 |
INTRODUCTION |
Prostacyclin (prostaglandin
(PG)1 I2) is a
labile metabolite of arachidonic acid, which is synthesized by the
sequential actions of PGH2 endoperoxide synthases 1 and 2 and prostacyclin synthase (1). The actions of prostacyclin generally
counteract those of thromboxane A2, and thus the relative
level of these two prostanoids in the circulation are important in the
local control of vascular tone and platelet aggregation (2, 3). The
main physiologic roles of thromboxane A2 and prostacyclin
are their contribution to the maintenance of vascular hemostasis:
thromboxane A2, synthesized primarily by platelets, induces
platelet shape change and aggregation and constriction of bronchial and
vascular smooth muscle, whereas prostacyclin, mainly synthesized by the
vascular endothelium, is a potent inhibitor of platelet aggregation and
induces vasodilation (4-7). Moreover, prostacyclin has been reported
to confer a cytoprotective effect against tissue injury in acute
myocardial ischemia or following hypoxic exposure of vascular
endothelial cells (8). Imbalances in thromboxane A2 or
prostacyclin have been reported to be a major contributing factor in
the development of a number of cardiovascular disorders including
thrombosis, myocardial infarction, unstable angina, stroke, and
atherosclerosis (9-13). In addition to its central role in the
cardiovascular system, prostacyclin may be important in the regulation
of renal blood flow (14); it also acts as a negative feedback regulator
of histamine secretion from mast cells (15) and acts as a lipolytic
agent, antagonizing the antilipolytic effect of PGE2, in
adipocytes (16).
The actions of prostacyclin are mediated via interaction with a
specific cell surface receptor, termed the prostacyclin receptor or IP
(17). The major intracellular signaling pathway used is stimulation of
adenylyl cyclase leading to increases in intracellular cAMP (18, 19), a
pathway thought to be relevant to inhibition of platelet aggregation
and vascular smooth muscle relaxation (3). However, recent evidence
indicates that IP agonists may couple to multiple signaling pathways
including activation and inhibition of adenylyl cyclase, via
Gs and Gi, respectively, stimulation of
phosphoinositide metabolism, and changes in
[Ca2+]i concentrations (20, 21). Evidence also
exists to indicate that iloprost, a stable carbacyclin analogue of
prostacyclin, can stimulate opening of ATP sensitive K+
channels resulting in hyperpolarization and relaxation of canine carotid artery (22).
Molecular cloning of the human (23, 24), mouse (25), and rat (26)
receptor confirmed that the IP is a member of the heterotrimeric G
protein-coupled receptor (GPCR) superfamily. Both the native and cloned
receptors, overexpressed in transfected cells, exhibit two classes of
binding sites, referred to as the high and low affinity binding sites,
for the radioligand [3H]iloprost (23, 27). The cloned
mouse (m) and human (h) IP can not only transduce stimulation of
adenylyl cyclase, leading to increases in cAMP, but can also induce
inositol 1,4,5-triphosphate (IP3) generation in response to
iloprost (23, 25, 28) through, as of yet undefined mechanism(s). The
effect of iloprost on IP3 generation was insensitive to
pertussis toxin treatment and was unaffected by loss of Gs
(25).
Members of the GPCR superfamily share many structural domains that
contribute to or act as determinants of receptor function (29-32). The
transmembrane domains and/or the extracellular loops provide the sites
for specific ligand binding within different receptor families, whereas
the intracellular regions, particularly the 3i loop, provide structural
determinants for specificity of receptor-G protein interaction. The
carboxyl-terminal cytoplasmic tail (C-tail) may also contribute to
receptor-G protein coupling specificity and contains many of the
sequence determinants important for agonist-induced receptor
desensitization (29, 33). Post-translational modification of GPCRs can
also play a primary role in regulating receptor function; receptor
phosphorylation on serines and threonines, located predominantly in the
C-tail, by specific protein kinases plays a key role in mediating
receptor desensitization following agonist stimulation (32). A number
of studies have indicated that the IP may undergo rapid
agonist-mediated receptor internalization and down-regulation in human
platelets and other cell types (27, 34, 35). It has been demonstrated
that the human IP is phosphorylated by protein kinase C but not by
cAMP-dependent protein kinase A (28, 36).
In eucaryotic cells, the process of isoprenylation results in the
post-translational lipid modification of proteins by either C-15
farnesyl or C-20 geranylgeranyl isoprenoids (37). These isoprenoids,
derived from the mevalonate/cholesterol biosynthetic pathway, are
attached via stable thioether linkages to specific carboxyl-terminal
cysteine residues located in distinct "isoprenylation motifs" of
proteins. Farnesyl transferase (FTase) attaches farnesyl to proteins
that end with the tetrapeptide consensus sequence CAAX, in
which C represents cysteine, A is an aliphatic amino acid, and
X is any amino acid except leucine or isoleucine (38, 39).
There are two geranylgeranyl transferases: geranylgeranyl protein
transferase I, which transfers geranylgeranyl to proteins containing a
CAAX sequence in which X is leucine or
isoleucine, and geranylgeranyl protein transferase II, which transfers
geranylgeranyl to proteins terminating in -CC or -CXC (40,
41). Among those proteins that are isoprenylated are the nuclear lamins
A and B, the 
subunits of the heterotrimeric G proteins, many
members of the Ha-Ras-related low molecular mass GTP-binding proteins, rhodopsin kinase, yeast mating factor a (37), and
inositol-1,4,5-trisphosphate 5-phosphatase (42). Recently, a novel
Ca2+-independent phospholipase A2,
cPLA2-, which shows homology to Ca2+-dependent cytoplasmic cPLA2
has been demonstrated to be isoprenylated (43). Most isoprenylated
proteins are found associated with distinct and specialized
intracellular membranes, which in many cases appears to be either
directly or partially mediated by the presence of the isoprenoid
moiety, which becomes inserted into the lipid bilayer of the targeted
intracellular membrane. It has also been proposed that the isoprenoid
group may also facilitate specific protein-protein interactions
in vivo (37); such interactions may be important in
regulating protein subcellular localization and protein function. In
the case of Ha-Ras, inhibition of isoprenylation, by use of selective
inhibitors of isoprenylation (e.g. in the presence of Statin
inhibitors of HMG-CoA reductase or FTase inhibitors) or by
site-directed mutagenesis of the CAAX motif, not only blocks p21ras membrane association but also completely prevents cell
transformation (44-47).
Analysis of the C-tail of the cloned IPs reveals the presence of an
identically conserved putative isoprenylation motif with the sequence
CSLC. In this study, we investigate whether the IP receptor is
isoprenylated and explore the functional significance of this
modification. We report that the IP is isoprenylated in mammalian
cells; studies involving in vitro assays confirm the nature
of the modification to be C-15 farnesyl. This represents the first
report of a G protein-coupled receptor to be isoprenylated. Disruption
of the isoprenylation motif by site-directed mutagenesis (C414SLC to S414SLC) yielded an IP receptor
that displayed ligand binding properties identical to those of the wild
type receptor but failed to couple to adenylyl cyclase and exhibited
diminished ability to couple to phospholipase C. In view of the central
role of prostacyclin in vascular hemostasis and the central yet
pleiotropic role of cholesterol and other metabolites of HMG-CoA
reductase in metabolism, the finding that the IP is isoprenylated and
that this modification is required for its function may have important
clinical implications relating to the use of the Statins or other
agents that interfere with isoprenylation on IP signaling.
| |
EXPERIMENTAL PROCEDURES |
Materials--
PGE2 and PGI2 were
purchased from Cayman Chemical Company. Iloprost and
[3H]iloprost (15.3 Ci/mmol) were purchased from Amersham
Pharmacia Biotech. Cicaprost was obtained from Schering AG (Berlin,
Germany). Lovastatin and Fura2/AM were purchased from Calbiochem.
Mevalonolactone and FPP were purchased from Sigma.
