Investigation of the binding site of human corticosteroid-binding globulin by affinity labeling. Demonstration of a cysteinyl residue in the binding site

Molecular and structural basis of steroid hormone binding and release from corticosteroid-binding globulin

Corticosteroid-binding globulin (CBG), a non-inhibitory member of the serine proteinase inhibitor (serpin) super-family, is the high-affinity transport protein for glucocorticoids in vertebrate blood. Plasma CBG is a glycoprotein with 30% of its mass represented by N-linked oligosaccharide chains. Its well-characterized steroid-binding properties represent a "bench-mark data set" used extensively for in silico studies of protein-ligand interactions and drug design. Recent crystal structure analyses of intact rat CBG and cleaved human CBG have revealed the precise topography of the steroid-binding site, and shown that cortisol-bound CBG displays a typical stressed (S) serpin conformation with the reactive center loop (RCL) fully exposed from the central beta-sheet A, while proteolytic cleavage of the RCL results in CBG adopting a relaxed (R) conformation with the cleaved RCL fully inserted within the protein core. These crystal structures have set the stage for mechanistic studies of CBG function which have so far shown that helix D plays a key role in coupling RCL movement and steroid-binding site integrity, and provided evidence for an allosteric mechanism that modulates steroid binding and release from CBG. These studies have also revealed how the irreversible release of steroids occurs after proteolysis and re-orientation of the RCL within the R conformation. This recent insight into the structure and function of CBG reveals how naturally occurring genetic CBG mutations affect steroid binding, and helps understand how proteolysis of CBG enhances the targeted delivery of biologically active steroids to their sites of action.

Corticosteroid-binding globulin: the clinical significance of altered levels and heritable mutations

Corticosteroid-binding globulin (CBG) is the specific high-affinity plasma transport glycoprotein for cortisol. Stress-induced falls in CBG levels may heighten hypothalamic-pituitary-adrenal axis responses and CBG:tissue interactions may allow targeted cortisol delivery. Three genetic variants of CBG have been identified that reduce cortisol binding affinity and/or CBG levels. These include the Leuven and Lyon mutations which reduce CBG:cortisol binding affinity 3- and 4-fold, respectively, and the null mutation resulting in a 50% (heterozygote) or 100% (homozygote) reduction in CBG levels. The three reported null homozygotes demonstrate that complete CBG deficiency is not lethal, although it may be associated with hypotension and fatigue. The phenotype of a CBG null murine model included fatigue and immune defects. One community-based study revealed that severe CBG mutations are rare in idiopathic fatigue disorders. The mechanisms by which CBG mutations may cause fatigue are unknown. There are preliminary data of altered CBG levels in hypertension and in the metabolic syndrome; however, the nature of these associations is uncertain. Further studies may clarify the functions of CBG, and clinical observations may validate and/or extend the phenotypic features of various CBG mutations.

Modeling of Human Corticosteroid Binding Globulin. Use of Structure–Activity Relations in Soft Steroid Binding to Refine the Structure

PURPOSE

We propose a model for human corticosteroid binding globulin that is capable of explaining at the molecular level the experimentally observed binding affinities of four ligands. A new method of analyzing data from docking studies is proposed.

METHODS

Displacement of radioactive ligand by competitive binding gives the experimentally determined binding affinities of the competitors. A theoretical model, based on homology with crystallographically determined structures, was studied in an automated docking procedure for the determination of theoretical affinities. The docking runs were analyzed by a hybrid principal component-clustering analysis.

RESULTS

Of the two binding sites considered, only one--that in the vicinity of Cys 60--can reproduce the experimentally observed order of binding affinities although the lowest energies are found at the site in the vicinity of Cys 228.

CONCLUSIONS

Models proposed for proteins should be always conditioned to take into account experimentally observed results. In the current work, we have shown that an informed analysis of theoretical docking studies can lead to a more logical model of the protein, one that can explain and give a deeper understanding of the stereochemical requirements of binding.

