INTRODUCTION

Data on hormonal regulation of myometrial contractility during the course of pregnancy implicate adenylyl cyclase (AC)¹ stimulatory pathways as a key component that may affect the degree of intracellular cAMP generation and consequently the contractile state of the uterus. Because one of the major sites of control of the biochemical events leading to uterine relaxation during normal pregnancy lies at the AC/cAMP system, the identification of AC isoforms in the pregnant myometrium is essential in understanding the influence exerted by the regulatory external signals (neurotransmitters and hormones) acting via G protein-coupled receptors. Hormonal control of AC activity is brought about by receptor-catalyzed activation of heterotrimeric G proteins that in turn regulate the cyclases by the release of  or subunits or kinase activation. Recent studies have revealed an unexpected diversity of G protein-regulated AC by identifying nine distinct AC cDNA from various mammalian tissues (1-5). All of these isoforms of AC differ in their tissue distribution and their regulatory properties, providing a mode for different cells to respond diversely to similar external stimuli. Among all of the AC identified so far, the highly similar types II, IV, and VII form the largest known subfamily. Types II and IV share the property of being highly stimulated by subunits of G_i/G_o inhibitory proteins in the presence of activated G_s (6, 7). These AC are also influenced by phosphorylation with protein kinase C (8-11). Types V and VI AC, a two-member subfamily, are inhibited directly by low levels of Ca²⁺ (2, 12), whereas AC I and VIII are regulated positively by Ca²⁺-calmodulin (13, 14). On the other hand, AC III can be phosphorylated by a calmodulin-dependent protein kinase II in response to the elevation of intracellular Ca²⁺ which results, in vivo, in a 50% inhibition of hormone-stimulated enzyme (15). The novel ninth AC is quite distinct from all of the other known AC subfamilies, and it is not affected by G proteins or Ca²⁺ (5). Thus, in vivo, when a cell type or tissue expresses various isoforms of AC one may expect that different physiological situations trigger different responses to the same external stimuli depending not only on the type of the receptor and G protein involved but also on the type of adenylyl cyclase to which they are coupled.

As an initial approach to this issue in human and rat myometrium, we first characterized the type(s) of AC involved in myometrial signaling in both species during pregnancy. Then, because AC types II and IV are present in human and rat myometria, we examined the possibility that the positive input to the myometrial AC system emanating from the activated ₂-adrenergic receptor (AR) coupled to Gi₂/Gi₃ proteins as described previously (16) involves G_i. This was performed by using transducin- (17) or a synthetic peptide corresponding to the binding site of in the AC II (18, 19) which tie up the G_i released by ₂-AR activation. Finally, because AC types V and VI are also strongly expressed in pregnant myometria, we investigated whether submicromolar concentrations of Ca²⁺ attenuate forskolin-stimulated AC activity. Elevation of intracellular Ca²⁺ is precisely the situation that occurs in response to external stimuli in the pregnant myometrium at the time of delivery. Data reported here provide an explanation for the switch in the stimulatory versus inhibitory input to myometrial AC reported at mid-pregnancy or term.

EXPERIMENTAL PROCEDURES

The [³H]cAMP assay system and -³⁵-S-dATP (3,000 Ci/mmol) were purchased from Amersham Corp. [³H]Forskolin (31 Ci/mmol) was from NEN Life Science Products. Isoproterenol, clonidine hydrochloride, and all other reagents of the highest grade available were obtained from Sigma. QEHA peptide with 95% purity was synthesized by Eurogentec (Belgium). Transducin- was kindly donated by Dr. D. T. Wieland. SKEE peptide and the cDNAs encoding for AC types V and VI were kindly provided by Drs. R. Iyengar and J. P. Pieroni. Other types of AC cDNAs were kindly provided as follows: type I, Dr. W. J. Tang; types II and III, Dr. R. R. Reed; type IV, Dr. A. Gilman; type VII, Dr. S. M. Lanier; type VIII, Dr. J. Krupinsky; type IX, Dr. R. Premont.

