Shbg) Expression

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Helsingin yliopiston digitaalinen arkisto

Regulation of Human Sex Hormone-Binding Globulin Gene (shbg) Expression

by Marja Jänne

Department of Biosciences, Division of Biochemistry, Faculty of Science, University of Helsinki, Finland

Academic Dissertation

To be presented for public criticism, with the permission of the Faculty of Science, University of Helsinki, in the auditorium 1041 at Viikki Biocenter, Viikinkaari 5, Helsinki, on April 14th, at 12 o’clock noon. Supervised by: Professor Geoffrey Hammond London Regional Cancer Centre and Departments of Obstetrics and Gynecology, and Pharmacology and Toxicology, University of Western Ontario, Canada

Reviewed by: Professor Pirkko Viliko Department of Biosciences, Division of Biochemistry, Faculty of Science, University of Helsinki, Finland

and

Docent Jorma Palvimo Department of Physiology, Institute of Biomedicine, University of Helsinki, Finland

Opponent: Professor llpo Hulitaniemi Department of Physiology, Institute of Biomedicine, University of Turku, Finland

ISBN 951-45-9170-4 (PDF version) Helsingin yliopiston verkkojulkaisut Helsinki 2000 omistettu maailman parhaalle ihmiselle, äidilleni Table of Contents

Original Publications...... 7

Abbreviations...... 8

1. Abstract...... 9

2. Review of the Literature...... 10 2.1 Sex hormones...... 10 2.1.1 Introduction...... 10 2.1.2 Biosynthesis...... 10 2.1.3 Mechanisms of action...... 12 2.1.3.1 Classical pathway...... 12 2.1.3.2 Non-genomic actions...... 13 2.2 Sex hormone-binding globulin (SHBG)...... 13 2.2.1 Introduction...... 13 2.2.2 SHBG/Androgen Binding Protein (ABP)...... 14 2.2.3 Species distribution...... 14 2.2.4 Quantitative measurements...... 15 2.2.5 Structure...... 15 2.2.5.1 Subunits and dimerization...... 15 2.2.5.2 Steroid binding...... 15 2.2.5.3 Glycosylation...... 16 2.2.5.4 Homology to other proteins...... 16 2.2.5.5 Crystal structure...... 17 2.2.6 Physiological functions...... 17 2.2.6.1 Plasma sex steroid carrier...... 17 2.2.6.2 Interaction with cell membranes...... 18 2.2.6.3 Function in the male reproductive tract...... 20 2.2.7 Gene structure and tissue-specific expression...... 20 2.2.8 Regulation of plasma SHBG levels...... 22 2.2.8.1 Sex hormones...... 22 2.2.8.2 Nutrition and body mass...... 22 2.2.8.3 Insulin...... 23 2.2.8.4 Thyroid hormone...... 23

3. Aims of the present study...... 24

4. Materials and methods...... 24

5. Results...... 26 5.1 Human shbg transgenic mice...... 26 5.1.1 Characterization of transgenic mouse lines...... 26 5.1.2 Tissue-specific expression of human shbg in mice...... 26

5 5.1.3 Human shbg transcription start sites in different mouse tissues...... 26 5.1.4 Alternative human shbg transcripts in mouse testis...... 27 5.1.5 Cellular localization of human SHBG mRNA and immunoreactive protein ...... 27 5.1.6 Human SHBG in mouse plasma and urine...... 27 5.1.7 Phenotypic analysis of human shbg transgenic mice...... 27 5.2 Regulation of human shbg expression in transgenic mice during development...... 28 5.2.1 Human shbg expression in mouse fetus...... 28 5.2.2 Serum SHBG levels during development...... 28 5.2.3 Correlation between serum SHBG and testosterone levels ...... 28 5.2.4 Testis-specific human shbg expression during development...... 28 5.2.5 Effect of androgens on human shbg expression in mouse kidney...... 29 5.3 In vitro characterization of human shbg liver-specific promoter ...... 29 5.3.1 Rodent and human shbg 5’-flanking regions...... 29 5.3.2 Possible targets for transcription factor binding within human shbg proximal promoter...... 29 5.3.3 Minimal promoter required for liver-specific transcription...... 29 5.3.4 Possible TATA-box within human shbg promoter...... 30 5.3.5 Orphan receptors COUP-TF and HNF-4 bind to human shbg promoter...... 30 5.3.6 Site-specific mutations at FP1 and FP3 reduce human shbg promoter activity.30 5.3.7 Transactivation of human shbg promoter by HNF-4 is antagonized by COUP- TF...... 31 5.3.8 Optimal COUP-TF/HNF-4 site within FP1...... 31 5.4 Putative silencer in the human shbg promoter...... 31 5.4.1 Characterization of human shbg region -803/-299...... 31 5.4.2 Human shbg region -536/-511 binds ubiquitous factor Sp1...... 32 5.4.3 Mutation within the Sp1 site partially recovers human shbg promoter activity.32 5.4.4 Analysis of human shbg -803/-299 deletion fragments...... 35 5.4.5 Removal of AT-rich sequence at -727/-697 partially recovers human shbg promoter activity...... 35

6. Discussion...... 36 6.1 Tissue-specific expression of human shbg ...... 36 6.1.1 Regulation of serum SHBG levels...... 36 6.1.2 Transgenic mouse testis produces mainly alternative human shbg transcripts..37 6.1.3 Human SHBG production in the transgenic mouse kidneys is regulated by androgens...... 37 6.2 Physiological consequences of human shbg expression in transgenic mice...... 38 6.3 Tissue-specificity of human and rodent shbg promoters...... 39 6.4 Human shbg expression is regulated by HNF-4 in liver cells...... 40 6.5 Silencing of human shbg promoter...... 41

7. Acknowledgements...... 43

8. References...... 44

6 Original Publications

This thesis is based on the following original articles, which are referred to by their roman numerals in the text.

I Jänne M, Deol HK, Power SGA, Yee S-Y, and Hammond, GL (1998). Human sex hormone- binding globulin gene expression in transgenic mice. Mol Endocrinol 12: 123-136

II Jänne M, Hogeveen K, Deol H, and Hammond GL (1999). Expression and regulation of human sex hormone-binding globulin transgenes in mice during development. Endocrinology 140: 4166-4174

III Jänne M and Hammond GL (1998). Hepatocyte nuclear factor-4 controls transcription from a TATA-less human sex hormone-binding globulin gene promoter. J Biol Chem 273: 34105-34114

In addition, some unpublished data are presented.

7 Abbreviations

ABP androgen binding protein AFP alpha-fetoprotein AR androgen receptor BCA binding capacity assay bp base pair BPH benign prostatic hyperplasia BSA bovine serum albumin COUP-TF chicken ovalbumin upstream promoter transcription factor DES diethylstilbestrol DHT dihydrotestosterone EMSA electrophoretic mobility shift assay FP footprint FSH follicle stimulating hormone GTFs general transcription factors hCG human chorionic gonadotrophin HDL high-density lipoprotein HNF-1 hepatocyte nuclear factor-1 HNF-4 hepatocyte nuclear factor-4 HRE hormone response element IRMA immunoradiometric assay KAP kidney androgen responsive protein kb kilo base pair kDa kilo Dalton LH luteinizing hormone LNCaP lymph node of prostatic cancer MCR metabolic clearance rate MODY maturity onset diabetes of the young NIDDM non insulin-dependent diabetes mellitus nt nucleotide ODC ornithine decarboxylase PSA prostate-specific antigen SHBG sex hormone-binding globulin Sp1 transcription factor TBP TATA-binding protein

T3 thyroid hormone THR thyroid hormone resistance

8 1. Abstract

Sex hormone-binding globulin (SHBG) is a plasma glycoprotein that binds androgens and estrogens with high affinity. It is produced by hepatocytes and secreted into the blood, where it regulates the biological activities of male and female sex steroids.

Unlike the human gene, shbg in rodents, such as rats and mice, is silent in the liver postnatally, and the gene expression is mainly restricted to the testes. The testicular gene product is referred to as the androgen binding protein (ABP), as it binds androgens with high affinity in the male reproductive tract. Human testis produces alternative shbg products, the function(s) of which is largely unknown.

To understand why shbg is expressed so differently in humans and rodents and to learn more about the function of the gene product, several lines of transgenic mice carrying human shbg in their genome were produced. These mice express the human gene in the liver and have appreciable amounts of circulating SHBG in the blood. Although it was anticipated that the presence of this high-affinity steroid binding protein would result in imbalances in free and bound sex hormone levels, the reproductive performance of these mice was not compromised in any way.

The comparisons of the expression patterns of two human shbg transgenes with different amounts of 5’-flanking sequences (~6 kb or ~0.85 kb) provided information about the regulatory mechanisms that govern human shbg expression and regulation in vivo. Specifically, upstream sequences present only in the larger human shbg transgene contributed to sexual dimorphism of SHBG levels in mouse plasma, while both transgenes were subject to the regulation by androgens in the kidney. Furthermore, testicular shbg expression required additional upstream sequences and was predominantly driven by a separate promoter.

In humans, plasma SHBG levels are regulated by different physiological and pathological conditions as well as nutrition and body mass. Hepatic SHBG production has also been studied in cultured cells, but little is known about the transcriptional regulation of the gene. In vitro studies demonstrated that the proximal human shbg promoter (-137/+7) is active in cultured liver cells and that it is regulated by an orphan steroid receptor, hepatocyte nuclear factor-4 (HNF- 4). In essence, HNF-4 seems to substitute for TATA-binding protein (TBP) in the recruitment of transcription initiating complex to the human shbg promoter. Detailed promoter analyses also revealed a putative silencer within the human shbg promoter. This silencer comprises, at least in part, a TA-rich stretch and it may co-operate with a DNA element that binds a ubiquitous transcription factor Sp1, or another related factor recognizing GC-rich sequences.

9 2. Review of the literature

2.1 Sex Hormones

2.1.1 Introduction

Sex steroids are responsible for normal sexual development, and the establishment of secondary sexual characteristics of male and female reproductive organs. They also induce sexual dimorphism in a number of other tissues, such as muscle, bone, liver, kidney, and brain. Being small, fat-soluble molecules, their effects are far-reached from the site of biosynthesis, and their actions are subject to strict regulation at multiple levels.

2.1.2 Biosynthesis

The biosynthesis of all steroid hormones begins with cleavage of the 6-carbon side chain of cholesterol to yield the 21-carbon precursor pregnenolone. The enzymatic steps of the two parallel routes that produce estrogens and androgens from pregnenolone have been well characterized and are shown in figure 1. The steroid A-ring can be oxidized at any point from pregnenolone to androstenediol, and the conversion of both androstenedione to testosterone and estrone to estradiol by 17β-dehydroxysteroid dehydrogenase is reversible. Although the steroidogenic enzymes are basically the same in the gonads of both genders, the steroid biosynthesis is regulated by pulsatile release of pituitary gonadotrophins in males, as opposed to cyclic release in females. The gonadotrophin secretion from the pituitary is also subject to negative feedback regulation by gonadal steroids.

Testosterone biosynthesis takes place in the Leydig cells located in the interstitial compartment between the seminiferous tubules in the testis. The synthesis and secretion of testosterone is stimulated by luteinizing hormone (LH) derived from the anterior pituitary. Although testosterone itself is responsible for several androgen effects, its 5α-reduced metabolite, 5α- dihydrotestosterone (DHT), is significantly more potent in some target organs, such as prostate and seminal vesicles. Also, the conversion of testosterone to estrogens by aromatase occurs in certain male tissues (i.e. brain and bone) to ensure a proper androgenic response.

In the ovary, thecal and granulosa cells synthesize testosterone and androstenedione, which are the precursors for estradiol and estrone, respectively. Estrogen synthesis from androgens in granulosa cells is catalyzed by aromatase. In addition to initiating the follicle development, the cyclic release of follicle stimulating hormone (FSH) from the pituitary increases the aromatase levels in the ovary. Subsequently, LH triggers the complete maturation of the follicle and induces ovulation. Large amounts of estrogen and progesterone are secreted by the corpus lu- teum derived from the ruptured follicle. Several estrogen responsive tissues, such as the breast adipose tissue, also contain aromatase probably to ensure high local concentration of estrogen in this tissue.

10 Figure 1. The Biosynthesis of sex steroid hormones. Enzymes involved in the numbered steps are: 20,22-desmolase (1), 17-hydroxylase (2), 17,20-desmolase (3), 17β-OH-steroid dehydrogenase (4), 3β-ol-dehydrogenase and ∆4,5-isomerase (5), 5α-reductase (6), 3α-reductase (7), and aromatase (8).

11 2.1.3 Mechanisms of action

2.1.3.1 Classical pathway

A large body of data is available on steroid hormone effects through genomic actions mediated by intracellular steroid receptors. This multistep process begins as the free steroid enters the target cell presumably by passive diffusion. Inside the cell, the steroid is recognized and bound by its receptor, and depending on the steroid and its cognate receptor, the binding occurs in the cytoplasm or nucleus. Upon complex formation with a proper ligand, the receptor undergoes conformational changes, which allow it to bind to specific DNA element and act as a transcription factor. Subsequent transcriptional activation or repression of the target gene reflects the effect of a given steroid hormone.

The nuclear receptor superfamily consists of structurally related proteins with distinct ligand preferences, i.e. steroid and thyroid hormones, vitamin D, and retinoids. Several nuclear receptors with unidentified ligands also exist, and these are referred to as orphan receptors. The receptor molecule is composed of three well-characterized functional domains; the N-terminal hypervariable region that contains the transactivation domain, the middle region that possesses the zinc-finger structure that recognizes and binds the DNA, and the C-terminal domain that is responsible for the ligand binding (Mangelsdorf et al., 1995; Tsai and O’Malley, 1994) (Figure 2).

The DNA binding by nuclear receptors is not entirely specific, and there is significant conser- vation among the hormone response elements (HRE). Generally the favored sequence contains a direct or inverted imperfect repeat of 6 bases with variable number of spacing nucleotides. While some of the nuclear receptors bind DNA exclusively as homodimers, several others can also form heterodimers with other family members (Mangelsdorf et al., 1995; Tsai and O’Malley, 1994).

The binding to the response element involves chromatin remodeling as the entire eucaryotic DNA is packaged in nucleosomes, which inhibit the initiation of transcription by general transcription factors (Kingston et al., 1996; Felsenfeld, 1992). Nuclear receptors not only interact with the general transcription machinery, but also with other sequence-specific transcription factors present in the nucleus as well as with numerous putative coactivators and repressors, which adds on to the diversity of nuclear receptor-mediated steroid hormone action (McKenna et al., 1999).

Hormone antagonists, which also bind to the nuclear receptor with high affinity, specifically block the modulation of target gene transcription by sex steroids. These include tamoxifen for estradiol and bicalutamide for testosterone. By contrast, a non-steroidal compound, diethylstilbestrol (DES), acts as a very potent estrogen by binding to and activating the estrogen receptor (Fuhrmann U et al., 1998).

Figure 2. Functional domains of nuclear receptors.

12 2.1.3.2 Non-genomic actions

As the classical model cannot explain all steroid effects, alternative mechanisms for steroid hormone action have been searched for. Perhaps the best-characterized system is the sperm acrosomal reaction (acrosomal membrane fuses with sperm membrane), which can be stimulated in vitro by human follicular fluid or progesterone alone. Blackmore et al. (1990, 1991) have shown that progesterone induces a very rapid rise in intracellular Ca2+-levels in human sperm. In addition to being an immediate response for progesterone, Ca2+-accumulation is also achieved with steroid coupled with bovine serum albumin (BSA), thus unable to diffuse into cell. There- fore, it seems likely that the effect is not mediated by a nuclear, but rather by a membrane receptor for progesterone. Furthermore, synthetic progesterone agonists, which readily activate the cognate nuclear receptor, fail to mimic the effect in full, which suggests that these two types of receptors differ also with respect to ligand preferences.

Sertoli cells, the male analog for granulosa cells, respond to testosterone by increasing the intracellular calcium levels. In support of plasma membrane action, similar to the progesterone effect on granulosa cells, even a BSA conjugated steroid is capable of triggering the calcium increase (Gorczynska and Handelsman, 1995).

Several reports have identified binding sites for estradiol on target cells (Pietras and Szego, 1977, 1980; Bression et al., 1986). Estradiol also induces a rapid increase in intracellular calcium in chicken and pig granulosa cells. An estrogen receptor antagonist, tamoxifen, does not block the effect, nor does a potent estrogen agonist, DES, mimic it. The estrogen-induced calcium increase is also resistant to RNA and protein synthesis inhibitors, actinomycin and cycloheximide, respectively, both of which would block any genomic steroid effect (Morley et al., 1992). In rat uterine cells, estrogen stimulates adenylate cyclase directly without prior RNA or protein synthesis. This leads to the activation of a reporter gene that contains a cAMP response element rather than an estrogen response element (Aronica et al., 1994). The effect is rather unique, since it is obtained with both estrogen agonist and antagonist.

In summary, several lines of evidence confirm that the classical pathway involving nuclear receptors does not mediate all steroid effects. First, some of the primary effects are extremely rapid, and they are independent of transcription and translation. Second, the molecules mediating these rapid effects have ligand preferences different from the nuclear receptors. Finally, the responses occur even without the steroid molecule entering the cell, which is the requirement for nuclear receptor binding.

2.2 Sex Hormone-binding Globulin (SHBG)

2.2.1 Introduction

In addition to being regulated at the levels of biosynthesis, metabolism, and their nuclear/ membrane receptors, steroid hormones are bound by specific transport proteins in plasma. Unlike serum albumin, which reversibly occupies a significant fraction of steroids in blood, these proteins bind steroids with high affinity and low capacity. Traditionally it was thought that extracellular binding proteins merely regulate the bioavailability of steroid and other hormones, but accumulating evidence suggests more complex mechanisms by which these molecules modulate hormone action (Hammond, 1990, 1995).