Identification of thyroxine-binding globulin-San Diego in a family from Houston and its characterization by in vitro expression using Xenopus oocytes

T4-binding globulin (TBG) is a liver glycoprotein that transports iodothyronines in serum. Several TBG variants with reduced T4 binding affinity have been described, all of which are also characterized by reduced serum TBG concentrations and reduced heat stability. Their loss of binding thus appears to be due to a general defect of the molecule. We now report the occurrence of a variant TBG, detected in a family from Houston, TX, with half the normal T4 binding affinity and heat stability but normal serum concentration and isoelectric focussing pattern. The propositus was identified by reduced total T4 and T3 serum levels. All family members were euthyroid, and inheritance followed an X-linked pattern. Sequence analysis of the TBG gene of the propositus and his heterozygous mother revealed two amino acid substitutions: serine 23 with threonine (S23T), and the known polymorphism leucine 283 with phenylalanine (L283F). These substitutions are identical to those of TBG-San Diego (TBG-SD), a variant with similar properties except for a reduced serum concentration. Expression of recombinant TBG-SD/H with the S23T substitution in Xenopus oocytes reproduced the binding defect and heat lability. The amount of TBG-SD/H synthesized and secreted by the oocytes was not different from that of normal TBG. The difference in serum TBG concentrations in affected members of the San Diego and Houston families thus does not appear to be due to an error in the measurement of TBG, but may be related to differences in the rates of degradation.

Modularity of serpins. A bifunctional chimera possessing alpha1-proteinase inhibitor and thyroxine-binding globulin properties

An exciting application of protein engineering is the creation of proteins with novel functions by the retrofitting of native proteins. Such attempts might be facilitated by the idea of a mosaic architecture of proteins out of structural units. Even though numerous theoretical concepts deal with the delineation of structural "modules," their potential in the design of proteins has not yet been sufficiently exploited. To address this question we used a gain of function approach by designing modular chimeric molecules out of two structurally homologous but functionally diverse members of the superfamily of serine-proteinase inhibitors, alpha1-proteinase inhibitor and thyroxine-binding globulin. Substitution of two of four alpha1-proteinase inhibitor modules (Lys222 to Leu288 and Pro362 to Lys394, respectively), identified by alpha-backbone distance analysis, with their thyroxine-binding globulin homologues resulted in a bifunctional chimera with inhibition of human leukocyte elastase and high affinity thyroxine binding. To our knowledge, this is the first report on a bifunctional chimera engineered from modules of homologous globular proteins. Our results demonstrate how a modular concept can facilitate the design of new functional proteins by swapping structural units chosen from members of a protein superfamily.

Substitutions of tryptophan residues in human corticosteroid-binding globulin: impact on steroid binding and glycosylation

Human corticosteroid-binding globulin (CBG) contains four tryptophan residues at positions 141, 185, 266 and 371; one of which is thought to be located in the steroid-binding site. These residues were substituted by site-directed mutagenesis and expression of mutant CBG cDNAs in Chinese hamster ovary cells. Analyses of the resulting mutants indicate that Trp371 is most likely located in the steroid-binding site, and that hydrophobic interactions between Trp141 and the steroid molecule or other amino-acids in the CBG polypeptide may also contribute to high-affinity interactions between CBG and its steroid ligands. In addition, substitution of Trp266 resulted in altered glycosylation of CBG, and this supports the concept that it participates in intra-molecular carbohydrate-polypeptide interactions which may influence the conformation and secretion of this glycoprotein.

Identification of the cysteine in the steroid-binding site of human corticosteroid binding globulin by site-directed mutagenesis and site-specific chemical modification

Binding of corticosteroids by human corticosteroid binding globulin (hCBG) is thought to involve interaction with one of its two cysteine residues (Cys60 and Cys228). To identify which of the two cysteine residues mediates steroid binding, we have produced mutant hCBGs containing serine or alanine in place of Cys228 by site-directed mutagenesis. Alteration of Cys228 to serine or alanine does not change the steroid binding affinity of hCBG, demonstrating that Cys228 is not involved in the binding interaction. This finding strongly suggests that Cys60 is the functionally important cysteine. By modifying the wild-type and mutant hCBGs with the sulfhydryl-specific reagents N-ethylmaleimide, iodoacetamide, and sodium tetrathionate, we have demonstrated that Cys60 is present at the steroid binding site, and that it may be directly involved in steroid binding. This result also identifies Cys60 as the accessible cysteine reported in previous studies.

Decreased cortisol-binding affinity of transcortin Leuven is associated with an amino acid substitution at residue-93

Genomic DNA was isolated from two related individuals who are homozygous for transcortin Leuven, a corticosteroid-binding globulin variant with decreased cortisol-binding affinity. This material was amplified using intron-specific oligonucleotide primers in a polymerase chain reaction to obtain the four exons that encode transcortin. Sequence analysis of these exons showed several mutations within the coding sequence of both individuals, but only one of these will result in an amino acid substitution. This mutation is located within exon 2 and alters the codon (CTC) normally associated with Leu-93 in the transcortin polypeptide to a codon (CAC) for histidine in the variant genes.