Sprague-Dawley rats (250-300 g) were obtained from Iffa Credo (L'Arbresle, France). The females were caged with males overnight, and successful mating was determined by the presence of spermatozoa in the vaginal smear (day 1 of pregnancy). In our breeding colony, parturition occurs between 12 and 19 h on day 22 for 80% of rats (20). Pregnant rats were sacrificed by cervical dislocation at mid-pregnancy (days 14-15) or term (day 22, 18 h). The uterine horns were quickly isolated, cut open lengthwise, and the fetoplacental units removed. The myometrium was then freed of adherent endometrium by scraping with a glass slide. Tissues were either used freshly or stored in liquid nitrogen until required.

Samples of myometrium from preterm (42 weeks gestation) nonlaboring women were taken from the upper border of the uterine incision during elective cesarian sections indicated for caephalo-pelvic disproportion. Tissue samples were frozen and stored in liquid nitrogen until required. This investigation had the approval of the Ethics Committee of INSERM.

Myometrial plasma membranes were prepared from freshly isolated tissues. Briefly, myometrial tissues were homogenized in 10 volumes of ice-cold 10 mM Tris, 250 mM sucrose (pH 7.4) by means of a Polytron (3 × 10 s). The homogenates were filtered through a double layer of gauze and centrifuged at 20,000 × g for 10 min at 4 °C. Pellets were then resuspended in 50 mM Tris-HCl, 10 mM MgCl₂ buffer (pH 7.4). Protein concentration was determined by the method of Schacterle and Pollack (21) with bovine serum albumin as the standard. 5-Nucleotidase was used as a plasma membrane marker enzyme. The specific activity of 5-nucleotidase was constant, indicating that all membrane samples were similarly enriched in plasma membranes.

The binding of [³H]forskolin to freshly prepared membranes of pregnant day 14 and day 22 myometria was determined by incubating membranes (1 mg of protein/tube) in a total volume of 200 µl of 50 mM Tris-HCl buffer (pH 7.5), 10 mM MgCl₂, and increasing concentrations (10-150 nM) of [³H]forskolin for 60 min at 0 °C, according to the methods of Nelson and Seamon (22) and Barber (23). Nonspecific binding was determined in the presence of 20 µM unlabeled forskolin. In some assays 0.1 mM Gpp(NH)p, the nonhydrolyzable guanosine triphosphate analog, was added to promote the formation of G_s·AC complex. Incubation was terminated by adding 5 ml of Tris-HCl ice-cold buffer followed by rapid filtration over GF/C glass fiber filter (Whatman, Clifton NJ). The filters were washed four times with 5 ml of incubation buffer and placed in scintillation vials with liquid scintillation. Radioactivity was counted in a spectrometer 1214 Rack-beta (LKB, Turku, Finland). Specific [³H]forskolin binding was calculated as the difference between total binding and nonspecific binding. B_max and K_d were determined from regression analysis of Scatchard plots.

For the first assays described below, AC activity in myometrial cell membranes was measured as described previously (24), using radioimmunoassay. Briefly, 20 µl of membrane fraction (2.5 mg of protein/ml) was incubated in 100 µl of a medium containing 50 mM Tris (pH 7.4), 5 mM MgSO₄, 5 mM creatine phosphate, 12 units of creatine phosphokinase, 0.5 mM ATP, 1 mM isobutylmethylxanthine, and 0.1% bovine serum albumin, for 10 min at 30 °C. Reactions were stopped by transferring the tubes to a boiling water bath for 3 min. After centrifugation at 2,500 × g for 30 min at 4 °C, cAMP accumulation in the supernatant was determined using the Amersham cAMP radioimmunoassay system. AC activity was assayed in response to 0.1 mM GTP in the absence or in the presence of 0.1 mM forskolin and to 10 µg/ml cholera toxin. When present in the assay, isoproterenol was used at a concentration of 0.1 mM which we have found to be optimal for promoting ₂-AR stimulation of AC activity in the presence of 0.1 mM GTP (24). Clonidine was used at a concentration of 0.1 mM which maximally potentiates stimulated isoproterenol AC activity (16). Purified  subunit of bovine retinal transducin, QEHA peptide, or SKEE peptide (18) at concentrations between 25 and 100 nM was incubated with membranes in the presence of isoproterenol plus clonidine at concentrations indicated above.

