The influence of N-glycosylation on biochemical properties of Amy1, an α-amylase from the yeast Cryptococcus flavus

The yeast Cryptococcus flavus secretes a glycosylated α-amylase (Amy1) when grown in a starch-containing medium. The effects of N-glycosylation on secretion, enzyme activity, and stability of this glycoprotein were studied. Addition of tunicamycin (TM) to the medium at a concentration higher than 0.5 μg mL⁻¹ affected C. flavus growth. Amy1 activity increased by 55% in the intracellular fraction after C. flavus growth in the presence of 0.5 μg mL⁻¹ TM. SDS–PAGE and gel activity detection showed that native enzyme and deglycosylated enzyme had apparent molecular mass of 68 and 64.5 kDa, respectively. The N-glycosylation process did not affect either optimum pH or optimum temperature. The K_M values of native and non-glycosylated α-amylases were 0.052 and 0.098 mg mL⁻¹, and V_max values were 0.038 and 0.047 mg min⁻¹, respectively. However, the non-glycosylated form was more sensitive to inactivation by both the proteolytic enzyme trypsin and high temperature. Furthermore, the activity of the non-glycosylated enzyme was affected by Hg²⁺ and Cu²⁺ suggesting that N-glycosylation is involved in the folding of Amy1.     

Characterization of adrenergic receptors and related transduction pathways in the liver of the rainbow trout

Epinephrine (EPI) is thought to act by stimulating adenylyl cyclase (ACase) and cAMP production through β-adrenoceptors in the liver of more primitive vertebrates. Recent observations, however, point to an involvement of α₁-adrenoceptors in EPI action, at least in some fish species. The role of the α₁- and β-adrenergic transduction pathways has been investigated in rainbow trout (Oncorhynchus mykiss) hepatic tissue. Radioligand-binding assays with the β-adrenergic antagonist ³H-CGP-12177 using hepatic membranes purified on a discontinuous sucrose gradient confirmed the presence of β-adrenoceptors (K_d0.36 nM, B_max 8.61 fmol · mg⁻¹ protein). We provide the first demonstration of α₁-adrenoceptors in these same membranes; analysis of binding data with the α₁-adrenergic antagonist ³H-prazosin demonstrated a single class of binding sites with a K_dof 15.4 nM and a B_max of 75.2 fmol · mg⁻¹ protein. There is a straight correlation between β-adrenoceptor occupancy, ACase activation and cAMP production. On the contrary, the role of inositol 1,4,5-trisphosphate (IP₃) has to be elucidated; in fact, despite the presence of specific microsomal binding sites for IP₃ (K_d 6.03 nM, B_max 90.2 fmol · mg⁻¹ protein), its cytosolic concentration was not modulated by EPI. On the other hand, we have previously shown in American eel and bullhead hepatocytes that α₁-adrenergic agonists are able to increase intracellular concentrations of IP₃ and Ca²⁺ and to activate glycogenolysis. These data suggest a marked variation in the liver of different fish both in terms of α₁-binding sites affinity and of α₁-adrenoceptor/IP₃/Ca²⁺ transduction systems.     

Purification and kinetic characterization of γ-aminobutyraldehyde dehydrogenase from rat liver

Oxidative deamination of putrescine, the precursor of polyamines, gives rise to γ-aminobutyraldehyde (ABAL). In this study an aldehyde dehydrogenase, active on ABAL, has been purified to electrophoretic homogeneity from rat liver cytoplasm and its kinetic behaviour investigated. The enzyme is a dimer with a subunit molecular weight of 51,000. It is NAD⁺-dependent, active only in the presence of sulphhydryl compounds and has a pH optimum in the range 7.3–8.4. Temperatures higher than 28°C promote slow activation and the process is favoured by the presence of at least one substrate. K_m for aliphatic aldehydes decreases from 110 μM for ABAL and acetaldehyde to 2–3 μM for capronaldehyde. The highest relative V-values have been observed with ABAL (100) and isobutyraldehyde (64), and the lowest with acetaldehyde (14). Affinity for NAD⁺ is affected by the aldehyde present at the active site: K_m for NAD⁺ is 70 μM with ABAL, 200 μM with isobutyraldehyde and capronaldehyde, and>800 μM with acetaldehyde. The kinetic behaviour at 37°C is quite complex; according to enzymatic models, NAD⁺ activates the enzyme (K_act 500 μM) while NADH competes for the regulatory site (K_in 70 μM). In the presence of high NAD⁺ concentrations (4 mM), ABAL promotes further activation by binding to a low-affinity regulatory site (K_act 10 mM). The data show that the enzyme is probably an E₃ aldehyde dehydrogenase, and suggest that it can effectively metabolize aldehydes arising from biogenic amines.     