13 2.2.2 SHBG/Androgen binding protein (ABP)

Sex hormone-binding globulin (SHBG) was first identified as a β-globulin that bound 17β- estradiol and testosterone in human plasma (Rosenbaum et al., 1966; Kato and Horton, 1968; Vermeulen and Verdonck, 1968; Rosner and Deakins, 1968; Lea and Stoa, 1972). By employ- ing various biochemical methods, such as charcoal adsorption of testosterone (Lea and Stoa, 1972), Sephadex gel filtration (Kato and Horton, 1968; Vermeulen and Verdonck, 1968; Pearlman and Crepy, 1967), equilibrium dialysis (Vermeulen and Verdonck, 1968; Lea and Stoa, 1972), and electrophoresis (Rosenbaum et al., 1966; Kato and Horton, 1968; Rosner and Deakins, 1968), several independent groups were able to demonstrate that this globulin in human serum binds testosterone and estradiol with much higher affinity than purified serum albumin, which was traditionally thought to be the main steroid transporter in plasma. It was also found that in normal females, but not males, testosterone is predominantly bound to SHBG (Rosner and Deakins, 1968; Lea and Stoa, 1972). Similar to several other plasma proteins, human SHBG is produced in the liver (Khan et al., 1981; Hammond et al., 1989; Gershagen et al., 1989) and secreted into blood where it serves as a modulator of sex steroid action.

An extracellular protein with high-affinity for androgens exists also in the male reproductive tract of several species and is referred to as the androgen binding protein (ABP). This androgen-binding activity, distinct from nuclear androgen receptor, was first identified in the cytosolic preparations from the rat epididymis (Ritzen et al., 1971; Hansson, 1972). Subsequent experiments revealed that the amount of epididymal ABP increases towards maturity (Weddington et al., 1974), and when the flow of testicular fluid to epididymis is blocked, the androgen- binding activity is significantly reduced in the epididymis with concomitant increase in the testis (French and Ritzen, 1973; Danzo et al., 1974). Immunocytochemical studies on male rats confirmed that ABP is made in the testis and is transported to and taken up by epithelial cells in the caput epididymis (Pelliniemi et al., 1981; French and Ritzen, 1973; Feldman et al., 1981).

Human SHBG and rat ABP are both dimeric glycoproteins showing significant homology in their structure as well as steroid-binding and immunological properties (Que and Petra, 1987; Musto et al., 1980; Cheng and Musto, 1982; Joseph et al., 1987, 1985; Hammond et al., 1987; Walsh et al., 1986; Reventos et al., 1988; Kotite et al., 1986), and it is now known that plasma SHBG and testicular ABP are the same protein with different tissue origin; SHBG is synthesized by hepatocytes (Khan et al., 1981) and ABP by Sertoli cells of the testis (Kierszenbaum et al., 1980; Steinberger et al., 1975; Hagenas et al., 1975). Moreover, within a given species, a single gene encodes both proteins (Que and Petra, 1987; Joseph et al., 1987, 1985; Walsh et al., 1986; Reventos et al., 1988; Gershagen et al., 1987). This review mainly focuses on plasma SHBG, and only some aspects concerning the testicular ABP are discussed.

2.2.3 Species distribution

An SHBG has been identified in the plasma of several species (Corvol and Bardin, 1973; Renoir et al., 1980; Wenn et al., 1977) with somewhat varying steroid binding specificity. Rodents, such as rats and mice do not have a circulating binding protein for androgens or estrogens (Corvol and Bardin, 1973; Renoir et al., 1980; Rosner and Deakins, 1968; Ritzen et al., 1971), and the SHBG in rabbit plasma binds only androgens (Rosner and Darmstadt, 1973). Apparently, only in primates, SHBG has a high affinity for both androgens and estradiol (Siiteri et al., 1982).

14 2.2.4 Quantitative measurements

A commonly applied procedure to compare relative affinities of different steroids for SHBG and to determine SHBG concentration in a plasma sample relies on dextran-coated charcoal to adsorb non SHBG-bound labeled steroid in solution (Hammond and Lahteenmaki, 1983). In such binding capacity assay (BCA), the competition for binding sites by different steroids and/ or non-specific tracer binding is assessed by adding excess of unlabelled steroid. A method devised by Scatchard is then utilized to determine the number of steroid binding sites and SHBG concentration in a given solution assuming that one SHBG dimer binds one steroid molecule.

The liquid-phase immunoradiometric assay (SHBG IRMA) employs radio-iodinated mono- clonal antibody to label SHBG that is also recognized by non-radioactive polyclonal antibody in solution. Antibody-bound SHBG is subsequently collected by means of an immobilized secondary antibody, and the SHBG concentration is calculated based on the recovered radioac- tivity. This assay is two orders of magnitude more sensitive and correlates remarkably well with BCA (Hammond et al., 1985).

2.2.5 Structure

2.2.5.1 Subunits and dimerization

The SHBG precursor polypeptide consists of a 29 amino acid hydrophobic leader sequence followed by 373 residues with two disulfide bridges (Hammond et al., 1986, 1987, 1989; Walsh et al., 1986). In plasma, human SHBG exists as a 90 kDa homodimer of two identical protomers. The dimerization is required for high-affinity binding of a steroid that further stabilizes the dimeric form of SHBG (Casali et al., 1990; Rosner et al., 1974). In addition to specific steroids, divalent cations, although probably acting independently of the steroid ligand, enhance efficient dimerization of SHBG (Rosner et al., 1974). The dimerization domain of SHBG has been mapped by site-directed mutations and spans the amino acids 138-148 (Bocchinfuso and Hammond, 1994).

2.2.5.2 Steroid binding

In spite of significant variations in steroid specificity between species, plasma SHBG generally binds DHT with highest affinity, followed by testosterone and estradiol (Hammond and Bocchinfuso, 1995; Cunningham et al., 1981). In essence, a 17β-OH group, and an electrone- gative group at C-3 within a planar C-19 steroid are required for optimal binding (Hammond and Bocchinfuso, 1995; Cunningham et al., 1981). The affinity of SHBG for testosterone and DHT is similar to that of the intracellular androgen receptor (AR), but the complex off-rate is much faster and SHBG does not bind several antiandrogens and antiestrogens, such as cyproterone or cyproterone-17-acetate and tamoxifen, respectively (Danzo and Eller, 1975). Rat testicular ABP differs from human plasma SHBG with respect to the DHT complex half-life; DHT dissociates approximately ten times faster from rat ABP than from human SHBG at 0oC, and this difference suggests a functional divergence between these proteins in different species (Schmidt et al., 1981).

Regions responsible for dimerization and steroid binding in SHBG overlap, and confirmation that the steroid-binding region resides in the amino-terminal half of SHBG came from studies

15 with chimeric proteins. The amino terminus of human SHBG was fused to the carboxyl terminus of rat ABP by means of a hybrid cDNA expressed in cultured cells. The fusion protein recovered from the culture medium had the same steroid binding properties as the wild-type human SHBG (Bocchinfuso et al., 1992a). Furthermore, several amino acid substitutions at the amino- terminal portion of human SHBG also interfere with the steroid binding. Specifically, methionine at 139 and lysine at 143 seem to play a central role within the steroid binding pocket of SHBG (Bocchinfuso and Hammond, 1994; Bocchinfuso et al., 1992a; Sui et al., 1992).

Affinity-labeling studies of SHBG have pinpointed amino acids that might be in direct contact with the steroid ligand. However, two groups, which used [14C]17β-bromoacetoxy-DHT to label the steroid-binding pocket in SHBG, identified different labeled residues. According to one group, lysine at 134 was exposed to the alkylating reagent within the DHT derivative (Namkung et al., 1990), whereas another group recovered histidine at 235 as a single labeled amino acid (Khan and Rosner, 1990). Grenot et al. (1992) used tritiated androgen and estrogen derivatives to photolabel SHBG, and their results pointed out methioneine at 139 as interacting with both estradiol and testosterone.

2.2.5.3 Glycosylation

Each SHBG protomer has two sites for N-glycosylation and one site for O- glycosylation (Que and Petra, 1987; Hammond et al., 1987; Avvakumov et al., 1983). The differential utilization of these sites results in heterogeneity of the two protomers, which is seen as a presence of heavy and light subunits in denaturing polyacrylamide gel electrophoresis (Danzo et al., 1989, 1991; Danzo and Bell, 1988; Hammond et al., 1986). The functional significance of the attached carbohydrates has been studied by site-directed mutagenesis and analysis of the glycosylation- deficient mutants secreted into cell culture medium (Bocchinfuso et al., 1992b). The lack of glycosylation does not influence SHBG’s steroid binding specificity or affinity for steroid ligands, which suggests that the attached carbohydrates are not essential for dimerization of the two subunits. When both N-linked carbohydrates are absent, less SHBG is secreted into culture medium, but it has the same steroid binding properties as the wild-type protein (Bocchinfuso et al., 1992b).

Power et al. (1992) have characterized a variant SHBG form due to polymorphism in the gene encoding SHBG. This polymorphism introduces an additional N-linked glycosylation site at residue 327, and when injected into rabbits, the variant SHBG has a reduced clearance rate as compared to the wild-type molecule (Cousin et al., 1998, 1999).

2.2.5.4 Homology to other proteins

Human SHBG and rat ABP show no structural similarity to other steroid binding proteins, such as the intracellular androgen and estrogen receptors (Bardin et al., 1988). They do, however, resemble protein S, a vitamin K dependent modulator of blood clotting in plasma (Hessing, 1991; Joseph, 1997a). Protein S binds to C4b-binding protein that is involved in the complement system (He et al., 1997). The most carboxyl-terminal part, which in most clotting factors is a serine protease-like domain, is replaced by an SHBG/ABP-like domain in protein S (Gershagen et al., 1987; Baker et al., 1987). This region is also homologous to the five globular domains of laminin A chain (Joseph, 1997a). Laminins are multidomain proteins with multiple functions; they participate in cell adhesion and differentiation, angiogenesis, tumor growth and metastasis, and neurite outgrowth (Kleinman et al., 1993; Sasaki et al., 1988; Engel, 1992). Exten-

16 sive sequence comparison has revealed domains in several other proteins that show significant homology to SHBG and ABP. These include the G1 domain of agrin in neuromuscular junctions (Patthy and Nikolics, 1993) and merosin, a laminin-like protein with more restricted tissue distribution (Ehrig et al., 1990).

2.2.5.5 Crystal structure

The amino-terminal portion of human SHBG has recently been crystallized in the presence of DHT. The crystal structure of residues 1-205, containing one of the two laminin G-like domains of SHBG, demonstrated the presence of two seven-stranded antiparallel β-sheets posi- tioned on top of each other (Grishkovskaya et al., 1999; 2000). The steroid-binding pocket was formed within this β-sandwich, as opposed to the α-helical ligand binding sites characterized in nuclear steroid receptors (Tanenbaum et al., 1998; Pike et al., 1999). More specifically, the contacts between the steroid ligand and SHBG are mediated mainly by hydrogen bonding between serine 42 and the oxygen at the C3 position in the steroid A ring as well as hydrogen bonds that may occur between the hydroxyl-group at C17 on the steroid D ring and asparagine 82 or aspartate 65. Other contact points with the steroid include phenylalanine 67, methionine 107, and methionine 139 (Grishkovskaya et al., 2000). The spherical G-domain of SHBG shows no similarity to albumin or other extracellular steroid-binding proteins, but it is structurally related to a set of functionally diverse proteins with known three-dimendional structure, such as pentraxin, neuramidases, glucanases, and lectins. (Grishkovskaya et al., 2000).

2.2.6 Physiological functions

2.2.6.1 Plasma sex steroid carrier

Since SHBG in human plasma has a high affinity and low capacity for androgens and estradiol, it is thought to regulate/influence the bioavailability and metabolic clearance of these steroids. The metabolic clearance rate (MCR) for most sex steroids is inversely related to their affinity for SHBG, i.e. both free and albumin bound steroids are irreversibly cleared from circulation, while SHBG bound steroids are resistant to metabolic clearance (Siiteri et al., 1982; Petra et al., 1985). For instance, androstenedione, which has a low affinity for SHBG, clears twice as fast as testosterone from the human circulation, while both androstenedione and testosterone have the same MCR within species with no plasma SHBG.

According to the free hormone hypothesis, only the unbound fraction of sex hormones is biologically active, as free steroids are able to enter the target cells by simple diffusion (Mendel, 1989). In support of this hypothesis, the uptake of testosterone and DHT is significantly reduced by rat prostate glands cultured in the presence of human pregnancy serum, which contains appreciable amounts of SHBG (Lasnitzki and Franklin, 1972, 1975). Also, Damassa et al. (1991) have shown that an androgen-stimulated proliferation of a prostate cell line, LNCaP, can be inhibited by adding SHBG to the culture medium. It has been further proposed that by having a significantly higher affinity for testosterone than for estradiol and by being largely occupied by testosterone in female plasma, SHBG could serve as an estrogen amplifier in normal women (Burke and Anderson, 1972).

17 Figure 3. Tissue exposure to free and protein-bound steroid hormones. The free hormone hypothesis (a) states that only the non protein-bound hormone is available to target cells, while according to bound hormone hypothesis (b) specific steroid-binding proteins allow tissue-specificity and amplification of hormonal responses (adapted from Pardridge et al., 1988).

The free hormone hypothesis with respect to extracellular steroid binding proteins has been repeatedly challenged by several authors. For instance, considering the SHBG concentration and the circulating amount of estradiol at any stage during the normal menstrual cycle in females, there would not be enough free steroid to fully occupy cellular estrogen receptors, which is the requirement for maximal estrogen response (Siiteri et al., 1982; Wu et al., 1976; Pardridge, 1988). Furthermore, if the pool of free hormone accounted for all the actions of sex steroids, all tissues would be exposed to the same hormone level (Pardridge, 1988). Pardridge has suggested that also the SHBG-bound steroid is available to cells; SHBG could release the steroid in the site of action and thereby amplify the steroid effect where needed (Pardridge, 1988) (Figure 3).

2.2.6.2 Interaction with cell membranes

Since it is highly unlikely that specific and high affinity steroid-binding proteins in plasma serve solely as reservoirs restricting hormone bioavailability and metabolic clearance, research has focused on identifying other possible mechanisms of actions also for SHBG. One proposed function involves specific receptors on cells of sex steroid target tissues. Estradiol-dependent binding of SHBG has been detected on the membranes of human decidual endometrium (Strel’chyonok et al., 1984; Avvakumov et al., 1986) indicating that SHBG selectively guides sex steroids into the site of action. The endometrium-specific receptor has been solubilized and partially characterized, and it seems to display two separate binding sites for SHBG (Fortunati et al., 1992). As according to earlier studies (Strel’chyonok et al., 1984; Avvakumov et al., 1986), the binding of SHBG to the solubilized receptor is specific and saturable, but the receptor fails to recognize SHBG bound to any steroid ligand, including estradiol (Fortunati et al., 1992). Specific receptors for SHBG have also been identified on other sex steroid dependent tissues, including testis and prostate (Porto et al., 1992; Hryb et al., 1985, 1989; Frairia et al., 1992).

In prostate cancer-derived LNCaP cells, the binding of a biologically active ligand to SHBG- receptor complex causes the accumulation of intracellular cAMP (Nakhla et al., 1990). The activation of adenylate cyclase within the cell is evoked by both estradiol and DHT, the strength of response corresponding to the relative affinity of each ligand for SHBG. However, the increase in cAMP levels is achieved only if SHBG is added to culture medium prior to the addition of the steroid, which indicates that it is the unliganded SHBG that recognizes the

18 receptor (Nakhla et al., 1990). A model proposes that only non steroid-bound SHBG recognizes its receptor, and this binding is inhibited by any steroid ligand regardless of its biological properties (as long as it binds SHBG). The binding of a proper ligand to receptor-SHBG complex then induces a conformational change in SHBG, which triggers a second messenger cas- cade eventually leading to elevated intracellular cAMP levels (Hryb et al., 1990; Nakhla et al., 1990).

Nakhla et al. (1994) have shown that estradiol is able to increase intracellular cAMP levels in the presence of SHBG also in human benign prostate hyperplasia (BPH) tissue. There are several observations suggesting that SHBG mediates the signal leading to the activation of a second messenger system. Firstly, antiestrogens do not block the response, which would be the case if it was mediated through intracellular estrogen receptor. Secondly, diethylstilbestrol (DES), a potent estrogen that does not bind to SHBG, fails to mimic the effects of estradiol. Thirdly, DHT, which has a significantly higher affinity for SHBG than estradiol, blocks the response probably by displacing the bound estradiol from SHBG. Thus, BPH tissue requires estradiol for the activation of the secondary messenger system, as opposed to LNCaP cells, which respond to both estrogens and androgens (Nakhla et al., 1990, 1994).

More detailed study of the SHBG receptor-mediated signaling system in prostate tissue suggested that estradiol could, through SHBG, activate AR via a ligand-independent mechanism. DHT stimulates the secretion of prostate specific antigen (PSA) by prostate tissue through classical transcriptional activation mediated by AR (Riegman et al., 1991; Lee et al., 1995; Henttu et al., 1992). Similar stimulation is achieved by estradiol in the presence of SHBG. As with DHT alone, the estradiol-SHBG mediated stimulation is inhibited by antiandrogens, but not by antiestrogens, which suggests that estrogen receptor (ER) does not mediate this response (Nakhla et al., 1997).