When measuring the effect of free Ca²⁺, AC activity was evaluated using the method described by Salomon et al. (25). Myometrial membranes were first washed three times in the presence of the cation chelator EGTA (1 mM) to remove Ca²⁺ and calmodulin from membrane preparations. 50 µg of membranes, resuspended in 50 mM Na-Hepes (pH 8.0), was then incubated for 10 min at 30 °C in a final volume of 100 µl containing 50 mM Na-Hepes (pH 8.0), 4 mM MgCl₂, 100 mM NaCl, 1 mM ATP, 0.2 mM cAMP, 0.2% bovine serum albumin, 60 mM creatine phosphate, 10 units/ml creatine phosphokinase, 1 mM isobutylmethylxanthine, [-³²P]ATP (3 × 10⁶ cpm/tube), 0.2 mM EGTA in the presence of 0.1 mM forskolin, 0.1 mM GTP, and increasing concentrations of CaCl₂ to yield 0.04-550 µM free Ca²⁺ (26). cAMP was isolated by sequential chromatography on columns containing Dowex AG 50W-X4 resin and alumina as described previously (25). The concentrations of free Ca²⁺ were calculated using an iterative computer program adapted from Fabiato and Fabiato (26).

Total RNA was extracted from myometrium and other tissues (brain, lung, kidney) by the cesium trifluoroacetate gradient method (27) as described previously (28). Poly(A)⁺ RNA was purified from total RNA using prepacked oligo(dt)-cellulose columns (Pharmacia Biotech Inc.). Ten µg of poly(A)⁺ RNA was then subjected to electrophoresis on 1% agarose, 3% formaldehyde gels followed by transfer to GeneScreen Plus membrane (NEN Life Science Products). The membrane was dried at 80 °C for 45 min, processed, and hybridized with radiolabeled probes as described previously (16). All of the cDNA probes were labeled by random priming with [-³²P]dATP to a specific activity averaging 10⁹ dpm/µg. Unbound radioactivity was separated by gel filtration using Sephadex G-50 DNA grade (Pharmacia). The membranes were prehybridized at 42 °C for 4 h in a buffer containing 45% formamide, 4 × SSC, 5 × Denhardt's solution, 100 mM NaH₂PO₄ (pH 6.6), 0.5% SDS, and 75 µg/ml denatured salmon sperm DNA and then hybridized overnight at 42 °C in the same medium containing 10% dextran sulfate and 10⁶ cpm/ml denatured cDNA probes. After hybridization, membranes were washed with a final stringency of 0.2 × SSC, 0.2% SDS at 55 °C and exposed to Kodak X-Omat AR films at 80 °C for 1-5 days. The intensities of the bands were determined by scanning densitometry. Size estimates of the RNA species were established by comparison with a RNA ladder (Life Technologies, Inc.). Hybridization with an oligo(dt)_12-18 probe (Pharmacia) was used as an internal control to estimate the amount of poly(A)⁺ RNA loaded in each well and transfer efficiency.

