Influence of pregnancy on G-protein coupling to cyclase activation in guinea-pig myometrium
adenylate
S. J. Arkinstall and C. T. Jones Laboratory of Cellular and Developmental Physiology, Institute of Molecular Medicine, University of Oxford, John Radcliffe Hospital, Headley Way, Headington, Oxford 0x3 9DU (Requests for offprints should be addressed to S. J. Arkinstall, Glaxo Institute for Molecular Biology, Route des Acacias 46, 1211 Geneva 24, Switzerland) REVISED MANUSCRIPT RECEIVED 23 April 1990 ABSTRACT
regulatory factors controlling uterine activity during pregnancy remain unclear in many species. Since myometrial relaxants raise intracellular cyclic AMP, modulation of signalling pathways coupling cell-surface receptors to adenylate cyclase activation could be an important site for control. To assess the functional activity of the stimulatory GTP-binding protein Gs we have measured adenylate cyclase activation by GTP, its non-hydrolysable analogue guanosine 5\m='\-(\g=b\-\g=g\-imido)triphosphate(Gpp(NH)p), The
fluoride, forskolin and manganese in a 50 000 g membrane fraction prepared from the myometrium of non-pregnant, mid-pregnant (30\p=n-\32days) and late\x=req-\ pregnant (62\p=n-\66days) guinea-pigs (full term 67 \m=+-\2 days). While forskolin- and manganese-dependent enzyme activation was unaltered by pregnancy, maximal stimulation by Gpp(NH)p and fluoride was
INTRODUCTION
enhanced by up to 200%. Recovery of adenylate cyclase activity in the 50 000 g fraction was essentially constant at 20\p=n-\24%of the total activity throughout pregnancy, and thus cannot explain the increases observed. Since guanine nucleotides and fluoride stimulate adenylate cyclase through activating Gs,
and forskolin and manganese act at the level of the catalytic unit, these data are consistent with a pregnancy-related increase in Gs functional coupling while adenylate cyclase activity is unaltered. These observations suggest a physiological regulation of myometrial Gs activity during pregnancy which could facilitate hormonal stimulation of adenylate cyclase and contribute to uterine quiescence by increasing uterine sensitivity to relaxants. Journal ofEndocrinology (1990) 127, 15\p=n-\21
a
family
The factors maintaining uterine quiescence during pregnancy and that trigger labour at term remain unclear, particularly in species deficient in placental 17a-hydroxylase activity (Liggins, 1983; Carsten & Miller, 1987). In view of the central importance of intracellular Ca2+, prostaglandins and cyclic AMP in regulating myometrial contractions (Carsten & Miller, 1987), a major site for control is probably the transduction pathways coupling cell-surface receptors to a range of effector and second-messenger systems including ion channels, phosphoinositide-specific phospholipase C, phospholipase A2 and adenylate
cyclase (Gilman, 1987). Receptors in the plasma membrane translate hormone and neurotransmitter signals to secondmessenger systems through pathways involving
of
GTP-binding proteins (G-proteins).
exemplified by transducin which couples light-activated rhodopsin to retinal cyclic GMPphosphodiesterase activation, and by Gs and G¡ which mediate stimulation and inhibition of adenylate cyclase in response to a wide range of receptor agonist (Gilman, 1987). Recent data suggest that uterine responsiveness to hormone/neurotransmitter stimu¬ lation can be controlled by sex steroids acting to modulate G-protein coupling to their effector targets. Such a regulation is emerging as an important mech¬ anism for cell control and can occur without a change in receptor number or affinity. Hence, in rabbit myometrium, oestrogen treatment increases a,adrenoceptor-mediated phosphoinositide hydrolysis and abolishes ß-adrenoceptor-dependent cyclic AMP formation in the absence of changes in receptor den¬ sity, effector enzyme activity or substrate availability These
are
(Roberts, Riemer, Bottari et al. 1989). Oestrogen treatment is also accompanied by decreased levels of myometrial Gs (Roberts et al. 1989). In addition, pro¬ gesterone uncouples a2-adrenoceptors from adenylate cyclase inhibition in the rabbit uterus, while in the guinea-pig this treatment increases myometrial sensi¬ tivity to ß-adrenoceptor-mediated relaxation; changes which appear independent of receptor number, agonist binding affinity, adenylate cyclase levels and ATP concentration (Mammen, Handberg, Story & Pennefather, 1987; Roberts et al. 1989). Furthermore, in ovine uterus both oestrogen and progesterone increase levels of a membrane protein demonstrating characteristics of a G-protein (Flint, 1988). These observations suggest that G-proteins represent important and specific targets for sex steroids, par¬ ticularly in their action to modulate uterine contrac¬ tile state and hormone responsiveness (Liggins, 1983; Roberts et al. 1989). Changes in G-protein activity could also regulate uterine activity during pregnancy. Thus, in human
