| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
subunits in the pregnant human myometrium is mimicked by elevated smoothelin expression, 1
Institut für Experimentelle und Klinische Pharmakologie und Toxikologie, Abteilung für Pharmakologie, Universitäts-Krankenhaus Eppendorf, 20246- Hamburg, Germany; and
* Kreiskrankenhaus Rendsburg, Gynäkologische Abteilung, 24768 Rendsburg, Germany
2Correspondence: Institute of Experimental and Clinical Pharmacology and Toxicology, Department of Clinical Pharmacology, Friedrich-Alexander-University Erlangen-Nuremberg, Fahrstr. 17, D-91054 Erlangen, Germany. E-mail: thomas.eschenhagen{at}pharmakologie.uni-erlangen.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
s has been
suggested as a mechanism that governs uterine quiescence during
pregnancy. We quantified several components of the AC pathway in
pregnant (P, n=21) and nonpregnant human myometria (NP,
n=10). AC activity was ~sevenfold higher in P than in
NP under basal and stimulated conditions (MnCl2/GTP/GTP +
isoproterenol). In addition, relative stimulation (% of basal) by
5'-guanosine-ß
-iminotriphosphate and forskolin was twofold higher
in P. ß-Adrenoceptor density was low and unaltered in P.
Gs mRNA splice variants did not differ in P. Using
antisera against different epitopes of Gs
(carboxyl-/more amino-terminal), we found unchanged expression of
Gs short and long (45, 47 kDa) in P. Two additional
proteins in P (51, 59 kDa) were detectable only by the
carboxyl-terminal antiserum and lacked GTP binding properties. The 59
kDa protein could be identified as a recently discovered cytoskeletal
protein, smoothelin, which was 10-fold increased in P. These data
indicate that the apparent up-regulation of large Gs
species in P is mimicked by elevated smoothelin. Therefore, the
increase in AC cannot be attributed to changes in Gs- or
ß-adrenoreceptors. Epitope sharing between Gs and
smoothelin should be considered in experiments on smooth muscle
tissues.—Gsell, S., Eschenhagen, T., Kaspareit, G., Nose, M., Scholz,
H., Behrens, O., Wieland, T. Apparent up-regulation of stimulatory
G-protein subunits in the pregnant human myometrium is mimicked by
elevated smoothelin expression.
Key Words: epitope sharing • smooth muscle • Gs mRNA expression • adenylyl cyclase
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3)
.
Adrenergic stimulation can induce both contractility and relaxation.
For instance, in rat and human myometrium, stimulation of
2-adrenoceptors leads to contraction by
inhibition of adenylyl cyclase (AC) through inhibitory G-proteins
(4
, 5)
. Ligand occupation of
ß2-adrenoceptors activates AC via stimulatory
G-proteins, increases cAMP and induces relaxation. In rabbit myometria
the effect of adrenergic stimulation is modulated by the gonadal
steroids estrogen and progesterone. Under estrogen predominance,
ß2-adrenoreceptor-mediated cAMP production and
Gs levels are reduced and contraction
prevails, whereas progesterone-predominance favors relaxation
(6)
. Similar data have been reported from rat myometrium
(7)
. In human myometria, Gs was
found to be increased during pregnancy and down-regulated during
preterm and term labor, whereas members of the pertussis-sensitive
Gi family remained unchanged during pregnancy
(8
, 9)
.
The resulting changes in Gs/Gi ratio could explain the change from a
relaxed to a contractile phenotype and led to the hypothesis that
expression levels of Gs may be crucial for
myometrial AC activity and thereby for regulation of the onset of
labor. The increase in total Gs during
pregnancy was due exclusively to the appearance of immunoreactive
proteins with an apparent molecular mass of 46 and 54 in addition to
45, 47, and 58 kDa bands also detected in nonpregnant tissues. The 46,
54 and 58 kDa proteins, however, failed to be ADP-ribosylated by
cholera toxin (8)
. The present study was undertaken to
follow up on these findings and test whether
Gs isoforms are indeed up-regulated in human
myometrium during pregnancy and whether corresponding alterations in
the Gs mRNA splice variant pattern occur.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
s (AA 385–394 in
Gs; RMHLRQYELL) was from Gramsch Laboratories
(Schwabhausen, Germany). Antibody K-20 against a more amino-terminal
epitope of Gs (AA 100–119;
KEAIETIVAAMSNLVPPVE) was from Santa Cruz Biotechnology Inc. (Santa
Cruz, Calif.). The mouse monoclonal immunoglobulin G1 antibody (R4A)
against smoothelin was a kind gift of Dr. G. van Eys (University of
Limburg, Maastricht, NL).