Pregnant rats were anesthetized, perfused with 0.9% NaCl, and fixed with 0.1 M phosphate buffer containing 1% picric acid and 4% paraformaldehyde (pH 7.4). Brain and uterus were quickly removed, fixed again for 90 min at room temperature, and then rinsed overnight at 4 °C in 0.1 M phosphate buffer (pH 7.4) containing 15% sucrose. Sections 20 µm thick were cut on a cryostat, thaw-mounted onto silane-coated glass slides, and stored at 80 °C until hybridization. The type II AC probe (5-GGATGCGTACAGCCAGGTCCTGTTTCAGAGACCAGGC-3) is an antisense oligonucleotide with a sequence complementary to bases 2352-2388 which has very low homology with other types of AC (29). Hybridization was performed as described by Houdeau and Boyer (30). Briefly, frozen sections were warmed, prehybridized at 40 °C for 15 and 45 min in a buffer containing 4 × SSC and 1 × Denhardt's solution, and washed twice in 4 × SSC. Slides were then treated sequentially with 0.25% acetic anhydride (in 1.3% triethanolamine, pH 8) and a graded alcohol series. Dried sections were hybridized at 40 °C in hybridization buffer (4 × SSC, 45% deionized formamide, 0.5% SDS, 1 × Denhardt's solution, 1% dextran) containing -³⁵S-dATP-labeled probe. To remove unhybridized probe, slides were washed twice in 2 × SSC and 1 × SSC at room temperature and then twice in 0.5 × SSC at 40 °C. Slides were dehydrated in a graded alcohol series, air dried, and exposed for 4 weeks to Ilford K-5 emulsion at 4 °C. They were then developed in D-19 solution and counterstained with cresyl violet acetate.

Data are presented as means ± S.E. Statistical analyses were performed by using Student's t test and analysis of variance for comparison of multiple means with a control group. A probability of less than 0.05 was assumed to denote a significant difference.

RESULTS

Types of AC expressed in rat and human myometria were determined by RNA blot analysis using specific cDNA probes for enzyme types I-IX. Brain, which expresses each enzyme type, was used as a positive control for hybridization experiments. Rat lung (AC types II-VI and type IX) or kidney (AC types III, IV, VI, and IX), NIH-3T3 fibroblasts (AC type VI), and DDT₁-MF2 smooth muscle cells (AC types VI-IX) (for review, see Refs. 3 and 11) were also used as negative or positive controls. Hybridization of poly (A)⁺ RNA with the different AC cDNA probes indicated that AC types II, III, IV, V, VI, VII, and IX are present in the pregnant rat myometrium with sizes in agreement with those reported previously in control tissues or cells (Fig. 1). Concerning AC type V, the four bands identified at 7.4, 6.4, 5.8, and 5 kb were also previously described in the brain (5). In the pregnant human myometrium, RNA blot analysis indicated similar expression, albeit to different levels, of enzyme types with the exception of AC type VII, which was undetectable (Fig. 1). None of the myometrial samples of either species expressed types I and VIII enzyme (Fig. 1). In rat myometrium, quantitative determinations of each enzyme type transcripts after normalization of data with [-³²P]oligo(dt) probe revealed only minor changes in the amount of specific AC mRNA from mid-pregnancy to term (data not shown). These observations suggest the possibility that the levels of the different AC types identified presently were not altered significantly during this period of pregnancy.

[View Larger Version of this Image (54K GIF file)]

In preliminary experiments we have established that specific binding of [³H]forskolin to myometrial membranes in the presence of MgCl₂ was saturable at 100 nM [³H]forskolin and reversible at high concentrations of unlabeled forskolin. Specific binding represents 50% of total binding. The time required for half-maximal binding was 10 min, and binding equilibrium at 0 °C was reached at 60 min (data not shown). In the absence of Gpp(NH)p, specific [³H]forskolin binding to myometrial membranes was similar at mid-pregnancy and term with no change of K_d values (Table I). Gpp(NH)p substantially increased both [³H]forskolin binding sites number (× 2-fold) and K_d value at mid-pregnancy, whereas no change was observed at term. These results strongly suggest differences in the number and/or the type of G_s/catalytic subunits complexes which could be formed at both stages of pregnancy.