[3H]FPP (15-30 Ci/mmol),
[3H]mevalonolactone (15-30 Ci/mmol), and
[3H]cAMP (15-30 Ci/mmol) were purchased from American
Radiochemicals Inc. Polyvinylidene difluoride (PVDF) filters,
Taq DNA polymerase, Chemiluminescence Western blotting kit,
mouse monoclonal 12Ca5 antibody, and rat monoclonal 3F10
anti-HA-peroxidase-conjugated antibody were purchased from Roche
Molecular Biochemicals. Peroxidase-conjugated anti-mouse and anti-rat
secondary antibodies and anti-v-Ha-Ras (259) rat monoclonal antibody
were purchased from Santa Cruz Laboratories. UltraspecTM
total RNA isolation system was purchased from Biotecx Laboratories. Mouse Moloney leukemia virus reverse transcriptase (RT), ribonuclease inhibitor RNasin, Klenow fragment, RQ DNase I, and deoxyribonucleotides were purchased from Promega. Thymus RNA was purchased from
CLONTECH. All oligonucleotides were synthesized by
Genosys Biotechnologies. RPMI 1640 and minimal essential medium were
purchased from HyClone Europe. Fetal bovine serum (FBS) was purchased
from PAA Laboratories Ltd. (Kingston-upon-Thames, UK).
Subcloning and Site-directed Mutagenesis of Mouse Prostacyclin
Receptor--
The full-length cDNA coding for the mIP was
subcloned from the plasmid pBluescript SK(
):mIP (25) into the
HindIII site of pcDNA3 (Invitrogen) to generate the
plasmid pcDNA3:mIP. Site-directed mutagenesis of nucleotide 1240 to
mutate Cys414 codon to Ser414 within the
C-tail of mIP was performed by PCR mutagenesis using pBluescript
SK():mIP as template and oligonucleotide primers: 5'-dTCAGAAGCTTATGAAGATGATGGCCAGC-3' (sense primer) and
5'-dCTCAAGCTTTCAGCAGAGGGAGGAGGCAGC-3' (antisense primer,
the single mutator base is underlined). The full-length cDNA
encoding the mutant mIPCSLC-SSLC, herein designated
mIPSSLC, was subcloned into the HindIII site of
pcDNA3 to generate the plasmid pcDNA3:mIPSSLC.
To facilitate amino-terminal epitope-tagging, the full-length cDNAs
encoding mIP and mIPSSLC were subcloned in-frame into the
HindIII site of the vector pHM6 (Roche Molecular
Biochemicals) to generate the plasmids pHM6:mIP and
pHM6:mIPSSLC, respectively; resulting proteins expressed in
mammalian cells contain the 9-amino acid HA tag (48) as an
amino-terminal extension.
The cDNA coding for the C-tail of the mIP (amino acids 326-417)
was subcloned in-frame into the BamHI-HindIII
sites of the vector pMAL-C (New England Biolabs) to generate the
plasmid pMAL-C:mIPT. Site-directed mutagenesis of
Cys414-Ser414 of mIP was carried out as
described previously using pBluescript SK():mIP as template; the
cDNA coding for the mutated C-tail (amino acids 322-417) was
subcloned into the BamHI-HindIII sites of pMAL-C
to generate the plasmid pMAL-C:mIPTSSLC. The sequences of
all the plasmids were verified by double-stranded DNA sequencing using
Sequenase version 2.0 (USB Corp.).
Transfections and Stable Cell Lines--
Human erythroleukemia
92.1.3 (HEL) cells and human embryonic kidney (HEK) 293 cells, obtained
from the American Type Tissue Culture Collection, were maintained at
37 °C, 5% CO2. HEL cells were routinely cultured in
RPMI, 10% FBS. HEK 293 cells were cultured in minimal essential medium
with Earle's salts and 10% FBS. To generate a stable mammalian cell
line overexpressing the 
2 adrenergic receptor (AR), an
EcoRI subfragment encoding the full-length cDNA for the
human 2 AR was subcloned from the plasmid pTF3 (American Type Culture Collection) into the EcoRI site of
pcDNA3.1(). The resulting plasmid pcDNA3.1:2
AR was then transfected into HEK 293 cells to generate stable cell
lines. Transfection of HEK 293 cells using the calcium phosphate/DNA
co-precipitation procedure and construction of stable cell lines
overexpressing 2 AR, mIP, HA:mIP, and
HA:mIPSSLC were carried out as described previously (49,
50).
RT-PCR--
Total RNA isolated from cell lines using the
UltraspecTM RNA isolation procedure was converted to first
strand cDNA with mouse Moloney leukemia virus RT as described
previously (51). Aliquots (3.5 µl) of each first strand cDNA were
then used as templates in subsequent PCR reactions (25 µl) using
primers designed to specifically amplify the mIPSSLC
cDNA sequence but not the mIP sequence: Primer A,
5'-TCAGAAGCTTATGAAGATGATGGCCAGC-3' (sense primer), and Primer B,
5'-CTCAAGCTTTCAGCAGAGGGAGG-3' (specific antisense primer).
Amplification of cDNA from the constitutively expressed
glyceraldehyde-3-phosphate dehydrogenase gene was used as an internal
control for each experiment. PCR reactions were carried out under
standard amplification conditions, and products were analyzed by
agarose gel electrophoresis.
Cloning of the cDNA Encoding the Carboxyl-terminal Tail of
the hIP into pMAL-C--
The cDNA coding for the C-tail of the hIP
(nucleotides 890-1158 encoding amino acids 297-386) was isolated from
total RNA from human thymus (CLONTECH) using RT-PCR
from random hexamer primed first strand cDNA and the PCR
primers:
5'-dGAGGGGATCCATCGAGGGTAGGAAGGCTGTCTTCCAGCG-3' (sense primer, BamHI site underlined) and
5'-dGTGGGGATCCAAGCTTTCAGCAGAGGGAGCAGGC-3' (antisense
primer, HindIII site underlined). The latter hIP cDNA was cloned in-frame into the BamHI-HindIII sites
of pMAL-C to generate the plasmid pMAL-C:hIPT. The sequence was
verified by double-stranded DNA sequencing using Sequenase version 2.0 (USB Corp.) and was confirmed to be identical to the previously cloned hIP cDNAs (23, 24).
Expression and Purification of Recombinant Fusion Proteins from
Escherichia coli--
The full-length cDNAs for Ha-Ras and Rac2
were subcloned into the expression vector pMAL-C (52). The
corresponding maltose-binding protein (MBP) fusions, MBP:Ha-Ras,
MBP:Rac2, and MBP nonfusion proteins, were expressed in E. coli TB1. Because the MBP fusion proteins based on IP were
vulnerable to proteolytic degradation, the MBP:hIPT, MBP:mIPT, and
MBP:mIPTSSLC were expressed in the protease-deficient
strain E. coli BL21(DE3). Induction of recombinant protein
expression was carried out essentially as described previously (52),
and proteins were purified by affinity chromatography on amylose resin
essentially as described by Ausubel et al. (53).
In Vitro Isoprenylation Assays--
In vitro
isoprenylation assays were carried out using the MBP fusion proteins as
substrates and the cytosolic fraction of HEL cell protein as source of
prenyl transferases. HEL cells were grown in RPMI, 10% FBS to a
density of 4 × 105 cells/ml and then incubated in
fresh medium containing 20 µM lovastatin for 16 h at
37 °C to deplete the endogenous pool of mevalonate and its
metabolites. Cells were harvested by centrifugation (500 × g at 4 °C for 5 min), washed three times in
phosphate-buffered saline, and resuspended in 600 µl of reaction
buffer (25 mM Tris-HCl, pH 7.5, 10 mM
dithiothreitol, 10 mM MgCl2, 50 µM leupeptin, 0.1 µM pepstatin). Cells were
homogenized and centrifuged (100,000 × g for 40 min at
4 °C). The resulting soluble fraction (S100) was used as
the source of prenyl transferase activity. In vitro isoprenylation assays (50 µl) contained reaction buffer supplemented with 100 pmol of recombinant fusion protein, 2.5 µM
[3H]FPP (1 µCi), 50 µg of S100 fraction
of HEL cells. Reactions were carried out at 37 °C for 1 h and
were terminated by addition of 500 µl of ice-cold acetone, and
precipitates were collected by centrifugation (13,000 rpm for 15 min).