Immobilized 4-aminophenyl 1-thio-β- -galactofuranoside as a matrix for affinity purification of an exo-β- -galactofuranosidase

An alternative and fast method for the purification of an exo-β- -galactofuranosidase has been developed using a 4-aminophenyl 1-thio-β- -galactofuranoside affinity chromatography system and specific elution with 10 mM -galactono-1,4-lactone in a salt gradient. A concentrated culture medium from Penicillium fellutanum was chromatographed on DEAE–Sepharose CL 6B followed by chromatography on the affinity column, yielding two separate peaks of enzyme activity when elution was performed with 10 mM -galactono-1,4-lactone in a 100–500 mM NaCl salt gradient. Both peaks behaved as a single 70 kDa protein, as detected by SDS-PAGE. Antibodies elicited against a mixture of the single bands excised from the gel were capable of immunoprecipitating 0.2 units out of 0.26 total units of the enzyme from a crude extract. The glycoprotein nature of the exo-β- -galactofuranosidase was ascertained through binding to Concanavalin A–Sepharose as well as by specific reaction with Schiff reagent in Western blots. The purified enzyme has an optimum acidic pH (between 3 and 6), and K_m and V_max values of 0.311 mM and 17 μmol h⁻¹ μg⁻¹ respectively, when 4-nitrophenyl β- -galactofuranoside was employed as the substrate.     

Bovine intermediate pituitary α-amidation enzyme: Preliminary characterization

A secretory granule-associated enzymatic activity that converts mono-[¹²⁵I]-D-Tyr-Val-Gly into mono-[¹²⁵I]-D-Tyr-Val-NH₂ has been studied. The activity is primarily soluble and shows optimal activity at pH 7 to pH 8. Amidation activity was stimulated 9-fold by addition of optimal amounts of copper (3 μM). In the presence of optimal copper, ascorbate stimulated the reaction 7-fold; none of the other reduced or oxidized cofactors tested was as effective. Taking into account the dependence of the reaction on ascorbate and molecular oxygen and the production of glyoxylate [2], it is suggested that the α-amidation enzyme is a monooxygenase. Lineweaver Burk plots with D-Tyr-Val-Gly as the varied substrate demonstrated Michelis-Menten type kinetics with the values of K_m and V_max increasing with the addition of ascorbate to the assay. A variety of peptides ending with a COOH-terminal Gly residue act as inhibitors of the reaction. Two synthetic peptides, γ₂MSH and ACTH(1–14), with carboxyl termini similar to the presumed physiological substrates for the enzyme, act as competitive inhibitors with similar K₁ values. It is likely that this secretory granule α-amidation activity is involved in the physiological biosynthetic α-amidation of a wide range of bioactive peptides.     

Gene cloning, purification, and characterization of α-methylserine aldolase from Bosea sp. AJ110407 and its applicability for the enzymatic synthesis of α-methyl-l-serine and α-ethyl-l-serine

The gene encoding α-methylserine aldolase was isolated from Bosea sp. AJ110407. Sequence analysis revealed that the predicted amino acid sequence encoded by the 1320-bp open reading frame was 65.0% similar to the corresponding sequence of the enzyme isolated from Ralstonia sp. AJ110405. The gene was expressed in Escherichia coli, and the recombinant enzyme was purified. Gel filtration revealed the molecular mass of the purified enzyme to be approximately 78 kDa, suggesting that the enzyme is a homodimer. The enzyme exhibited a specific peak at 429 nm in the spectrum and contained 1 mol pyridoxal 5′-phosphate per mole of the subunit. The V_max value was 1.40 μmol min⁻¹ mg⁻¹, and the K_m value was 1.5 mM for the reaction wherein formaldehyde was released from α-methyl-l-serine. This enzyme could also catalyze the reverse reaction, i.e., the synthesis of α-methyl-l-serine from l-alanine and formaldehyde. This activity was inhibited in the excess of formaldehyde; however, α-methyl-l-serine was efficiently produced from l-alanine in the presence of formaldehyde. This method was also applicable for producing α-ethyl-l-serine from l-2-aminobutyric acid.     

