
resulting raft disruption appears to uncouple the GnRHR from signaling
intermediates that lie upstream of PKC isozymes and may, in fact, reflect
uncoupling of the GnRHR from its cognate heterotrimeric G-protein complex.
Consistent with this possibility, we find that, like the GnRHR, cholesterol
depletion of T3-1 cells leads to an attenuation in the amounts of
immunodetectable Gq/ll localized to low density
fractions (Fig. 10B)
suggesting that raft organization may be critical for GnRH coupling to
Gq. If correct, disruption of raft organization
should be revealed as a loss of GnRH signaling to phospholipase C and, as a
consequence, attenuation in the ability of GnRH to liberate IP3
from membrane phospholipid. In accordance with this we find that CD treatment
virtually eliminates the GnRH-induced IP3 response in T3-1
cells (Fig. 10C).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
T3-1 cell line. Raft localization of the GnRHR was
equally evident in heterologous cells (CHO) suggesting that this constitutive
segregation into lipid rafts is not unique to T3-1 cells but rather is
intrinsic to the GnRHR itself.
As yet we do not know what structural features of the GnRHR account for
raft localization. However, perhaps the most unique structural feature of this
GPCR is the virtual absence of an intracellular C terminus
(24). In more prototypical
GPCRs, this region is quite extensive and important for coupling to
G-proteins, agonist-induced receptor internalization, and
phosphorylation-mediated desensitization
(48). Phosphorylation of the C
terminus is typically thought of as requisite for subsequent interaction with
-arrestin, which hinders further G-protein activation and targets the
deactivated receptors for internalization
(26,
49). Due to the absence of an
intracellular C terminus, the mammalian GnRHR is neither phosphorylated nor
internalized via a -arrestin-dependent mechanism
(29,
31,
50). Finally, in several
GPCRs, cysteine residues located in the intracellular C terminus are targets
for reversible palmitoylation
(51–53),
a particularly intriguing observation as palmitoylation has been implicated in
increasing protein affinity with lipid rafts and caveolae
(54,
55). In short, given this
fundamental difference in the structural and functional properties of the
GnRHR vis a vis prototypical GPCRs, it is tempting to speculate that the
unusual behavior of the GnRHR in the plasma membrane may partially reflect the
absence of an intracellular C terminus. In this regard, it should be of
particular interest to examine the inmembrane behavior of non-mammalian and
Type II GnRH receptors, which possess intracellular C-terminal tails
(56–58).
Regardless of the structural features of the GnRHR that direct raft
association it appears that localization of the GnRHR to lipid rafts is
functionally significant. In support of this notion, raft disruption resulting
from the removal of cholesterol leads to a loss of GnRH activation of
ERK and c-fos gene expression. Importantly, this loss of
signaling was not due to a loss of cell-surface binding and was reversible by
cholesterol replenishment and reconstitution of lipid rafts. It is also
important to underscore that the lesion in GnRH signaling resulting from
cholesterol depletion was not reflected as a generalized loss of signaling.
Specifically, while GnRH activation of ERK was lost in CD-treated cells, this
same paradigm had little effect on the efficacy of PMA induced ERK
phosphorylation. Thus, the ERK signaling cascade in cholesterol-depleted
T3-1 cells is sufficiently intact to transduce a PMA signal but not a
GnRH signal. As such, the primary lesion in the GnRH response likely lies
upstream of PKC. Based on several lines of evidence we suggest that this
lesion may partially reflect uncoupling of the GnRHR from its cognate
heterotrimeric G-protein complex. First, consistent with others
(68), we find that the
presence of Gq in low density membrane fractions is
susceptible to disruption by cholesterol depletion. Second, cholesterol
depletion significantly attenuates the ability of GnRH to increase
intracellular levels of IP3.
GnRH activation of ERK proceeds through c-raf kinase
(22,
59). Consistent with this, we
find that, like the GnRHR, c-raf kinase also segregates into low density
sucrose fractions prepared from T3-1 cells, whole mouse pituitaries,
and several clonal cell lines. Thus, the localization of at least a portion of
c-raf kinase to low density membrane compartments is not cell type specific.
It is also clear, however, that the localization of c-raf to low density
microdomains is not exclusive. The limited presence of c-raf kinase in lipid
rafts relative to the larger pool of c-raf kinase associated with cytosol or
other membrane compartments (higher density fractions) may reflect the
rate-limiting presence of a putative platform or scaffolding protein(s)
reminiscent of the yeast protein Ste5p that is known to bind multiple members
of the ERK signaling cascade
(60). Similarly, several
mammalian scaffold proteins have also been characterized including 14-3-3
proteins and multiple isoforms of JNK inhibitory protein (JIP)
(7,
8,
61). If correct, then
compartmentalization of such a scaffold or platform into lipid rafts
associated with GnRHR would reflect a key mechanism for the initiation and
organization of GnRH signaling in gonadotropes. Finally, we should point out
that caveolin itself has been implicated as a scaffolding protein
(7); however, the absence of
expression of caveolin in T3-1 cells would suggest that this protein is
not requisite for organizing the GnRHR and c-raf kinase into a signaling
platform or scaffold.