Reactions were analyzed by SDS-PAGE followed by fluorography.
Isoprenylation of Prostacyclin Receptor in Transfected HEK 293 Cells--
To optimize cellular uptake of mevalonolactone, stably
transfected HEK 293 cells overexpressing the HA-tagged mIP, and
mIPSSLC were co-transfected with the plasmid pMEV, a
mammalian expression vector encoding a membrane-bound mevalonolactone
transporter (54). As a positive control, HEK 293 cells were also
co-transfected with pCMV5:Ha-Ras (44) plus pMEV or, as a negative
control, with pCMV5 plus pMEV. At 36 h post-transfection,
lovastatin (20 µM) was added to deplete the intracellular
pool of mevalonate and its metabolites, and cells were further
incubated for 12 h. Thereafter, the medium was replaced with fresh
medium (3 ml of minimal essential medium, 10% FBS/10 cm dish)
containing 40 µM lovastatin and 150 µCi of
[3H]mevalonolactone (15 Ci/mmol). Following 16 h of
incubation, the cells were harvested, and aliquots (150 µg) of whole
cell protein were resolved by SDS-PAGE and electroblotted onto PVDF membranes. The remaining whole cell protein (500 µg) from each transfection was subjected to immunoprecipitation using the anti-HA 12Ca5 (1: 250 dilution) or anti-v-Ha-Ras (1: 200 dilution) antibodies as described previously (44). Immunoprecipitates were resolved by
SDS-PAGE followed by electroblotting onto PVDF membranes. Blots were
then soaked in Amplify for 30 min and exposed to Kodak Xomat XAR film
for up to 30 days to detect 3H-labeled proteins.
Thereafter, membranes were screened by immunoblot analysis using either
the anti-HA 3F10 peroxidase conjugate (to detect HA-tagged proteins) or
using the anti-v-Ha-Ras rat monoclonal antibody followed by
peroxidase-conjugated anti-rat IgG (to detect Ha-Ras); immunoreactive
proteins were visualized using the chemiluminescence detection system,
as described by the supplier.
Radioligand Binding Studies--
Transfected HEK 293 cells were
harvested by centrifugation at 500 × g at 4 °C for
5 min, washed three times in phosphate-buffered saline, and homogenized
in homogenization buffer (25 mM Tris-HCl, pH 7.5, 0.25 M sucrose, 10 mM MgCl2, 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride).
Homogenates were centrifuged (100,000 × g for 40 min at 4 °C), and the pellet fraction (P100) was resuspended
in resuspension buffer (10 mM MES-KOH, pH 6.0, 10 mM MnCl2, 1 mM EDTA, 10 mM indomethacin). Radioligand binding assays were carried
out at 30 °C for 1 h in a final assay volume of 100 µl using
35-100 µg of membrane P100 fraction/assay in the
presence of 4 nM [3H]iloprost (15.3 Ci/mmol),
for saturation binding studies, or in the presence of 0.1-160
nM [3H]iloprost for Scatchard analysis.
Nonspecific binding was determined in the presence of 0.2 mM iloprost for saturation binding studies or in the
presence of 500-fold molar excess of nonlabeled iloprost for Scatchard
binding studies. Reactions were terminated by the addition of 4 ml of
ice-cold resuspension buffer followed by filtration through Whatman
GF/C filters; filters were washed three times with resuspension buffer
(3 ml) and then subjected to liquid scintillation counting in
scintillation fluid (5 ml/filter). Radioligand binding data were
analyzed using the PRISM 2 computer program (GraphPad Software Inc.,
San Diego, CA) to determine the Kd and Bmax values. The suitability of both one-site
and two-site binding models was examined using the F-test. The level of
significance of the results of the F-test were tested to
p < 0.05.
Measurement of cAMP--
cAMP assays were carried out as
described previously (50). Briefly, cells were washed three times in
ice-cold phosphate-buffered saline; cells (approximately 1-2 × 106 cells) were resuspended in 200 µl of HEPES-buffered
saline (140 mM NaCl, 4.7 mM KCl, 2.2 mM CaCl2, 1.2 mM
KH2PO4, 11 mM glucose, 15 mM HEPES-NaOH, pH 7.4) containing 1 mM
3-isobutyl-1-methylxanthine and preincubated at 37 °C for 10 min.
Thereafter ligands (50 µl) were added; cells were stimulated at
37 °C for 10 min in the presence of 1 µM ligand
(iloprost or cicaprost) or, for dose response studies, in the presence
of 10
11 to 106 M iloprost or
cicaprost. As controls, cells were incubated in the presence of 50 µl
of HEPES-buffered saline in the absence of ligand. Reactions were
terminated by heat inactivation (at 100 °C for 5 min), and the level
of cAMP produced was quantified by radioimmunoassay (50). In separate
experiments, to examine the effect of co-transfection of
G
s on cAMP generation, stably transfected HEK293 cells
were transiently co-transfected with a plasmid encoding the subunit
of GS (pCMV:Gs; 25 µg/10-cm dish) and
pADVA (10 µg/10 cm dish). To investigate the effect of inhibition of
isoprenylation on cAMP generation, HEK 293 cells stably overexpressing
the mIP or the 2 AR were preincubated for 16 h in
the presence of 10 µM lovastatin prior to harvesting for cAMP assays. Thereafter, cells were stimulated in the presence of 1 µM iloprost, 1 µM cicaprost (HEK 293 mIP
cells), or 10 µM alprenolol (HEK 293 2 AR
cells) for 10 min. Levels of cAMP produced by ligand-treated cells over
basal stimulation, in the presence of HEPES-buffered saline, were
determined as pmol cAMP/mg cell protein ± S.E. Results are
expressed as fold stimulation relative to basal (fold increase ± S.E.). Data were analyzed using the unpaired Student's t
test. p values of less than or equal to 0.05 were considered
to indicate a statistically significant difference.
Measurement of Intracellular Ca2+
Mobilization--
[Ca2+]i measurements in HEK
293 cells were made by monitoring changes in the intensity of Fura2
fluorescence, as described previously (55). For measurement of
[Ca2+]i, cells that had been loaded with 5 µM Fura2/AM for 45 min at 37 °C were diluted to
0.825 × 106 cells/ml in modified
Ca2+/Mg2+-free Hanks' buffered salt solution
containing 20 mM HEPES, pH 7.67, 0.1% bovine serum albumin
plus 1 mM CaCl2. Fura2 fluorescence from gently
stirred cells (2-ml aliquots) was recorded using a Perkin-Elmer LS50-B
spectrofluorimeter at excitation wavelengths of 340 and 380 nm and an
emission wavelength of 510 nm. Cells were stimulated with
108 to 105 M iloprost or
cicaprost for dose response experiments or, in the case of cells
co-transfected with pCMV:Gq or pCMV:G11
(55), with 106 M iloprost or cicaprost.