Methyllycaconitine: a selective probe for neuronal α-bungarotoxin binding sites

The ability of methyllycaconitine (MLA) to inhibit the binding of [¹²⁵I]α-bungarotoxin to rat brain membranes, frog and human muscle extracts and the human muscle cell line TE671 has been measured. MLA showed a markedly higher affinity for the rat brain site (K_i 1.4 × 10⁻⁹ M) than for the muscle receptors (K_i; 10⁻⁵-10⁻⁶ M). Structure modelling techniques were used to fit the structure of MLA to a nicotinic pharmacophore model. MLA is the first low molecular weight ligand to be shown to discriminate between muscle nicotinic receptors and their α-bungarotoxinbinding counterpart in the brain, and as such may be a useful structural probe for pursuing the structural and functional properties of the neuronal protein.     

Purification and properties of a novel raw starch degrading-cyclodextrin glycosyltransferase from Klebsiella pneumoniae AS- 22

A novel raw starch degrading α-cyclodextrin glycosyltransferase (CGTase; E.C. 2.4.1.19), produced by Klebsiella pneumoniae AS-22, was purified to homogeneity by ultrafiltration, affinity and gel filtration chromatography. The specific cyclization activity of the pure enzyme preparation was 523 U/mg of protein. No hydrolysis activity was detected when soluble starch was used as the substrate. The molecular weight of the pure protein was estimated to be 75 kDa with SDS-PAGE and gel filtration. The isoelectric point of the pure enzyme was 7.3. The enzyme was most active in the pH range 5.5–9.0 whereas it was most stable in the pH range 6–9. The CGTase was most active in the temperature range 35–50°C. This CGTase is inherently temperature labile and rapidly loses activity above 30°C. However, presence of soluble starch and calcium chloride improved the temperature stability of the enzyme up to 40°C. In presence of 30% (v/v) glycerol, this enzyme was almost 100% stable at 30°C for a month. The K_m and k_cat values for the pure enzyme were 1.35 mg ml⁻¹ and 249 μM mg⁻¹ min⁻¹, respectively, with soluble starch as the substrate. The enzyme predominantly produced α-cyclodextrin without addition of any complexing agents. The conditions employed for maximum α-cyclodextrin production were 100 g l⁻¹ gelatinized soluble starch or 125 g l⁻¹ raw wheat starch at an enzyme concentration of 10 U g⁻¹ of starch. The α:β:γ-cyclodextrins were produced in the ratios of 81:12:7 and 89:9:2 from gelatinized soluble starch and raw wheat starch, respectively.     

Stable isotope dilution analysis of human urinary metabolites of 17α-methyltestosterone

A method based on gas chromatography–mass spectrometry–selected-ion monitoring was developed to measure the main metabolites of 17α-methyltestosterone, 17α-methyl-5α-androstan-3α,17β-diol and 17α-methyl-5β-androstan-3α,17β-diol, in human urine. 17α-Methyl-[²H₃]-5α-androstan-3α,17β-diol and 17α-methyl-[²H₃]-5β-androstan-3α,17β-diol were used as internal standards. The methods involved purification using a Sep-Pak C₁₈ cartridge, hydrolysis by β-glucuronidase from Ampullaria and derivatization with N-methyl-N-trimethylsilyl-trifluoroacetamide/dithioerythriol/ammonium iodide. Quantitation was achieved by selected-ion monitoring of the characteristic fragment ions ([(M+H)−2×TMSOH]⁺) of the di-TMS derivatives on the chemical ionization mode. The method provides a specific, sensitive and reliable technique to determine the urine levels of 17α-methyl-5α-androstan-3α,17β-diol and 17α-methyl-5β-androstan-3α,17β-diol, and can be applied to pharmacokinetic studies of 17α-methyltestosterone.     