myometrium supersensitivity mediated
relaxation
to
ß-adrenoceptor-
in late pregnancy Wood & Burnstock, 1969) (Nakanishi, McLean, when receptor number is unchanged (Falkay & Kovacs, 1985), and in guinea-pig myometrium both relaxation and adenylate cyclase activation in response to ß-adrenoceptor stimulation is increased near term (Elmer, Aim & Thorbert, 1980; Hatjis, 1985). This is consistent with a well-established role for cyclic AMP as a mediator of uterine relaxation (Do Khac, Mokhtari & Harbon, 1986; Carsten & Miller, 1987), and an enhanced responsiveness to occurs
stimulation by tocolytic hormones during pregnancy could reflect an increased functional coupling by Gs. Such a modulation of Gs could be an important mech¬ anism regulating myometrial sensitivity to a range of relaxants acting through adenylate cyclase and play a key role in maintaining uterine quiescence during pregnancy. Despite this, there are only limited and apparently contradictory data on the functional activity of myometrial Gs during pregnancy. Hence, while GTP-stimulated adenylate cyclase is increased two- to threefold in myometrial membranes from late-pregnant guinea-pig, fluoride-dependent enzyme activation by Gs falls by approximately 20% at this time (Hatjis, 1985). In contrast, while activation of adenylate cyclase by guanine nucleotides is dimin¬ ished in near-term human myometrium (Litime, Pointis, Breuiller et al. 1989), in rat myometrium, adenylate cyclase activation by GTP, fluoride and cholera toxin falls sharply during the first half of ges¬ tation, followed by a full restoration of this response before parturition (Tanfin & Harbon, 1987). These observations in the rat indicate an altered functional coupling by Gs and are consistent with measurements
of cholera toxin-catalysed ADP ribosylation which suggest that uterine Gs levels are diminished at mid-gestation and subsequently increased late in pregnancy (Tanfin & Harbon, 1987). Although such apparently discrepant observations may be due, in part, to differences in species and/or gestational age, a further likely interpretation is that they reflect a pregnancy-dependent regulation of Gs. Given that the 'progesterone block' fails to account for uterine quiescence in many species during pregnancy (Porter, 1970; Liggins, 1983) and that regulation of G-protein function could be a key factor maintaining myometrial relaxation, this paper re-evaluates a potential pregnancy-dependent modu¬ lation of Gs functional coupling to adenylate cyclase activation in the guinea-pig myometrium. MATERIALS AND METHODS
Tissue
preparation Myometrium was obtained from virgin oestrous (300-500 g) or pregnant (30-32 days and 62-66 days) Dunkin-Hartley guinea-pigs (full term 67 + 2 days) which were killed by cervical dislocation and bleeding from the neck. Whole uteri were excised, trimmed free of fat and connective tissue, cut open lengthwise and stripped of fetoplacental tissues and endometrium. Whole myometria (non-pregnant) or myometrial strips running from the tubai end to the cervix and from opposite the placental implantation site (preg¬ nant) were minced with scissors and homogenized in 9 vol. ice-cold Tris-HCl buffer containing 0-5 mmol dithiothreitol (DTT)/1 at pH 7-4 with six 30-s bursts of an ultra-Turrax TP 18/10 homogenizer (Scientific Instrument Centre, London, U.K.) at 20000r.p.m. interspersed with 1 -min cooling periods on ice. After filtration through two layers of cheesecloth the crude homogenate was centrifuged for 5 min at 500 g and
2°C and then for 10 min at 10 000 # and 2°C. The supernatant was then centrifuged three times for 30 min at 50 000 g and 2°C with intermediate washing in 9 vol. and final resuspension in 1 vol. ice-cold Tris-HCl (10 mmol/1) containing 0-5 mmol DTT/1 at pH 7-4 using one 30-s burst of the homogenizer. This fraction contained 3-6 mg protein/ml and was used immediately for assay of adenylate cyclase activity.