Calibration of the fluorescence signal was performed in 0.2% Triton
X-100 to obtain the maximal fluorescence (Rmax)
and 1 mM EGTA to obtain the minimal fluorescence
(Rmin). The ratio of the fluorescence at 340 and
380 nm is a measure of [Ca2+]i assuming a
Kd of 225 nM Ca2+ for
Fura2/AM (56).
| |
RESULTS |
Isoprenylation of the Prostacyclin Receptor--
Comparison of the
amino acid sequences of the bovine, human, mouse, and rat prostacyclin
receptor (23-26) reveals the presence of identically conserved
putative isoprenylation CAAX motifs, each with the sequence
CSLC. To investigate whether the IP is isoprenylated, we initially
constructed stable HEK 293 cells overexpressing the mIP. To facilitate
subsequent immunochemical identification and isolation of IP, stable
cell lines overexpressing an epitope-tagged IP, containing the 9-amino
acid HA epitope tag (48) fused to the amino terminus of IP were also
generated (HEK 293 HA:mIPWT). Stable cell lines were
initially screened for IP expression using the selective radioligand
[3H]iloprost under saturating ligand concentrations (4 nM); individual isolates were then selected for further
characterization. Scatchard analysis confirmed that the presence of the
HA amino-terminal extension did not affect the ability of the receptor
to bind [3H]iloprost (Table
I). Both the native mIP and HA:mIP
exhibited two affinity binding sites for iloprost, termed the high and
low affinity binding sites, which were almost identical for both
receptors (Table I). Data obtained for both the high and low affinity
Kd values compared well with values of 1 and 44 nM previously reported for the human IP (57). However,
Namba et al. (25) previously reported a single affinity
binding site (Kd = 4.6 nM iloprost) for
the cloned mIP overexpressed in Chinese hamster ovary cells. The reason
for this apparent discrepancy between our data and the results of Namba
et al. (25) is unclear but may be due to the different host
cells used.
View this table:
[in this window]
[in a new window]
|
Table I
Radioligand binding properties of prostacyclin receptors expressed
in HEK 293 cells
Scatchard analysis was carried out on membrane fractions of HEK 293 cells stably transfected with the native prostacyclin receptor (mIP) or
with the epitope tagged wild type (HA:mIP) or mutant receptors
(HA:mIPSSLC). Data are presented as the means ± S.E.
(n = 4). The suitability of the two affinity state
binding site model was determined using the F-test to a significance
level of 0.05.
|
|
To investigate whether the IP is isoprenylated in mammalian cells, HEK
293 cells overexpressing HA:mIP were co-transfected with the vector
pMEV prior to metabolic labeling of cells in the presence of
[3H]mevalonolactone. The plasmid pMEV codes for a
mevalonate transporter and thereby enhances cellular uptake of
mevalonate in transfected cells (54). Moreover, to deplete the
intracellular pool of mevalonate and its metabolites, cells were
preincubated with and metabolically labeled in the presence of the
HMG-CoA reductase inhibitor lovastatin. As a positive control for
isoprenylation, metabolic labeling was also investigated in HEK 293 cells transiently transfected with the plasmid encoding Ha-Ras, a known
substrate for isoprenylation (58). Nontransfected HEK 293 cells or
cells transfected with pMEV alone served as negative controls. In each
case, isoprenylated proteins in the nonfractionated cells (Fig.
1A) or cells
immunoprecipitated with the anti-Ha-Ras or anti-HA antibodies used to
detect IPs (Fig. 1B) were visualized by fluorography. The
presence of the pMEV plasmid greatly enhanced the efficiency of
metabolic labeling of proteins with [3H]mevalonolactone
(Fig. 1A, compare lane 1 versus lanes 2-5) and in each case cells showed uniform metabolic labeling of endogenous isoprenylated proteins in the presence of pMEV. Efficient metabolic labeling of Ha-Ras was evident in the nonfractionated cells (Fig. 1A, lane 2) and, more especially, following
immunoprecipitation using anti-Ha-Ras antibody (Fig.
1B, lane 2).
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 1.
Isoprenylation of the prostacyclin receptor
in mammalian cells. A and B, HEK 293 cells
stably overexpressing the epitope-tagged HA:mIP (lane 4) or
HA:mIPSSLC (lane 5) were transiently
co-transfected with pMEV, encoding a mevalonate transporter. As a
positive control, HEK 293 cells were transiently co-transfected with
both pMEV and pCMV5:Ha-Ras (lane 2). Nontransfected HEK 293 cells (lane 1) or cells transfected with pMEV only
(lane 3) served as negative controls. At 36 h post
transfection, cells were incubated in the presence of 20 µM lovastatin for 12 h and were subsequently labeled
with 150 µCi of [3H]mevalonolactone (16 Ci/mmol) for
16 h in the presence of 40 µM lovastatin. Cells were
processed as described under "Experimental Procedures," and 150 µg of total cell protein was resolved by SDS-PAGE and electroblotted
onto PVDF membrane (A). The remainder of the protein (500 µg) was immunoprecipitated with either the anti-Ha-Ras antibody
(lanes 1 and 2) or the anti-HA 12Ca5 antibody
(lanes 3-5). Immunoprecipitates were resolved by SDS-PAGE
and electroblotted onto PVDF membrane (B). Membranes were
soaked in Amplify for 30 min and exposed to Xomat XAR-5 film (Kodak)
for 30 days. C and D, following fluorography,
lanes 1 and 2 from the protein blot in
B were screened with the anti-Ha-Ras antibody
(C), whereas lanes 3-5 were screened with the
anti-HA 3F10-peroxidase-conjugated antibody (D). In each
case (C and D) immunoreactive bands were detected
using the chemiluminescence detection system. The presence of the heavy
(55 kDa) and light chain (doublet of 25.7 and 26.6 kDa) of the
anti-Ha-Ras antibody were also detected because the same primary
antibody was used in both the immunoprecipitation and Western blot
analysis (C, lanes 1 and 2); however,
an additional doublet of 21.4 and 22.9 kDa was evident in Ha-Ras
transfected cells (C, lane 2). In the case of the
HA:mIP or HA:mIPSSLC (D, lanes 4 and
5, respectively), two major immunoreactive bands, of 43-45
and 66 kDa corresponding to nonglycosylated and glycosylated mIPs,
respectively, were immunoprecipitated in each case. The positions of
the molecular mass markers (kDa) are indicated to the left
of each panel, and the position of the isoprenylated 21.4-kDa protein
corresponding to Ha-Ras in A is indicated by an
arrow.
|
|
The results also clearly indicated (Fig. 1B) that an
additional isoprenylated protein with an estimated molecular mass of 44 kDa was immunoprecipitated from cells overexpressing HA:mIP (Fig.
1B, lane 4) that was absent in cells not
overexpressing IP. The cloned mIP has a predicted molecular mass of
44.7 kDa (25). Following fluorography, the identities of the
isoprenylated proteins in Fig. 1B to be those of Ha-Ras
(lane 2) and HA:mIP (lane 4), respectively, were
further confirmed by immunoscreening of those blots with the
anti-Ha-Ras and anti-HA specific antibodies (Fig. 1, C and
D, respectively). The presence of the heavy (55 kDa) and
light chain (doublet of 25.7 and 26.6 kDa) of the anti-Ha-Ras antibody
were also detected because the same antibody was used in both the
immunoprecipitation and subsequent immunoblot analysis (Fig.
1C, lanes 1 and 2); however, an
additional doublet, with approximate molecular masses of 21.4 and 22.9 kDa, representing the isoprenylated and nonisoprenylated forms of
Ha-Ras, respectively (44), was clearly evident in those cells
transfected with Ha-Ras (Fig. 1C, lane 2). In the
case of the HA:mIP (Fig. 1D, lane 4), 2 major
immunoreactive bands of 43-45 and 66 kDa corresponding to
nonglycosylated and glycosylated mIPs, respectively, were detected in
the immunoprecipitates by immunoblot analysis. Thus, because only the
44-kDa form of mIP appeared to be isoprenylated following 30 days
exposure of the [3H]fluorogram (Fig. 1B,
lane 4), it is possible that only the nonglycosylated form
of mIP may be isoprenylated. Alternatively, it is possible that failure
to detect isoprenylation of the glycosylated form of IP may be due to
the fact that the mIP overexpressed in HEK 293 cells are not fully or
efficiently processed to the mature, glycosylated, or indeed
isoprenylated form, and therefore, as indicated in Fig. 1D
(lane 4), a substantial portion of mIP is in the
nonglycosylated form. Thus, failure to detect isoprenylated forms of
the glycosylated mIP may be due to the fact that the majority of mIP
expressed is in the nonglycosylated form rather than due to the fact
that the glycosylated receptor is not isoprenylated per se.