SHBG can also mediate the regulation of cell growth by estrogens and androgens. DHT and estradiol stimulate, in the presence of SHBG, the growth of prostate carcinoma cells, and this increased cell proliferation is associated with elevated cAMP levels inside the cell (Nakhla and Rosner, 1996). By contrast, estrogen induced growth of MCF-7 breast cancer cells is inhibited by SHBG (Fortunati et al., 1996). The reduction in growth rate is accompanied by intracellular cAMP accumulation, which is absolutely dependent on both estradiol and SHBG. Therefore, it seems unlikely that SHBG simply sequesters estradiol and restricts its availability to cells, but may actually transmit the growth inhibitory signal to the cell nucleus.

Although SHBG seems to regulate intracellular signaling and cancer cell growth, such results should be interpreted with caution. The LNCaP cells, for instance, are considered to be a poor model for sex steroid action, as these cells contain a mutated androgen receptor. This mutated receptor is also responsive to estrogens and anti-androgens resulting in significantly altered genomic sex steroid effects (Veldscholte, et al. 1990).

SHBG does not exist exclusively in the plasma, but the interaction with a specific receptor leads to internalization of SHBG by some cells. In Macaca nemestrina, immunoreactive SHBG is localized in several cell types other than where the gene is expressed. These include prostate, seminiferous tubules of the testis, and parenchymal cells in the liver (Bordin and Petra, 1980). SHBG also enters cultured breast cancer and prostate cancer cells, MCF-7 and LNCaP, respectively, via a mechanism that resembles receptor-mediated endocytosis (Bordin and Petra, 1980; Porto et al., 1991; Nakhla et al., 1990). Moreover, i.v. injection of human SHBG into adult rats,

19 which do not have circulating SHBG, leads to accumulation of estradiol and SHBG into uterus and the epithelium of oviducts, (Noe et al., 1992).

Although there has been an inconsistency as to whether the steroid ligand is required in receptor binding, it is widely accepted that the recognition of SHBG by cells is specific and saturable, which is characteristic of receptor binding, and that these receptors are present only on the cells of sex steroid responsive tissues (Frairia et al., 1992). The domain, which is responsible for receptor binding in SHBG has been identified and spans the residues 48-57 (Khan et al., 1990). Interestingly, this is the same region in SHBG that exhibits greatest sequence similarity to the protein S carboxyl-terminus (Gershagen et al., 1987; Baker et al., 1987) that binds the C4b-binding protein of the complement system (He et al., 1997).

2.2.6.3 Function in the male reproductive tract

Testicular ABP has been best characterized in rats because these rodents do not have plasma SHBG (Corvol and Bardin, 1973), which could easily contaminate tissue preparations. ABP is used as a marker of Sertoli cell function as its concentration varies upon hormonal stimulation of the seminiferous tubules (Steinberger et al., 1975; Hansson et al., 1973; Kotite et al., 1978; Sanborn et al., 1975; Ritzen et al., 1975; Hagenas et al., 1975) and at different spermatogenic stages (Kierszenbaum et al., 1980; Anthony et al., 1984; Ritzen et al., 1982). The majority of the ABP produced by Sertoli cells is secreted to the lumen of seminiferous tubules and is transported to the epididymis where it is taken up by the epithelial cells (Gunsalus et al., 1981). In the epididymis, testosterone-bound ABP probably serves to provide a proper androgenic environment for maturing spermatozoa (Turner et al., 1984). One fifth of the synthesized ABP is secreted into the blood, and during puberty, the levels of circulating ABP can be relatively high in male rats (Gunsalus et al., 1981), but the significance of this circulating protein is not known.

2.2.7 Gene structure and tissue-specific expression

The gene encoding human SHBG (shbg) is located in p12-p13 region of chromosome 17. It contains eight exons separated by unusually small introns, with the entire coding region spanning only 3.2 kb of genomic DNA (Hammond et al., 1989).

Although an SHBG cDNA was initially cloned from a human liver library (Que and Petra, 1987), several testis-specific transcripts have since been identified (Hammond et al., 1989; Gershagen et al., 1989). However, these testicular cDNAs are unusual reflecting alternative splicing of the primary transcript in this tissue. One of the cDNAs cloned had a unique 5’- sequence that corresponds to genomic shbg sequence found approximately 1.9 kb upstream of the translation initiation codon for SHBG and encodes an alternative exon one. In the liver, the first exon encodes the hydrophobic signal sequence for secretion. The testicular cDNA also lacks 208 bp corresponding to exon 7, which results in premature termination of translation, i.e. the last 118 amino acids are replaced by 9 different residues (Hammond et al., 1989). In fact, none of the testicular cDNAs identified corresponded the full length SHBG secreted by the liver. Interestingly, Alu sequences typical of human genomic DNA are found immediately preceding the alternatively spliced exon sequences (Hammond et al., 1989), but whether these repetitive sequences contribute to the alternative splicing is not currently known (Figure 4).

20 A B C D A B 1 2 3 4 5 6 7 8

brain testis exons encoding SHBG

Figure 4. Alternative splicing of shbg. The eight exons that code for plasma SHBG and testicular ABP are numbered. Differentially utilized exons are shown as solid boxes, and alternative first exons utilized in human testis and rat brain are alphabetized (adapted from Hammond and Bocchinfuso, 1996).

Human shbg expression has also been demonstrated in several other tissues, including placenta (Larrea et al., 1993), ovary (Misao et al., 1998a), uterine endometrium (Misao et al., 1994) and myometrium, (Misao et al., 1997a) as well as in several breast cancer cell lines (Moore et al., 1996). In addition to the normal SHBG mRNA, most of the tissues analyzed contained the variant SHBG form lacking exon 7 sequence with a truncated 3’-end (Misao et al., 1997a,b,c; 1998a,b; Moore et al., 1996) identical to those characterized in human testis. The analyses of the specific transcripts in normal and malignant tissues have shown that as the state of histo- logical dedifferentiation advances, the relative amount of the variant SHBG mRNA increases (Misao et al., 1998a,b). Also, in the ovary, the normal mRNA seems to be less abundant in women after menopause (Misao et al., 1998a). In light of these studies, it has been proposed that the balance between the normal and variant SHBG production is essential for the healthy function and estrogen responsiveness of female reproductive organs. It is not known, however, whether any of these alternatively spliced transcripts are translated to produce protein, and it is thus difficult to speculate how the corresponding gene products might function.

Unlike the human gene, the rat shbg is not expressed in the liver postnatally (Corvol and Bardin, 1973; Renoir et al., 1980). However, rodents have circulating liver-derived SHBG in the fetus, around the time when Wolffian ducts differentiate to male-specific reproductive organs, epididymis and seminal vesicles (Sullivan et al., 1991; Becker and Iles, 1985). Immunoreactive SHBG is also found in several other tissues of prenatal rat, and it has been proposed that fetal SHBG not only delivers androgens to the site of action, but also protects other tissues from excess androgen exposure (Becchis et al., 1996; Sullivan et al., 1991). The cDNAs for ABP/ SHBG have also been isolated from a rat brain library (Figure 4). However, similar to the human testicular cDNAs, these represent alternatively spliced variants with alternative exon 1 sequences, and it is possible that separate promoters are responsible for shbg expression in different rat tissues (Wang et al., 1990).

The sequences that flank human and rat shbg coding regions at the 5’-end, probably contribute to differential expression patterns between species, but the exact regions responsible are yet to be identified. Consistent with the expression pattern of the endogenous gene, mice transgenic for rat shbg have abundant production of testicular ABP, and transcripts also exist in the brain, but not in the liver (Joseph et al., 1997b; Larriba et al., 1995). Approximately 1.5 kb upstream of the transcription start site are sufficient to maintain this tissue-specific expression pattern of rat shbg in mice.

21 2.2.8 Regulation of plasma SHBG levels

2.2.8.1 Sex hormones

In humans, plasma SHBG levels are very low in both sexes at birth. During the first weeks of life, they start to increase rapidly reaching a constant level around second or third month. At prepuberty, even before the plasma levels of male and female sex steroids start to rise, SHBG concentrations decline markedly in males and to a lesser extent in females. This sexual dimorphism is maintained until the sixth decade in men and postmenopause in women, after which the levels increase slightly in males and further decrease in females (Anderson, 1974; Vermeulen, 1988).

During pregnancy, up to 7-fold increases in plasma SHBG levels occur. Also, oral administration of estrogens in both sexes leads to significantly increased plasma SHBG levels, which return to the basal state after the treatment is stopped (Rosner and Deakins, 1968; Lea and Stoa, 1972; Anderson, 1974; Toscano et al., 1992; Vermeulen, 1988). The administration of human chorionic gonadotrophin (hCG), which stimulates Leydig cell testosterone production, decreases plasma SHBG in prepubertal boys, and pharmacological doses of pure androgen to both adult males and females has similar but more subtle effects (Forest et al., 1988). However, SHBG levels generally correlate poorly with androgens and estrogens in most physiological and pathological conditions (Toscano et al., 1992).

Studies on human hepatoblastoma cells (HepG2), which synthesize and secrete SHBG into the culture medium (Khan et al., 1981), also argue against the direct role of sex steroids in SHBG production by the liver. In these cells, both androgens and estrogens increase (Loukovaara et al., 1995a,b; Plymate et al., 1988) or, depending on the study, have no effect (Rosner et al., 1984) on SHBG production. Furthermore, Loukovaara et al. (1995a,b) have shown that the estrogen-induced increase in SHBG levels is more pronounced inside the cell than in the culture medium. They hypothesized that this increase might be mostly due to the enhanced production of alternative transcripts and gene products that lack the secretion signal.

Although it seems unlikely that sex steroids affect plasma SHBG concentration by a direct transcriptional mechanism, they may modify the carbohydrate composition of the two subunits. Indeed, there is a variation with respect to the number of bi- and tri-antennary type N- linked oligosaccharides attached to SHBG subunits in males, females and pregnant females, i.e. individuals with different estrogenic status (Danzo et al., 1991; Tardivel-Lacombe and Degrelle, 1991). Ain et al. (1987) have also shown that estrogen-induced modulation of the sialic acid content in thyroxine-binding globulin prolongs the half-life of this molecule (Ain et al., 1987). Thus, sex steroids may exert their effects on plasma SHBG by regulating its clearance from plasma (Joseph, 1994).

2.2.8.2 Nutrition and body mass

Fiber-enriched nutrition is known to increase SHBG levels in humans possibly by increasing the amount of phytoestrogens (Adlercreutz et al., 1987). These isoflavonoid compounds have shown to increase SHBG production also in HepG2 cells, but the effect is highly compound dependent (Loukovaara et al., 1995c). By contrast, high fat diet has the opposite effect, and SHBG levels correlate inversely with body mass and plasma triglyceride concentration (Reed et al., 1987; Lapidus et al., 1986; Vermeulen, 1988). In men, there is also a positive correlation

22 between SHBG levels and serum HDL as well as apolipoprotein A-1, which is the predominant apoprotein associated with HDL (Hämäläinen et al., 1987). In a 12-year follow-up study, Lapidus et al. (1986) demonstrated that among women with low plasma SHBG levels there is a higher occurrence of cardiovascular disease and premature death. However, several other reports have significant discrepancy with respect to the role of low SHBG as a risk factor in the development of cardiovascular disease (Pugeat et al., 1995).

2.2.8.3 Insulin

Generally, adiposity also increases the pancreatic insulin production, and upper body obesity increases the amount of free testosterone and decreases the clearance of insulin in the liver. These increases in insulin and testosterone levels correlate negatively with plasma SHBG concentration (Peiris et al., 1989). However, the negative correlation between fasting plasma insulin and SHBG levels in females (Peiris et al., 1989; Preziosi et al., 1993; Lindstedt et al., 1991) is independent of androgens and body weight (Peiris et al., 1989). Unlike the sex steroids, insulin correlates remarkably well with SHBG also in puberty and in several physiological conditions (Toscano et al., 1992). This notion is employed to predict the occurrence of noninsulin dependent diabetes (NIDDM) in females. In males, the situation is more complex and apparently reverse in terms of the correlation of insulin and SHBG levels (Simon et al., 1992; Haffner et al., 1993). In vitro studies on HepG2 cells also support insulin’s direct role in regulating liver SHBG production. In these cells, insulin added to culture medium increases both secreted and intracellular SHBG (Loukovaara et al., 1995b; Plymate et al., 1988).

2.2.8.4 Thyroid hormone

The relationship between SHBG and thyroid hormone levels in plasma is well-established (Anderson, 1974). At a cellular level, thyroid hormone stimulates SHBG production in cultured HepG2 cells in vitro (Rosner et al., 1984; Loukovaara et al., 1995b; Plymate et al., 1988). Patients suffering from thyrotoxicosis have abnormally high thyroid hormone levels and accordingly enhanced hepatic SHBG production (Anderson, 1974). By contrast, in thyroid hormone resistance (THR), in spite of abnormally high thyroid hormone levels, SHBG levels remain normal and fail to rise when additional T3 is administered (de Nayer et al., 1986; Sarne et al., 1988). SHBG is therefore a good marker for peripheral responses for thyroid hormone and a valuable tool for distinguishing THR from thyrotoxicosis to avoid maltreatment (Sarne et al., 1988).

23 3. Aims of the present study

The expression of shbg is regulated differently in humans and rodents. While humans have circulating plasma SHBG in adults, the predominant androgen-binding activity is found in the male reproductive tract in rodents. This may reflect the functional differences between these protein homologs, i.e. plasma SHBG and testicular ABP. In the present study, the following questions were placed to elucidate the physiological function and regulation of plasma SHBG:

1) What are the requirements for the liver- and testis-specific expression of human shbg in vivo? 2) What are the physiological consequences of SHBG production by adult rodent liver? 3) How is human shbg expression regulated in vivo during development and by hormones in different tissues? 4) What are the factors that contribute to the hepatocyte-specific gene expression and thereby govern the differences between the expression patterns of rodent and human genes?

4. Materials and Methods

Materials and methods that are described in detail in the original articles are listed in Table 1. Only the procedures employed in additional unpublished studies are given below.

Antibodies. Antiserum against transcription factor Sp1 was purchased from Santa Cruz.

Plasmid constructions. To introduce human shbg sequence from +7 to +60 into -803/+7shbg and -299/+60shbg fragments, a region spanning nucleotides from -41 to +60 was PCR-amplified using an 11 kbp genomic DNA fragment as the template (III). The forward primer spanned human shbg nts -45/-22 and the reverse primer was complementary to nts +43/+60 with a sequence (GAATGAAGCTT) for a HindIII site at its 5’-end. The PCR-product was digested with SmaI and HindIII and linked to nts -803/-41 of the shbg promoter in pSP72 (III). The -803/+60shbg and -299/+60shbg fragments were then subcloned into the XhoI/HindIII and XbaI/HindIII sites of pGL2 Basic (Promega), respectively. Several 5’-deletion fragments of the shbg promoter were also PCR-amplified from the 11 kbp DNA fragment. The forward primers for these constructs corresponded to the following sequences relative to the shbg transcription start site in the liver: nts -696/-679, -670/-650, -617/-596, -552/-531, -509/-489 and -458/-438.

Table 1. Materials and methods described in the original articles (I-III).

Experimental procedure Described in

Generation and characterization of transgenic mouse lines I RNA analyses: northern blotting and primer extension I, II in situ hybridization and immunohistochemistry I, II Plasma SHBG and testosterone measurements I, II Western blotting I, II Surgical procedures and hormone treatments on mice II in vitro DNase I footprinting and EMSA III shbg-luciferase plasmid constructions III Cell culture and transient transfection assays III

24 In each case, a sequence (TGAACTCGAG) for an XhoI site was added to these primers for subcloning. The reverse primer spanned human shbg nts -317/-290. PCR-products were digested with XhoI and XbaI and subcloned into pBluescript KS- (Stratagene) for sequencing. Human shbg region -803/-299 in -803/+60pSP72 was replaced with each deletion fragment followed by subcloning into pGL2 Basic as XhoI-HindIII fragments. As an exception, the shbg fragments containing nts -509/-299 and -458/ -299 were subcloned directly into -803/+60pSP and sequenced after subcloning the -509/+60 and -458/+60 fragments into pGL2 Basic. Human shbg region -803/-299 was also excised from a 4 kbp genomic XhoI fragment in pBluescript KS- and subcloned as a HindIII/XbaI fragment into pSELECT (Promega) for site-directed mutagenesis. The AT-rich sequence within human shbg -803/-696 was removed using an oligonucleotide complementary to nts -750/-675- excluding nts -727/-697. Sp1 site at position -526/-517 was altered using the anti-sense Sp1 mutant EMSA oligonucleotide (see below).

In vitro DNaseI footprinting of human shbg -803/-299. Nuclear proteins from adult mouse liver were extracted as described (III). Several human shbg region promoter probes were gen- erated by digesting the -299/-803shbg fragment with HaeIII or HinfI. The following internal fragments were chosen as probes: -803/-656, -735/-585, -587/-360 and -541/-300. Fragments were blunted and subcloned into EcoRV site of pBluescript KS- in such orientation that when plasmid was linearized with HindIII, and end-labelled with the Klenow fragment of DNA polymerase I and [α-32P]dCTP, the coding DNA strand contained the radioactive nucleotides. After labelling, fragments were released from the plasmid and purified (III). Approximately 10 000 cpm of the labelled probe was used in an in vitro footprinting assay (III).