Table I. [3H]Forskolin binding parameters in rat myometrial plasma membranes on day 14 of pregnancy and term Membranes were incubated with Tris-MgCl2 buffer and increasing concentrations of [3H]forskolin in the absence or presence of 0.1 mM Gpp(NH)p as described under "Experimental Procedures." Binding parameters of [3H]forskolin were determined from regression analysis of the Scatchard plots. Data are the means ± S.E. from the number of experiments indicated in parentheses. [3H]Forskolin bound Kd values fmol/mg protein nM  Gpp(NH)p   Day 14 31  ± 5 (7) 68  ± 11 (7)   Term 41  ± 8 (5) 71  ± 10 (5) +Gpp(NH)p   Day 14 62  ± 7a (4) 177  ± 45a (4)   Term 34  ± 5 (4) 62  ± 10 (4) a Statistically significant difference from Gpp(NH)p (p < 0.05; nonpaired Student's t test).

As shown in Table II, basal AC activities in mid-pregnant and term myometria were similar. Although GTP increased myometrial AC activity at both stages in a dose-dependent manner, there was significantly more stimulation in myometria taken at mid-pregnancy than from myometria at term (see V_max in Table II and Fig. 2). Sensitivity of AC, evaluated by half-maximal stimulation, was not different between these two stages (mean value for EC₅₀ = 69.0 ± 1.5 nM). Stimulation with cholera toxin or forskolin also elevated myometrial AC activity but again to a lesser degree at term (Table II). Similar results were found with Gpp(NH)p (data not shown). In contrast, when the activity of the catalytic subunit of AC was evaluated in the presence of Mn²⁺ (10 mM) and GDPS (300 µM) no decrease was observed at the last stages of pregnancy.² Altogether, these results indicate that the reduced AC stimulability at term is probably caused, at least in part, by a lower level of functional G_s rather than by changes in AC expression. Stimulation of AC was also investigated after activation of the ₂-AR/G_s signaling cascade. Maximal response to the agonist isoproterenol was observed at 0.1 mM and was substantially higher in myometria taken at mid-pregnancy (Table II).

Table II. Effects of GTP, forskolin, cholera toxin, and isoproterenol on adenylyl cyclase activity in myometrial plasma membranes of day 14 and term pregnant rat AC activity was assayed by radioimmunoassay as described under "Experimental Procedures" in the absence or presence of 0.1 mM GTP, 0.1 mM forskolin, 0.1 mM isoproterenol, or 10 µg/ml cholera toxin. Results are expressed as total pmol of cAMP/mg of protein/10 min. Data are the means ± S.E. of the number of determinations indicated in parentheses. Stimulators Day 14 of pregnancy Term None (basal activity) 71  ± 4 (8) 69  ± 3 (3) GTP 137  ± 3 (19) 103  ± 2a (19) Forskolin + GTP 254  ± 7 (8) 192  ± 9a (6) Cholera toxin + GTP 280  ± 10 (10) 213  ± 9a (10) Isoproterenol + GTP 175  ± 15 (10) 136  ± 6a (8) a Statistically significant difference between the two stages of pregnancy (p < 0.05; nonpaired Student's t test).

[View Larger Version of this Image (15K GIF file)]

Effects of Transducin-, QEHA Peptide, and SKEE Peptide on Clonidine Potentiation of Isoproterenol-stimulated Adenylyl Cyclase Activity of Myometrial Membranes at Mid-pregnancy

We reported previously (16) that activation of ₂-AR/G_i signaling cascade by a micromolar concentration of clonidine resulted in a potentiation of ₂-AR-stimulated AC activity in myometrial membranes of mid-pregnant rat. Because myometrium expresses AC II and IV, we then asked whether this positive input emanating from the activated ₂-AR involves G_i complexes. To test this hypothesis, we first incubated the myometrial membrane fraction taken at mid-pregnancy, costimulated with isoproterenol and the ₂-AR-agonist clonidine, in the presence of transducin-, a scavenger of G. As illustrated in Fig. 3, activation of ₂-AR by 0.1 mM clonidine significantly potentiates the isoproterenol-stimulated AC activity by 1.6-fold. In the presence of increasing concentrations of transducin- this potentiation is blocked in a dose-dependent manner, suggesting a role for G complexes. To confirm this result, we examined the capability of the synthetic peptide QEHA, encoding residues 956-982 in the C2a region of AC II, which interacts with G (18, 19), to tie up endogenous G and, thus, to prevent potentiation of AC mediated by ₂-AR pathway. As demonstrated in Fig. 4, QEHA peptide was also efficient to block in a dose-dependent manner the ₂-AR-mediated potentiation of isoproterenol-stimulated AC activity of mid-pregnant myometrium. To explore further the specificity of the effect of the QEHA peptide, we performed similar experiments in the presence of the SKEE peptide (18) corresponding to the cognate region of AC III. As shown in Fig. 4, the SKEE peptide did not significantly affect AC activity in response to ₂-AR stimulation. Moreover, basal and isoproterenol-stimulated activity of myometrial AC were not altered by the various concentrations of the QEHA peptide (Table III). Altogether these results confirm the G involvement in ₂-AR-mediated potentiation of AC.