Patients and procurement of tissue
Myometrial biopsies of 21 pregnant women with a gestation age
between 27 to 41 wk (xn=37±3.3) were obtained
during Caesarian section. Nonpregnant samples (n=10) were
taken at hysterectomies performed for benign gynecological disorders
(e.g., menorrhagia and dysmenorrhea of premenopausal women) and served
as control. The samples were frozen in liquid nitrogen and stored at
-80°C until needed. This investigation had the approval of the
research ethic committee of Hamburg.
Crude membrane preparation
Preparation was carried out on ice. About 500 mg frozen tissue
was homogenized in 5 ml ice-cold (mM) 600 sucrose/10 imidazol-HCl, pH
7.0 with a Polytron homogenizer five times for 20 s, followed by
centrifugation for 30 min at 3000 x g at 4°C in a
Beckman JA-20 rotor. The supernatant was mixed 1:2 with (mM) 160 KCl
and 20 MOPS (pH 7.4) and centrifuged 1 h at 100,000 x
g at 4°C in a Kontron TFT 28.38 rotor. Pellets were
resuspended in 300 µl membrane buffer (mM: 50 Tris, 5
MgCl2, 5 EDTA, 5 EGTA) containing aprotinin 2
µg/ml and stored at -80°C.
Adenylyl cyclase assay
AC activity was determined in crude membranes (20 µg) from
individual patients according to Salomon (10)
as described
previously (11)
, with the following modifications:
reaction mixture contained (in mM): 0.4 IBMX, (3 µCi/ml; 800
Ci/mmol), 10 creatine phosphate, 0.1 cAMP, 0.1 EGTA, 50 NaCl, 40 Tris,
5 MgCl2, 0.1 ATP, 0.2% bovine serum albumin, 0.4
mg/ml creatine kinase, and 0.1 nM [-32P]ATP
(3 µCi/ml; 800 Ci/mmol). [3H]-cAMP 35 pM (1
µCi/µl; 28.4 Ci/mmol; NEN-DuPont, Boston, Mass.) was added to
determine the recovery of [-32P]cAMP. For
Mn2+ (5 mM MnCl2)
stimulation, membranes were incubated in reaction mixture without
Mg2+. Stimulation by forskolin or isoproterenol
was performed in the presence of GTP (100 µM). The reaction was
performed at 30°C for 20 min. The assay was linear between 5 and 50
µg protein and incubation times between 5 and 30 min (not shown).
Recovery rates determined by [3H]-cAMP were
63 ± 0.3% (n=96 columns). Protein concentration was
determined according to Bradford (12)
using bovine IgG as
standard.
Radioligand binding
The density of ß-adrenoceptors was determined in crude
membranes (200 µg) by saturation binding experiments (12.5–3000 pM)
with the nonspecific
ß1/ß2 antagonist CGP
12177
[(+)-3H-4-(3-t-butylamino-2-hydroxy-propoxy)
benzimidazole-2-one, 45.4 Ci/mmol; NEN-DuPont) in the absence and
presence of 1 µM propranolol, as described previously
(13)
.
5'-Nucleotidase activity
5'-Nucleotidase activity was determined in crude membranes with
a commercially available kit according to the manufacturer’s protocol
(Sigma Diagnostics, Deisenhofen, Germany). To standardize all other
measurements except smoothelin to the plasma membrane content, the
ratio of mean 5'-nucleotidase activity of all samples in the respective
assay and the individual 5'-nucleotidase activity was used as an a
correction factor and multiplied with the respective individual values
of AC, ß-adrenoceptors, and Gs.
Immunoblotting
For determination of Gs, crude membrane
proteins were resolved on 9% Long Ranger (FMC Bioproducts, Rockland,
Minn.) polyacrylamide gels containing 6 M urea (50 µg/lane for
Gs; 100 µg/lane for smoothelin) and
electrotransferred onto nitrocellulose membranes (Schleicher/Schuell,
Dassel, Germany). Antisera 3A-150 and K-20 were used at 1:75,000 and
1:500, respectively. Anti-smoothelin antibody R4A (14)
was
used at 1:2.5 (according to personal communication of Dr. van Eys).
Membranes were incubated with an alkaline phosphatase-coupled goat anti
rabbit IgG (1:2500, Dianova, Hamburg, Germany) and developed with
NBT/BCIP (Gibco BRL, Life Technologies, Eggenstein, Germany). Intensity
of bands was quantified using Zerodescan (CSP).
Homology screening
Searches for sequence homology were performed through EMBL
database (primates library) using FASTA version 3.0t71, November 1996
(15)
.