Nuclear extract preparation from cultured cells. Human hepatoblastoma HepG2 cells were cultured as described (III). Nuclear extracts were prepared by scraping the cells from the plates in PBS. All steps were performed on ice. After low-speed centrifugation, the cell pellets were resuspended in 10 mM HEPES (pH 7.6), 1.5mM MgCl2, 10mM KCl, 1 mM DTT, and lysed by passing them 10 times through a 22 gauge needle attached to a syringe. Nuclei were spun down and resuspend in the same buffer followed by a centrifugation in an eppendorf centrifuge at full speed. Nuclei were then resuspended in 20 mM HEPES (pH 7.6), 1.5 mM MgCl2, 420 mM NaCl, 0.2 mM EDTA, 0.5 mM DTT, 0.5 mM PMSF, 25% glycerol and stirred at 0oC for 30 min. Lysed nuclei were spun down and the supernatant containing nuclear proteins was dialyzed twice 2 hours in cold against 20 mM HEPES (pH 7.6), 1.5 mM MgCl2, 100 mM KCl, 0.2 mM EDTA, 0.5 mM DTT and frozen in aliquots at -70 oC.

EMSA oligonucleotides. The following double-stranded oligonucleotides were also used as radiolabeled probes and/or unlabelled competitors:

Human shbg -536/-511: 5’-gAAAAGACTGGGGGAGGAGTTCAGAC-3’ 3’-TTTTCTGACCCCCTCCTCAAGTCTGg-5’

Mutant shbg -536/-511: 5’-gAAAAGACTGGGTTAGGAGTTCAGAC-3’ 3’-TTTTCTGACCCAATCCTCAAGTCTGg-5’

Sp1 consensus: 5’-GCTGAGGGCGGGACCGCATC-3’ 3’-CGACTCCCGCCCTGGCGTAG-5’

25 5. Results

5.1 Human shbg transgenic mice

5.1.1 Characterization of transgenic mouse lines (I and unpublished results)

Transgenes comprising the entire coding region of human shbg and either ~850 bp or ~6 kb of 5’-flanking sequences were introduced into the mouse genome (I: Figure 1a). Microinjection of these shbg fragments into oocytes yielded four DNA-positive founder animals when analyzed by PCR and subsequently by Southern blotting.

One of the founders carried a 11-kb shbg transgene incorporated into two distinct sites within the genome as evidenced by three different genotypes transmitted to the progeny. F1 mice with one of the two integration events were maintained as separate lines, and line shbg11-a (I: Figure 1b) was predominantly used for further analysis.

One founder carrying a 4 kb human shbg transgene also transmitted two distinct genotypes to the progeny, but strictly according to Mendelian inheritance typical of a single integration event. Only a subset of the F1 animals carried the same genotype as the founder male, others seemed to have lost copies of the transgene. More extensive Southern blot analysis revealed a rearrangement of the most 3’ transgene copy in these animals resulting in the presence of only one intact copy (not shown). The two distinct genotypes seemed to remain stable in F1 animals, which were then used to establish two separate lines for further analyses (shbg4-b and shbg4- c; I: Figure 1b).

The remaining two founders carrying 11 kb and 4 kb shbg transgenes, shbg11-b and shbg4-a, respectively, had multiple transgene copies in one integration site. A Southern blot of all five lines that were used for further analyses is shown in I: Figure 1b.

5.1.2 Tissue-specific expression of human shbg in mice (I)

Northern blot analysis of RNA extracted from five human shbg transgenic male mice, each representing a separate line, demonstrated that both transgenes were active in the liver. In addition, kidney displayed a strong expression of human shbg with the exception of line shbg4- b, in which there was apparent loss of transgene copies and rearrangement in one of the two remaining copies (I: Figure 2).

The 5’-sequences included only in the 11 kb human shbg transgenes were required for testis- specific expression, as evidenced by almost undetectable levels of SHBG mRNA in the testes of mice carrying the 4 kb transgenes (I: Figure 2).

5.1.3 Human shbg transcription start sites in different mouse tissues (I)

A radiolabelled oligonucleotide specific for exon 1 was used in primer extension analyses to map the transcription start sites for human shbg in transgenic mice. The extension products from liver and kidney RNA placed the major transcription start site 60 bp upstream of translation initiation codon in exon 1, while the same primer failed to anneal to shbg transcripts from transgenic mouse testis (I: Figure 4). These results indicated that testicular shbg transcripts

26 contained an alternative exon 1 sequence, and that an alternative promoter was possibly responsible for transcription in this tissue.

5.1.4 Alternative human shbg transcripts in mouse testis (I)

The lack of liver-specific exon 1 sequences in testicular human shbg transcripts was confirmed by a northern blot analysis. As expected, these testicular transcripts failed to hybridize with a probe consisting exclusively of exon 1 sequences, while mRNA isolated from transgenic mouse liver and kidney readily hybridized with the same probe (I: Figure 3). Furthermore, while an oligonucleotide complementary to exon 1 failed to anneal to the the testicular transcripts, exon 2 specific primer produced three products of different sizes in primer extension analysis (I: Figure 5). This confirmed that a separate promoter located further upstream drives the transcription of human shbg in transgenic mouse testis, and that the liver-specific exon 1 sequence is replaced by an alternative exon.

5.1.5 Cellular localization of human SHBG mRNA and immunoreactive protein (I)

The cellular localization of human SHBG in various transgenic mouse tissues was studied by immunohistochemistry. In the liver, both human SHBG mRNA and protein localized in hepatocytes, the signal being more pronounced in periportal lobes than in the close proximity to the central vein (I: Figure 6a, b).

Immunoreactive SHBG was also found in the luminal surface of the epithelial cells that line the proximal convoluted tubules in the kidney (I: Figure 6c) as well as in the interstitial compartment in the testes of both shbg11 and shbg4 animals (I: Figure 6e,g). While the SHBG mRNA co-localized with the protein in the kidney, it was present only in the Sertoli cells of transgenic mice carrying the 11 kb human shbg transgene, consistent with the northern blot analyses (I: Figure 6f,h). Moreover, human shbg expression in transgenic mouse testis appeared to vary with respect to stage of spermatogenesis in each seminiferous tubule, being highest during stages X-XII (I: Figure 7).

5.1.6 Human SHBG in mouse plasma and urine (I)

Production of SHBG by liver resulted in considerable amounts of steroid binding activity in transgenic mouse plasma. This lead to significantly increased amounts of testosterone in both transgenic male and female mouse blood when compared to wild type littermates (I: Table 1).

Immunoreactive SHBG could also be detected in transgenic mouse urine with the exception of shbg4-b mice that do not express the transgene in the kidney. Western blot and steroid binding capacity assays demonstrated that SHBG in urine was structurally similar to the plasma protein and possessed an intact steroid binding site (I: Figure 8).

5.1.7 Phenotypic analysis of human shbg transgenic mice (unpublished)

In spite of the presence of relatively very high concentrations of a binding protein for sex steroids in their blood, shbg transgenic mice showed no obvious abnormalities with respect to their reproductive performance. Also, male sperm counts and morphology appeared to be normal, and no gross abnormalities were observed in any other tissues examined by routine histol- ogy. All five lines described have been bred to homozygosity with normal litter sizes and sex

27 distribution within one litter.

5.2 Regulation of human shbg expression in transgenic mice during development (II)

5.2.1 Human shbg expression in mouse fetus

In situ hybridization analyzes of 17.5 days post coitus embryonic sections demonstrated shbg transcripts in shbg11 mouse liver and gut (II: Figure 1). Northern blotting was also used to assess the liver-specific fetal shbg expression in wild-type and transgenic mice. Endogenous mouse shbg expression was evident at the end of gestation with SHBG mRNA levels decreas- ing towards perinatal period. By contrast, human shbg transgene expression increased in the lines between embryonic day 17 and birth (II: Figure 2a). Immunoreactive mouse and human SHBG could also be detected in the serum of both wild type and human shbg transgenic fetuses (II: Figure 2b).

5.2.2 Serum SHBG levels during development

Liver-specific gene expression was mainly monitored on the basis of SHBG measured from blood samples. Serum SHBG levels appeared to rise in both males and females during the first 20 days of life, after which the levels declined in females, as opposed to a further rise in males (II: Figure 3a). This sexual dimorphism was accompanied by significant variability in blood SHBG levels between individual animals. It appeared that by day 40, the sexual dimorphism was achieved in full, and the variations among individuals in both sexes leveled off (II: Figure 3a).

Transgenic mice carrying the 4 kb human shbg transgene showed no such sexual dimorphism. In shbg4-a mice, plasma SHBG levels rose steadily until day 60, after which they decreased slightly (II: Figure 3b).

5.2.3 Correlation between serum SHBG and testosterone levels

In general, all transgenic mice had 10-100 times higher serum testosterone concentrations than their wild-type littermates. In both shbg11-a and shbg4-a males, serum testosterone levels started to decrease after birth and remained low until 30 days of age. In females, serum testosterone levels also rose only after weaning age, i.e. in animals approaching puberty (II: Figure 4a-d). After puberty, significant variations in serum testosterone levels were obvious between individual shbg mice (35nM-1.75µM in males), being consistently higher than in their wild type littermates and correlating remarkably well with serum SHBG levels in both males and females.

5.2.4 Testis -specific human shbg expression during development

In human shbg11-a mice, transgene expression in the testis seemed to begin already before 10 days of age, and increased significantly as the animals reached sexual maturity by 30 days of age (II: Figure 5). Therefore, it seems that human shbg is active in the mouse testis even before blood testosterone levels start to rise.

28 5.2.5 Effect of androgens on human shbg expression in mouse kidney

Kidney RNA was isolated from male and female mice of various ages in order to study kidney- specific human shbg expression in transgenic mice. Northern blot analyses showed increase in SHBG mRNA by day 20 in both sexes. Around day 40 in shbg4-a mice and day 20 in shbg11- a mice, sexual dimorphic expression of shbg was observed (II: Figure 6). The higher amounts of SHBG mRNA in male kidneys suggested that human shbg transgene expression was regulated by male sex steroids.

To further elucidate the possible androgen dependence of human shbg expression in mouse kidney, shbg11-a and shbg4-a male transgenic mice were castrated, and the SHBG mRNA levels were compared to those of intact littermates. As expected, castration reduced the amount of human SHBG mRNA in mouse kidney (II: Figure 7a), while no changes were observed in plasma SHBG (not shown) and accordingly in liver-specific gene expression (II: Figure 7a).

To confirm that the SHBG induction in transgenic mouse kidney was a pure androgen response, castrated males and ovariectomized females were treated with DHT. The hormone was administered by subcutaneous injection of 100 µg daily dose of DHT in sesame oil for up to 5 days. Following steroid treatments, SHBG mRNA levels were increased in both sexes (II: Figure 7 b-c).

5.3 In vitro characterization of human shbg liver-specific promoter (III)

5.3.1 Rodent and human shbg 5’-flanking regions

Despite obvious similarities between human and rat shbg 5’-flanking regions immediately upstream of liver and testis-specific exon 1, respectively, sequences with significant differences also exist. Generally these differences represent DNA stretches in the human gene that are not found in the rat gene (III: Figure 1).

5.3.2 Possible targets for transcription factor binding within the human shbg proximal promoter

To map important regions for liver-specific expression, the human shbg proximal promoter spanning nucleotides from –299 to +7, relative to the transcription start site, was studied in DNaseI footprinting assays. Several sites were specifically protected by nuclear proteins from DNaseI-digestion and were likely to represent DNA regions that interact with transcription factors involved in transcriptional activation or repression (III: Figure 2). The most pronounced protections were observed in regions spanning nucleotides -41/-18 (FP1), -61/-52 (FP2), -88/ -66 (FP3), -114/-96 (FP4), -152/-134 (FP5), and -200/-177 (FP6).

5.3.3 Minimal promoter required for liver-specific transcription

To assess which of the DNA elements contribute to the basal expression of shbg in hepatocytes, series of deletion mutants were analyzed in human HepG2 hepatoblastoma cells. The activity of a given deletion mutant was determined as its potential to drive transcription from a luciferase reporter gene linked downstream of each mutant promoter. Transient transfection assays in HepG2 cells revealed that the basal activity of human shbg promoter was achieved

29 when four most proximal footprints were present, and that sequences upstream of -299 re- pressed transcription (III: Figure 3).

5.3.4 Possible TATA-box within human shbg promoter

Both human and rat shbg genes have been characterized as TATA-less genes, since neither contains a canonical TATA-box within a region immediately upstream of the tissue-specific transcription start sites (Joseph et al., 1988). However, the sequence TTTAAC at –30 within human shbg footprint 1 could serve as an atypical target for the TATA-binding protein, TBP. To assess this possibility, the TTTAAC sequence was changed to TCCCAC, because this was expected to abolish any binding of TBP to this element and to reduce the promoter activity. The activity of the mutated –137/+7shbg fragment was indeed significantly lower than that of the wild type promoter (III: Figure 4a), which suggested that the TTTAAC sequence is crucial for efficient transcription initiation.

On the other hand, if TBP binds to human shbg footprint 1 (FP1), by converting the TTTAAC to an optimal TATA-sequence, TATAAA, one should be able to increase the promoter activity (Wiley et al., 1992). Interestingly, this mutation had a favorable effect on the promoter activity in HeLa, but not in HepG2 cells (III: Figure 4a), and it was concluded that a factor other than TBP recognizes the shbg promoter element, which is necessary for basal transcription in hepatoma cells.

5.3.5 Orphan receptors COUP-TF and HNF-4 bind to human shbg promoter

In addition to a TATA-like sequence, FP1 contains stretches that resemble binding sites for several liver-specific transcription factors. These include hepatocyte nuclear factors 1 and 4, HNF-1 and HNF-4, respectively. Since an unlabelled competitor oligonucleotide corresponding consensus HNF-1 element failed to displace the bound complex from labeled FP1 oligo in an EMSA (not shown), the possible involvement of HNF-1 in shbg regulation was not studied further.

While FP1 (5’-GGGCAACCTTTAACCCT-3’) showed a rather weak consensus sequence for orphan receptor HNF-4, a site further upstream within FP3 (GGGTCAAGGGTCA), displayed a significant homology to the chicken ovalbumin upstream promoter transcription factor, COUP- TF, target sequence (Leng et al., 1996).

DNA-protein interactions at FP1 and FP3 were studied in EMSAs, which showed that the same protein complexes recognize both FP1 and FP3 (III: Figure 5). Furthermore, antibody supershift analyses demonstrated that these complexes consist of HNF-4 and COUP-TF in different ra- tios. It appeared that FP1 is predominantly occupied by COUP-TF in vitro, while FP3 is more or less equally bound by both factors (III: Figure 6).

5.3.6 Site-specific mutations at FP1 and FP3 reduce human shbg promoter activity

Since elements within the human shbg proximal promoter bind a liver-specific factor, HNF-4, it was anticipated that mutations at these sites would severely compromise the promoter activity in hepatoma cells, as was already shown for FP1. The effect of site-specific mutations within FP1 and FP3 (III: Figure 7a) was analyzed in DNaseI protection and transient transfection assays. When the mutations were studied individually and in combination, it was evident

30 that the DNA-protein interactions were severely impaired at FP1 and FP3 as compared to the wild-type shbg fragment (III: Figure 7b). It was also evident that the COUP-TF/HNF-4 site within FP1 had a major role in transcriptional activation when compared to the one within FP3. The disruption of both sites led to reduction of promoter activity by more than 70% (III: Figure 7c).

5.3.7 Transactivation of human shbg promoter by HNF-4 is antagonized by COUP- TF

The potential of COUP-TF and HNF-4 to induce transactivation of human shbg promoter was further studied in cotransfection assays. While significant transactivation of human shbg wild- type promoter accompanied HNF-4 overexpression, the effect of COUP-TF was very moderate (III: Figure 8). In fact, COUP-TF overexpression efficiently blocked the transactivation by HNF-4.

The transactivation was efficiently abolished by mutations within COUP-TF/HNF-4 sites, again FP1 showing a more crucial role in mediating HNF-4 effects. The cotransfection experiments were also performed in HeLa cells, which have endogenous COUP-TF, but not HNF-4. The effect of HNF-4 on the human shbg promoter activity was reproduced also in this cell line, but was somewhat less pronounced than that observed in HepG2 cells (III: Figure 8).

5.3.8 Optimal COUP-TF site within FP1

To further study the role of TBP and the liver-specific factor, HNF-4, in mediating transcriptional activation of shbg, the TTTAAC within FP1 was changed by site directed mutagenesis to an optimal COUP-TF/HNF-4 binding site identified within FP3 (GGGTCAAGGGTCA). The mutation distorted any resemblance of this site to TATA-box and should prevent the binding of TBP while enhancing the binding of HNF-4 to FP1. When studied in the context of a reporter gene in HepG2 cells, the optimal COUP-TF/HNF-4 site within FP1 did not reduce the promoter activity (III: Figure 9). This suggested that the binding of HNF-4 or COUP-TF to the region around -30, where the TATA-box is generally located, is favorable for the activation of human shbg in liver cells.

5.4 Putative silencer in the human shbg promoter (unpublished)

Since a 4 kb human shbg transgene comprising 803 bp upstream of transcription start site was shown to be very active in transgenic mouse liver and kidney, the repression by the region upstream of –299 observed in vitro was somewhat surprising. The effect of this region on the human shbg promoter activity was confirmed in the context of fragments that extended down to translation start site (+60, III). Surprisingly, the 5’-UTR increased the promoter activity when -299 nts upstream of human shbg liver-specific transcription start site were included, whereas in the context of the -803/+60shbg promoter, reporter gene expression was dramatically reduced (Figure 5). As a result, the activity of -803/+60shbg was lower than -803/+7shbg and only approximately 2-fold higher than the promoterless pGL2 Basic.

