The binding of methylsulfonyl-polychloro-biphenyls to uteroglobin

The binding of methylsulfonyl-polychloro-biphenyls (methylsulfonyl-PCBs) to purified uteroglobin was studied by a dextran-coated charcoal assay using 4,4'-bis([ 3H]methylsulfonyl)-2,2',5,5'-tetrachlorobiphenyl [(3H-MeSO2)2TCB] as radiolabeled ligand. The specific binding of this ligand to uteroglobin was enhanced by the presence of dithiothreitol, and the optimal concentration of dithiothreitol for binding was 20 mM. The specific [(3H-MeSO2)2TCB] binding was inhibited by 4-methylsulfonyl-2,2',4',5,5'-pentachlorobiphenyl in a concentration-dependent manner. The molecular structures of methylsulfonyl-PCBs, and progesterone, were fitted into the X-ray crystallographic structure of uteroglobin using the molecular graphics program TOM. In these simulations the water-accessible surfaces of the ligands appeared quite similar, and fitted nicely in the internal water-accessible surface of uteroglobulin. Moreover, it appeared from the computer-supported ligand-binding studies that the sulfone oxygens of the studied methylsulfonyl-PCBs, as well as the carbonyl (C20) of progesterone, may form hydrogen bonds with the hydroxyl group of TYR 21 of uteroglobulin. These findings may explain why both steroids and methylsulfonyl-PCBs interact with the same protein, although these two types of ligands are structurally dissimilar.     

Identification of residues essential for progesterone binding to uteroglobin by site-directed mutagenesis

In order to identify amino acids directly involved in progesterone binding to rabbit uteroglobin we have mutated Phe 6, Tyr 21 and Thr 60 by site-directed mutagenesis of the uteroglobin cDNA. These residues have been postulated previously to participate in progesterone binding. High-level expression of the mutated uteroglobin cDNAs in Escherichia coli yields recombinant protein mutants that, like natural uteroglobin, form stable dimers, suggesting that the tertiary structure of the protein has not been altered. Substitution of Phe 6 by Ser or Ala does not change the progesterone binding characteristics. In contrast, replacement of Tyr 21 by Phe or Ala, drastically decreases progesterone binding. In addition, replacement of Thr 60 by Ala reduces the affinity for progesterone by a factor of three. These data suggest a direct interaction of progesterone with these two amino acids and support the idea of direct hydrogen bonding of the carbonyl (C3 and C20) of progesterone with the hydroxyl groups of Tyr 21 and Thr 60, respectively.     

Binding of progesterone to the proteins of the uterine luminal fluid. Identification of uteroglobin as the binding protein.

1. The uterine luminal fluid of rabbits treated with estradiol and progesterone contains a protein factor with high affinity for [3-H] progesterone which is not present in the uterine secretion of control rabbits treated with estradiol. 2. This progesterone dependent factor is shown by gel filtration and polyacrylamide gel electrophoresis to be identical with the uterus specific protein uteroglobin, which seems to be required during the preimplantation phase. Uteroglobin specific antiserum, prepared in guinea pigs, completely inhibits the progesterone binding activity of the proteins of the uterine fluid. 3. Progesterone binding to uteroglobin is dependent upon millimolar concentrations of dithioerythritol. At saturation, one molecule of progesterone binds per uteroglobin molecule and the apparent association constant is 2 x 10-6 M-1 at 0 degrees C. 4. The progesterone binding species of uteroglobin exhibits a molecular weight of around 12 000 on polyacrylamide gels containing dodecylsulfate, and of 15 000 upon gel filtration, indicating a non-globular shape. This molecule is compased of two subunits of similar molecular size which are held together by a disulfide bridge among other forces.     

Hormonal control of uteroglobin secretion in rabbit uterus. Inhibition of uteroglobin synthesis and messenger ribonucleic acid accumulation by oestrogen and anti-oestrogen administration