Immunoprecipitation
Crude membranes (200 µg/sample) were pelleted by
centrifugation (10 min at 14,000 x g) and solubilized
for 80 min in 60 µl RIPA-D buffer (mM: 150 NaCl, 25 Tris/HCl, 4 EDTA,
1 PMSF, 0.1% sodium dodecyl sulfate (SDS), 1% Triton X 100, 0.5%
DOC) at 4°C. 110 microliters of RIPA-B buffer (RIPA-D without SDS)
was added and the sample was centrifuged at 10,000 x g
and 4°C for 10 min. The supernatant was incubated without and with
antibodies (3A-150, 1:1,000; K-20, 1:1,000) for 3.5 h at 4°C.
Protein A Sepharose beads [200 µl of 10% (v/v), Pharmacia Biotech,
Uppsala, Sweden] in RIPA-A buffer (mM: 150 NaCl, 25 Tris/HCl, 4 EDTA,
1 PMSF) were added, incubated (while gently shaken) overnight at 4°C,
then pelleted and washed twice with RIPA-A buffer. Precipitates were
eluted by adding sample buffer (Laemmli buffer containing: 50% [v/v]
glycerin, 10% [v/v] ß-mercaptoethanol, 7.5% [v/v] SDS, 300 mM
Tris/HCl pH 6.8, 0.25% bromphenol blue), separated on a 9% acrylamide
gel containing 6 M urea, and electrotransferred onto nitrocellulose
membranes. The following immunoblots were performed as described above.
Photoaffinity labeling of GTP binding proteins with
[-32P]GTP-azido-anilide
Synthesis and purification of the photoreactive GTP analog
[-32P]GTP-azido-anilide were performed
according to Offermanns et al. (16)
in darkness or with
red light, with the following modifications: 1 mCi
[-32P]GTP (6000 Ci/mmol; NEN-DuPont) was
lyophilized and then dissolved in 40 µl EDAC (30 mg/ml in 0.15 M MES,
pH 5.6) and incubated for 10 min at room temperature. 10 µl of
4-azidoaniline (Fluka, Buchs, Switzerland; 40 mg/ml in 1,4-dioxane) was
added and the reaction was incubated for 3 h at room temperature
with continuous shaking. Purification was performed using Waters Oasis
HLB extraction cartridges (Waters Corp., Milford, Mass.) and a
discontinuous gradient (2.8, 9, 20, and 90%) of solvent B (100 mM
triethylamine in ethanol) in solvent A (100 mM triethylamine in water).
About 25–30 fractions of 500 µl each were collected. Purity was
chromatographically controlled in aliquots (200,000 dpm) of each
fraction on PEI-cellulose using 1M LiCl as solvent, and exposure to
X-ray film for 4–8 h. Fractions containing purified
[-32P]GTP-azido-anilide were lyophilized and
dissolved in water, giving a final concentration of 1–2 µCi/µl.
Photoaffinity labeling of G-proteins was performed as follows: 100
microgram crude membranes were incubated in safe light for 3 min at
30°C in a volume of 50 µl containing 29 µl membrane buffer B (mM:
0.2 EDTA, 20 MgCl2, 60 HEPES, 20 NaCl; pH 7.4),
1µl adenosine deaminase (5µg/µl), and 10 µl isoproterenol (60
µM). 10 µl (1.5 µCi)
[-32P]GTP-azido-anilide was added and the
samples were incubated for 10 min at 30°C. Reaction was stopped on
ice, followed by centrifugation at 11,600 x g for 5
min at 4°C. The pellets were resuspended in 60 µl membrane buffer B
containing 2 mM DTT, irradiated with 100 J UV light (254 nm), and
centrifuged at 11,600 x g for 5 min at 4°C. The
pellets were solubilized in RIPA-D buffer and immunoprecipitated as
described above.
Preparation of RNA
Total RNA was extracted with the commercially available kit
RNAzol (Biotec Lab., Houston, Tex.), according to the manufacturer’s
protocol, and assessed for purity and integrity by spectroscopy and
agarose gel electrophoresis.
Generation and amplification of Gs-cDNAs by
reverse transcription-polymerase chain reaction (RT-PCR)
RT reaction was performed with 1 µg total myometrial RNA,
oligo(dT)25–30 primer (Pharmacia) and RNase
H- reverse transcriptase Superscript II 200
units (Gibco BRL). Two specific 20-mer oligodeoxynucleotide primers
(fw: 5'-AGAAGCAGCTGCAGAAGGAC-3'; rev: 5'-ACAATGGTTTCAATC
GCCTC-3') were designed for amplification of all known four
Gs splice variants
(Gs-large: 241 nt (-CAG)
[Gs1] and 244 nt (+CAG)
[Gs-2]; Gs-small:
196 nt (- CAG) [Gs-3] and 199 nt (+CAG)
[Gs-4]) of human Gs
gene (17
, 18)
, containing the sites of alternative
splicing associated with exon 3. PCR amplification and subcloning were
accomplished by standard protocols, sequences were controlled by
sequencing (Medigene, Martinsried, Germany).