Molecular studies of corticosteroid binding globulin structure, biosynthesis and function

Phylogenetic comparisons of the primary structure of corticosteroid binding globulin (CBG) have revealed several conserved domains that include sites for N-glycosylation and a region which probably represents a portion of the steroid binding site. The major site of CBG biosynthesis in adults is clearly the liver, and the human CBG gene promoter contains sequence elements that interact with liver-specific transcription factors. Low levels of CBG gene expression have been detected in other tissues, and these may be important for fetal development during late gestation when hepatic CBG mRNA levels are low. Studies of the ontogeny of CBG biosynthesis in the rat have also indicated that plasma CBG levels may be influenced by a more rapid clearance of the protein during pubertal development. Analyses of the structural organization and chromosomal location of the human CBG gene have further confirmed its close relationship with the serine proteinase inhibitors, and suggests that CBG, alpha 1-proteinase inhibitor and alpha 1-antichymotrypsin evolved relatively recently by gene duplication. The functional significance of this relationship has been examined and our studies suggest that a specific interaction between CBG and elastase on the surface of neutrophils may represent a physiologically important event that promotes the delivery of glucocorticoids to these cells at sites of inflammation.

Corticosteroid binding globulin, testosterone-estradiol binding globulin, and androgen binding protein belong to protein families distinct from steroid receptors

The cDNA nucleotide sequences and the deduced amino acid sequences of human corticosteroid binding globulin (hCBG), human testosterone-estradiol binding globulin (hTeBG), and rat androgen binding protein (rABP) were determined. Studies of the steroid binding sites suggest they are toward the carboxy-terminus in hTeBG and rABP and more central in hCBG. hCBG has remarkable sequence homology with members of a superfamily whose functions have diverged; these include thyroxine-binding protein, serine protease inhibitors, egg white proteins, and angiotensinogen. hTeBG and rABP have a 68% amino acid sequence identity. Hybridization studies suggest that hTeBG is probably even more closely related, if not identical, to hABP. The carboxy-terminal sequences of hTeBG and rABP are also similar to that of protein S, a vitamin-K-dependent clotting factor. There were no nucleotide or amino acid sequence homologies between hCBG, hTeBG, or rABP and other steroid binding proteins such as steroid receptors, albumin, alpha-fetoprotein, and vitamin D binding protein. We conclude that the "extracellular steroid binding proteins" and steroid receptors do not appear to have descended from a common ancestor.

Physico-chemical properties and evidence for electrophoretic variants of rat transcortin

Isolation of rat plasma transcortin was carried out by affinity chromatography, as previously described for human. The protein was shown to be pure by PAGE and one single N-terminal amino acid was identified (Ser), which suggested that the protein molecule has a single polypeptide chain. This assumption is supported by SDS-PAGE. The amino acid composition was reported and compared with the one of human transcortin. The purified protein always migrated in PAGE (with or without SDS) as a double band; the faster component being more intense than the slower one. Whether transcortin was free or bound to corticosterone, the same aspect was observed. Molecular weight of these two variants were determined by SDS-PAGE as 65,900 and 75,800. Polymers only appeared after irreversible denaturation of the protein, as previously described for human transcortin. Various other physical parameters were determined: a sedimentation coefficient of 3.71 S +/- 0.18 was calculated by ultracentrifugation in sucrose gradient, association constants at 4 degrees C for corticosterone and cortisol (2.7 X 10(9) M-1 and 4.2 X 10(8) M-1, respectively).

Reversible dissociation of cortisol-transcortin complex by sodium para-chloromercuribenzoate

Mercurials are considered as sulphydryl group specific reagents and one of them, sodium para-chloromercuribenzoate (PCMB), is currently used for SH titration. It has been shown that cellular steroid receptors are reversibly inactivated by mercurials even when the binding site is occupied by the steroid (Coty, W.A. (1980) J. Biol. Chem. 255, 8035-8037). This is a striking difference with alkylating SH reagents such as iodoacetic acid or N-ethylmaleimide, since these reagents inactivate only steroid-free receptors. In order to explain this discrepancy, we tested, in the present study, the specificity of PCMB on a blood plasma steroid binding protein: human transcortin. This protein presents the advantage, over cellular receptors, of being well characterized and to be available in a pure state. The transcortin-cortisol complex was also reversibly inactivated by PCMB when the reaction was carried out at a high excess of reagent over protein; such conditions are those previously used with steroid receptors. The reversibility was obtained not only with a reducing agent (dithiothreitol) but also with EDTA, which suggests a poor stability of the protein mercurial bond and therefore a nonspecific action. The decrease of activity was the result of a loss of binding sites and Scatchard plot analysis did not reveal any detectable decrease of the affinity constant for cortisol. Transcortin possesses two SH groups per molecule, one of these being buried in native conformation. After blockage of the accessible SH group by aminoethylation, transcortin kept the same activity, but when this aminoethylated transcortin was incubated with PCMB a loss of activity was obtained, although the residual buried SH group was again titrable with Ellman's reagent. Therefore, we can conclude that the action of PCMB on proteins must be interpreted with precaution, since it can induce an inactivation that is SH-independent.