Effects of transducin- on clonidine potentiation of isoproterenol-stimulated adenylyl cyclase activity of myometrial membranes at mid-pregnancy.

[View Larger Version of this Image (40K GIF file)]

[View Larger Version of this Image (43K GIF file)]

Table III. Effects of increasing concentrations of QEHA peptide on basal and isoproterenol-stimulated adenylyl cyclase activities of myometrial membranes at mid-pregnancy AC activity, in the presence or absence of QEHA peptide, was measured by radioimmunoassay as described under "Experimental Procedures." Values are expressed as total activity (pmol/mg of protein/10 min). Results are the means ± S.E. of three independent determinations. Adenylyl cyclase activity QEHA peptide Control 50 nM 100 nM 200 nM Basal 93  ± 4 99  ± 3 110  ± 6 109  ± 5 Isoproterenol (0.1 mM) 121  ± 3 130  ± 2 119  ± 6 123  ± 2

Among the seven isoforms of AC identified in pregnant myometrium, types V and VI are characterized by an activity regulated by submicromolar concentrations of Ca²⁺ (2, 31, 32). Thus, to evaluate the functional dependence of myometrial AC population on Ca²⁺, we tested the effects of cation chelator EGTA on basal and stimulated AC activity in particulate membrane preparations. Basal AC activity did not change significantly during the second half of pregnancy, and mean basal value was 115 ± 10 pmol of cAMP/mg of protein/10 min. At the optimal concentration of 0.2 mM, EGTA increased with the same amplitude basal, GTP- and forskolin-stimulated AC activities at mid-pregnancy and term (respectively by 3.5 ± 0.2 fold and 2.4 ± 0.2 fold, n = 11), thus revealing an inhibitory effect of endogenous Ca²⁺ on AC activity at both stages of pregnancy. Further, increasing the free Ca²⁺ concentrations from 0.047 to 550 µM inhibited AC stimulated by 0.1 mM forskolin as shown in Fig. 5. Complete inhibition was achieved at a 550 µM concentration of free Ca²⁺. The Ca²⁺ inhibition curve was shallow and best fit to a two-site model compared with a one-site model (p < 0.05). The inset of Fig. 5 compares the means of IC₅₀ estimates for either the high or low affinity site and maximal AC activity in response to forskolin in the presence of EGTA in myometrial membranes from mid-pregnancy and term. In the plateau region, 3.9 µM free Ca²⁺ accounted for approximately 40% inhibition of forskolin-stimulated AC, revealing a Ca²⁺-dependent inhibition of myometrial AC activity in the submicromolar range.

[View Larger Version of this Image (27K GIF file)]