5.4.1 Characterization of the human shbg region -803/-299.

To map the putative silencer located upstream of –299 within the human shbg promoter, four internal probes covering the entire human shbg region -803/-299 were used in in vitro footprinting

31 Figure 5. Identification of a putative silencer in the human shbg promoter. HepG2 cells were transfected with human shbg promoter-reporter constructs and a pCMVlacz plasmid as an internal control for transfection efficiency. The transcriptional activity (means ± S.D.) of each construct is indicated relative to the activity of -299/+7shbg. assays, and in these assays, numerous possible binding sites for mouse liver nuclear proteins were observed (Fig 6a-d). Transcription factor search database MatInspector version 2.1 (Quandt et al., 1995) identified several putative binding sites for C/EBPβ as well as a putative Sp1 site within footprinted region -536/-511.

5.4.2 Human shbg region –536/-511 binds ubiquitous factor Sp1

Of the footprints shown in figure 5, the one spanning nucleotides -536/-511 contains a sequence that resembles the consensus binding site (-526/-517: 5’-GGGGAGGAGT-3’) for a ubiquitous transcription factor Sp1 (Quandt et al., 1995). EMSA showed the complexes that form between human shbg region -536/-511 and nuclear proteins isolated from HepG2 cells (Figure 7a; lane 2), and these complexes could be competed for by both specific competitor (Figure 7a; lanes 3-4) and a consensus Sp1 oligonucleotide (Figure 7a; lanes 5-6), but not with a COUP-TF oligonucleotide (III) (Figure 7a; lanes 7-8). Furthermore, an antibody against Sp1 supershifted a complex that formed between human shbg region -536/-511 and nuclear proteins (Figure 7b, lane 3).

5.4.3 Mutation within the Sp1 site partially recovers human shbg promoter activity

To study the function of the Sp1 binding site identified in figure 6, mutations were introduced to this region and tested in an EMSA (Figure 7c). The sequence GGGGAGGAGT was altered to GGTTAGGAGT since this abolished the binding of Sp1 (Xie et al., 1997). The mutated oligonucleotide did not compete for the wild-type oligonucleotide (Figure 7c; lanes 5-6) and when labeled, its mobility was not retarded by nuclear proteins (Figure 7c; lane 7). The same mutations were then introduced into the -803/+60shbg fragment and the effect on transcription analyzed in a transient transfection in HepG2 cells. When compared to the wild type -803/ +60shbg, the mutation within the Sp1 site partially recovered the promoter activity (Figure 7d). This suggests that the Sp1 site does not repress the human shbg promoter activity on its own, but is likely to be required for the suppression by region -803/-299.

32 Figure 6. DNaseI footprinting of transcriptional regulatory elements within human shbg 803/-299. An end- labelled probe spanning human shbg -803/-656 (a), -735/-585 (b), -587/-360 (c), or -541/-300 (d) was incubated with increasing amounts of mouse liver nuclear extract (lanes 3 and 4) followed by digestion with DNaseI. Lane 2 is without the addition of nuclear proteins. Maxam-Gilbert A/G-sequencing (lane1) is shown as a marker to define the boundaries of each footprinted region (shown in brackets).

33 Figure 7. Sp1-binding site within human shbg promoter represses transcription. a, double-stranded, end-labelled oligonucleotide spanning human shbg -536/-511 was incubated with 2 µg HepG2 nuclear proteins in the absence (lane 2) or presence of specific (lane 3-6) or non-specific (lanes 7-8) competitors. Free probe was separated from the protein-DNA complexes by non-denaturing polyacrylamide gel electrophoresis. b, antibody supershift was performed by adding antiserum against Sp1 (lane 3) or normal rabbit serum (lane 2). c, double-stranded labelled human shbg -536/-511 oligonucleotide (lanes 1-6) was incubated with 4 µg HepG2 nuclear proteins in the absence (lane 2) or presence of excess wild-type (lanes 3-4) or mutant (lanes 5-6) competitor oligonucleotide. Lanes 1 in a, b, and c, and lane 7 in c are without the addition of nuclear proteins. d, HepG2 cells were transfected with wild-type and mutant -803/+60shbg-luciferase constructs together with a pCMVlacz plasmid as an internal control for transfection efficiency. The activity of the mutant fragment (mean ± S.D.) is indicated relative to the activity of the wild-type promoter.

34 5.4.4 Analysis of human shbg -803/-299 deletion fragments

To identify the human shbg region responsible for the strong repression of transcription within -803/+60shbg, several promoter deletions with 5’-ends based on the possible protein binding sites were constructed. Promoter deletion fragments in front of a luciferase reporter gene were analyzed in HepG2 cells, and the first suppressor appeared to be located between nts -299 and -458, while further deletions seemed to have only moderate effects if any on the promoter activity (Figure 8). The most predominant repression occurred when sequences -803/-670 were added. This region contains a part of an Alu sequence (Hammond et al., 1989) including a TAAAA-sequence which is repeated five times.

5.4.5 Removal of AT-rich sequence at -727/-697 partially recovers human shbg promoter activity

To study the effect of the six TAAAA repeats within -727/-697 on human shbg promoter activity, the -803/+60shbg promoter was modified by site-directed mutagenesis system using an oligonucleotide lacking the AT-rich stretch. Deleted sequence was analyzed by MatInspector (Quandt et al., 1995), to confirm that the change did not introduce potential binding sites for transcriptional regulators that might affect the promoter activity. Transient transfection in HepG2 cells (Figure 9) showed that the activity of -803/+60shbg was partially recovered and corresponded to the activity of -670/+60shbg (Fig 8). Thus, the AT-rich sequence within -803/-696 is at least partially responsible for the reduced activity of the -803/+60shbg fragment.

Figure 8. Detailed mapping of the silencer within human shbg promoter. The human shbg promoter deletion mutants were fused to a luciferase reporter gene. The positions of the footprints identified in figure 6 are indicated as shaded and the TA-rich strech as black boxes, respectively. HepG2 cells were transfected with each promoter-reporter plasmid together with a pCMVlacz plasmid as an internal control for transfection efficiency. The transcriptional activity (mean ± S.D.) of each promoter deletion mutant is indicated relative to the luciferase activity of -803/ +60shbg.

35 Figure 9. The effect of the TA-rich strech at 727/-697 on the transcriptional activity of human shbg promoter. HepG2 cells were transfected with wild-type and mutant human shbg -803/+60luciferase constructs together with a pCMVlacz plasmid as an internal control for transfection efficiency. The activity of the mutant fragment (mean ± S.D.) is indicated relative to the activity of the wild-type promoter.

6. Discussion

6.1 Tissue-specific expression of human shbg

6.1.1 Regulation of serum SHBG levels

In humans, SHBG is produced by the liver and secreted into the blood, where it binds androgens and estrogens with high affinity. Unlike the human gene, shbg in rats and mice is silent in the liver postnatally (Joseph, 1994). However, the incorporation of the human gene into the mouse genome results in SHBG production by adult mouse liver and kidney. Furthermore, the present study shows that all the elements required for the liver-specific gene expression are included in the 4 kb human shbg transgene that contains the entire coding region together with less than 1 kb upstream of the translation start site.

In both humans and human shbg transgenic mice, blood SHBG levels are low at birth, but increase steadily with age until puberty, when a significant discrepancy occurs. While human serum SHBG levels are generally higher in women than in men (Anderson, 1974; Vermeulen, 1988), the transgenic mice carrying a 11 kb human shbg fragment show a completely opposite sexual dimorphism. This inconsistency could be explained by the absence of other regulatory sequences not included in either shbg transgene, but might also be due to physiological differences between rodents and humans.

The fact that serum SHBG levels are not sexually dimorphic in mice carrying the 4 kb human shbg transgene indicates that sequences upstream of the 4 kb fragment respond to repressor signals in the female mouse liver. Whether this repression involves ovarian hormones waits for further studies, but it appears that higher blood SHBG levels in transgenic males are not due to direct effects of androgens, because castration has no effect on liver mRNA or serum SHBG levels. However, any effects that sex hormones might have on SHBG production in these mice can certainly be influenced by the unusual presence of very high amounts of SHBG in their plasma.

36 6.1.2 Transgenic mouse testis produces mainly alternative human shbg transcripts

Similar to the mouse endogenous shbg and rat shbg transgene (Joseph et al., 1997b; Larriba et al., 1995), the 11 kb human shbg transgene is also active in the transgenic mouse testis. How- ever, the transcription of human shbg in the testis of transgenic mice is under different regulation than that in the liver, as evidenced by almost nonexistent testis-specific shbg expression in mice carrying the 4 kb human shbg transgene. It appears that an entirely separate upstream promoter controls shbg transcription in the testis and that the resulting transcripts lack exon 1 sequences including the translation start site utilized in the liver and kidney. The transcripts derived from this tissue are therefore not likely to produce SHBG or ABP, as the first exon, which is replaced by the alternative exon sequences in the testis, corresponds to the signal sequence for secretion.

Variant shbg transcripts have been identified previously among the cDNAs isolated from a human testis library (Hammond et al., 1989). Since the majority of these cDNAs represent alternatively spliced gene products, it is not surprising that the expression of human shbg in the transgenic mouse testis predominantly gives rise to alternative spliced variants.

The analysis of testicular human shbg gene products in transgenic mice is hampered by the presence of immunoreactive SHBG derived from plasma, which probably accounts for the signal seen in the interstitial compartment of both shbg11 and shbg4 animals. It is also possible that available SHBG antibodies fail to recognize protein derived from these alternative transcripts. Therefore, it is not yet known whether alternative transcripts produce functional protein, but it is tempting to speculate that such gene products serve a specific function in human testis and several other tissues including female reproductive organs (Misao et al., 1997a,b,c, 1998a,b; Moore et al., 1996).

The lack of exon 1 sequences encoding the signal peptide probably results in the production of an intracellular protein. Since these alternative transcripts often also lack exon 7, the steroid binding property of the possible protein may deviate from that of plasma SHBG. An shbg gene product with an alternate amino terminus has been characterized also in rats. Joseph et al. demonstrated that such alternative mRNA encodes for an intracellular protein with a leader sequence capable of targeting the gene product into the cell nucleus (Joseph et al., 1996). It was hypothesized that this type of variant ABP/SHBG might locally regulate the sex steroid action mediated by nuclear receptors.

Similar to the gene encoding rat ABP (Kierszenbaum et al., 1980; Anthony et al., 1984; Ritzen et al., 1982), human shbg expression changes upon different stages of seminiferous epithelial cycle. The rat gene can also be stimulated with androgens and FSH in vivo (Danzo, 1990, 1995), but whether human shbg is under similar control in the testis awaits further studies. In any case, the alternative human shbg promoter seems to be activated even before the animals reach sexual maturity and prior to increases in plasma testosterone, and is therefore most likely regulated by mechanisms that do not necessarily involve sex steroids.

6.1.3 Human SHBG production in the transgenic mouse kidneys is regulated by androgens

Human shbg expression in mouse kidney was somewhat surprising, because this tissue has not

37 been reported to express shbg in humans. However, mouse endogenous shbg seems to be active also in this tissue at low level (Jänne, Hogeveen and Hammond, unpublished data). Being a very androgen-responsive tissue, kidney may represent an important, but so far ig- nored site for shbg expression in humans.

The obvious rearrangement and lack of multiple copies of the transgene in shbg4-c mice compared to other lines carrying the same human shbg fragment resulted in the lack of transgene expression in kidney. Preliminary Southern blot analyses suggest that the possible rearrangement occurred at the 3’-end of the transgene. Since the livers of these mice efficiently produce SHBG, this region could be of interest as an enhancer required for kidney-specific human shbg expression.

Similar to several other genes (Catterall et al., 1986), human shbg expression displays a significant sexual dimorphism in the kidney, and unlike in the liver, this dimorphism was directly linked to male sex steroids by castration and DHT administration experiments. Should the human kidney also produce SHBG, it is possible that the androgen environment at a given time determines the level of shbg expression to ensure locally the proper balance between free and bound hormone.

6.2 Physiological consequences of human shbg expression in transgenic mice

In spite of appreciable amounts of SHBG in transgenic mouse plasma, no obvious abnormalities were observed in their reproductive performance. This was somewhat unexpected as it was anticipated that the presence of SHBG would lead to imbalance in free and unbound steroids in these rodents that normally do not have such large amounts of circulating sex steroid binding protein in their plasma. Since the amount of free hormone was not measured, it is not known whether SHBG occupies a significant fraction of sex steroids in the transgenic mouse serum. However, considering the level of circulating SHBG, especially females might be expected to have only trace amounts of free, biologically active estradiol in the blood. It seems therefore doubtful that the sole function of SHBG is to passively restrict the access of sex steroids to target tissues (Mendel, 1989). It could, for instance, release the bound steroids where needed or act through membrane contacts that do not lead to specific signal propagation in mice. Unpublished observations (Hodgert-Jury and Hammond) indicate that human SHBG accumu- lates in certain steroid responsive mouse tissues such as uterine stroma, and that this accumulation is steroid-dependent.

As more data is available about the possible actions of SHBG, it becomes evident that this system regulating the actions of sex steroids is very diverse and yet more remains to be discovered. Without doubt, cloning of the SHBG receptor is essential to fully confirm that the intracellular signaling cascades triggered by a steroid are mediated by an SHBG-receptor complex (Fortunati et al., 1992, 1993; Nakhla et al., 1994, 1997; Frairia et al., 1992). These putative actions have questioned the free hormone hypothesis, which may explain only a subset of the actual functions of extracellular steroid binding proteins. Also, considering the similarity of the receptor-binding region of SHBG with domains of several other proteins, novel and unexpected functions for SHBG may be discovered in the future.

Although the shbg transgenic mice present an apparently normal phenotype, circulating SHBG may cause subtle changes in physiological responses. Significantly elevated testosterone levels

38 in transgenic animals may represent such situation, as the SHBG-bound testosterone in shbg mice is probably resistant to metabolic clearance. It is also possible that the sex steroid-regulated pituitary gonadotrophin secretion is modified in these mice due to an altered feedback regulation by gonadal steroids.

Since kidney is an androgen-responsive tissue, renal SHBG production may influence the local regulation of androgen action. However, studies on ornithine decarboxylase (ODC) and kidney androgen-responsive protein (KAP) genes indicated that SHBG does not restrict the androgen access to kidney, as these androgen-responsive genes are expressed at equal level in both human shbg transgenic and wild-type mouse kidneys (Bardin and Catterall, 1981; Crozat et al., 1992) (Jänne and Hammond, unpublished results).

Any effects that possible testicular human shbg products may have in transgenic mice are very subtle and do not compromise their reproductive performance. Mice transgenic for rat shbg produce large amounts of ABP that is secreted from Sertoli cells and transported to epididymis (Joseph et al., 1997b; Larriba et al., 1995). Infertility of these mice is probably due to increased amounts of functional ABP in male reproductive organs, and may account for abnormalities in the seminiferous tubule morphology and germ cell content (Joseph et al., 1997b; Larriba et al., 1995). Since the majority of human shbg testicular transcripts in transgenic mice lack the secretion signal associated with the SHBG precursor polypeptide, the functions of any resulting gene products are likely to be intracellular.

6.3 Tissue-specificity of human and rodent shbg promoters.

Since the human shbg transgenes are active in the mouse liver, the lack of mouse endogenous shbg expression by this organ is not due to the lack of necessary factors to support transcription. It is also evident that promoter elements that respond to liver transcriptional activators are present within a region immediately upstream of translation start site in exon 1 included in the 4 kb human shbg transgene. However, although the rodent and human genes are very homologous in their sequence immediately upstream of the translation start site in exon 1, a rat shbg transgene containing 1.5 kb of 5’-flanking sequences is active in the testis but not in adult mouse liver (Reventos et al., 1993).

In light of currently available data, it seems that homologous promoter regions in rodent and human shbg drive the gene expression in different tissues. Interestingly, although adult mice and rats do not have circulating SHBG, the gene is active in the fetal liver (Sullivan et al., 1991; Becker and Iles, 1985). Since human shbg seems to be expressed in both fetal and adult mouse liver, it is tempting to explain this paradox with a silencer mechanism such as that demonstrated for the alpha-fetoprotein (AFP) gene. Comparable to rodent shbg, the AFP gene is active in the fetal liver and its expression declines towards birth (Camper and Tilghman, 1989). The promoter element responsible for this developmental silencing has been mapped upstream of the transcription initiation site (Vacher and Tilghman, 1990). In support of this type of re- pressive mechanism in terms of adult liver-specific shbg transcription in rodents, the mouse proximal promoter is able to drive transcription from a luciferase reporter gene in cultured hepatoma cells (Jänne and Hammond, unpublished results). This suggests that the region immediately upstream of exon 1 serves as a liver-specific promoter in both humans and rodents. In fact, the high-affinity HNF-4 binding site within human shbg FP1 is perfectly conserved in rat and mouse genes. Whether the rat proximal shbg promoter is also able to support transcription in liver cells is not known, but it is possible that a specific silencer represses the liver-

39 specific promoter and prevents the production of circulating SHBG in adult rodents.

Unlike the human transgene comprising less than 1 kb of 5’-flanking sequences, a rat shbg transgene is expressed at a high level in mouse testis (Joseph et al., 1997b; Larriba et al., 1995). These transgenic mice also secrete ABP into epididymis, which indicates that 5’-flanking sequence and transcription unit corresponding to the human shbg liver-specific promoter is actually responsible for rat shbg expression in transgenic mouse testis. Characterization of the rat shbg proximal promoter in vitro has demonstrated that it is active in cultured Sertoli cells, but the evidence of testis-specific enhancer elements is still lacking (Fenstermacher and Joseph, 1997).