RNase protection assay (RPA)
RPAs were performed with the RPA II kit (Ambion, Austin, Tex.)
as described previously (19)
with a few modifications:
antisense cRNA was synthesized in a total volume of 20 µl, using 100
µCi [32P]-UTP (800 Ci/mmol; NEN) per
reaction, giving the following sizes: Gs-1:
318 nt [241 nt cDNA+77 nt MCS]; Gs-2: 266 nt
[171 nt cDNA+95 nt MCS]; Gs-4: 276 nt [199
nt cDNA+77nt MCS]. Each sample contained 3 µg total RNA and 200,000
dpm 32P-labeled cRNA. Dried gels were exposed on
imaging plates (BAS-IP MP 2040 P; Fuji) for 18–36 h and scanned by
PhosphoImager (BAS 2000; Fuji). Spots were quantified using software
TINA 2.0 (Raytest) and Zerodescan (CSP). Quantification in absolute
values was performed as described (19)
.
Statistical analysis
All values presented are arithmetic means ±
SE. Statistical significance was estimated using Student’s
t test for unpaired observations. For AC assays and
smoothelin expression, statistical significance was estimated using
Mann-Whitney U test for unpaired nonparametric observations.
A P value of less than 0.05 was considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
; P
vs. NP: 55±5 and 130±13 U/g membrane protein), most likely reflecting
the expected reduction in sarcolem sarcoplasm relation in hypertrophied
pregnant myometria.
|
Adenylyl cyclase activity
In pregnant myometria, basal AC activity was increased by
~sevenfold (P vs. NP: 16.7±2.9 and 2.3±0.41 pmol·mg
protein-1·min-1;
Fig. 2
). In membranes from P
and NP, AC activity could be stimulated independent of
Gs by MnCl2 (5 mM) or
dependent on Gs (forskolin 3 µM; ref
20
) by the GTP analog
5'-guanosine-ß-iminotriphosphate (GppNHp,10 µM; Fig. 2
) and, to
a lesser extent, by GTP (100 µM; Fig. 3
). In general, the values obtained
with either stimulus reflected the higher AC activity in pregnant
myometria with similar interindividual variabilities. The
ß-adrenoreceptor agonist isoproterenol (10 µM) did not further
increase GTP-elevated AC activity in either NP or P (Fig. 3)
.
|
|
ß-Adrenoreceptor density
Myometrial ß-adrenoreceptors were quantified in saturation
experiments (12.5–3000 nM) with [3H]CGP 12177.
Similar receptor densities (Bmax) of 5.8±0.9 and
6.2±1.1 fmol/mg protein were obtained in nonpregnant (n=5)
and pregnant (n=10) myometrial membranes. Scatchard analysis
indicated a KD of ~80 pM (two experiments).
Gs mRNA
An RNase protection assay with two different antisense
[32P]cRNA probes was designed that allows for
quantification of the four known Gs mRNA
splice variants (Fig. 4
). A
266 nt probe (Gs-cRNA-266) differentiated
between the two large splice variants Gs-1 and
Gs-2, whereas fragments of the short
Gs-3 and Gs-4 (60+63
and 60+66 nt, respectively) were not distinguishable from a short
fragment of Gs-1 (Fig. 5A
). A 276 nt probe
(Gs-cRNA-276) detected
Gs-4 as a single protected fragment of 199 nt
(Fig. 4
and Fig. 5B
). Gs-1 mRNA was
~18-fold and 2.3-fold more abundant than
Gs-2 and Gs-4,
respectively. Gs-1- and
Gs-2 mRNA levels were identical in NP
(n=7–10) and P (n=19–21), whereas the small
Gs-4 mRNA was modestly increased in P
(6.9±0.2 vs. 5.8±0.3 amol/µg; P=0.037; Fig. 5C
).
|
|
Gs-3 could not be directly detected in
these assays. However, riboprobe Gs-cRNA-276
produced a 133 nt protected fragment that could derive from protection
by either Gs-1-,
Gs-2-, or Gs-3-mRNA.
Since the ratio of (Gs-1 +
Gs-2 +
Gs-3)/Gs-4 (199 nt
fragment) was almost identical to that of
(Gs-1 +
Gs-2)/Gs-4 (2.42 vs.