Moreover, because of the heterogeneous nature of glycosylated proteins,
it is also possible that failure to observe isoprenylation of the
glycosylated forms of the receptor following metabolic labeling may be
due to diffusion of the weak 3H-labeled signal, further
reducing the chances of its detection by fluorography.
The Mouse and Human Prostacyclin Receptors Are Modified by C-15
Farnesyl Isoprenoids--
To establish whether the IP is modified by a
C-15 farnesyl or a C-20 geranylgeranyl isoprenoid moiety, a recombinant
form of the mIP in which its carboxyl-terminal cytoplasmic tail region (amino acids 326-417; mIPT) fused to the MBP was expressed in E. coli, and the resulting MBP:mIPT fusion protein was utilized as
substrate in in vitro isoprenylation assays. MBP or MBP
fusion proteins of Ha-Ras or Rac2 low molecular mass GTP-binding
proteins served as control substrates. Proteins were purified by
affinity chromatography on amylose resin and were used in in
vitro isoprenylation assays using [3H]FPP as
isoprenyl donor and the cytosolic fraction prepared from lovastatin-treated HEL cells as source of prenyl transferases. Similar
to Ha-Ras (Fig. 2, lanes 2),
MBP:mIPT was isoprenylated in the presence of [3H]FPP
(Fig. 2A, lane 4), thereby confirming the nature
of its modification to be C-15 farnesyl. On the other hand, neither the MBP nonfusion protein or MBP:Rac2 protein, a known substrate for geranylgeranyl protein transferase (59), was isoprenylated in the
presence of [3H]FPP (Fig. 2, lanes 1 and
3, respectively). It should be noted that additional
proteins labeled by [3H]FPP present in all lanes (Fig. 2)
correspond to endogenous FTase substrates that had accumulated in HEL
cells in their nonisoprenylated state because of treatment of those
cells with lovastatin.
View larger version (37K):
[in this window]
[in a new window]
|
Fig. 2.
Isoprenylation of the prostacyclin receptor
in vitro. A, the MBP (lane
1) or the MBP fusion proteins: MBP:Ha-Ras (lane 2),
MBP:Rac2 (lane 3), MBP:mIPT (lane 4) and
MBP:mIPTSSLC (lane 5) were incubated at 37 °C
for 1 h with 2.5 µM [3H]FPP (15-30
Ci/mmol; 1 µCi) in the presence 50 µg of S100 protein
fraction from lovastatin-treated HEL cells, as a source of prenyl
transferases. B, in separate experiments, in
vitro isoprenylation assays were carried out using the MBP
(lane 1) or the MBP fusion proteins: MBP:Ha-Ras (lane
2), MBP:Rac2 (lane 3), and MBP:hIPT (lane
4). Reactions were terminated by acetone precipitation and were
analyzed by SDS-PAGE followed by fluorography (2-day exposures). The
positions of the molecular mass markers (kDa) are indicated to the
left of each panel; the positions of the isoprenylated
proteins corresponding to: MBP:Ha-Ras (A and B,
lane 2; 62.5 kDa), MBP:mIPT (A, lane
4; 51.5 kDa), and MBP:hIPT (B, lane 4; 50 kDa) are indicated by an arrow to the right of
the protein in each case.
|
|
We also investigated whether the hIP undergoes isoprenylation. An MBP
fusion protein containing amino acids 297-386 of the C-tail of hIP
expressed and purified from E. coli served as specific protein substrate for in vitro isoprenylation assays.
MBP:hIPT was an efficient substrate for FTase, thereby establishing
that the hIP is also farnesylated (Fig. 2B, lane
4).
Disruption of Isoprenylation by Site-directed Mutagenesis--
To
establish whether the isoprenylation motif of the IP behaves as a
typical CAAX type substrate for FTase, site-directed mutagenesis was carried out in which the critical Cys residue of mIP,
within the sequence C414SLC, was mutated to a Ser
(S414 SLC). Either mutant receptors were expressed in
E. coli as a fusion protein to MBP (containing amino acids
326-417 of mIP) or the full-length, mutant mIP was expressed in
mammalian cells. The MBP:mIPTSSLC protein, expressed and
purified from E. coli, was used as substrate in in
vitro isoprenylation assays but, unlike that of MBP:mIPT, it
failed to undergo isoprenylation (Fig. 2A, lane
5). Hence, the Cys414 residue of mIP, within the
sequence CSLC, is the acceptor for the C-15 farnesyl group, and
site-directed mutagenesis of this site abolished isoprenylation of
IP.
To investigate the functional significance of isoprenylation of the IP,
stable HEK 293 cell lines overexpressing the full-length epitope-tagged
HA:mIPSSLC were constructed. Confirmation of the identity
of the stable cell lines was initially carried out by RT-PCR using
oligonucleotide primers that should detect expression of the mutant
(mIPSSLC) but not the wild type (mIP) receptor mRNA
(data not shown). Thereafter, stable cell lines were initially screened
for IP expression using the selective radioligand
[3H]iloprost under saturating ligand concentrations (4 nM); individual isolates were then selected for further
characterization. Approximately 10 individual isolates of
HA:mIPSSLC stable cell lines were screened in this way, and
no significant differences in radioligand binding between the mutant
and wild type receptors were observed (data not shown).
Comparison of metabolic labeling of HEK 293 cells overexpressing the
wild type (HA:mIPWT) and mutant (HA:mIPSSLC)
receptors in the presence of [3H]mevalonolactone
confirmed that site-directed mutagenesis of the Cys414 to
Ser414 abolished isoprenylation of mIP (Fig. 1B,
compare lanes 4 and 5). The presence and identity
of the HA:mIPSSLC protein, not detected in the
[3H]fluorogram in Fig. 1B (lane 5),
was confirmed by screening the fluorogram by immunoblot analysis using
the anti-HA monoclonal antibody (Fig. 1D, lane
5). Similar to that previously observed for the mIPWT,
two major immunoreactive bands, of 43-45 and 66 kDa, corresponding to
native and glycosylated mIPs, respectively, were detected in the
immunoprecipitates by immunoblotting.
Isoprenylation of the Prostacyclin Receptor Does Not Influence
Ligand Binding--
To investigate the potential role of
isoprenylation on IP function, we initially investigated the ligand
binding properties of the mutant HA:mIPSSLC using
[3H]iloprost as radioligand. Scatchard analysis, carried
out using membranes from HEK 293 cells stably transfected with
HA:mIPSSLC confirmed that, similar to the native (mIP) or
epitope-tagged (HA:mIP) wild type receptors, the mutant
HA:mIPSSLC contains two affinity binding sites for
iloprost. Disruption of isoprenylation had no effect on either the
affinity (Kd) or ability
(Bmax) of the IP to bind iloprost (Table I).
Values obtained for the Kd and
Bmax determinants for both the high and low
affinity sites of HA:mIPSSLC were not significantly
different than those of the wild type receptors (Table I), thereby
confirming that the presence of the isoprenoid moiety does not
influence ligand binding by IP.
Isoprenylation of the Prostacyclin Receptor Is Required for
Functional Coupling to Adenylyl Cyclase--
To explore the functional
significance of isoprenylation on receptor signaling, we investigated
the effect of site-directed mutagenesis of the CAAX motif of
IP (C414SLC to S414SLC) on its ability to
couple to adenylyl cyclase by analyzing ligand-mediated cAMP
generation. HEK 293 cells stably overexpressing the wild type
HA:mIPWT or mutant HA:mIPSSLC receptors, or as
controls, nontransfected cells were stimulated with the selective IP
agonists iloprost or cicaprost using concentrations ranging from
1011 to 106 M ligand (Fig.