TGFbeta1 and TGFbeta3 are partially redundant effectors in brain vascular morphogenesis

Gene deletion experiments have shown that the three TGFβ isoforms regulate distinct developmental processes. Recent work by our group and others showed that the integrins αvβ6 and αvβ8 activate latent forms of TGFβ1 and TGFβ3. This raises the possibility that TGFβ1 and TGFβ3 act redundantly in developmental processes where both isoforms are expressed and activation is by integrins. To investigate this issue, we generated mice with defective integrin-mediated TGFβ1 activation (Tgfb1^RGE/RGE) that were also homozygous for a null mutation in the TGFβ3 gene. Tgfb1^RGE/RGE; Tgfb3^−/− mice have severely perturbed development of the brain vasculature that is highly similar to that in mice lacking αvβ8. Some Tgfb1^RGE/RGE; Tgfb3^+/− and Tgfb1^RGE/RGE; Tgfb3^+/+ mice have milder, background-dependent versions of the phenotype. In addition, we found that Tgfb3 gene status influences embryonic lethality due to TGFβ1 deficiency after limited backcrossing to the BALB/c background. Conversely, Tgfb1 gene status modifies the extent of palate fusion in Tgfb3^−/− mice after limited backcrossing to the ICR background. Our results are consistent with a functional connection between TGFβ1 and TGFβ3 during development based on a shared mechanism of activation.     

Properties of γ-aminobutyraldehyde dehydrogenase from Escherichia coli

γ-Aminobutyraldehyde dehydrogenase from Escherichia coli K-12 has been purified and characterized from cell mutants able to grow in putrescine as the sole carbon and nitrogen source. The enzyme has an M_r of 195 000±10 000 in its dimeric form with an M_r of 95 000±1000 for each subunit, a pH optimum at 5.4 in sodium citrate buffer, and does not require bivalent cations for its activity. K_m values are 31.3±6.8 μM and 53.8±7.4 μM for Δ-1-pyrroline and NAD⁺, respectively. An inhibitory capacity for NADH is also shown using the purified enzyme.     

Dehydroepiandrosterone 7α-hydroxylation in human tissues: Possible interference with type 1 11β-hydroxysteroid dehydrogenase-mediated processes

Dehydroepiandrosterone (DHEA) is 7α-hydroxylated by the cytochome P450 7B1 (CYP7B1) in the human brain and liver. This produces 7α-hydroxy-DHEA that is a substrate for 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) which exists in the same tissues and carries out the inter-conversion of 7α- and 7β-hydroxy-DHEA through a 7-oxo-intermediary. Since the role of 11β-HSD1 is to transform the inactive cortisone into active cortisol, its competitive inhibition by 7α-hydroxy-DHEA may support the paradigm of native anti-glucocorticoid arising from DHEA. Therefore, our objective was to use human tissues to assess the presences of both CYP7B1 and 11β-HSD1. Human skin was selected then and used to test its ability to produce 7α-hydroxy-DHEA, and to test the interference of 7α- and 7β-hydroxy-DHEA and 7-oxo-DHEA with the 11β-HSD1-mediated oxidoreduction of cortisol and cortisone. Immuno-histochemical studies showed the presence of both CYP7B1 and 11β-HSD1 in the liver, skin and tonsils. DHEA was readily 7α-hydroxylated when incubated using skin slices. A S9 fraction of dermal homogenates containing the 11β-HSD1 carried out the oxidoreduction of cortisol and cortisone. Inhibition of the cortisol oxidation by 7α-hydroxy-DHEA and 7β-hydroxy-DHEA was competitive with a K_i at 1.85 ± 0.495 and 0.255 ± 0.005 μM, respectively. Inhibition of cortisone reduction by 7-oxo-DHEA was of a mixed type with a K_i at 1.13 ± 0.15 μM. These findings may support the previously proposed native anti-glucocorticoid paradigm and suggest that the 7α-hydroxy-DHEA production is a key for the fine tuning of glucocorticoid levels in tissues.     

Mitochondrial dihydrolipoyl succinyltransferase deficiency accelerates amyloid pathology and memory deficit in a transgenic mouse model of amyloid deposition