2.63 [NP] and 2.21 [P]), we conclude that
Gs-3-mRNA is not expressed in human myometrium
at considerable amounts.
Gs protein
Immunoblotting with an antibody (3A-150) directed against the
extreme carboxyl-terminal decapeptide of Gs
detected two prominent bands with an apparent molecular mass of ~45
and 47 kDa in both NP and P. In addition, antiserum 3A-150 recognized
two proteins with an apparent molecular mass of ~51 kDa and 59 kDa
(Fig. 6A
, lanes
1–4). The 59 kDa protein was the most prominent signal in P, but was
barely detectable in NP. In contrast, antibody K-20 directed against a
more amino-terminal epitope in the Gs protein
(AA 100–119), recognized only the 45 kDa
(Gs-short) and 47 kDa proteins
(Gs-long; Fig. 6A
, lanes 5–8).
When related to individual 5'-nucleotidase activity, their levels were
identical in NP and P, regardless of the antibody used (Fig. 6B
).
|
The identity of the various protein bands was further
investigated by binding experiments with the photoreactive GTP analog
[-32P]-GTP-azido-anilide and subsequent
immunoprecipitation with antibodies 3A-150 or K-20 and
SDS-polyacrylamide gel electrophoresis (PAGE). This analysis revealed
only two labeled Gs at 45 and 47 kDa in both
NP and P (Fig. 7
) and no
labeling of higher molecular mass subtypes.
|
Identification and quantification of the 59 kDa protein recognized
by antiserum 3A-150
Since a number of antisera that were all directed against the
extreme carboxyl terminus of Gs (3A-150, RM/1,
C267) had similarly recognized the higher molecular weight proteins in
pregnant myometrial tissue (9)
, an EMBL database screen
was performed for proteins with sequence homology with the
carboxyl-terminal decapeptide RMHLRQYELL of
Gs. This analysis revealed that a recently
described 59 kDa cytoskeletal glycoprotein, smoothelin (EMBL accession
number: Z49989) shares a 71% homology in a carboxyl-terminal stretch
of eight amino acids (Fig. 8A
). Accordingly,
antiserum 3A-150 immunoprecipitated smoothelin as demonstrated by a
single band of 59 kDa detected by a smoothelin antiserum (Fig. 8B
, lane 2). In contrast, immunoprecipitations without
antiserum or with antiserum K-20 were negative for smoothelin (Fig. 8B
, lanes 1, 3). In addition, the predominant 59 kDa band in
membranes from P that was recognized by antiserum 3A-150 comigrated
with the protein detected by the anti-smoothelin antibody
(rf = 0.46). Since antiserum 3A-150 had detected
the 59 kDa protein at much higher levels in P (34.0±8.4)
(n=6) vs. 3.3±1.1 AU (NP; n=7); Fig. 6A
, lanes 1–4), we quantified smoothelin levels in NP and P
by immunoblotting (Fig. 9A
). Smoothelin levels
were increased ~10-fold in pregnant myometria (P vs. NP: 11.1±0.3
and 1.1±0.1 AU; n=4; Fig. 9B
).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). An increase in cAMP formation induces
relaxation of the uterus, a decrease favors contraction. Several
prostanoids and catecholamines stimulate AC activity via stimulatory
G-proteins, and changes of the latter have been proposed as important
factors controlling uterine relaxation during pregnancy
(9)
. Studies in animals and humans revealed substantial
species differences. In guinea pig, basal and GTP-dependent stimulation
of AC activity was increased with unchanged forskolin stimulation in
pregnant myometria (21
, 22)
. A study of human myometria
found increased GppNHp, most likely enhanced
Gs-dependent stimulation of AC in pregnant
myometria, but an unchanged basal AC activity (9)
. In
contrast, our data predominantly point to an overall increase in AC
activity in pregnant myometrium. Although we observed substantial
interindividual variability, the increase in AC activity was clearly
evident under every condition tested: AC activity in pregnant
myometrium was ~sevenfold higher. AC activity in the presence of
MnCl2 (no GTP present) was also significantly
increased. Since MnCl2 is believed to stimulate
AC independent of G-proteins, these data argue for changes in the
catalytic subunit of the AC itself.
We would like to offer three fundamental differences between our study
and the previous work of Europe-Finner and co-workers (9) that may
contribute to this obvious discrepancy. First, as outlined before, we
related all measurements in membrane preparations to the plasma
membrane marker 5'-nucleotidase. We consider this to be necessary
because 5'-nucleotidase activity/mg protein was ~2.4-fold
lower in membranes from pregnant myometria, which as we believe
reflects dilution of the plasma membrane fraction due to hypertrophy
during pregnancy. In crude membrane preparations, as used in the
present and in the former study (9)
, the increase in other
particulate protein-containing material caused by the hypertrophy will
reduce the fraction of plasma membranes. Moreover, the observed
dilution factor correlates exactly with the hypertrophy factor
determined by electrophysiological measurement of membrane capacity (M.