Steroid-protein interaction: from past to present

In this review, the association between biologically active compounds and proteins is seen in the light of axioms expressed by Paracelsus and Paul Ehrlich long ago, and the physiological significance of the interactions is pointed out. Of the various types of proteins that form noncovalent complexes with steroid hormones, only the serum proteins will be discussed. Recent results, obtained in several laboratories, on the physicochemical properties of the human corticosteroid-binding globulin (CBG, transcortin) makes this glycoprotein perhaps the best known one among the steroid-binding serum proteins. Influence of pH on stability of the complexes, kinetics of the associations and their significance, as well as thermodynamic parameters of complex formation are being discussed. Characteristics of the binding sites are deduced from specificity studies. Influence of the entrance of hydrophilic or hydrophobic substituents into the steroid molecule illuminates the difference between typically hydrophobic binders such as the progesterone-binding globulin (PBG) of the pregnant guinea-pig and typically hydrophilic binders such as CBG. Complete elucidation of the steroid-binding proteins awaits the determination of the amino acid sequence and the X-ray crystallographic analysis of the steroid-protein complex.

Identification of histidine and methionine residues in the active site of the human uterine progesterone receptor with the affinity labels 11 alpha- and 16 alpha-(bromoacetoxy)progesterone

The affinity labels 11 alpha- and 16 alpha-(bromo[2'-3H]acetoxy)progesterone (BAP) react covalently with amino acids present in the progesterone binding site of the human uterine progesterone receptor. Hydrolysis of the affinity labeled receptor followed by separation and analysis of the amino acid products demonstrated the sites of affinity labeling. The 11 alpha-BAP alkylates the 1-position of a histidine residue. The 16 alpha-BAP alkylates the 3-position of histidine, and a methionine residue. Affinity labeling did not occur in the presence of excess progesterone, and under the optimum conditions for affinity labeling of the receptor, heat-denatured receptor, bovine serum albumin, and 20 beta-hydroxysteroid dehydrogenase were not affinity labeled. This is the first report of the identification of specific amino acid residues in the binding site of a steroid hormone receptor.

Steroid-protein interactions. Human corticosteroid binding globulin: some physicochemical properties and binding specificity

Reducing agents (dithiothreitol and beta-mercaptoethanol) significantly decrease the affinity constants of the human corticosteroid-binding globulin (CBG)-cortisol complex in proportion to their concentration; the resulting Ka values are more consistent than those obtained in the absence of the reductants. The effect is reversible. The equilibrium association constants of the CBG complexes with cortisol and progesterone show a relatively broad pH maximum between pH 8 and 11. In this pH range, cortisol was found to be bound more strongly than progesterone; this relationship is reversed around pH 6. The van't Hoff plot of the temperature effect on Ka of the CBG-cortisol complex (4-41 degrees C) exhibits a nonlinear, possibly biphasic temperature dependency. The shape of the van't Hoff plot was similar in the presence of mercaptoethanol. The association of cortisol and progesterone to human CBG at 4 and 37 degrees C is enthalpy driven, compensating for the unfavorable change in entropy. Studies with 47 steroids served to elucidate the influence on binding affinity of polar and nonpolar groups and other structural alterations. The contribution of specific structural changes in the steroid molecule to the free energy of binding can be calculated from the results. Important structures for optimal binding are the 20-oxo group, a 10 beta-methyl group, and a double bond at the 4 position. A complementary image of the binding site with respect to the nature of binding at various locations is proposed.