The constitutive presence of c-raf kinase within lipid rafts is consistent with studies of the insulin signaling cascade and, more recently, the use of artificial rafts to examine the interaction of c-raf kinase with lipid microdomains. In regard to the latter, U. Rapp and co-workers (62) demonstrate c-raf kinase preferentially associated with artificial lipid rafts with high affinity. In the absence of conditions that promote lipid raft formation, raf kinase displayed moderate binding affinity directly with cholesterol; however, the binding affinity increased markedly in the context of lipid rafts. Similar observations with c-raf kinase were demonstrated for binding of ceramides within lipid rafts. In bulk plasma membrane not associated with rafts, c-raf kinase associated with phospholipids such as phosphotidylserine and phosphatidic acid, a product of phospholipase D catalytic activity. In addition to these studies, c-raf kinase has been shown to associate with lipid rafts in the context of insulin signaling via recruitment into internalized endosomes (63, 64). In this system, c-raf kinase appears to be recruited to membrane rafts by an interaction with raft-associated phosphatidic acid generated by insulin activation of phospholipase D. This interaction appears to be necessary for activation of the ERK cascade in a ras-dependent manner. This mechanism stands in contrast to the present studies in which c-raf kinase is constitutively present in low density fractions independent of any recruitment by signaling intermediates downstream of GnRHR activation.
In the insulin signaling system, cholesterol depletion and, presumably,
raft disruption blocks the recruitment of ras to lipid rafts and ERK
activation but does not appear to affect c-raf kinase membrane association or
activation (63). In the case
of T3-1 cells, cholesterol depletion to levels 50% of control
cells was sufficient to disrupt GnRHR association with lipid rafts. These same
conditions, however, did not affect the association of c-raf kinase with lipid
rafts. Thus, it is possible that the association of c-raf kinase with lipid
rafts in T3-1 cells may be independent of cholesterol content within
the membrane. Alternatively, if raft association of c-raf kinase in
T3-1 cells is cholesterol dependent then it is possible that the
cholesterol depletion was not sufficient to disrupt this association.
Unfortunately, a marked reduction in T3-1 cell viability precludes more
extensive cholesterol depletion. Finally, it is interesting to note that the
association of c-raf kinase within membrane rafts was sensitive to increasing
levels of non-ionic detergent. In relatively low detergent conditions, c-raf
kinase was constitutively localized to membrane rafts independent of GnRH
administration. However, while increasing detergent concentrations effectively
reduced c-raf localization to membrane rafts in unstimulated T3-1
cells, GnRH receptor activation restored c-raf kinase to this low density
membrane compartment. This increased resistance to detergent solubilization
raises the possibility that GnRH action resulting in modification of c-raf
kinase (such as phosphorylation) may increase the stability of the association
of c-raf kinase with lipid microdomains.
Since the isolation of the first cDNA encoding the mammalian GnRHR much
progress has been made in identifying structure-function relationships in this
unique GPCR (23,
24). Additionally, the use of
epitope- or fluorophore-tagged GnRH receptors has allowed direct observations
of the "in-membrane" behavior of unoccupied, agonist occupied, and
antagonist-occupied receptors. For example, we and others
(65–67)
have utilized resonance energy transfer methods to demonstrate that agonist
but not antagonist leads to self-association of GnRH receptors in the plasma
membrane. Based on the present studies, we suggest that this self-association
occurs in the context of discrete lipid microdomains that contain not only the
GnRHR but also downstream signaling components, such as Gq
and c-raf kinase, that serve to transduce a GnRH signal to the level of ERK
activation. As such, disrupting these microdomains effectively lesions the
ability of GnRH to activate ERK and increase c-fos gene expression.
In this model then, the GnRHR exists as a resident protein in a pre-formed
signaling platform that is poised to both receive and efficiently transmit the
GnRH signal. Confirmation of this model awaits definition of the complement of
proteins organized within this platform including identification of potential
scaffolding proteins and definition of the structural features of the GnRHR
that direct its association with lipid microdomains.
|
FOOTNOTES |
|---|
¶ To whom correspondence should be addressed. Tel.: 970-491-7571; Fax: 970-491-3557; E-mail: colin.clay{at}colostate.edu.
1 The abbreviations used are: GPCR, G-protein-coupled receptor; GnRH,
gonadotropin-releasing hormone; DMEM, Dulbecco's modified Eagle's medium; MES,
4-morpholineethanesulfonic acid; HA, hemagglutinin; HRP, horseradish
peroxidase; PBS, phosphate-buffered saline; RIPA, radioimmune precipitation
assay buffer; ERK, extracellular signal-regulated kinase; TEM, transmission
electron microscopy; PKC, protein kinase C; CHO, chinese hamster ovary; CD,
methyl--cyclodextrin; IP3, inositol 1,4,5-trisphosphate;
MAPK, mitogen-activated protein kinase.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