Korth, personal communication). Second, in contrast to the previous
studies, we determined AC activity in individual membrane preparations
and not in membrane pools obtained from several patients. Taking the
substantial interindividual differences in AC activity into account,
such a pooling procedure may disguise differences otherwise detected.
Third, the former study measured AC activity in the presence of high
MgCl2 (12 mM) compared to a medium concentration
in our study (5 mM). AC activity is markedly stimulated by free
Mg2+ ions (23
, 24)
. This could
theoretically obscure differences in ‘basal’ activity.
In accordance with Europe-Finner and co-workers (9), we detected a
higher efficacy of GppNHp to stimulate AC in pregnant myometria
(1.1-fold in NP vs. 2.1-fold in P). Similarly, relative stimulation of
AC by forskolin was increased in pregnant myometrium (2.4-fold in NP
vs. 4.8-fold in P). Because the effects of GppNHp and forskolin depend
(partially) on Gs, these increases could be
taken as an argument for the role of Gs. Thus,
increased cAMP formation in pregnant myometria may be due to various
reasons. Increased activity of the AC itself appears to be involved,
but the input by Gs may be enhanced also.
We observed no stimulatory effect of isoproterenol on AC activity in
either pregnant or nonpregnant myometria. These data, which are in
apparent contrast to data reported from rat (25)
,
correlate well with the low ß-adrenoreceptor density in human
myometrium. An earlier report indicated a density of 9.4 fmol/mg
(26)
. We found 5.8 and 6.2 fmol/mg protein in nonpregnant
and pregnant myometria, respectively, which is in line with Engelhardt
et al. (27)
, who reported a density of ~6 fmol/mg. These
authors also could not detect stimulation by isoproterenol (personal
communication). The amount of ß-adrenoreceptors in pregnant human
myometrium is therefore 15- to 20-fold lower than that found in the rat
(28)
; in contrast to the rat, their expression is
apparently not up-regulated in human myometrium during pregnancy. We
know, however, that in electrophysiological recordings of cAMP-mediated
effects in smooth muscle cells of pregnant human myometrium,
isoproterenol evokes moderate cAMP-mediated effects (data not shown),
proving coupling to AC stimulation. But due to their low expression,
this effect appears to escape detection in the membrane AC assay. In
any case, the low ß-adrenoreceptor density in pregnant human
myometrium together with their down-regulation during tocolytic therapy
with ß-adrenoreceptor agonists (27)
may be one reason
for the often unsatisfactory effectiveness of these drugs in the
treatment of preterm labor (29)
.
In animal models, there are several lines of evidence for changes in
the myometrial expression of G-proteins during pregnancy. In the rat,
Gi3 decreases significantly toward term of
pregnancy, and Gi2 and Gß increase at
midgestation and return to normal levels at term (30)
. In
the same species, Gs increases at midgestation
and decreases at term (4)
. In guinea pig,
Gs is higher in nonpregnant uterus than in the
pregnant uterus near term; the opposite changes occur with
Gi and Go, whereas
Gq/11 is low in early pregnancy but increases
near term (31)
. In contrast, the expression of
Gi subtypes and Gq/11
is not altered in humans during pregnancy (9
, 32)
.
Furthermore, Go subtypes are expressed in
pregnant human myometria, but are absent in pregnant rat myometria
(32)
. Thus, caution must be taken in extrapolating these
results to humans.
Immunoblot experiments on human myometrial membranes with RM/1, an
antiserum raised against the extreme carboxyl terminus of
Gs, showed that several proteins (45, 46, 47,
54, and 58 kDa) were expressed in pregnant myometria (8)
.
The 54 and 46 kDa proteins, which were barely detectable in nonpregnant
samples, were the most prominent signals in pregnant myometria, but
only the 45 and 47 kDa proteins were ADP-ribosylated by cholera toxin.
Despite lacking this typical property of all
Gs isoforms, including
XL-Gs (33)
, all proteins
recognized by antiserum RM/1 were regarded as
Gs isoforms, and thus a marked increase of
Gs expression in the human uterus during
pregnancy was reported. The results were confirmed with additional
antisera (3A-150, C267), which were, however, raised against the same
extreme carboxyl-terminal epitope of Gs
(9)
. Analysis of the four known mRNA splice variants of
Gs 1–4 (16
, 17
, 34)
revealed an
increase in splice variants Gs-2 and
Gs-4 by ~100%, but did not match the
apparent increase in the large protein bands (35)
.