Investigations were conducted to quantify activity of uteroglobin mRNA and secretion of uteroglobin in rabbit uterus after administration of progesterone and 5alpha-dihydrotestosterone, either alone or concomitantly with oestradiol-17beta and tamoxifen, a non-steroidal anti-oestrogen. Poly(A)-containing mRNA was isolated from the uterine tissue by extraction with phenol/chloroform, precipitation with ethanol and chromatography on oligo(dT)-cellulose. Cell-free translation in vitro of the poly(A)-containing mRNA was carried out in a wheat-germ lysate, and the product isolated by specific immuno-precipitation with anti-uteroglobin antiserum purified by affinity chromatography. Radioimmunoassay was utilized to determine uteroglobin content in the uterine flushings and tissue preparations. When given for 5 days, both progesterone (1mg/kg per day) and 5alpha-dihydrotestosterone (25mg/kg per day) elicited a marked induction of uteroglobin secretion, which was accompanied with accumulation of uteroglobin mRNA in the tissue. Concomitant administration of oestradiol-17beta (50mug/kg per day) or tamoxifen (12.5mg/kg per day) significantly decreased both progesterone- and 5alpha-dihydrotestosterone-induced uteroglobin secretion, with a parallel decrease in the uteroglobin-mRNA activity. The decline in the uteroglobin content of the uterine flushes brought about by oestradiol-17beta or tamoxifen administration was not due to inhibition of secretion of this protein by the endometrial cells, since a simultaneous decrease occurred in the tissue uteroglobin content. After a 5-day pretreatment with progesterone (1mg/kg per day), administration of oestradiol-17beta (50mug/kg per day) during the ensuing 4 days greatly accelerated the decay of the uteroglobin content in the uterine fluid.     

Prolactin enhances uteroglobin gene expression by uteri of immature rabbits

The effect of prolactin on uteroglobin production by immature rabbits was studied with neonatal (1 day old) and juvenile (14 days old) does. The animals were divided into 11 treatment groups for each age category and exposed to a 9-day injection protocol. Each day the animals received a subcutaneous injection of oestradiol-17 beta and/or ovine prolactin and/or progesterone, or were sham-injected. Juvenile animals, which received 100 micrograms oestradiol/kg 24 h-1, plus progesterone or plus prolactin and progesterone, produced detectable amounts of uteroglobin in the uterine secretions (0.034 +/- 0.010 mg uteroglobin/mg total protein and 0.098 +/- 0.03 l mg uteroglobin/mg total protein, respectively). None of the animals in the other juvenile treatment groups or any of the neonatal groups produced uteroglobin. From this survey it was apparent that uteroglobin secretion could be induced by exogenous oestradiol and progesterone in rabbits treated as early as 14 days of age, and that the added supplementation of prolactin enhanced the response to the ovarian steroids. As a result, additional juvenile animals were injected with 100 micrograms oestradiol +/- prolactin + progesterone and the effects of these two treatments were quantitated as follows: uteroglobin mRNA levels by slot-blot hybridization; endometrial surface area by computerized image analysis; and oestrogen, progesterone and prolactin receptors by immunocytochemistry. Prolactin modified the response of the juvenile rabbit uterus to oestradiol + progesterone for all parameters tested.     

Interchain cysteine bridges control entry of progesterone to the central cavity of the uteroglobin dimer.

The progesterone-binding protein uteroglobin has been expressed in Escherichia coli in an unfused, soluble form. Like mature uteroglobin from rabbit endometrium (UG), the E.coli produced uteroglobin (UG1) dimerizes in vitro, forms an antiparallel dimer with Cys3-Cys69' and Cys69-Cys3' disulfide bonds and binds progesterone under reducing conditions. In order to analyze the dimerization and the reduction dependence of progesterone binding in more detail, we separately replaced cysteine 3 and cysteine 69 by serines. Under reducing conditions, both uteroglobin variants (UG1-3Ser and UG1-69Ser) bind progesterone with the same affinity as the wild-type suggesting that both cysteine residues are not directly involved in progesterone binding. In contrast to the wild-type protein, both cysteine variants also bind progesterone with high affinity in the absence of reducing agents. In addition, UG1-3Ser and UG1-69Ser both form covalently linked homodimers. Thus, unnatural Cys69-69' and Cys3-3' disulfide bonds exist in UG1-3Ser and UG1-69Ser, respectively. These data together with computer models based on X-ray diffraction data strongly support the idea that progesterone reaches its binding site located in an internal hydrophobic cavity via a hydrophobic tunnel along helices 1 and 4. Under non-reducing conditions the tunnel is closed by two disulfide bridges (Cys3-Cys69' and Cys69-Cys3') that lie in the most flexible region of the dimer. Reduction or replacement of a cysteine residue enables conformational changes that open the channel allowing progesterone to enter.     

Evidence for the identity of anti-proteinase pulmonary protein CCSP and uteroglobin

Purified Clara cell secretory protein (CCSP) from rabbit lung was analyzed by SDS gel electrophoresis, and immunoblotting with a specific anti-uteroglobin antibody as well as for its ability to bind [3H]progesterone. The results obtained indicate that proteins CCSP and uteroglobin are identical.     