Adenylate cyclase assay Assays were performed in triplicate
at 30 °C for 10 min in a final volume of 60 µ containing 20 µ mem¬ brane preparation (60-120 µg protein/tube) or filtered crude homogenate ( 150-200 µg protein/tube), 50 mmol Tris-HCl/1,2 mmol [a-32P] ATP/1 (15-25 d.p.m./pmol), 2-5 mmol cyclic AMP/1, 16 mmol MgCl2/l, 0-4 mmol
3-isobutylmethylxanthine/l, lOmmolKCl/l, 10 mmol phosphoenolpyruvic acid (PEP)/1, 7 pg pyruvate kinase and 3 pg myokinase. Ion chelation by ATP and PEP was calculated to result in a final Mg concen¬ tration of 8 mmol/1. MnCl2 and NaF at 10 mmol/1 (in the presence of 50 pmol/l A1C13) and guanosine 5'-(ß, imido)triphosphate (Gpp(NH)p) and forskolin at 100 pmol/l were employed for maximal activation of adenylate cyclase where indicated. Forskolin was dis¬
solved in 95% (v/v) ethanol which resulted in a final assay concentration of up to 0-95% (v/v) which failed to influence enzyme activity. Reactions were termi¬ nated by adding 120 pi 1 % (w/v) sodium lauryl sulphate containing 20 mmol ATP/1 and 20 mmol [3H]cyclic AMP/1 (100 000-500 000 d.p.m.; added as internal standard to monitor product recovery), and the mixture was then heated at 100 °C for 3 min, diluted with 770 pi water and centrifuged for 15 min at 1500g. The supernatant was decanted onto plastic mini chromatography columns (Biorad, Watford, Herts, U.K.) containing 2 ml Dowex 50 (H+) cation exchange resin (Sigma, Poole, Dorset, U.K.) which were eluted with water to obtain the [32P]cyclic AMP fraction. This was further purified by adding suc¬ cessively 0-3 ml each of 0-2 mol ZnS04/l and 0-2 mol Ba(OH)2/l, vortex mixing and centrifugation for 5 min at 1500 g and 20 °C. The Ba-Zn precipitate step was repeated once and the fraction counted for 3H and 32P by liquid scintillation spectrometry. Enzyme activity was corrected for recovery of [3H]cyclic AMP (55-60%) and is expressed as pmol
cyclic AMP/min per mg protein or per g myometrium. The limit for detection of adenylate cyclase activity was 10-15 pmol cyclic AMP/min per mg protein.
Reactions were linear with time for 20 min and with membrane protein levels up to 300 pg/assay at an
activity of 1500 pmol cyclic AMP/min per mg protein.
Newman-Keul's test between individual groups. Values were considered statistically different when P
adenylate
S. J. Arkinstall and C. T. Jones Laboratory of Cellular and Developmental Physiology, Institute of Molecular Medicine, University of Oxford, John Radcliffe Hospital, Headley Way, Headington, Oxford 0x3 9DU (Requests for offprints should be addressed to S. J. Arkinstall, Glaxo Institute for Molecular Biology, Route des Acacias 46, 1211 Geneva 24, Switzerland) REVISED MANUSCRIPT RECEIVED 23 April 1990 ABSTRACT
regulatory factors controlling uterine activity during pregnancy remain unclear in many species. Since myometrial relaxants raise intracellular cyclic AMP, modulation of signalling pathways coupling cell-surface receptors to adenylate cyclase activation could be an important site for control. To assess the functional activity of the stimulatory GTP-binding protein Gs we have measured adenylate cyclase activation by GTP, its non-hydrolysable analogue guanosine 5\m='\-(\g=b\-\g=g\-imido)triphosphate(Gpp(NH)p), The
fluoride, forskolin and manganese in a 50 000 g membrane fraction prepared from the myometrium of non-pregnant, mid-pregnant (30\p=n-\32days) and late\x=req-\ pregnant (62\p=n-\66days) guinea-pigs (full term 67 \m=+-\2 days). While forskolin- and manganese-dependent enzyme activation was unaltered by pregnancy, maximal stimulation by Gpp(NH)p and fluoride was
INTRODUCTION
enhanced by up to 200%. Recovery of adenylate cyclase activity in the 50 000 g fraction was essentially constant at 20\p=n-\24%of the total activity throughout pregnancy, and thus cannot explain the increases observed. Since guanine nucleotides and fluoride stimulate adenylate cyclase through activating Gs,
and forskolin and manganese act at the level of the catalytic unit, these data are consistent with a pregnancy-related increase in Gs functional coupling while adenylate cyclase activity is unaltered. These observations suggest a physiological regulation of myometrial Gs activity during pregnancy which could facilitate hormonal stimulation of adenylate cyclase and contribute to uterine quiescence by increasing uterine sensitivity to relaxants. Journal ofEndocrinology (1990) 127, 15\p=n-\21
a
family
The factors maintaining uterine quiescence during pregnancy and that trigger labour at term remain unclear, particularly in species deficient in placental 17a-hydroxylase activity (Liggins, 1983; Carsten & Miller, 1987). In view of the central importance of intracellular Ca2+, prostaglandins and cyclic AMP in regulating myometrial contractions (Carsten & Miller, 1987), a major site for control is probably the transduction pathways coupling cell-surface receptors to a range of effector and second-messenger systems including ion channels, phosphoinositide-specific phospholipase C, phospholipase A2 and adenylate
cyclase (Gilman, 1987). Receptors in the plasma membrane translate hormone and neurotransmitter signals to secondmessenger systems through pathways involving
of
GTP-binding proteins (G-proteins).