In the present investigation, we designed an RNase protection assay
that allowed quantification of Gs mRNA splice
variants in absolute amounts. We found that
Gs-1 (Gs-long without
CAG) is the splice variant predominantly expressed in human myometria.
In accordance with the data obtained by Europe-Finner et al.
(34)
, Gs-2
(Gs-long with CAG) showed a much weaker
expression in myometrial tissue than Gs-1.
Gs-4 (Gs-short with
CAG) was expressed at about half of the amount of
Gs-1, and Gs-3 mRNA
levels were below the detection limit of the RPA. Apart from the small,
but mathematically significant, increase in
Gs-4, we detected no alterations in
Gs-mRNA expression in pregnant myometria.
Since Gs-4 encodes a small
Gs protein, we conclude that the apparent
increase in large Gs isoforms has no
counterpart on the mRNA level.
There is often no quantitative correlation between mRNAs and the
respective protein(s) translated (35)
. In addition,
apparently larger isoforms of Gs may arise
from posttranslational modification, such as covalent linkage of fatty
acids (e.g., palmitoylation) to Gs proteins.
Nevertheless, all Gs isoforms should
1) be recognized by both antiserum 3A-150 (raised against
the extreme carboxyl terminus of Gs) and
antiserum K-20 (raised against a more amino-terminal epitope in
Gs that is unaffected by the known
splicing mechanisms) and 2) should be labeled by the
photoreactive GTP-analog
[-32P]GTP-azido-anilide. Only the 45 and 47
kDa bands, which represent the ubiquitously expressed isoforms
Gs-short and Gs-long,
respectively (36
37
38)
, fulfilled these criteria. When
referred to total membrane protein, levels of
Gs 45 and 47 kDa were significantly lower in
pregnant myometria. When related to the plasma membrane marker,
5'-nucleotidase activity (see above), no difference in the content of
Gs 45 and 47 kDa between nonpregnant and
pregnant myometria was observed. Taken together, we did not obtain
evidence for up-regulation of Gs either on the
mRNA or protein level.
The question arose as to the identity of the two additional strong
signals in membranes from pregnant myometria that migrated in our gel
system at 51 and 59 kDa. Taking into consideration that the
electrophoretic mobility of proteins is altered in SDS-PAGE containing
urea (39
, 40)
and that they were detected by antiserum
3A-150 in pregnant myometria in a pattern similar to that reported
before for other antisera raised against the extreme carboxyl terminus
of Gs, these proteins are most likely
identical to the 46 and 54 kDa proteins described earlier (8
, 9)
. Through a database search for proteins with an epitope
similarity with the carboxyl-terminal peptide of
Gs, we found that a recently identified human
cytoskeletal glycoprotein, smoothelin, exhibits a 71% homology in its
carboxyl terminus. Smoothelin consists of 371 amino acids and has an
apparent molecular mass of 59 kDa (14)
.
Immunohistochemistry of avian gizzard smooth muscle revealed
codistribution of smoothelin with desmin and filamen and association
with the actin cytoskeleton (41)
. Smoothelin is highly
specific for the contractile phenotype of smooth muscle. We obtained
threefold evidence that smoothelin is indeed the 59 kDa protein
recognized by the Gs antiserum 3A-150.
1) Using the monoclonal anti-smoothelin antibody R4A, we
could show that in contrast to antiserum K-20, antiserum 3A-150
immunoprecipitated smoothelin. 2) In normal immunoblots, the
59 kDa band recognized by 3A-150 strictly comigrated with the single
band recognized by the smoothelin antiserum R4A. 3)
Smoothelin was found to be ~10-fold up-regulated during pregnancy, a
value very similar to that obtained for the 59 kDa band recognized by
3A-150. In accordance with our data, several other cytoskeletal
proteins are up-regulated in human myometrium during pregnancy, such as
desmin (42)
, -actin, and myosin light chain
(43)
. Therefore, the data suggest that the elevated
smoothelin expression accounts for the additional 59 kDa protein
recognized by all antisera that are raised against the extreme carboxyl
terminus of Gs in pregnant human myometria.
The exclusive expression of smoothelin in smooth muscle and the fact
that it has only recently been identified may explain why the problem
with cross-reactivity of these antisera has not been observed before
and why smoothelin has been misinterpreted as a large isoform of
Gs. We did not yet find a correlate for the 51
kDa band also recognized by antiserum 3A-150, but it is reasonable to
suggest that this protein is another, so far unknown protein that
shares an epitope similarity with Gs.