Isolation and characterization of uteroglobin from the lung of the hare (Lepus capensis)

Uteroglobin has been purified from hare lung by gel filtration and chromatography on carboxymethyl-cellulose. Hare uteroglobin appears homogeneous by electrophoresis under both denaturing and nondenaturing conditions. Its chemical and immunological properties as well as its ability to bind progesterone are compared to those of rabbit uteroglobin. The two proteins have the same N-terminal residue (glycine) and both lack tryptophan but differ in amino acid composition. Sodium dodecyl sulfate-gel electrophoresis shows that hare uteroglobin is composed of two subunits of identical Mr (about 7000) held together by disulfide bridges. The amino acid composition indicates a subunit composed of 65-67 residues, which is compatible with the apparent Mr observed. Thus, hare uteroglobin appears to be slightly smaller than the rabbit protein. Hare uteroglobin partially reacts with anti-rabbit uteroglobin in a radioimmunoassay and also binds progesterone, although this binding is relatively unaffected by dithiothreitol. The synthesis of hare uteroglobin in the uterus appears to be rather insensitive to ovarian steroid hormones.     

Changes in cytosol and nuclear progesterone receptor concentrations in the rabbit uterus and their relation to induction of progesterone-regulated uteroglobin.

To study the relation between steroid receptor concentrations and biological response, we measured cytosol and nuclear progesterone receptors from rabbit uterus under different experimental conditions, and compared receptor values with induction of uteroglobin, a progesterone-regulated protein. A 5-day progesterone treatment (1 mg/kg per day) reduced the nuclear receptor content by 40%, slightly elevated cytosol receptor levels and increased uteroglobin content 3000-fold. Estradiol and tamoxifen altered progesterone-induced changes in the receptor and uteroglobin values: cytosol and nuclear receptors rose significantly, but uteroglobin induction declined markedly, when progesterone treatment was supplemented with estradiol or tamoxifen. A 50% inhibition of progesterone action on uteroglobin synthesis was achieved with 1 μg/kg of estradiol per day. Thus under certain conditions, there is a clear disparity between steroid receptor levels and biological response.     

Clara cell 10 kDa protein (CC10): comparison of structure and function to uteroglobin

The cellular localization, functional activities and structures of rat and human Clara cell 10 kDa proteins (CC10) are compared to rabbit uteroglobin. CC10 is present exclusively in the non-ciliated cells of the surface epithelium of the pulmonary airways, whereas uteroglobin is reported to be present in the lung and reproductive organs. There is about 55% identity between the amino acid sequences of rat CC10 and either rabbit uteroglobin or human CC10. The latter two have 61% identity. Using the known structure of uteroglobin as the model, correlations between the structure and function for this group of proteins are made. Substitution of the residues for the rat and human CC10 into the structure of uteroglobin suggests that these proteins may be members of a structurally homologous family. Some of the functional differences may be due to distortion of the hydrophobic pocket in the dimeric protein and a surface hypervariability located on one contiguous helix and beta turn. Rat CC10 and rabbit uteroglobin both, nearly equally, inhibit papain and bind progesterone. Human CC10 does not inhibit papain and has markedly lower progesterone binding (4.6% of rabbit uteroglobin). Antiinflammatory activity of synthetic peptides corresponding to a homologous sequence region of uteroglobin and the two Clara cell proteins was tested. The region chosen has sequence similarity to lipocortin I. The peptides not only failed to inhibit carrageenan-induced foot pad swelling but exacerbated it. All three proteins inhibit pancreatic phospholipase A2. The phospholipase A2 inhibitory effect of CC10 may be important in regulating the inflammatory responses in the lung.     

Cloning of rabbit genomic fragments containing the uteroglobin gene

Hybrid phages containing the uteroglobin gene were isolated from a library of Charon 4A lambda phages carrying fragments of rabbit genome. Two phages which contain overlapping rabbit DNA covering approximately 35 kilobase pairs (kb) on both sides of the uteroglobin gene have been analyzed. One of the phages contains the complete gene. Digestion of rabbit liver DNA by restriction endonucleases and Southern hybridization showed a pattern identical with that obtained with the DNA of recombinant phages suggesting that no rearrangement has taken place during cloning. These experiments also showed the presence of a single gene for uteroglobin in the rabbit genome. Isolation of this gene is of interest for the study of the mechanism of action of progesterone.     