DISCUSSION

This study reports for the first time that pregnant human and rat myometria express transcripts encoding distinct populations of AC. Indeed, using Northern blot analysis, we detected AC II-VI and IX in both human and rat myometria, whereas type VII was present only in rat. The absence of detectable levels of types I and VIII mRNA is consistent with previous findings showing that these two enzymes are brain-specific. Although it is generally considered that all the structures that expressed more than one AC type contain several cell types, it is important to note that in the myometrium, smooth muscle cells present a great abundance and are distributed evenly throughout the longitudinal and circular layers. Only few other cell types (vascular or conjunctive) are observed. AC mRNA detected in this study using in situ hybridization derived mainly from myometrial cells as exemplified in Fig. 6, A and B, illustrating the specific expression of AC type II in smooth muscle cells. Furthermore, in agreement with previous results (29, 33), control slides of rat brain hybridized in the same conditions indicated a strong expression of AC II in pyramidal cells and dentate gyrus of hippocampus (Fig. 6C). All of these observations suggested to us that G-activated (types II and IV and presumably VII) and Ca²⁺-inhibitable (types III, V, and VI) AC expressed in human and rat myometrium might contribute to physiological processes sensitive to G_i proteins and Ca²⁺-dependent regulation of the cAMP generation in this tissue.

[View Larger Version of this Image (128K GIF file)]

To estimate the total levels of AC in myometria from mid-pregnant and delivering rats, we measured the maximal levels of high affinity [³H]forskolin binding sites (B_max). In the presence of Gpp(NH)p, our results demonstrated an increase of the [³H]forskolin binding sites at mid-pregnancy but not at term. Because it was proposed that the guanine nucleotide-stimulated high affinity binding of [³H]forskolin represents binding to the G_s-activated form of AC, one may suggest that the G_s/catalytic unit complexes are reduced at term. Potential differences in the affinity of this ligand (K_d) observed in mid-pregnant rat myometrium could mean either differences in the interaction of [³H]forskolin with some isoforms of AC present in the myometrium or differential expression of isoforms according to the stage of pregnancy. Although our results show that the expression of AC transcripts in rat myometrium over the second half of pregnancy remains stable, we cannot exclude possible changes at the protein levels. At the present time, this issue cannot be settled because no specific antibody that allows to quantify directly each AC type is available. Interestingly, the decrease of [³H]forskolin binding sites number reported at term in the presence of Gpp(NH)p correlates well with the decline of forskolin-stimulated AC activity. Furthermore, the reduced ability of guanine nucleotides or cholera toxin (acting at the G_s level) to stimulate myometrial AC activity at term and the fact that Mn²⁺/GDPS (acting solely at the catalytic unit of the AC, 34) has no different regulatory effects, points to a defect in the availability or functionality of regulatory G_s proteins rather than a reduction in the catalytic subunits of AC. This conclusion supports previous data of our laboratory establishing that amounts of cholera toxin-catalyzed ADP-ribosylated G_s subunits diminished by 3-fold at delivery versus pregnancy (35). In addition, as shown by radioligand binding experiments, the reduced response to isoproterenol at term is not caused by changes in total -AR number (28) but is the consequence of G_s protein uncoupling (24).