Mitochondrial dysfunction and oxidative stress are involved in Alzheimer disease (AD) pathogenesis. In human AD brains, the activity of the α-ketoglutarate dehydrogenase enzyme complex (α-KGDHC) is reduced. KGDHC is mostly involved in NADH production. It can also participate in oxidative stress and reactive oxygen species (ROS) production. The mitochondrial dihydrolipoyl succinyltransferase enzyme (DLST) is a key subunit specific to the α-KGDHC. In cultured cells, reduction of DLST increased H₂O₂-induced ROS generation and cell death. Thus, we asked whether partial genetic deletion of DLST could accelerate the onset of AD pathogenesis, using a transgenic mouse model of amyloid deposition crossed with DLST^+/− mice. Tg19959 mice, which carry the human amyloid precursor protein with two mutations, develop amyloid deposits and progressive behavioral abnormalities. We compared Tg19959 mice to Tg19959-DLST^+/− littermates at 2–3 months of age and studied the effects of DLST deficiency on amyloid deposition, spatial learning and memory, and oxidative stress. We found that α-KGDHC activity was reduced in DLST^+/− mice. We also found that DLST deficiency increased amyloid plaque burden, Aβ oligomers, and nitrotyrosine levels and accelerated the occurrence of spatial learning and memory deficits in female Tg19959 mice. Our data suggest that α-KGDHC may be involved in AD pathogenesis through increased mitochondrial oxidative stress.     

5α,6α-Epoxy-cholestan-3β-ol (cholesterol α-oxide): A specific substrate for rat liver glutathione transferase B

A semi-micro assay was developed for the conjugation of 5α,6α-epoxy-cholestan-3β-ol (cholesterol α-oxide) with glutathione. The soluble supernatant of rat liver homogenate catalysed the reaction at a rate of 0.2–0.5 pmol.min⁻¹ .mg protein⁻¹ with 4μM cholesterol α-oxide, while the reaction in the presence of GSH alone was barely detectable. Enzymic activity in the soluble supernatant was due equally to the two forms of glutathione transferase B (100 pmol.min⁻.mg protein⁻¹), glutathione transferases AA, A, C and E being unreactive. The activity of purified glutathione transferase B was about 5-times that expected from the activity of the soluble supernatant. Complex enzyme kinetics were obtained suggestive of substrate inhibition.     

Simulation analysis of intracellular Na and Cl homeostasis during β1-adrenergic stimulation of cardiac myocyte

To quantitatively understand intracellular Na⁺ and Cl⁻ homeostasis as well as roles of Na⁺/K⁺ pump and cystic fibrosis transmembrane conductance regulator Cl⁻ channel (I_CFTR) during the β1-adrenergic stimulation in cardiac myocyte, we constructed a computer model of β1-adrenergic signaling and implemented it into an excitation-contraction coupling model of the guinea-pig ventricular cell, which can reproduce membrane excitation, intracellular ion changes (Na⁺, K⁺, Ca²⁺ and Cl⁻), contraction, cell volume, and oxidative phosphorylation. An application of isoproterenol to the model cell resulted in the shortening of action potential duration (APD) after a transient prolongation, the increases in both Ca²⁺ transient and cell shortening, and the decreases in both Cl⁻ concentration and cell volume. These results are consistent with experimental data. Increasing the density of I_CFTR shortened APD and augmented the peak amplitudes of the L-type Ca²⁺ current (I_CaL) and the Ca²⁺ transient during the β1-adrenergic stimulation. This indirect inotropic effect was elucidated by the increase in the driving force of I_CaL via a decrease in plateau potential. Our model reproduced the experimental data demonstrating the decrease in intracellular Na⁺ during the β-adrenergic stimulation at 0 or 0.5 Hz electrical stimulation. The decrease is attributable to the increase in Na⁺ affinity of Na⁺/K⁺ pump by protein kinase A. However it was predicted that Na⁺ increases at higher beating rate because of larger Na⁺ influx through forward Na⁺/Ca²⁺ exchange. It was demonstrated that dynamic changes in Na⁺ and Cl⁻ fluxes remarkably affect the inotropic action of isoproterenol in the ventricular myocytes.     