Taken together, the present study showed AC activity to be markedly
increased in pregnant human myometrium and confirmed alterations in
Gs-dependent AC regulation. It refutes,
however, the hypothesis that this is due to alterations in the
expression (either quantity or pattern) of Gs
protein isoforms. Instead, up-regulation of the
cytoskeletal glycoprotein smoothelin was identified that, by epitope
sharing, mimics up-regulation of the formerly described 56–59 kDa
[larger Gs isoforms]. It is likely that
smoothelin is involved in myometrial hypertrophy, but unlikely that it
relates to the increase in AC activity. The reasons for the latter
remain unknown, but given that the largest difference was the overall
increase in AC activity, our results point to alterations in the
adenylyl cyclase (subtypes) itself. This would be in line with recent
data showing that 2-adrenergic agonists that
normally inhibit AC via Gi can actually stimulate
adenylyl cyclase via Gi-released ß subunits
during pregnancy (25
, 44)
.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Some of the results reported here have been published in abstract form in Naunyn Schmiedebergs Arch.
Pharmacol., Vol. 357 (Suppl.), p. R58 (1998) and Vol. 359 (Suppl.), p. R55 (1999). 
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2-adrenergic receptor-adenylate cyclase complex in the pregnant human myometrium. J. Clin. Endocrinol. Metab. 70,1299-1304[Abstract]
s in pregnancy. Endocrinology 132,2484-2490[Abstract]
s in human myometrium in term and preterm labor: a mechanism for parturition. J. Clin. Endocrinol. Metab. 79,1835-1839[Abstract]
gene. Proc. Natl. Acad. Sci. USA 85,2081-20852/ß2-adrenergic receptor-coupled adenylyl cyclase system in rat myometrium during pregnancy. J. Biol. Chem. 270,11012-11016 and Gi3 during gestation. FEBS Lett. 278,4-8[Medline]
and Go2 isoforms in human myometrium. Eur. J. Pharmacol. 288,9-15[Medline]
s -small (with serine) and Gs-large (with serine) messenger ribonucleic acid splice variants promote the increased expression of 46 and 54 kilodalton Gs protein isoforms in pregnancy and their down-regulation during labor. J. Clin. Endocrinol. Metab. 81,1069-1075[Abstract]
-subunit of Gs messenger ribonucleic acid in rat brown adipose tissue. Endocrinology 125,2328-2335[Abstract]
-subunit of the human liver stimulatory G (Gs) regulatory component of adenylyl cyclase. FEBS Lett 206,36-42[Medline]
This article has been cited by other articles:
|
|
 |
|
  T. Croci, R. Cecchi, P. Marini, C. Rouget, N. Viviani, G. Germain, F. Guagnini, Y. Fradin, L. Descamps, M. Pascal, et al. In Vitro and in Vivo Pharmacological Characterization of Ethyl-4-{trans-4-[((2S)-2-hydroxy-3-{4-hydroxy-3[(methylsulfonyl)amino]-phenoxy}propyl) Amino]cyclohexyl}benzoate Hydrochloride (SAR150640), a New Potent and Selective Human beta3-Adrenoceptor Agonist for the Treatment of Preterm Labor J. Pharmacol. Exp. Ther., June 1, 2007; 321(3): 1118 - 1126. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Frambach, T. Muller, S. Freund, S. Engelhardt, M. Sutterlin, M. J. Lohse, and J. Dietl Self-Limitation of Intravenous Tocolysis with {beta}2-Adrenergic Agonists Is Mediated through Receptor G Protein Uncoupling J. Clin. Endocrinol. Metab., May 1, 2005; 90(5): 2882 - 2887. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Rouget, M. Bardou, M. Breuiller-Fouche, C. Loustalot, H. Qi, E. Naline, T. Croci, D. Cabrol, C. Advenier, and M. J. Leroy {beta}3-Adrenoceptor Is the Predominant {beta}-Adrenoceptor Subtype in Human Myometrium and Its Expression Is Up-Regulated in Pregnancy J. Clin. Endocrinol. Metab., March 1, 2005; 90(3): 1644 - 1650. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Gogarten, K. S. Lindeman, C. A. Hirshman, and C. W. Emala Tumor Necrosis Factor Alpha Stimulates Adenylyl Cyclase Activity in Human Myometrial Cells Biol Reprod, March 1, 2003; 68(3): 751 - 757. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Schlabritz-Loutsevitch, N. Hellner, R. Middendorf, D. Muller, and J. Olcese The Human Myometrium as a Target for Melatonin J. Clin. Endocrinol. Metab., February 1, 2003; 88(2): 908 - 913. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Mittmann, C. H. Chung, G. Hoppner, C. Michalek, M. Nose, C. Schuler, A. Schuh, T. Esche |