exemplified by transducin which couples light-activated rhodopsin to retinal cyclic GMPphosphodiesterase activation, and by Gs and G¡ which mediate stimulation and inhibition of adenylate cyclase in response to a wide range of receptor agonist (Gilman, 1987). Recent data suggest that uterine responsiveness to hormone/neurotransmitter stimu¬ lation can be controlled by sex steroids acting to modulate G-protein coupling to their effector targets. Such a regulation is emerging as an important mech¬ anism for cell control and can occur without a change in receptor number or affinity. Hence, in rabbit myometrium, oestrogen treatment increases a,adrenoceptor-mediated phosphoinositide hydrolysis and abolishes ß-adrenoceptor-dependent cyclic AMP formation in the absence of changes in receptor den¬ sity, effector enzyme activity or substrate availability These
are
(Roberts, Riemer, Bottari et al. 1989). Oestrogen treatment is also accompanied by decreased levels of myometrial Gs (Roberts et al. 1989). In addition, pro¬ gesterone uncouples a2-adrenoceptors from adenylate cyclase inhibition in the rabbit uterus, while in the guinea-pig this treatment increases myometrial sensi¬ tivity to ß-adrenoceptor-mediated relaxation; changes which appear independent of receptor number, agonist binding affinity, adenylate cyclase levels and ATP concentration (Mammen, Handberg, Story & Pennefather, 1987; Roberts et al. 1989). Furthermore, in ovine uterus both oestrogen and progesterone increase levels of a membrane protein demonstrating characteristics of a G-protein (Flint, 1988). These observations suggest that G-proteins represent important and specific targets for sex steroids, par¬ ticularly in their action to modulate uterine contrac¬ tile state and hormone responsiveness (Liggins, 1983; Roberts et al. 1989). Changes in G-protein activity could also regulate uterine activity during pregnancy. Thus, in human
myometrium supersensitivity mediated
relaxation
to
ß-adrenoceptor-
in late pregnancy Wood & Burnstock, 1969) (Nakanishi, McLean, when receptor number is unchanged (Falkay & Kovacs, 1985), and in guinea-pig myometrium both relaxation and adenylate cyclase activation in response to ß-adrenoceptor stimulation is increased near term (Elmer, Aim & Thorbert, 1980; Hatjis, 1985). This is consistent with a well-established role for cyclic AMP as a mediator of uterine relaxation (Do Khac, Mokhtari & Harbon, 1986; Carsten & Miller, 1987), and an enhanced responsiveness to occurs
stimulation by tocolytic hormones during pregnancy could reflect an increased functional coupling by Gs. Such a modulation of Gs could be an important mech¬ anism regulating myometrial sensitivity to a range of relaxants acting through adenylate cyclase and play a key role in maintaining uterine quiescence during pregnancy. Despite this, there are only limited and apparently contradictory data on the functional activity of myometrial Gs during pregnancy. Hence, while GTP-stimulated adenylate cyclase is increased two- to threefold in myometrial membranes from late-pregnant guinea-pig, fluoride-dependent enzyme activation by Gs falls by approximately 20% at this time (Hatjis, 1985). In contrast, while activation of adenylate cyclase by guanine nucleotides is dimin¬ ished in near-term human myometrium (Litime, Pointis, Breuiller et al. 1989), in rat myometrium, adenylate cyclase activation by GTP, fluoride and cholera toxin falls sharply during the first half of ges¬ tation, followed by a full restoration of this response before parturition (Tanfin & Harbon, 1987). These observations in the rat indicate an altered functional coupling by Gs and are consistent with measurements