Refinement of the C222(1) crystal form of oxidized uteroglobin at 1.34 A resolution

The structure of uteroglobin, a progesterone binding protein from rabbit uterine fluid, was determined and refined at 1.34 A resolution to a conventional R-factor of 0.229. The accuracy of the co-ordinates is estimated to be 0.15 A. The isotropic temperature factor of individual atoms was refined and its average value is 11.9 A2 for the 548 non-hydrogen atoms of the protein monomer. A total of 83 water molecules was located in difference electron density maps and refined, first using a constant occupancy factor of 1 and then variable occupancy, the final (Q) being 0.63. The mean temperature factor of the water oxygen atoms is 26.4 A2. Uteroglobin is a dimer and its secondary structure consists of four alpha-helices per monomer that align in an anti-parallel fashion. There is one beta-turn between helix 2 and helix 3 (Lys26 to Glu29); 76% of the residues are part of the alpha-helices. In the core of the dimeric protein molecule, between the two monomers that are held together by two disulfide bridges, we have observed a closed cavity. Its length is 15.6 A and its width is 9 A; 14 water molecules could be positioned inside. In the "bottom" part of the protein, near the C terminus, we have observed a smaller cavity, occupied by two water molecules. The calculation of the molecular surface revealed four surface pockets whose possible functional implications are discussed below.     

Variable expression of the uteroglobin gene following the administration of norethisterone and its A-ring reduced metabolites

Enzyme-mediated A-ring reduction of norethisterone (NET) results in the transformation of a molecule with potent intrinsic progestational activity into neutral derivatives with estrogen-like effects. To ascertain whether these structural modifications of NET are able to modify the uteroglobin (U) gene (G) expression, a series of experiments assessing the UG products after the administration of NET and its reduced A-ring metabolites were conducted in prepubertal female rabbits. Synthesis of endometrial uteroglobin and its specific mRNA were studied in animals following the administration of NET, 5 alpha-dihydro NET,3 beta,5 alpha-tetrahydro NET and progesterone. Animals treated with either estradiol or vehicle alone served as controls. The uteroglobin content in uterine flushings and cytosols was determined by immunodiffusion and polyacrilamide gel electrophoresis techniques and by a specific double-antibody radioimmunoassay, while the U mRNA synthesis was assessed by its molecular hybridization to [alpha 32P]d-ATP uteroglobin cDNA. NET induced a significant increase of the uterine content of uteroglobin similar to that observed with progesterone with a simultaneous increase on U mRNA synthesis. On the contrary, 5 alpha-NET and 3 beta,5 alpha-NET induced very little, if any uteroglobin synthesis with a concomitantly low U mRNA production as compared with NET; thus exhibiting a similar effect to that observed in estradiol-treated animals. The overall results were interpreted as demonstrating that the enzyme mediated structural changes of NET which occur at the target organs induce variable expression of the uteroglobin gene. The data indicate that the rabbit uteroglobin gene products are suitable molecular markers to evaluate the hormonal potency of contraceptive synthetic progestins and their derivatives.     

Glucocorticoid-dependent uteroglobin synthesis and uteroglobulin mRNA levels in rabbit lung explants cultured in vitro

In rabbit lung explants cultured in vitro in a synthetic medium, the synthesis of the protein uteroglobin decayed progressively becoming virtually undetectable between 24-48 h of culture. Addition of glucocorticoids to the medium maintained the synthesis of uteroglobin. This glucocorticoid effect was dose-dependent with optima at about 0.1 microM and 1 microM for dexamethasone and cortisol respectively. Estradiol, progesterone, triiodothyronine, insulin or 10% calf serum added to the medium were ineffective in maintaining uteroglobin synthesis. Actinomycin D (10 micrograms/ml) added to the medium inhibited the effect of cortisol on uteroglobin synthesis. After 24 h of culture, both the relative levels of uteroglobin mRNA, measured by molecular hybridization, and uteroglobin synthesis were correlatively higher (up to 10-fold) in glucocorticoid-treated than in control explants.     

Evidence that progesterone binding uteroglobin is similar to myosin alkali light chain

Using a computer program designed to detect evolutionary relationships between proteins, I find that exon 2 of rabbit uteroglobin, a progesterone binder, and part of myosin alkali light chain have a comparison score that is 7.2 standard deviations higher than that obtained with a comparison of randomized sequences of these proteins. The probability (p) of getting this score by chance is less than 10(-12). This theoretical finding that these sequences are similar has led to the experimental finding that copper, calcium and the tranquilizer trifluoperazine, a calmodulin binding ligand, affect progesterone binding to uteroglobin.