Previously, we have shown that catecholamines acting through ₂-AR/G_i2-3 proteins signaling pathway potentiate -AR-stimulated AC activity in rat myometrium at mid-pregnancy (16). These data suggested a tight interplay between the two adrenergic pathways in the regulation of AC types II and IV through complexes according to the findings of Tang and Gilman (6), Gao and Gilman (7), and Federman et al. (36). Here, the demonstration that the positive input emanating from the activated ₂-AR involves complexes of G_i proteins expressed in rat myometrium (37) was achieved using transducin- and QEHA peptide: both factors abolished the potentiation of AC activity induced by the ₂-AR agonist clonidine. In the same work (16), we reported a switch in the stimulatory versus inhibitory input to AC population from the ₂-AR/G_i protein-signaling pathway at late term. The idea that changes in AC isoform expression might underlie such a versatility of the ₂-AR signal is highly speculative inasmuch as it is not possible to demonstrate that some AC types (specially types II and IV) are the most prevalent isoforms in the mid-pregnant myometrium compared with late term myometrium. At this latter period, no substantial modification in the amounts of specific types of AC transcripts as well as no alteration in the basal activity of the AC system have been observed. So, in the parturient rat myometrium, when the balance among G_i3, G_i2, and G_s proteins was changed (35, 37) we suggest the possibility that distinct patterns of responsiveness of the myometrial AC population may account for the switch between the two types of input (positive and negative) initiated by ₂-AR activation. Interestingly, a decline of immunodetected  subunits of G proteins which is expected to reflect the status of the complexes (38) paralleled the 1.7-fold decrease of Gi₃ subunit in the late pregnant rat myometrium (37, 38). It is then conceivable to propose that the reduced availability of Gi complexes together with the reduced levels of functional G_s protein may underlie the loss of AC potentiation observed at term in response to ₂-/₂-AR activation. On the other hand, the role of G_i protein-coupled ₂-AR in mediating AC inhibition in the late pregnant myometrium has been well documented in rat (16) and human (39). To date, the inhibitory regulation of AC type II through G_i proteins still remains controversial since the data reported by Chen and Iyengar in COS7 cells transfectants (40) and the recent work of Lanier's laboratory in collaboration with us on DDT₁-MF2 cells AC II transfectants (11) are not in agreement with other findings in Sf9 cell membranes (4). Nevertheless, AC types III, V, and VI expressed in rat and human myometrium are highly sensitive to inhibition by G_i proteins (4, 40) and could then present possible targets to ₂-AR/G_i protein-mediated inhibition at term. The presence of AC types V and VI is also consistent with the observation that physiological relevant concentrations of Ca²⁺ inhibit basal and stimulated AC population in pregnant myometrial membranes of rat (our present work) and human (41). This is in line with preliminary evidences showing that these enzymes are regulated directly by submicromolar concentrations of Ca²⁺ (2, 31, 32). As demonstrated for cardiac adenylyl cyclases (42), the inability of maximally stimulating concentrations of Mg²⁺ to overcome Ca²⁺ inhibition (data not shown) indicates that Ca²⁺ concentrations in the µM range are not competitive with Mg²⁺ (i.e. high affinity effect) by contrast with millimolar concentrations of Ca²⁺ (43). Half-maximal inhibition value reported in this work is also in agreement with values found for cardiac muscle (42, 44) or the aorta smooth muscle (45) AC system. Interestingly, these Ca²⁺ concentrations at which AC inhibition is reported are in the range of those required to activate myosin light chain phosphorylation (46) and correspond also to Ca²⁺ concentrations measured in myometrial membranes from rats close to term treated with the uterotonic agent oxytocin (47). In addition, Ca²⁺ can affect also indirectly AC activity because inhibition of AC type III activity through its phosphorylation by calmodulin-casein kinase II is mediated by intracellular increase of Ca²⁺ (15). In conditions prevailing in vivo in the pregnant myometrium at term when intramyometrial submicromolar concentrations of Ca²⁺ rise, such regulation may also provide a mechanism for attenuation of hormone-stimulated cAMP generation. Alternatively, Lanier's group together with our laboratory (11) have established that the stimulatory/inhibitory input to the AC types II and IV initiated through ₂-AR activation is modified differentially by protein kinase C-mediated effects. In particular, activation of protein kinase C in AC IV-transfected DDT₁-MF2 cells eliminates the stimulatory input due to ₂-AR activation. All of these observations indicate that the rise of intracellular Ca²⁺ occurring at term in response to external stimuli (oxytocin, prostaglandin F₂, endothelins, acetylcholine through M₃-muscarinic receptors, norepinephrine through ₁-AR or _2B-AR) would reduce the synthesis of the smooth muscle relaxation mediator, i.e. cAMP, thereby allowing myometrial cells to contract.

In conclusion, analysis of the functional properties of the AC population identified in the myometrium demonstrates for the first time that complexes and Ca²⁺ are two main regulators of the AC/cAMP cascade during pregnancy in the rat. Because of similarities between human and rat myometrium with regard to AC types, as well as G proteins and membrane receptors, we postulate similar physiological regulations of the AC/cAMP system in human. Thus, cross-talks between various hormonal signals that are routed differentially through G_s, G_i, or G_q proteins according to the stage of pregnancy may dictate the final response of the AC/cAMP system and consequently the contractile state of the uterus.