of cholera toxin-catalysed ADP ribosylation which suggest that uterine Gs levels are diminished at mid-gestation and subsequently increased late in pregnancy (Tanfin & Harbon, 1987). Although such apparently discrepant observations may be due, in part, to differences in species and/or gestational age, a further likely interpretation is that they reflect a pregnancy-dependent regulation of Gs. Given that the 'progesterone block' fails to account for uterine quiescence in many species during pregnancy (Porter, 1970; Liggins, 1983) and that regulation of G-protein function could be a key factor maintaining myometrial relaxation, this paper re-evaluates a potential pregnancy-dependent modu¬ lation of Gs functional coupling to adenylate cyclase activation in the guinea-pig myometrium. MATERIALS AND METHODS
Tissue
preparation Myometrium was obtained from virgin oestrous (300-500 g) or pregnant (30-32 days and 62-66 days) Dunkin-Hartley guinea-pigs (full term 67 + 2 days) which were killed by cervical dislocation and bleeding from the neck. Whole uteri were excised, trimmed free of fat and connective tissue, cut open lengthwise and stripped of fetoplacental tissues and endometrium. Whole myometria (non-pregnant) or myometrial strips running from the tubai end to the cervix and from opposite the placental implantation site (preg¬ nant) were minced with scissors and homogenized in 9 vol. ice-cold Tris-HCl buffer containing 0-5 mmol dithiothreitol (DTT)/1 at pH 7-4 with six 30-s bursts of an ultra-Turrax TP 18/10 homogenizer (Scientific Instrument Centre, London, U.K.) at 20000r.p.m. interspersed with 1 -min cooling periods on ice. After filtration through two layers of cheesecloth the crude homogenate was centrifuged for 5 min at 500 g and
2°C and then for 10 min at 10 000 # and 2°C. The supernatant was then centrifuged three times for 30 min at 50 000 g and 2°C with intermediate washing in 9 vol. and final resuspension in 1 vol. ice-cold Tris-HCl (10 mmol/1) containing 0-5 mmol DTT/1 at pH 7-4 using one 30-s burst of the homogenizer. This fraction contained 3-6 mg protein/ml and was used immediately for assay of adenylate cyclase activity.
Adenylate cyclase assay Assays were performed in triplicate
at 30 °C for 10 min in a final volume of 60 µ containing 20 µ mem¬ brane preparation (60-120 µg protein/tube) or filtered crude homogenate ( 150-200 µg protein/tube), 50 mmol Tris-HCl/1,2 mmol [a-32P] ATP/1 (15-25 d.p.m./pmol), 2-5 mmol cyclic AMP/1, 16 mmol MgCl2/l, 0-4 mmol
3-isobutylmethylxanthine/l, lOmmolKCl/l, 10 mmol phosphoenolpyruvic acid (PEP)/1, 7 pg pyruvate kinase and 3 pg myokinase. Ion chelation by ATP and PEP was calculated to result in a final Mg concen¬ tration of 8 mmol/1. MnCl2 and NaF at 10 mmol/1 (in the presence of 50 pmol/l A1C13) and guanosine 5'-(ß, imido)triphosphate (Gpp(NH)p) and forskolin at 100 pmol/l were employed for maximal activation of adenylate cyclase where indicated. Forskolin was dis¬
solved in 95% (v/v) ethanol which resulted in a final assay concentration of up to 0-95% (v/v) which failed to influence enzyme activity. Reactions were termi¬ nated by adding 120 pi 1 % (w/v) sodium lauryl sulphate containing 20 mmol ATP/1 and 20 mmol [3H]cyclic AMP/1 (100 000-500 000 d.p.m.; added as internal standard to monitor product recovery), and the mixture was then heated at 100 °C for 3 min, diluted with 770 pi water and centrifuged for 15 min at 1500g. The supernatant was decanted onto plastic mini chromatography columns (Biorad, Watford, Herts, U.K.) containing 2 ml Dowex 50 (H+) cation exchange resin (Sigma, Poole, Dorset, U.K.) which were eluted with water to obtain the [32P]cyclic AMP fraction. This was further purified by adding suc¬ cessively 0-3 ml each of 0-2 mol ZnS04/l and 0-2 mol Ba(OH)2/l, vortex mixing and centrifugation for 5 min at 1500 g and 20 °C. The Ba-Zn precipitate step was repeated once and the fraction counted for 3H and 32P by liquid scintillation spectrometry. Enzyme activity was corrected for recovery of [3H]cyclic AMP (55-60%) and is expressed as pmol
cyclic AMP/min per mg protein or per g myometrium. The limit for detection of adenylate cyclase activity was 10-15 pmol cyclic AMP/min per mg protein.
Reactions were linear with time for 20 min and with membrane protein levels up to 300 pg/assay at an
activity of 1500 pmol cyclic AMP/min per mg protein.
Newman-Keul's test between individual groups. Values were considered statistically different when P