Whether an alternative promoter solely drives shbg expression in human testis is still unclear, but it is possible that ABP found in rodent testis and epididymis does not play a significant role in human reproductive biology. The fact that rabbits have both plasma SHBG and testicular ABP suggests, however, that one of the protein homologs does not exclude the presence of the other (Cheng et al., 1984).

6.4 Human shbg expression is regulated by HNF-4 in liver cells

In humans, plasma SHBG levels are regulated by different physiological and pathological conditions as well as by nutritional status and body mass (Pugeat et al., 1995; Reed et al., 1987; Lapidus et al., 1986; Vermeulen, 1988). Hepatic SHBG production has also been studied in cultured cells (Rosner et al., 1984; Loukovaara et al., 1995a,b,c; Plymate et al., 1988), but little is known about the transcriptional regulation of the gene.

Human and rat shbg genes have been referred to as TATA-less genes, because neither contains a consensus TATA-sequence in close proximity to the tissue-specific transcription start sites (Joseph et al., 1988). Although the human shbg promoter contains a TATA-like sequence at -30 relative to the transcription start site in the liver, the present data suggest that TBP does not mediate transcriptional activation through this element. Instead, the promoter region spanning the TA-rich proximal sequence binds two orphan members of the nuclear receptor superfamily present in mouse liver nuclear extract. Of the two orphan receptors, only HNF-4 appears to transactivate the human shbg promoter, and this transactivation can be blunted by overexpression of COUP-TF.

Both COUP-TF and HNF-4 bind DNA as dimers, and while COUP-TF has several dimerization partners (Berrodin et al., 1992), HNF-4 forms exclusively homodimers (Jiang et al., 1995). Therefore, antagonism between these factors at human shbg FP1 could be mediated through a common binding site in the DNA motif. However, since both COUP-TF and HNF-4 also interact with general transcription factors (Malik and Karathanasis, 1996; Hadzopoulou-Cladaras et al., 1997), these interactions could determine the relative effectiveness of each orphan receptor to activate human shbg promoter. For instance, ATF transactivation domain of HNF-4 has been shown to recruit TFIIB to the ApoA1 promoter, thereby influencing the entry of the preinitiation complex followed by transcriptional activation (Malik and Karathanasis, 1996). The FP1 sequence within the human shbg promoter is well conserved in rat shbg promoter, and could therefore represent an important element for liver-specific transcription in both species (Figure 10).

Although COUP-TF and HNF-4 also bind to FP3 within the human shbg proximal promoter,

40 Figure 10. Proposed mechanism for transcriptional intiation from the human shbg promoter. HNF-4 controls transcription initiation by substituting TATA-binding protein (TBP) and recruiting the transcription initiation complex (TAFs, TFIIB (IIB), TFIIF (IIF), TFIIE (IIE), and TFIIH (IIH)).

FP1 seems to be more important in mediating basal transcriptional activity in hepatoma cells in vitro. FP3, which shows a relatively weak responsiveness to HNF-4 overexpression in cells, could play a yet unidentified role in human shbg expression during specific situations. The availability of a natural ligand, through influencing the conformational state of COUP-TF and HNF-4, may further modulate the promoter activity in different tissues. Long chain fatty acyl- CoA thioesters have been shown to enhance the dimerization and transactivation capacity of HNF-4 (Hertz et al., 1998), but a ligand activating COUP-TF is yet to be found. The involvement of fatty acids in the regulation of shbg could explain the variations of plasma SHBG levels in certain human physiological and pathological conditions, such as obesity (Reed et al., 1987; Lapidus et al., 1986; Vermeulen, 1988).

HNF-4 regulates several other liver-specific genes, such as the one encoding factor X. A single nucleotide change within the factor X promoter disrupts the binding of HNF-4 and leads to decreased plasma factor X levels and hemophilia B (Reijnen et al., 1992). Plasma SHBG has also been linked to some diseases. For instance, low plasma SHBG has been recognized as an early indicator of maturity onset diabetes of the young (MODY) (Rudberg and Persson, 1995). Recently, the DNA mutation that causes this disease was mapped to HNF-4 locus (Yamagata et al., 1996), and this mutation was shown to indirectly cause pancreatic beta cell dysfunction and diabetes. The identification of HNF-4 as a regulator of liver-specific shbg expression could thus form a link between diabetes and low plasma SHBG. Moreover, the gene encoding HNF- 4 is also expressed in the kidney and intestine (Sladek, 1994) consistent with the human shbg expression in adult and fetal mice, respectively. These data provide evidence for the regulation of human shbg expression by HNF-4 also in vivo.

6.5 Silencing of human shbg promoter

Although the human shbg transgene that extends up to -803 nts from the transcription start site is active in the transgenic mouse liver, the sequences upstream of -299 repress transcription when placed in front of a luciferase reporter gene in vitro.

It is evident that the Alu repeats at -727/-697 in the human shbg promoter bind a factor that represses transcription, as the removal of these sequences partially recovers the promoter activity. The characterization of this factor is under way, but preliminary analyses indicate that there are multiple complexes involved in this silencing.

41 In support of a physiological role for the more distal TA-rich sequence within the human shbg promoter, there is a polymorphism with respect to the number of TAAAA repeats among genomic DNA samples from different individuals (Hogeveen and Hammond, unpublished results). However, since the DNA samples that were screened were derived from patients with variable health status, i.e. obesity and hirsutism (Pugeat et al., 1996), it has not yet been possible to evaluate the significance of this polymorphism.

In the context of other genes, Sp1 has been shown to participate in specific recruitment of other transcription factors (Wood et al., 1996; Kim et al., 1999). It is therefore of interest that the mutation within Sp1 binding site increases the activity of the human shbg promoter without completely rescuing it. However, even though Sp1 binds to shbg promoter in vitro, it is more likely that the protein interacting with this site in intact cells is indeed a repressor, such as Sp3 or WT-1 also belonging to the same protein family (Philipsen S. and Suske G., 1999).The factor binding to the TA-rich silencer further upstream could then rely on Sp1-related factor on the recruitment to the promoter, but the significance of such interactions is yet to be proved.

42 7. Acknowledgements

This work was carried out at the London Regional Cancer Centre, Ontario, and completed at the Department of Biosciences in the University of Helsinki during years 1994-2000. I wish to thank Professor Carl Gahmberg, the head of the Division of Biochemistry.

I also wish to express my warmest gratitude to

My supervisor, Professor Geoff Hammond for his enthusiasm, huge faith, and understanding attitude towards my temper.

Professor Heikki Rauvala for providing me with supportive surroundings for completing my studies and the manuscript of this thesis.

Professor Pirkko Vihko and Docent Jorma Palvimo for constructive criticism to improve my thesis.

My friends and collegues at LRCC and the Hammond lab for co-operation and help, especially David and Allen for friendship and excellent technical assistance, SPY for collaboration, feedback, and long night time discussions, Kevin for being my friend and for sharing the ups and downs of “shag” research.

My friends and fellow grad students in Canada, especially Cathy, Michelle, Colleen, and Huay- Leng for always being there for me.

Members of the Rauvala lab for friendly and stimulating atmosphere in Viikki. Anni, Carole, and Henri for helping me with the outward appearance of this thesis.

My uncle, Professor Olli Jänne for his invaluable help and advice in various situations through- out my “scientific career”.

My dad for introducing science to me at young age and for being a scientist to look up to. Sami, Päivi, and Cenneth for love and encouragement. Anton och Niki för glädje och många roliga stunder.

All my friends, especially Seija for her friendship and support here and overseas. Nunnu and Hanna for being blond.

This study has been supported by the Medical Research Council of Canada and the Finnish Cultural Foundation.

Helsinki, March 2000,

43 8. References

Adlercreutz, H., Hockerstedt, K., Bannwart, C., Bloigu, S., Hamalainen, E., Fotsis, T., and Ollus, A. (1987). Effect of dietary components, including lignans and phytoestrogens, on enterohepatic circulation and liver metabolism of estrogens and on sex hormone binding globulin (SHBG). Journal of Steroid Biochemistry 27, 1135-1144.

Ain, K.B., Mori, Y., and Refetoff, S. (1987). Reduced clearance rate of thyroxine-binding globulin (TBG) with increased sialylation: a mechanism for estrogen-induced elevation of serum TBG concentration. Journal of Clinical Endocrinology & Metabolism 65, 689-696.

Anderson, D.C. (1974). Sex-hormone-binding globulin. Clinical Endocrinology 3, 69-96.

Anthony, C.T., Danzo, B.J., and Orgebin-Crist, M.C. (1984). Investigations on the relationship between sperm ferti- lizing ability and androgen-binding protein in the restricted rat. Endocrinology 114, 1413-1418.

Aronica, S.M., Kraus, W.L., and Katzenellenbogen, B.S. (1994). Estrogen action via the cAMP signaling pathway: stimulation of adenylate cyclase and cAMP-regulated gene transcription. Proceedings of the National Academy of Sciences of the United States of America 91, 8517-8521.

Avvakumov, G.V., Matveentseva, I.V., Akhrem, L.V., Strel’chyonok, OA, and Akhrem, A.A. (1983). Study of the carbohydrate moiety of human serum sex hormone-binding globulin. Biochimica et Biophysica Acta 760 , 104-110.

Avvakumov, G.V., Zhuk, N.I., and Strel’chyonok, O.A. (1986). Subcellular distribution and selectivity of the protein-binding component of the recognition system for sex-hormone-binding protein-estradiol complex in human decidual endometrium. Biochimica et Biophysica Acta 881, 489-498.

Baker, M.E., French, F.S., and Joseph, D.R. (1987). Vitamin K-dependent protein S is similar to rat androgen- binding protein. Biochemical Journal 243, 293-296.

Bardin, C.W. and Catterall, J.F. (1981). Testosterone: a major determinant of extragenital sexual dimorphism. Sci- ence 211, 1285-1294.

Bardin, C.W., Gunsalus, G.L., Musto, N.A., Cheng, C.Y., Reventos, J., Smith, C., Underhill, D.A., and Hammond, G. (1988). Corticosteroid binding globulin, testosterone-estradiol binding globulin, and androgen binding protein be- long to protein families distinct from steroid receptors. Journal of Steroid Biochemistry 30, 131-139.

Beato, M., Chavez, S., and Truss, M. (1996). Transcriptional regulation by steroid hormones. Steroids 61, 240-251.

Beato, M. and Sanchez-Pacheco, A. (1996). Interaction of steroid hormone receptors with the transcription initiation complex. Endocrine Reviews 17, 587-609.

Becchis, M., Sullivan, P.M., Ordronneau, P., Petrusz, P., Joseph, and DR (1996). Distribution of immunoreactive androgen-binding protein/sex hormone-binding globulin in tissues of the fetal rat. Steroids 61 , 392-400.

Becker, R.R. and Iles, D.J. (1985). Developmental pattern of androgen-binding protein secretion during the critical period of sexual differentiation. Archives of Andrology 14, 107-114.

Berrodin, T.J., Marks, M.S., Ozato, K., Linney, E., and Lazar, M.A. (1992). Heterodimerization among thyroid hormone receptor, retinoic acid receptor, retinoid X receptor, chicken ovalbumin upstream promoter transcription factor, and an endogenous liver protein. Molecular Endocrinology 6, 1468-1478.

Blackmore, P.F., Beebe, S.J., Danforth, D.R., and Alexander, N. (1990). Progesterone and 17 alpha- hydroxyprogesterone. Novel stimulators of calcium influx in human sperm. Journal of Biological Chemistry 265 , 1376-1380.

44 Blackmore, P.F., Neulen, J., Lattanzio, F., and Beebe, S.J. (1991). Cell surface-binding sites for progesterone mediate calcium uptake in human sperm. Journal of Biological Chemistry 266, 18655-18659.

Bocchinfuso, W.P., Warmels-Rodenhiser, S., and Hammond, G.L. (1992a). Structure/function analyses of human sex hormone-binding globulin by site-directed mutagenesis. FEBS Letters 301, 227-230.

Bocchinfuso, W.P., Ma, K.L., Lee, W.M., Warmels-Rodenhiser, S., and Hammond, G.L. (1992b). Selective removal of glycosylation sites from sex hormone-binding globulin by site-directed mutagenesis. Endocrinology 131, 2331-2336.

Bocchinfuso, W.P. and Hammond, G.L. (1994). Steroid-binding and dimerization domains of human sex hormone-binding globulin partially overlap: steroids and Ca2+ stabilize dimer formation. Biochemistry 33, 10622-10629.

Bordin, S. and Petra, P.H. (1980). Immunocytochemical localization of the sex steroid-binding protein of plasma in tissues of the adult monkey Macaca nemestrina. Proceedings of the National Academy of Sciences of the United States of America 77, 5678-5682.

Bression, D., Michard, M., Le Dafniet, M., Pagesy, P., and Peillon (1986). Evidence for a specific estradiol binding site on rat pituitary membranes. Endocrinology 119, 1048-1051.

Burke, C.W. and Anderson, D.C. (1972). Sex-hormone-binding globulin is an oestrogen amplifier. Nature 240, 38-40.

Camper, S.A. and Tilghman, S.M. (1989). Postnatal repression of the alpha-fetoprotein gene is enhancer independent. Genes & Development 3, 537-546.

Casali, E., Petra, P.H., and Ross, J.B. (1990). Fluorescence investigation of the sex steroid binding protein of rabbit serum: steroid binding and subunit dissociation. Biochemistry 29, 9334-9343.

Catterall, J.F., Kontula, K.K., Watson, C.S., Seppanen, P.J., Funkenstein, B., Melanitou, E., Hickok, N.J., Bardin, C.W., and Janne, O.A. (1986). Regulation of gene expression by androgens in murine kidney. Recent Progress in Hormone Research 42, 71-109.

Cheng, S.L. and Musto, N.A. (1982). Purification and characterization of androgen binding protein from rabbit epididymis. Biochemistry 21, 2400-2405.

Cheng, S.L., Kotite, N., and Musto, N.A. (1984). Comparison of rabbit androgen binding protein with testosterone estradiol binding globulin—I. Physical and chemical properties. Journal of Steroid Biochemistry 21, 669-676.

Corvol, P. and Bardin, C.W. (1973). Species distribution of testosterone-binding globulin. Biology of Repro- duction 8, 277-282.

Cousin, P., Dechaud, H., Grenot, C., Lejeune, H., and Pugeat, M. (1998). Human variant sex hormone-binding globulin (SHBG) with an additional carbohydrate chain has a reduced clearance rate in rabbit. Journal of Clinical Endocrinology & Metabolism 83, 235-240.

Cousin, P., Dechaud, H., Grenot, C., Lejeune, H., Hammond, G.L., and Pugeat, M. (1999). Influence of glycosylation on the clearance of recombinant human sex hormone-binding globulin from rabbit blood. Journal of Steroid Bio- chemistry and Molecular Biology 70, 115-121.

Crozat, A., Palvimo, J.J., Julkunen, M., and Janne, O.A. (1992). Comparison of androgen regulation of ornithine decarboxylase and S-adenosylmethionine decarboxylase gene expression in rodent kidney and accessory sex organs. Endocrinology 130, 1131-1144.

45 Cunningham, G.R., Tindall, D.J., Lobl, T.J., Campbell, J.A., Means, and AR (1981). Steroid structural requirements for high affinity binding to human sex steroid binding protein (SBP). Steroids 38, 243-262.

Damassa, D.A., Lin, T.M., Sonnenschein, C., and Soto, A.M. (1991). Biological effects of sex hormone- binding globulin on androgen-induced proliferation and androgen metabolism in LNCaP prostate cells. Endocrinology 129, 75-84.

Danzo, B.J., Eller, B.C., and Orgebin-Crist, M.C. (1974). Studies on the site of origin of the androgen binding protein present in epididymal cytosol from mature intact rabbits. Steroids 24, 107-122.

Danzo, B.J. and Eller, B.C. (1975). Steroid-binding proteins in rabbit plasma: Separation of testosterone-binding globulin (TeBG) from corticosteroid-binding globulin (CBG), preliminary characterization of TeBG, and changes in TeBG concentration during sexual maturation. Molecular & Cellular Endocrinology 2, 351-368.

Danzo, B.J. and Bell, B.W. (1988). The microheterogeneity of androgen-binding protein in rat serum and epididymis is due to differences in glycosylation of their subunits. Journal of Biological Chemistry 263, 2402-2408.

Danzo, B.J., Bell, B.W., and Black, J.H. (1989). Human testosterone-binding globulin is a dimer composed of two identical protomers that are differentially glycosylated. Endocrinology 124, 2809-2817.

Danzo, B.J., Pavlou, S.N., and Anthony, H.L. (1990). Hormonal regulation of androgen-binding protein in the rat. Endocrinology 127, 2829-2838.

Danzo, B.J., Black, J.H., and Bell, B.W. (1991). Analysis of the oligosaccharides on androgen-binding proteins: implications concerning their role in structure/function relationships. Journal of Steroid Biochemistry & Molecular Biology 40, 821-831.

Danzo, B.J. (1995). The effects of a gonadotropin-releasing hormone antagonist on androgen-binding protein distribution and other parameters in the adult male rat. Endocrinology 136, 4004-4011. de Nayer, P., Lambot, M.P., Desmons, M.C., Rennotte, B., Malvaux, and Beckers, C. (1986). Sex hormone-binding protein in hyperthyroxinemic patients: a discriminator for thyroid status in thyroid hormone resistance and familial dysalbuminemic hyperthyroxinemia. Journal of Clinical Endocrinology & Metabolism 62, 1309-1312.

Ehrig, K., Leivo, I., Argraves, W.S., Ruoslahti, E., and Engvall, E. (1990). Merosin, a tissue-specific basement membrane protein, is a laminin-like protein. Proceedings of the National Academy of Sciences of the United States of America 87, 3264-3268.