3). Whereas the mIPWT cells
produced a significant, dose-dependent induction in cAMP generation in response to iloprost, the levels of cAMP production by
mIPSSLC were not significantly different from those
produced by the control, nontransfected HEK 293 cells (Fig.
3A). Similar results were obtained with the highly selective
IP agonist cicaprost (Fig. 3B). Therefore, isoprenylation of
IP is required for its functional coupling to adenylyl cyclase because
disruption of its isoprenylation by site-directed mutagenesis
significantly impairs ligand-mediated cAMP generation.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 3.
Analysis of cAMP generation. HEK 293 cells stably transfected with the epitope-tagged HA:mIP
(IPWT), HA:mIPSSLC (IPSSLC), or
nontransfected cells (HEK 293) were stimulated in the presence of
1011 to 106 M iloprost
(A) or cicaprost (B) at 37 °C for 10 min. In
each case, basal cAMP levels were determined by exposing the cells to
the vehicle under identical reaction conditions. Levels of cAMP
produced in ligand-stimulated cells relative to basal cAMP levels were
expressed as fold stimulation of basal (fold increase in cAMP ± S.E.). The data presented are the mean values of four independent
experiments, carried out in duplicate.
|
|
To investigate whether failure of the mutant IPSSLC to
stimulate adenylyl cyclase activity might be due to its altered
ability, or lack thereof, to interact with the transducing G protein,
HEK 293 cells stably overexpressing the wild type HA:mIPWT
or mutant HA:mIPSSLC receptors, or as controls, HEK 293 cells, were co-transfected with the cDNA encoding the subunit
of Gs. cAMP generation was quantified in response to
stimulation with 1 µM iloprost or cicaprost, and the
ability of the co-transfected G protein to augment ligand-induced cAMP
levels was interpreted as enhanced receptor-G protein interaction (Fig.
4, A and B,
respectively). HEK 293 HA:mIPWT cells co-transfected with
Gs showed a 1.75-fold (p 
0.001) augmentation in cAMP generation following stimulation with iloprost compared with those same cells not transfected with Gs
(Fig. 4A). On the other hand, control cells co-transfected
with Gs failed to generate significant increases in
cAMP, whereas HA:mIPSSLC cells showed only a marginal
increase in cAMP levels in response to stimulation with 1 µM iloprost relative to those cells not transfected with
Gs (p 0.05). Similar data were
obtained using cicaprost as ligand (Fig. 4B), where
co-transfection of Gs resulted in a 1.5-fold
augmentation in ligand-stimulated cAMP levels in cells expressing
mIPWT (p 0.001) but not in cells
expressing mIPSSLC (p = 0.1) or control HEK
293 cells (p = 0.1). Therefore, these studies indicate
that isoprenylation of the IP may be required for efficient interaction
of IP with Gs.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 4.
Effect of co-expression of
Gs on cAMP generation. To
investigate the effect of co-expression of Gs on
ligand-induced cAMP generation, HEK 293 cells stably transfected with
the epitope-tagged HA:mIP (IPWT), HA:mIPSSLC
(IPSSLC), or nontransfected HEK 293 cells (control)
co-transfected with the cDNA coding for Gs were
stimulated with 1 µM iloprost (A) or cicaprost
(B) at 37 °C for 10 min. Levels of cAMP generated in the
presence (+) or absence () of Gs were measured. In
each case, basal cAMP levels were determined by exposing the cells to
the vehicle under identical reaction conditions. Levels of cAMP
produced in ligand-stimulated cells relative to basal cAMP levels were
expressed as fold stimulation of basal (fold increase in cAMP ± S.E.). The data presented are the mean values of four independent
experiments, carried out in duplicate.
|
|
Isoprenylation of the Prostacyclin Receptor Is Required for
Efficient Coupling to Phospholipase C--
The main effector system
believed to be activated by the IP is adenylyl cyclase. However, recent
studies indicate that IP can also activate PLC isoforms leading to
increases in IP3 (21, 23, 25). We extended our studies to
investigate the potential role of isoprenylation of IP on its ability
to couple to PLC. Activation of PLC in HEK 293 cells stably
overexpressing the wild type (HA:mIPWT) or mutant
(HA:mIPSSLC) IPs or in control nontransfected cells was
assessed by measurement of intracellular calcium
([Ca2+]i) mobilization because of
receptor-mediated IP3 generation. Cells were stimulated
with 1 µM iloprost or cicaprost (Fig. 5, A and C) or, for
dose response experiments, with 108 to 105
M ligand (Fig. 5, B and D). Cells
expressing the wild type mIP or the mutant mIPSSLC yielded
mobilization of [Ca2+]i in response to 1 µM iloprost (Fig. 5A); however, mobilization of [Ca2+]i by mIPSSLC
(
[Ca2+]i = 183 ± 44.9 nM)
was significantly reduced compared with mIPWT
([Ca2+]i = 399 ± 75.1 nM).
In dose response experiments using iloprost (108 to
105 M), mIPWT mobilized
significantly more [Ca2+]i at the various
concentrations compared with the mIPSSLC (Fig.
5B). The control HEK 293 cells failed to mobilize
[Ca2+]i in response to any concentration of
iloprost, implying that these cells lack endogenous IPs that couple to
PLC activation. Similar results were obtained using the highly
selective IP agonist cicaprost (Fig. 5, C and D).
At 1 µM cicaprost, IPWT mobilized 145 ± 22.3 nM [Ca2+]i compared with
IPSSLC, which mobilized 66.6 ± 6.6 nM
[Ca2+]i. In dose response experiments,
mIPWT mobilized significantly more
[Ca2+]i at the various cicaprost concentrations
in comparison with the mIPSSLC (Fig. 5D). The
control HEK 293 cells did not mobilize [Ca2+]i in
response to cicaprost stimulation (Fig. 5, C and D).
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 5.
Analysis of [Ca2+]i
mobilization. HEK 293 cells stably transfected with HA:mIP
(IPWT), HA:mIPSSLC (IPSSLC), or
nontransfected cells (HEK 293) were loaded with Fura2/AM prior to
stimulation with iloprost (A and B) or cicaprost
(C and D). A and C show
changes in intracellular calcium mobilized (nM) in response
to stimulation with 1 µM ligand, which was added at the
times indicated by arrows. The data presented are representative of at
least four independent experiments. B and D show
changes in intracellular calcium mobilized (nM) following
dose response stimulation of cells using 108 to
105 M ligand. Data presented represent the
mean values ± S.E. (n = 4). Fura2 fluorescence
was recorded on a Perkin-Elmer LS50-B spectrofluorimeter at excitation
wavelengths of 340 and 380 nm and an emission wavelength of 510 nm. The
ratio of fluorescence at 340-380 nm is a measure of
[Ca2+]i mobilization assuming a
Kd of 225 nM Ca2+ for
Fura2/AM.
|
|
To further investigate the observed reduction in
[Ca2+]i mobilized by the IPSSLC
compared with the IPWT in response to IP selective
agonists, we examined the effect of co-transfecting the subunits of
the heterotrimeric G proteins Gq and G11 on
receptor-mediated signaling. Gq and G11 have
been established to mediate receptor activation of PLC isoforms (60, 61). Whereas IPWT showed mobilization of
[Ca2+]i in response to 1 µM
iloprost, there were 1.9- and 1.5-fold augmentations of
[Ca2+]i mobilization in cells co-transfected with
Gq and G11, respectively (Fig.