Photoaffinity labeling of steroid binding proteins with unmodified ligands

Photoactivation of the alpha,beta-unsaturated ketones of natural and synthetic steroid molecules by light of lambda greater than or equal to 330 nm allows their covalent attachment to steroid-binding proteins. The general validity of this method is demonstrated with two steroid hormone receptors and the steroid-binding protein uteroglobin. Progesterone can be covalently attached to the partially purified progesterone receptor and to uteroglobin, and comigrates with the binding proteins upon electrophoresis in polyacrylamide gels containing sodium dodecyl sulfate. Similarly the synthetic glucocorticoid triamcinolone acetonide can be covalently bound to the partially purified glucocorticoid of rat liver. This method allows the identification of steroid hormone receptors after electrophoresis in polyacrylamide gels containing sodium dodecyl sulfate. Labeling with radioactive steroids is specific since it can be prevented by the addition of an excess of non-radioactive ligand. Digestion of the labeled binding proteins with trypsin or chymotrypsin yields a defined pattern of radioactive peptides, demonstrating that covalent attachment takes place at specific binding sites.

Electron spin resonance study of human transcortin: Thiol groups and binding site topography

A series of cortisol analogs bearing a nitroxide free radical on C-17 side chains with a variation of distance between the steroid D-ring and the spin label from 7.4 to 17.6 A has been synthesized. These analogs were found to retain a good affinity for the specific corticosteroid binding site of purified human transcortin. The spin-labeled cortisol analogs were used to probe the human transcortin binding site structure by electron spin resonance (ESR) spectroscopy. A total depth of approx. 25 A was estimated for the binding site crevice. Use of sulfhydryl reagents (N-ethylmaleimide, p-chloromercuribenzoate) showed that a maximum of two sulfhydryl groups were titratable after reduction and denaturation of the protein. One of these thiol groups appeared to be involved in the cortisol binding site and could not be detected in the presence of bound steroid. ESR study of its environment, using spin-labeled N-ethylmaleimide reagents of various side-chain lengths, led to the conclusion that this thiol was at a depth of approx. 15 A or more in the binding site cavity. The second sulfhydryl group may be present in an oxidized form in the purified native transcortin, since it became titratable only after reductive treatment of the protein. ESR study showed that this thiol may be located in a crevice at approx. 15 A from the protein surface. These findings are compatible with a structural organization of the transcortin cortisol binding site, taking into account tentative models previously proposed by others.

Kinetic and equilibrium studies on steroid interaction with human corticosteroid-binding globulin

Kinetic and equilibrium studies on the interaction of steroids with human corticosteroid-binding globulin (CBG, transcortin) were performed with pH, temperature, and steroid structure as variables. Dissociation rate constants were determined fluorometrically; the values for cortisol, corticosterone, deoxycorticosterone, and progesterone are 0.031, 0.047, 0.10, and 0.16 s-1, respectively, at 20 degrees C, pH 7.4. The pH dependence of the dissociation rate constant for the corticosterone complex below pH 10.5 at 20 degrees C is given by koff = 0.043 (1 + [H+]/10(-6.50)) s-1; above pH 11, koff = 0.030 (1 + 10(-12.15/[H+] s-1. A temperature-dependence study of koff for the cortisol and progesterone complexes gave values of 0.0028 s-1 and 0.012 s-1 at 4 degrees C, respectively, and 0.88 s-1 and 4.5 s-1 at 37 degrees C, with progesterone dissociating about four to five times faster over the entire temperature range. The affinity constants, determined by equilibrium dialysis, for the binding of cortisol, corticosterone, and progesterone at 4 degrees C were 7.9, 7.2, and 7.0 X 10(8) M-1; values of 0.40 and 0.26 X 10(8) M-1 were determined at 37 degrees C for cortisol and progesterone. The close similarity of the affinity constants of the three steroids combined with differing dissociation rates implies that the association rate changes with steroid structure, in contrast to our earlier findings with progesterone-binding globulin.

Affinity labelling of human transcortin

The binding site of transcortin has been studied by using bromoacetyltestosterone and bromoacetylated derivatives of progesterone which were monohydroxylated at different positions of the steroid nucleus. Specificity of affinity labelling was demonstrated by the displad cortisol analog was added to a [3H]cortisol-transcortin complex solution. The binding site crevice was found to be very narrow in the vicinity of the A and B rings of steroid since 2alpha-hydroxyprogesterone, 6alpha- or 6beta-bromoacetoxyprogesterone and dexamethasone could not displace bound cortisol. A specific affinity labelling was obtained with 11alpha-bromoacetoxyprogesterone, 16alpha-bromoacetoxyprogesterone and 17beta-bromoacetyltestosterone. The results of the affinity labelling by these hormone analogs suggested that one methionine and one histidine residues were located within the active site:methionine might interact with the 11beta-hydroxyl group and histidine with the 20 keto group of cortisol.