Engel, J. (1992). Laminins and other strange proteins. Biochemistry 31, 10643-10651.

Feldman, M., Lea, O.A., Petrusz, P., Tres, L.L., Kierszenbaum, A.L., and French, F.S. (1981). Androgen-binding protein. Purification from rat epididymis, characterization, and immunocytochemical localization. Journal of Bio- logical Chemistry 256, 5170-5175.

Felsenfeld, G. (1992). Chromatin as an essential part of the transcriptional mechanism. Nature 355, 219-224.

Fenstermacher, D.A. and Joseph, D.R. (1997). DNA sequences and their binding proteins required for Sertoli cell- specific transcription of the rat androgen-binding protein gene. Molecular Endocrinology 11, 1387-1400.

Forest, M.G., Lecoq, A., David, M., and Pugeat, M. (1988). Effects of human chorionic gonadotropin, androgens, adrenocorticotropin hormone, dexamethasone and hyperprolactinemia on plasma sex steroid-binding protein. An- nals of the New York Academy of Sciences 538, 214-234.

Fortunati, N., Fissore, F., Fazzari, A., Berta, L., Varvello, L., and Frairia, R. (1992). Receptor for sex steroid-binding protein of endometrium membranes: solubilization, partial characterization, and role of estradiol in steroid-binding protein-soluble receptor interaction. Steroids 57, 464-470.

46 Fortunati, N., Fissore, F., Fazzari, A., Berta, L., Benedusi-Pagliano, E., and Frairia, R. (1993). Biological relevance of the interaction between sex steroid binding protein and its specific receptor of MCF-7 cells: effect on the estradiol-induced cell proliferation. Journal of Steroid Biochemistry & Molecular Biology 45, 435-444.

Fortunati, N., Fissore, F., Fazzari, A., Becchis, M., Comba, A., Catalano, M.G., Berta, L., and Frairia, R. (1996). Sex steroid binding protein exerts a negative control on estradiol action in MCF-7 cells (human breast cancer) through cyclic adenosine 3',5'-monophosphate and protein kinase A. Endocrinology 137, 686-692.

Frairia, R., Fortunati, N., Fissore, F., Fazzari, A., Zeppegno, Varvello, L., Orsello, M., and Berta, L. (1992). The membrane receptor for sex steroid binding protein is not ubiquitous. Journal of Endocrinological Investigation 15, 617-619.

French, F.S. and Ritzen, E.M. (1973). A high-affinity androgen-binding protein (ABP) in rat testis: evidence for secretion into efferent duct fluid and absorption by epididymis. Endocrinology 93, 88-95.

Fuhrmann, U., Parczyk, K., Klotzbucher, M., Klocker, H., and Cato, A.C.B. (1998). Recent developments in molecular action of antihormones. Journal Molecular Medicine 76, 512-524.

Gershagen, S., Fernlund, P., and Lundwall, A. (1987). A cDNA coding for human sex hormone binding globulin. Homology to vitamin K-dependent protein S. FEBS Letters 220, 129-135.

Gershagen, S., Lundwall, A., and Fernlund, P. (1989). Characterization of the human sex hormone binding globulin (SHBG) gene and demonstration of two transcripts in both liver and testis. Nucleic Acids Research 17, 9245-9258.

Gorczynska, E. and Handelsman, D.J. (1995). Androgens rapidly increase the cytosolic calcium concentration in Sertoli cells. Endocrinology 136, 2052-2059.

Grenot, C., de Montard, A., Blachere, T., de Ravel, M.R., Mappus, and Cuilleron, C.Y. (1992). Characterization of Met-139 as the photolabeled amino acid residue in the steroid binding site of sex hormone binding globulin using delta 6 derivatives of either testosterone or estradiol as unsubstituted photoaffinity labeling reagents. Biochemistry 31, 7609-7621.

Grishkovskaya, I., Skelenar, G., Avvakumov, G.V., Dales, D., Behlke, J., Hammond, G.L., and Muller Y.A. (1999). Crystallization of the N-terminal domain of human sex hormone-binding globulin, the major sex steroid carrier in blood. Acta Crystallographica D55, 2053-2055.

Grishkovskaya, I., Avvakumov, G.V., Sklenar, G., Dales, D., Hammond, G.L., and Muller, Y.A. (2000). Crystal structure of human sex hormone-binding globulin: steroid transport by a laminin G-like domain. EMBO Journal 19, 504-512.

Gunsalus, G.L., Larrea, F., Musto, N.A., Becker, R.R., Mather, J.P., and Bardin, C.W. (1981). Androgen binding protein as a marker for Sertoli cell function. Journal of Steroid Biochemistry 15, 99-106.

Hadzopoulou-Cladaras, M., Kistanova, E., Evagelopoulou, C., Zeng, S., Cladaras, C., and Ladias, J.A. (1997). Func- tional domains of the nuclear receptor hepatocyte nuclear factor 4. Journal of Biological Chemistry 272, 539-550.

Haffner, S.M., Valdez, R.A., Morales, P.A., Hazuda, H.P., and Stern, M.P. (1993). Decreased sex hormone-binding globulin predicts noninsulin-dependent diabetes mellitus in women but not in men. Journal of Clinical Endocrinology & Metabolism 77, 56-60.

Hagenas, L., Ritzen, E.M., Plooen, L., Hansson, V., French, F.S., and Nayfeh, S.N. (1975). Sertoli cell origin of testicular androgen-binding protein (ABP). Molecular & Cellular Endocrinology 2, 339-350.

Hammond, G.L. and Lahteenmaki, P.L. (1983). A versatile method for the determination of serum cortisol binding globulin and sex hormone binding globulin binding capacities. Clinica Chimica Acta 132, 101-110.

47 Hammond, G.L., Langley, M.S., and Robinson, P.A. (1985). A liquid-phase immunoradiometric assay (IRMA) for human sex hormone binding globulin (SHBG). Journal of Steroid Biochemistry 23, 451-460.

Hammond, G.L., Robinson, P.A., Sugino, H., Ward, D.N., and Finne, J. (1986). Physicochemical characteristics of human sex hormone binding globulin: evidence for two identical subunits. Journal of Steroid Biochemistry 24, 815-824.

Hammond, G.L., Underhill, D.A., Smith, C.L., Goping, I.S., Harley, MJ, Musto, N.A., Cheng, C.Y., and Bardin, C.W. (1987). The cDNA-deduced primary structure of human sex hormone-binding globulin and location of its steroid-binding domain. FEBS Letters 215, 100-104.

Hammond, G.L., Underhill, D.A., Rykse, H.M., and Smith, C.L. (1989). The human sex hormone-binding globulin gene contains exons for androgen-binding protein and two other testicular messenger RNAs. Molecular Endocrinology 3, 1869-1876.

Hammond, G.L. (1990). Molecular properties of corticosteroid binding globulin and the sex-steroid binding proteins. Endocrine Reviews 11, 65-79.

Hammond, G.L. (1995). Potential Functions of Plasma Steroid-Binding Proteins. Trends in Endocrinology and Metabolism 6, 298-304.

Hammond, G.L. and Bocchinfuso, W.P. (1995). Sex hormone-binding globulin/androgen-binding protein: steroid- binding and dimerization domains. Journal of Steroid Biochemistry & Molecular Biology 53, 543-552.

Hammond, G.L. and Bocchinfuso, W.P. (1996). Sex Hormone-Binding Globulin: Gene Organization and Structure/ Function Analyses. Hormone Research 45, 197-201.

Hansson, V. (1972). Further characterization of the 5 -dihydrotestosterone binding protein in the epididymal cytosol fraction. In vitro studies. Steroids 20, 575-596.

Hansson, V., Reusch, E., Trygstad, O., Torgersen, O., Ritzen, E.M., and French, F.S. (1973). FSH stimulation of testicular androgen binding protein. Nature - New Biology 246, 56-58.

He, X., Shen, L., Malmborg, A.C., Smith, K.J., Dahlback, B., and Linse, S. (1997). Binding site for C4b-binding protein in vitamin K-dependent protein S fully contained in carboxy-terminal laminin-G-type repeats. A study using recombinant factor IX-protein S chimeras and surface plasmon resonance. Biochemistry 36, 3745-3754.

Henttu, P., Liao, S.S., and Vihko, P. (1992). Androgens up-regulate the human prostate-specific antigen messenger ribonucleic acid (mRNA), but down-regulate the prostatic acid phosphatase mRNA in the LNCaP cell line. Endocrinology 130, 766-772.

Hertz, R., Magenheim, J., Berman, I., and Bar-Tana, J. (1998). Fatty acyl-CoA thioesters are ligands of hepatic nuclear factor-4alpha. Nature 392, 512-516.

Hessing, M. (1991). The interaction between complement component C4b-binding protein and the vitamin K-dependent protein S forms a link between blood coagulation and the complement system. Biochemical Journal 277, 581-592.

Hryb, D.J., Khan, M.S., and Rosner, W. (1985). Testosterone-estradiol-binding globulin binds to human prostatic cell membranes. Biochemical & Biophysical Research Communications 128, 432-440.

Hryb, D.J., Khan, M.S., Romas, N.A., and Rosner, W. (1989). Solubilization and partial characterization of the sex hormone-binding globulin receptor from human prostate. Journal of Biological Chemistry 264, 5378-5383.

Hryb, D.J., Khan, M.S., Romas, N.A., and Rosner, W. (1990). The control of the interaction of sex hormone-binding globulin with its receptor by steroid hormones. Journal of Biological Chemistry 265, 6048-6054.

48 Hämäläinen, E., Tikkanen, H., Harkonen, M., Naveri, H., and Adlercreutz, H. (1987). Serum lipoproteins, sex hormones and sex hormone binding globulin in middle-aged men of different physical fitness and risk of coronary heart disease. Atherosclerosis 67, 155-162.

Jiang, G., Nepomuceno, L., Hopkins, K., and Sladek, F.M. (1995). Exclusive homodimerization of the orphan receptor hepatocyte nuclear factor 4 defines a new subclass of nuclear receptors. Molecular & Cellular Biology 15, 5131-5143.

Joseph, D.R., Hall, S.H., and French, F.S. (1985). Identification of complementary DNA clones that encode rat androgen binding protein. Journal of Andrology 6, 392-395.

Joseph, D.R., Hall, S.H., and French, F.S. (1987). Rat androgen-binding protein: evidence for identical subunits and amino acid sequence homology with human sex hormone-binding globulin. Proceedings of the National Academy of Sciences of the United States of America 84, 339-343.

Joseph, D.R., Hall, S.H., Conti, M., and French, F.S. (1988). The gene structure of rat androgen-binding protein: identification of potential regulatory deoxyribonucleic acid elements of a follicle-stimulating hormone-regulated protein. Molecular Endocrinology 2, 3-13.

Joseph, D.R. (1994). Structure, function, and regulation of androgen-binding protein/sex hormone-binding globulin. Vitamines and Hormones 49:197-280.

Joseph, D.R., Becchis, M., Fenstermacher, D.A., and Petrusz, P. (1996). The alternate N-terminal sequence of rat androgen-binding protein/sex hormone-binding globulin contains a nuclear targeting signal. Endocrinology 137, 1138-1143.

Joseph, D.R. (1997a). Sequence and functional relationships between androgen-binding protein/sex hormone- binding globulin and its homologs protein S, Gas6, laminin, and agrin. Steroids 62, 578-588.

Joseph, D.R., O’Brien, D.A., Sullivan, P.M., Becchis, M., Tsuruta, JK, and Petrusz, P. (1997b). Overexpression of androgen-binding protein/sex hormone-binding globulin in male transgenic mice: tissue distribution and phenotypic disorders. Biology of Reproduction 56, 21-32.

Kato, T. and Horton, R. (1968). Studies of testosterone binding globulin. Journal of Clinical Endocrinology & Metabolism 28, 1160-1168.

Khan, M.S., Knowles, B.B., Aden, D.P., and Rosner, W. (1981). Secretion of testosterone-estradiol-binding globulin by a human hepatoma-derived cell line. Journal of Clinical Endocrinology & Metabolism 53, 448-449.

Khan, M.S. and Rosner, W. (1990a). Histidine 235 of human sex hormone-binding globulin is the covalent site of attachment of the nucleophilic steroid derivative, 17 beta-bromoacetoxydihydrotestosterone. Journal of Biological Chemistry 265, 8431-8435.

Khan, M.S., Hryb, D.J., Hashim, G.A., Romas, N.A., and Rosner, W. (1990b). Delineation and synthesis of the membrane receptor-binding domain of sex hormone-binding globulin. Journal of Biological Chemistry 265, 18362- 18365.

Kierszenbaum, A.L., Feldman, M., Lea, O., Spruill, W.A., Tres, L.L., Petrusz, P., and French, F.S. (1980). Localiza- tion of androgen-binding protein in proliferating Sertoli cells in culture. Proceedings of the National Academy of Sciences of the United States of America 77, 5322-5326.

Kim, M.K., Lee, J.S., and Chung, J.H. (1999). In vivo transcription factor recruitment during thyroid hormone receptor-mediated activation. Proceedings of the National Academy of Sciences of the United States of America 96, 10092-10097.

Kingston, R.E., Bunker, C.A., and Imbalzano, A.N. (1996). Repression and activation by multiprotein complexes that alter chromatin structure. Genes & Development 10, 905-920.

49 Kleinman, H.K., Weeks, B.S., Schnaper, H.W., Kibbey, M.C., Yamamura, and Grant, D.S. (1993). The laminins: a family of basement membrane glycoproteins important in cell differentiation and tumor metastases. Vitamins and Hormones 47, 161-186.

Kotite, N.J., Nayfeh, S.N., and French, F.S. (1978). FSH and androgen regulation of Sertoli cell function in the immature rat. Biology of Reproduction 18, 65-73.

Kotite, N.J., Cheng, S.L., Musto, N.A., and Gunsalus, G.L. (1986). Comparison of rabbit epididymal androgen binding protein and serum testosterone estradiol binding globulin—II. Immunological properties. Journal of Steroid Biochemistry 25, 171-176.

Lapidus, L., Lindstedt, G., Lundberg, P.A., Bengtsson, C., and Gredmark, T. (1986). Concentrations of sex-hormone binding globulin and corticosteroid binding globulin in serum in relation to cardiovascular risk factors and to 12-year incidence of cardiovascular disease and overall mortality in postmenopausal women. Clinical Chemistry 32, 146- 152.

Larrea, F., Diaz, L., Carino, C., Larriva-Sahd, J., Carrillo, L., Orozco, H., and Ulloa-Aguirre, A. (1993). Evidence that human placenta is a site of sex hormone-binding globulin gene expression. Journal of Steroid Biochemistry & Molecular Biology 46, 497-505.

Larriba, S., Esteban, C., Toran, N., Gerard, A., Audi, L., Gerard, and Reventos, J. (1995). Androgen binding protein is tissue-specifically expressed and biologically active in transgenic mice. Journal of Steroid Biochemistry & Mo- lecular Biology 53, 573-578.

Lasnitzki, I. and Franklin, H.R. (1972). The influence of serum on uptake, conversion and action of testosterone in rat prostate glands in organ culture. Journal of Endocrinology 54, 333-342.

Lasnitzki, I. and Franklin, H.R. (1975). The influence of serum on the uptake, conversion and action of dihydrotestosterone in rat prostate glands in organ culture. Journal of Endocrinology 64, 289-297.

Lea, O.A. and Stoa, K.F. (1972). The binding of testosterone to different serum proteins: a comparative study. Jour- nal of Steroid Biochemistry 3, 409-419.

Lee, C., Sutkowski, D.M., Sensibar, J.A., Zelner, D., Kim, I., Amsel, I., Shaw, N., Prins, G.S., and Kozlowski, J.M. (1995). Regulation of proliferation and production of prostate-specific antigen in androgen-sensitive prostatic cancer cells, LNCaP, by dihydrotestosterone. Endocrinology 136, 796-803.

Leng, X., Cooney, A.J., Tsai, S.Y., and Tsai, M.J. (1996). Molecular mechanisms of COUP-TF-mediated transcriptional repression: evidence for transrepression and active repression. Molecular & Cellular Biology 16, 2332-2340.

Lindstedt, G., Lundberg, P.A., Lapidus, L., Lundgren, H., Bengtsson, C., and Bjorntorp, P. (1991). Low sex-hormone-binding globulin concentration as independent risk factor for development of NIDDM. 12-yr follow-up of population study of women in Gothenburg, Sweden. Diabetes 40, 123-128.

Loukovaara, M., Carson, M., and Adlercreutz, H. (1995a). Regulation of production and secretion of sex hormone- binding globulin in HepG2 cell cultures by hormones and growth factors. Journal of Clinical Endocrinology & Metabolism 80, 160-164.

Loukovaara, M., Carson, M., and Adlercreutz, H. (1995b). Regulation of sex-hormone-binding globulin production by endogenous estrogens in vitro. Biochemical & Biophysical Research Communications 206, 895-901.

Loukovaara, M., Carson, M., Palotie, A., and Adlercreutz, H. (1995c). Regulation of sex hormone-binding globulin production by isoflavonoids and patterns of isoflavonoid conjugation in HepG2 cell cultures. Steroids 60, 656-661.

Malik, S. and Karathanasis, S.K. (1996). TFIIB-directed transcriptional activation by the orphan nuclear receptor hepatocyte nuclear factor 4. Molecular & Cellular Biology 16, 1824-1831.

50 Mangelsdorf, D.J., Thummel, C., Beato, M., Herrlich, P., Schutz, Umesono, K., Blumberg, B., Kastner, P., Mark, M., and Chambon, P. (1995). The nuclear receptor superfamily: the second decade. Cell 83, 835-839.