6A;
[Ca2+]i = 389 ± 91.1 nM by
mIPWT; [Ca2+]i = 739 ± 93 nM by mIPWT + Gq; and
[Ca2+]i = 574 ± 113 nM by
mIPWT + G11). Stimulation of cells
expressing IPSSLC with 1 µM iloprost resulted
in [Ca2+]i mobilization; however, co-transfection
of Gq and G11 did not significantly
augment mobilization of [Ca2+]i in these cells
(Fig. 6B; [Ca2+]i = 217 ± 44.9 nM by mIPSSLC;
[Ca2+]i = 213 ± 24.5 nM by
mIPSSLC + Gq; and
[Ca2+]i = 175 ± 63.1 nM by
mIPSSLC + G11). Similar results were
obtained with the selective IP agonist cicaprost (Fig. 6, C
and D) where co-transfection of G11 and
Gq resulted in 2-3-fold augmentations of
[Ca2+]i mobilized by cells expressing
IPWT but not in cells expressing IPSSLC. Thus,
these studies indicate that isoprenylation of IP may be required for
its efficient coupling to Gq and G11.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 6.
Effect of co-expression of Gq
or G11 intracellular
Ca2+ mobilization. To investigate the effect of
co-expression of a member of the Gq family of
heterotrimeric G proteins on ligand-induced intracellular calcium
mobilization, HEK 293 cells stably transfected with HA:mIP
(IPWT; A and C) or
HA:mIPSSLC (IPSSLC; B and
D) were transiently co-transfected with the cDNAs for
either Gq or G11. In each case, levels of
[Ca2+]i mobilized in the presence of
Gq or G11 were compared with those
mobilized in the absence of co-transfecting the G protein. Cells were
loaded with Fura2/AM prior to stimulation with 1 µM
iloprost (A and B) or cicaprost (C and
D), which were added at the times indicated by the
arrows. Data were calculated as changes in intracellular
Ca2+ mobilized ([Ca2+]i ± S.E., nM) and are representative of at least three
independent experiments.
|
|
Lovastatin, an Inhibitor of HMG-CoA Reductase, Reduces Prostacyclin
Receptor-mediated cAMP Generation--
To investigate whether
inhibition of isoprenylation in the presence of a member of the Statin
family of HMG-CoA reductase inhibitors had any effect on IP signaling,
HEK 293 cells overexpressing either the wild type IP (mIP), or as a
control for a nonisoprenylated receptor, the 2 AR were
preincubated in the presence (10 µM) or absence of
lovastatin for 16 h. Receptor-mediated cAMP generation was
measured in response to stimulation with the IP agonists iloprost or
cicaprost (1 µM) or with the 2 AR agonist
alprenolol (10 µM). In the absence of lovastatin, HEK 293 (mIP) cells produced between 24.4-fold (iloprost) and 23.6-fold
(cicaprost) increases in cAMP generation in response to ligand
stimulation (Fig. 7A).
Treatment of these cells with lovastatin significantly reduced specific ligand-mediated cAMP generation by 86% (iloprost; p 0.001) and 85% (cicaprost; p 0.001). On the other
hand, stimulation of HEK 293 (2 AR) cells with
alprenolol produced a 7.7-fold increase in cAMP generation that was not
significantly reduced following lovastatin treatment (p > 0.05; Fig. 7B). Thus, the signal transducing properties
of the IP were significantly more sensitive to lovastatin treatment
in comparison with the nonisoprenylated 2 adrenergic receptor.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 7.
Effect of lovastatin on IP-mediated cAMP
generation. HEK 293 cells stably transfected with the
epitope-tagged HA:mIP (A) or with the 2
adrenergic receptor (B) were preincubated with (+) or
without () lovastatin (10 µM) for 16 h prior to
harvesting. Thereafter, cells were stimulated in the presence of 1 µM iloprost, 1 µM cicaprost (A)
or 10 µM alprenolol (B) at 37 °C for 10 min. In each case, basal cAMP levels were determined by exposing the
cells to the vehicle under identical reaction conditions. Levels of
cAMP produced in ligand-stimulated cells relative to basal cAMP levels
were expressed as fold stimulation of basal (fold increase in cAMP ± S.E.). The data presented are the mean values of three or four
independent experiments, carried out in duplicate.
|
|
| |
DISCUSSION |
In this study we have established that IP is isoprenylated.
Specifically, metabolic labeling of cells in the presence of isoprenoid precursor [3H]mevalonolactone, followed by
immunoprecipitation of proteins, confirmed the presence of a 44-kDa
isoprenylated protein in HEK 293 cells stably expressing the
epitope-tagged HA:mIPWT, which was absent in cells not
overexpressing IP. The estimated molecular mass of 44 kDa for the mIP
expressed in HEK 293 cells compared well with that predicted for the
cloned mIP (25). The identity of the latter isoprenylated protein to be
that of the HA-mIP was further confirmed by secondary screening of the
immunoprecipitates by immunoblot analysis using anti-HA specific
monoclonal antibodies. An additional, diffuse protein band with a
molecular mass of 66 kDa, which was not detected in the
3H-labeled immunoprecipitate, was detected by immunoblot
analysis. The IP from a number of species each contain two closely
conserved putative N-linked glycosylation sites (23-26),
and we have confirmed that the larger 66-kDa protein represents the
glycosylated form of the mIP. Failure to observe isoprenylation of the
larger glycosylated form of IP may indicate that this form of the
receptor is not actually isoprenylated. Alternatively, it is possible
that failure to detect isoprenylation of the glycosylated form of IP is
due to the fact that the mIP overexpressed in HEK 293 cells are not fully or efficiently processed to the mature, glycosylated form, and
therefore, as indicated in Fig. 1D, lane 4, a
substantial portion of mIP is in the nonglycosylated form. Thus,
failure to detect isoprenylated forms of the glycosylated mIP may be
due to the fact that the majority of mIP expressed is in the
nonglycosylated form rather than due to the fact that the glycosylated
receptor is not isoprenylated per se.
The carboxyl-terminal four amino acids of IP, which are absolutely
conserved in the bovine, human, mouse, and rat receptors, resemble a
typical CAAX type isoprenylation motif (37). Moreover, the
presence of a Cys residue in the X position of the
CAAX predicts IP to be modified by a C-15 farnesyl group
rather than a C-20 geranylgeranyl group (38, 39). Studies involving
in vitro assays confirm that, like Ha-Ras, recombinant forms
of both the mouse and human IPs are modified by C-15 farnesyl groups.
Site-directed mutagenesis of C414 SLC to S414
SLC abolished isoprenylation of mIP both in vitro and in
transfected cells, confirming that it behaves as a typical
CAAX substrate of FTase and that Cys414 is the
acceptor amino acid.
The IP has been reported to have two affinity binding sites for the
selective ligand iloprost, known as the high and low affinity sites,
with respective Kd values reported to be 1 and 40 nM iloprost (57). However, the molecular basis of the two affinity states is currently unknown. In terms of functional
characterization of the role of isoprenylation on IP signaling, we
initially investigated its potential role in radioligand binding.
Stable HEK 293 cell lines overexpressing a HA epitope-tagged
mIPWT or mutant mIPSSLC, in which
Cys414 within the isoprenylation motif CSLC was mutated to
Ser414, were established; the presence of the HA epitope
tag had no effect on ligand binding or cell signaling by IP. Both the
IPWT and IPSSLC exhibited two affinity binding
sites for [3H]iloprost, a high affinity binding site
(Kd for HA:mIP, 0.66 ± 0.12 nM
versus Kd for HA-mIPSSLC,
0.42 ± 0.13 nM) and a low affinity binding site
(Kd for HA:mIP, 40.8 ± 2.18 nM
versus Kd for HA-mIPSSLC,
38.3 ± 2.98 nM). Values obtained for both the high
and low affinity binding sites compare well with those previously
reported for the hIP (57). In addition, there was no appreciable change
in the ability of the IPSSLC receptor to bind the ligand
[3H]iloprost relative to the IPWT.
Characterization of these properties was carried out on a number of
independent isolates of HEK 293 cells stably transfected with IPSSLC and IPWT, and no significant differences
in the respective receptor types to bind ligand were observed. Hence,
we conclude that isoprenylation does not influence the ability of IP to
bind ligand and does not represent the molecular basis of the two
affinity states of IP for iloprost.