Stachyose synthesis in leaves of Cucurbita pepo

An enzyme synthesizing stachyose, galactinol-raffinose galactosyltransferase (EC2.4.1.67), has been purified ca 40-fold from mature leaves of Cucurbita pepo using ammonium sulphate precipitation, Sephadex gel filtration and DEAE-Sephadex gel chromatography. The purified enzyme fraction was separated from all but 2 % of the total,α-galactosidase activity extracted from the tissue. The enzyme was optimally active at pH 6.9 and was stable for at least a month at 4° in the presence of 20 mM 2-mercaptoethanol. The enzyme displayed high specificity for the donor galactinol (K_m 7.7 mM) and the acceptor raffinose (K_m 4.6 mM) and was unable to effect synthesis of any other member of the raffinose series of galactosyl-sucrose oligosaccharides. Co²⁺, Hg²⁺, Mn²⁺ and Ni²⁺ ions were particularly inhibitory; no metal ion promotion was observed and 5 mM EDTA was ineffective. Myo-inositol was strongly inhibitory (K_i 2 mM), melibiose weakly so. Tris buffer (0. 1 M) was also inhibitory. Galactinol hydrolysis occurred in the absence of the acceptor raffinose but there was no hydrolysis of either raffinose or stachyose in the absence of the donor galactinol. The reaction was readily reversible and exchange reactions were detected between substrates and products. It is proposed that the synthesis of stachyose in mature leaves ofC. pepo proceeds via this galactosyltransferase and not via α-galactosidase.     

Microfluorometric analysis of Cl permeability and its relation to oscillatory Ca signalling in glucose-stimulated pancreatic β-cells

The cytoplasmic concentrations of Cl⁻([Cl⁻]_i) and Ca²⁺ ([Ca²⁺]_i) were measured with the fluorescent indicators N-(ethoxycarbonylmethyl)-6-methoxyquinilinum bromide (MQAE) and fura-2 in pancreatic β-cells isolated from ob/ob mice. Steady-state [Cl⁻]_i in unstimulated β-cells was 34 mM, which is higher than expected from a passive distribution. Increase of the glucose concentration from 3 to 20 mM resulted in an accelerated entry of Cl⁻ into β-cells depleted of this ion. The exposure to 20 mM glucose did not affect steady-state [Cl⁻]_i either in the absence or presence of furosemide inhibition of Na⁺, K⁺, 2 Cl⁻ co-transport. Glucose-induced oscillations of [Ca²⁺]_i were transformed into sustained elevation in the presence of 4,4′ diisothiocyanato-dihydrostilbene-2,2′-disulfonic acid (H₂DIDS). A similar effect was noted when replacing 25% of extracellular Cl⁻ with the more easily permeating anions SCN⁻, I⁻, NO₃⁻ or Br⁻. It is concluded that glucose stimulation of the β-cells is coupled to an increase in their Cl⁻ permeability and that the oscillatory Ca²⁺ signalling is critically dependent on transmembrane Cl⁻ fluxes.     

α2-macroglobulin ‘fast’ forms inhibit superoxide production by activated macrophages

Mouse peritoneal macrophages activated by bacillus Calmette-Guerin (BCG) were incubated with human α₂-macroglobulin converted to its ‘fast’ form with either trypsin or methylamine before being stimulated with phorbol myrystate acetate. Both α₂-macroglobulin-trypsin and α₂-macroglobulin-methylamine inhibited macrophage production of superoxide anion (O₂⁻) while native α₂-macroglobulin had little effect except at high concentration. The α₂-macroglobulin ‘fast’ forms, which bind with a K_d of about 8 nM, inhibited 50% generation of O₂⁻(ID₅₀) at a concentration of 7 nM while α₂-macroglobulin inhibited O₂⁻ production with an ID₅₀ of 141 nM. The ‘fast’ forms of α₂-macroglobulin may play a role in the feedback regulation of inflammatory reactions.     

Interleukin-1β-stimulated PGE2 production from early first trimester human decidual cells is inhibited by dexamethasone and progesterone

Cytokines are known to increase the production of prostaglandins by human decidual cells, but negative regulators have not been identified. We have examined the effects of dexamethasone and progesterone on prostaglandin (PG) E2 synthesis by cultured human first trimester decidual cells. The numbers of cyclooxygenase (COX) enzyme positive cells were visualised by immunocytochemistry, using antibodies specific for COX-1 and COX-2. Interleukin-1β stimulated the production of prostaglandins E2 and F2α dose-dependently, and this was associated with increased numbers of COX-2 positive cells. Progesterone (10⁻⁷−10⁻⁶ M) and dexamethasone (10⁻⁷−10⁻⁶ M) inhibited basal and interleukin-1β-stimulated prostaglandin production, and decreased the numbers of COX-2 positive cells. Neither interleukin-1β nor the steroids affected numbers of COX-1 positive cells. COX-2 seems to be the main enzyme controlling the synthesis of PGE2 by human decidual cells, and may be negatively regulated by progesterone.