We then examined the role of isoprenylation on IP effector coupling.
mIP demonstrated significant generation of cAMP in response to iloprost
and cicaprost compared with the control, nontransfected HEK 293 cells.
Disruption of isoprenylation by site-directed mutagenesis, generating
mIPSSLC, significantly impaired the ability of the receptor
to couple to adenylyl cyclase activation in response to the selective
ligands iloprost and cicaprost. Similar data were obtained when
PGI2 was used as selective ligand (data not shown). These
data were validated in separate experiments where cells were
co-transfected with the subunit of the heterotrimeric G protein
Gs and the ability of Gs to augment
ligand-induced increases in cAMP levels was measured. Co-transfection
of Gs resulted in a significant augmentation of cAMP
generation in cells expressing mIPWT but not in cells
expressing mIPSSLC in response to the IP ligands iloprost
or cicaprost. Thus, isoprenylation of IP is required for its functional
coupling to the effector adenylyl cyclase, and impairment of this
modification may be associated with reduced interaction of IP with
Gs.
Increasing evidence indicates that IP may couple to multiple effector
systems (21). This has been confirmed in recent studies whereby the
cloned human and mouse IPs overexpressed in a number of cell types
produced ligand (iloprost)-stimulated inositol phosphate generation
(25, 36). In our studies, we examined receptor-mediated activation of
phospholipase C by measuring mobilization of
[Ca2+]i released from intracellular stores in
response to inositol phosphate generation. IP demonstrated significant,
dose-dependent increases in mobilization of
[Ca2+]i in response to iloprost and cicaprost
compared with the control nontransfected HEK 293 cells.
Disruption of IP isoprenylation by site-directed mutagenesis
significantly reduced agonist-mediated [Ca2+]i
mobilization by mIPSSLC cells in response to
dose-dependent stimulation by iloprost or cicaprost. To
investigate whether this reduction in [Ca2+]i
mobilization was due to an altered ability of the receptor to interact
with the coupling G protein, cells expressing either the
mIPWT or mIPSSLC were co-transfected with the
subunit of Gq or G11 as
representative members of the Gq family of heterotrimeric G
proteins. Co-transfection of either Gq or
G11 resulted in a significant augmentation of [Ca2+]i mobilization in cells expressing
mIPWT in response to stimulation with the selective IP
ligands, iloprost, or cicaprost. However, when compared with the
IPWT, no significant augmentation of
[Ca2+]i mobilization by IPSSLC was
observed in cells co-transfected with Gq and
G11. Thus, these studies confirm that Gq
and G11, members of the heterotrimeric Gq family, mediate IP activation of PLC. Our data also
indicate that the presence of the isoprenoid moiety is not only
required for agonist-induced stimulation of adenylyl cyclase by IP but is also required for the efficient coupling to PLC activation. Moreover, the fact that in the case of either effector system, co-transfection of the coupling G proteins in the presence of IPSSLC, but not IPWT, failed to augment second
messenger generation indicates that the isoprenoid moiety may in some
way be involved in regulating interaction, either directly or
indirectly, with the coupling G protein. Contrary to our findings,
Smyth et al. (28) reported that deletion of a substantial
portion of the C-tail of hIP (hIP
313), including the
CAAX motif, generated a receptor that exhibited identical
coupling to adenylyl cyclase, relative to the wild type IP, but
exhibited diminished ability to couple to PLC activation. Full
understanding of this apparent discrepancy will require further investigation.
To our knowledge, this represents the first report of a G
protein-coupled receptor to be isoprenylated. Lipid modification of the
protein components of a diverse array of eucaryotic signal transduction
cascades is widely documented and has been shown to play essential
roles in conferring specific cellular function on those proteins, often
by promoting or facilitating their association with specific cellular
membranes (62). Protein lipidation appears at numerous steps in the
process of signaling through heterotrimeric G proteins. Many GPCRs are
subject to palmitoylation within their carboxyl-terminal cytoplasmic
C-tail region (63-66); it has been proposed that the presence of the
carbon 16 saturated fatty acyl group may anchor a portion of the C-tail
to the plasma membrane, thereby forming a fourth intracellular loop in
this region (29). In at least one case, that of the 2
adrenergic receptor, the palmitoylation state of the receptor is
influenced by agonist occupation (67). This finding suggests a dynamic
role for palmitoylation that has been implicated in receptor coupling
to G proteins and in receptor down-regulation, by influencing GPCR
sequestration from the plasma membrane following agonist activation
(68). Palmitoylation of GPCRs may also play a role in agonist-mediated desensitization of receptors by phosphorylation. Specific G
protein-coupled receptor kinases (GRKs) are involved in phosphorylating
the agonist-occupied receptor at or near the carboxyl terminus;
palmitoylation of GPCRs in this region can influence recognition by
GRKs (69). Whereas members of the GPCR superfamily have not been
established to be isoprenylated, several important downstream
components of GPCR signal transduction paths are, notably the subunits of the heterotrimeric G proteins and rhodopsin kinase or GRK 1 (37). Most isoprenylated proteins are localized at cell membranes and
the presence of the lipophilic isoprenoid moiety, be it a C-15 farnesyl
or a C-20 geranylgeranyl group, is generally essential for this
membrane association. Thus, in terms of the IP, the presence of the
S-farnesyl cysteine may anchor a portion of the C-tail
region in the plasma membrane, thereby forming a putative fourth
intracellular loop. However, unlike palmitoylation, which is believed
to be a transient or dynamic modification, isoprenylation is a stable,
permanent modification and thus, should the 4th intracellular loop be
formed, it would most likely represent a stable structural domain on
the receptor. Our studies indicate that the presence of the isoprenoid group, or the putative fourth intracellular loop, is required for
functional coupling of the receptor to the effector system but does not
influence ligand binding. Whether isoprenylation influences receptor
phosphorylation by the second messenger kinases (28) or receptor
desensitization by the agonist activated GRKs or receptor sequestration
(27, 34, 35) remains to be established.
For many isoprenylated proteins, in particular those containing a
CAAX motif, isoprenylation is the first in a sequential series of additional processing steps. Following isoprenylation, most
CAAX proteins are proteolytically processed by removal of the AAX residues by a microsomal endoprotease specific for
isoprenylated peptides (70). Thereafter, methylation occurs on the
carboxyl group of the isoprenylated cysteine by a specific methyl
transferase (37), increasing hydrophobicity and, thereby, further
enhancing membrane association. In addition to the carboxyl-terminal
processing initiated by isoprenylation, many CAAX proteins
require an additional signal for stable membrane association. This
second signal can be provided by palmitoylation in some proteins, as
exemplified by Ha-Ras and N-ras (58) or by a stretch of basic amino
acid residues immediately upstream of the isoprenylated cysteines, as
in the case of Ki-Ras (71). Whether the IP undergoes these additional
modifications is currently unknown. However, clarification of these
issues will be important as, should they occur, these modifications may
clearly influence receptor signaling.
In addition to their influences on promoting membrane associations, in
many well documented cases, the isoprenyl group is also believed to be
important in mediating protein-protein interactions. Members of the Rab
family of low molecular mass guanine nucleotide-binding proteins cycle
between a membrane associated GTP bound form, in their activated state,
and a nonmembrane or cytosolic GDP bound form, in their nonactivated
state. Specific interactions between the isoprenoid group of the Rab
protein and its GDP dissociation inhibitor protein maintains the Rab
protein in its inactive state until the activation signal is
transmitted, at which point the GDP dissociation inhibitor complex
delivers the G protein to the membrane (72). Farnesylation of yeast
Ras2 increases its affinity for adenylyl cyclase 100-fold (73), whereas
farnesylation of Ki-Ras is required for its efficient interac
TOP
ABSTRACT
INTRODUCTION