McKenna, N.J., Xu, J., Tsai, S.Y., Tsai, M.J., and O’Malley, B.W. (1999). Nuclear receptor coactivators: multiple enzymes, multiple complexes, multiple functions. Journal of Steroid Biochemistry & Molecular Biology 69, 3-12.

Mendel, C.M. (1989). The free hormone hypothesis: a physiologically based mathematical model. Endocrine Reviews 10, 232-274.

Misao, R., Itoh, N., Mori, H., Fujimoto, J., and Tamaya, T. (1994). Sex hormone-binding globulin mRNA levels in human uterine endometrium. European Journal of Endocrinology 131, 623-629.

Misao, R., Nakanishi, Y., Fujimoto, J., and Tamaya, T. (1997a). Expression of sex hormone-binding globulin exon VII splicing variant mRNA in human uterine myometrium and leiomyoma. Journal of Steroid Biochem- istry & Molecular Biology 63, 269-274.

Misao, R., Fujimoto, J., Nakanishi, Y., and Tamaya, T. (1997b). Expression of sex hormone-binding globulin exon VII splicing variant mRNA in human uterine endometrium. Journal of Steroid Biochemistry & Molecu- lar Biology 62, 385-390.

Misao, R., Nakanishi, Y., Fujimoto, J., and Tamaya, T. (1997c). Expression of sex hormone-binding globulin exon VII splicing variant messenger RNA in human uterine endometrial cancers. Cancer Research 57, 5579- 5583.

Misao, R., Nakanishi, Y., Fujimoto, J., Iwagaki, S., and Tamaya, T. (1998a). Dominant expression of sex- hormone-binding-globulin exon-7 splicing variant over wild-type mRNA in human ovarian cancers. Interna- tional Journal of Cancer 77, 828-832.

Misao, R., Nakanishi, Y., Fujimoto, J., and Tamaya, T. (1998b). Expression of sex hormone-binding globulin exon VII splicing variant messenger ribonucleic acid in human ovarian endometriosis. Fertility & Sterility 69, 324-328.

Moore, K.H., Bertram, K.A., Gomez, R.R., Styner, M.J., and Matej, L.A. (1996). Sex hormone binding globulin mRNA in human breast cancer: detection in cell lines and tumor samples. Journal of Steroid Biochem- istry & Molecular Biology 59, 297-304.

Morley, P., Whitfield, J.F., Vanderhyden, B.C., Tsang, B.K., and Schwartz, J.L. (1992). A new, nongenomic estrogen action: the rapid release of intracellular calcium. Endocrinology 131, 1305-1312.

Musto, N., Gunsalus, G.L., and Bardin, C.W. (1980). Purification and characterization of androgen binding protein from the rat epididymis. Biochemistry 19, 2853-2860.

Nakhla, A.M., Khan, M.S., and Rosner, W. (1990). Biologically active steroids activate receptor-bound human sex hormone-binding globulin to cause LNCaP cells to accumulate adenosine 3',5'-monophosphate. Journal of Clinical Endocrinology & Metabolism 71, 398-404.

Nakhla, A.M., Khan, M.S., Romas, N.P., and Rosner, W. (1994). Estradiol causes the rapid accumulation of cAMP in human prostate. Proceedings of the National Academy of Sciences of the United States of America 91, 5402-5405.

Nakhla, A.M. and Rosner, W. (1996). Stimulation of prostate cancer growth by androgens and estrogens through the intermediacy of sex hormone-binding globulin. Endocrinology 137, 4126-4129.

Nakhla, A.M., Romas, N.A., and Rosner, W. (1997). Estradiol activates the prostate androgen receptor and prostate-specific antigen secretion through the intermediacy of sex hormone-binding globulin. Journal of Biological Chemistry 272, 6838-6841.

51 Namkung, P.C., Kumar, S., Walsh, K.A., and Petra, P.H. (1990). Identification of lysine 134 in the steroid- binding site of the sex steroid-binding protein of human plasma. Journal of Biological Chemistry 265, 18345- 18350.

Noe, G., Cheng, Y.C., Dabike, M., and Croxatto, H.B. (1992). Tissue uptake of human sex hormone-binding globulin and its influence on ligand kinetics in the adult female rat. Biology of Reproduction 47, 970-976.

Pardridge, W.M. (1988). Selective delivery of sex steroid hormones to tissues in vivo by albumin and by sex hormone-binding globulin. Annals of the New York Academy of Sciences 538, 173-192.

Patthy, L. and Nikolics, K. (1993). Functions of agrin and agrin-related proteins. Trends in Neurosciences 16, 76-81.

Pearlman, W.H. and Crepy, O. (1967). Steroid-protein interaction with particular reference to testosterone binding by human serum. Journal of Biological Chemistry 242, 182-189.

Peiris, A.N., Sothmann, M.S., Aiman, E.J., and Kissebah, A.H. (1989). The relationship of insulin to sex hormone-binding globulin: role of adiposity. Fertility & Sterility 52, 69-72.

Pelliniemi, L.J., Dym, M., Gunsalus, G.L., Musto, N.A., Bardin, C.W., and Fawcett, D.W. (1981). Immunocy- tochemical localization of androgen-binding protein in the male rat reproductive tract. Endocrinology 108, 925-931.

Petra, P.H., Stanczyk, F.Z., Namkung, P.C., Fritz, M.A., and Novy, M.J. (1985). Direct effect of sex steroid- binding protein (SBP) of plasma on the metabolic clearance rate of testosterone in the rhesus macaque. Journal of Steroid Biochemistry 22, 739-746.

Philipsen, S., and Suske, G. (1999). A tale of three fingers: the family of mammalian Sp/XKLF transcription factors. Nucleic Acids Reseach 27, 2991-3000.

Pietras, R.J. and Szego, C.M. (1977). Specific binding sites for oestrogen at the outer surfaces of isolated endometrial cells. Nature 265, 69-72.

Pietras, R.J. and Szego, C.M. (1980). Partial purification and characterization of oestrogen receptors in subfractions of hepatocyte plasma membranes. Biochemical Journal 191, 743-760.

Pike, A.C.W., Brzozowski, A.M., Hubbard, R.E., Bonn, T., Thorsell, A.-G., Engström, O., Ljunggren, J., Gustavsson, J.-A., and Carlsquist, M. (1999). Structure of the ligand-binding domain of oestrogen receptor β in the presence of a partial agonist and a full antagonist. EMBO Journal 18, 4608-4618.

Plymate, S.R., Matej, L.A., Jones, R.E., and Friedl, K.E. (1988). Inhibition of sex hormone-binding globulin production in the human hepatoma (Hep G2) cell line by insulin and prolactin. Journal of Clinical Endocrinology & Metabolism 67, 460-464.

Porto, C.S., Gunsalus, G.L., Bardin, C.W., Phillips, D.M., and Musto, N.A. (1991). Receptor-mediated endocytosis of an extracellular steroid-binding protein (TeBG) in MCF-7 human breast cancer cells. Endocrinology 129, 436-445.

Porto, C.S., Abreu, L.C., Gunsalus, G.L., and Bardin, C.W. (1992). Binding of sex-hormone-binding globulin (SHBG) to testicular membranes and solubilized receptors. Molecular & Cellular Endocrinology 89, 33-38.

Power, S.G., Bocchinfuso, W.P., Pallesen, M., Warmels, R., Van Baelen, H., and Hammond, G.L. (1992). Molecular analyses of a human sex hormone-binding globulin variant: evidence for an additional carbohydrate chain. Journal of Clinical Endocrinology & Metabolism 75, 1066-1070.

Preziosi, P., Barrett-Connor, E., Papoz, L., Roger, M., Saint-Paul, M., Nahoul, K., and Simon, D. (1993). Interrelation between plasma sex hormone-binding globulin and plasma insulin in healthy adult women: the

52 telecom study [see comments]. Journal of Clinical Endocrinology & Metabolism 76, 283-287.

Pugeat, M., Moulin, P., Cousin, P., Fimbel, S., Nicolas, M.H., Crave, J.C., and Lejeune, H. (1995). Interrelations between sex hormone-binding globulin (SHBG), plasma lipoproteins and cardiovascular risk. Journal of Steroid Biochemistry & Molecular Biology 53, 567-572.

Pugeat, M., Crave, J.C., Tourniaire, J., and Forest, M.G. (1996). Clinical utility of sex hormone-binding globulin measurement. Hormone Research 45, 148-155.

Quandt, K., Frech, K., Karas, H., Wingender, E., and Werner, T. (1995). MatInd and MatInspector: new fast and versatile tools for detection of consensus matches in nucleotide sequence data. Nucleic Acids Research 23, 4878-4884.

Que, B.G. and Petra, P.H. (1987). Characterization of a cDNA coding for sex steroid-binding protein of human plasma. FEBS Letters 219, 405-409.

Reed, M.J., Cheng, R.W., Simmonds, M., Richmond, W., and James, V.H. (1987). Dietary lipids: an additional regulator of plasma levels of sex hormone binding globulin. Journal of Clinical Endocrinology & Metabolism 64, 1083-1085.

Reijnen, M.J., Sladek, F.M., Bertina, R.M., and Reitsma, P.H. (1992). Disruption of a binding site for hepatocyte nuclear factor 4 results in hemophilia B Leyden. Proceedings of the National Academy of Sciences of the United States of America 89, 6300-6303.

Renoir, J.M., Mercier-Bodard, C., and Baulieu, E.E. (1980). Hormonal and immunological aspects of the phylogeny of sex steroid binding plasma protein. Proceedings of the National Academy of Sciences of the United States of America 77, 4578-4582.

Reventos, J., Hammond, G.L., Crozat, A., Brooks, D.E., Gunsalus, GL, Bardin, C.W., and Musto, N.A. (1988). Hormonal regulation of rat androgen-binding protein (ABP) messenger ribonucleic acid and homology of human testosterone-estradiol-binding globulin and ABP complementary deoxyribonucleic acids. Molecular Endocrinology 2, 125-132.

Reventos, J., Sullivan, P.M., Joseph, D.R., and Gordon, J.W. (1993). Tissue-specific expression of the rat androgen- binding protein/sex hormone-binding globulin gene in transgenic mice. Molecular & Cellular Endocrinology 96, 69- 73.

Riegman, P.H., Vlietstra, R.J., van der Korput, J.A., Brinkmann, A.O., and Trapman, J. (1991). The promoter of the prostate-specific antigen gene contains a functional androgen responsive element. Molecular Endocrinology 5, 1921- 1930.

Ritzen, E.M., Nayfeh, S.N., French, F.S., and Dobbins, M.C. (1971). Demonstration of androgen-binding components in rat epididymis cytosol and comparison with binding components in prostate and other tissues. Endocrinology 89, 143-151.

Ritzen, E.M., Hagenas, L., Hansson, V., and Frensh, F.S. (1975). In vitro synthesis of testicular androgen binding protein (ABP): stimulation by FSH and androgen. Current Topics in Molecular Endocrinology 2 , 353-366.

Ritzen, E.M., Boitani, C., Parvinen, M., French, F.C., and Feldman, M. (1982). Stage-dependent secretion of ABP by rat seminiferous tubules. Molecular & Cellular Endocrinology 25, 25-33.

Rosenbaum, W., Christy, N.P., and Kelly, W.G. (1966). Electrophoretic evidence for the presence of an estrogen- binding beta-globulin in human plasma. Journal of Clinical Endocrinology & Metabolism 26, 1399-1403.

Rosner, W. and Deakins, S.M. (1968). Testosterone-binding globulins in human plasma: studies on sex distribution and specificity. Journal of Clinical Investigation 47, 2109-2116.

53 Rosner, W. and Darmstadt, R.A. (1973). Demonstration and partial characterization of a rabbit serum protein which binds testosterone and dihydrotestosterone. Endocrinology 92, 1700-1707.

Rosner, W., Toppel, S., and Smith, R.N. (1974). Testosterone-estradiol-binding globulin of human plasma: denaturation and protection. Biochimica et Biophysica Acta 351, 92-98.

Rosner, W., Aden, D.P., and Khan, M.S. (1984). Hormonal influences on the secretion of steroid-binding proteins by a human hepatoma-derived cell line. Journal of Clinical Endocrinology & Metabolism 59, 806- 808.

Rudberg, S. and Persson, B. (1995). Indications of low sex hormone binding globulin (SHBG) in young females with type 1 diabetes, and an independent association to microalbuminuria. Diabetes Medicine 12, 816-822.

Sanborn, B.M., Elkington, J.S., Chowdhury, M., Tcholakian, R.K., and Steinberger, E. (1975). Hormonal influences on the level of testicular androgen binding activity: effect of FSH following hypophysectomy. Endocrinology 96, 304-312.

Sarne, D.H., Refetoff, S., Rosenfield, R.L., and Farriaux, J.P. (1988). Sex hormone-binding globulin in the diagnosis of peripheral tissue resistance to thyroid hormone: the value of changes after short term triiodothyronine administration. Journal of Clinical Endocrinology & Metabolism 66, 740-746.

Sasaki, M., Kleinman, H.K., Huber, H., Deutzmann, R., and Yamada, Y. (1988). Laminin, a multidomain protein. The A chain has a unique globular domain and homology with the basement membrane proteoglycan and the laminin B chains. Journal of Biological Chemistry 263, 16536-16544.

Schmidt, W.N., Taylor, C.A.J., and Danzo, B.J. (1981). Use of a photoaffinity ligand to compare androgen- binding protein (ABP) present in rat Sertoli cell culture media with ABP present in epididymal cytosol. Endocrinology 108, 786-794.

Siiteri, P.K., Murai, J.T., Hammond, G.L., Nisker, J.A., Raymoure, WJ, and Kuhn, R.W. (1982). The serum transport of steroid hormones. Recent Progress in Hormone Research 38, 457-510.

Simon, D., Preziosi, P., Barrett-Connor, E., Roger, M., Saint-Paul, M., Nahoul, K., and Papoz, L. (1992). Interrelation between plasma testosterone and plasma insulin in healthy adult men: the Telecom Study. Diabetologia 35, 173-177.

Sladek, F.M. (1994). Orphan receptor HNF-4 and liver-specific gene expression. Receptor 4, 64

Steinberger, A., Heindel, J.J., Lindsey, J.N., Elkington, J.S., Sanborn, B.M., and Steinberger, E. (1975). Isola- tion and culture of FSH responsive Sertoli cells. Endocrine Research Communications 2, 261-272.

Strel’chyonok, O.A., Avvakumov, G.V., and Survilo, L.I. (1984). A recognition system for sex-hormone- binding protein-estradiol complex in human decidual endometrium plasma membranes. Biochimica et Biophysica Acta 802, 459-466.

Sui, L.M., Cheung, A.W., Namkung, P.C., and Petra, P.H. (1992). Localization of the steroid-binding site of the human sex steroid-binding protein of plasma (SBP or SHBG) by site-directed mutagenesis. FEBS Letters 310, 115-118.

Sullivan, P.M., Petrusz, P., Szpirer, C., and Joseph, D.R. (1991). Alternative processing of androgen-binding protein RNA transcripts in fetal rat liver. Identification of a transcript formed by trans splicing. Journal of Biological Chemistry 266, 143-154.

Tanenbaum, D.M., Wang, Y., Williams, S.P., and Sigler, P.B. (1998). Crystallographic comparison of the estrogen and progesterone receptor’s ligand binding domains. Proceedings of the National Academy of Sci- ences of the United States of America 95, 5998-6003.

54 Tardivel-Lacombe, J. and Degrelle, H. (1991). Hormone-associated variation of the glycan microheterogeneity pattern of human sex steroid-binding protein (hSBP). Journal of Steroid Biochemistry & Molecular Biology 39, 449-453.

Toscano, V., Balducci, R., Bianchi, P., Guglielmi, R., Mangiantini, A., and Sciarra, F. (1992). Steroidal and non- steroidal factors in plasma sex hormone binding globulin regulation. Journal of Steroid Biochemistry & Molecular Biology 43, 431-437.

Tsai, M.J. and O’Malley, B.W. (1994). Molecular mechanisms of action of steroid/thyroid receptor superfamily members. Annual Review of Biochemistry 63, 451-486.

Turner, T.T., Jones, C.E., Howards, S.S., Ewing, L.L., Zegeye, B., and Gunsalus, G.L. (1984). On the androgen microenvironment of maturing spermatozoa. Endocrinology 115, 1925-1932.

Vacher, J. and Tilghman, S.M. (1990). Dominant negative regulation of the mouse alpha-fetoprotein gene in adult liver. Science 250, 1732-1735.

Veldscholte, J., Ris-Stalpers, C., Kuiper, G.G., Jenster, G., Berrevoets C., Claassen, E., van Rooij, H.C., Trapman, J., Brinkmann, A.O., and Mulder, E. (1990). A mutation in the ligand-binding domain of the androgen receptor of human LNCaP cells affects steroid binding characteristics and response to anti-androgens. Biochemical and Biophysical Research Communications 173, 534-540.

Vermeulen, A. and Verdonck, L. (1968). Studies on the binding of testosterone to human plasma. Steroids 11, 609-635.

Vermeulen, A. (1988). Physiology of the testosterone-binding globulin in man. Annals of the New York Academy of Sciences 538, 103-111.

Walsh, K.A., Titani, K., Takio, K., Kumar, S., Hayes, R., Petra, and PH (1986). Amino acid sequence of the sex steroid binding protein of human blood plasma. Biochemistry 25, 7584-7590.

Wang, Y.M., Bayliss, D.A., Millhorn, D.E., Petrusz, P., and Joseph, D.R. (1990). The androgen-binding protein gene is expressed in male and female rat brain. Endocrinology 127, 3124-3130.

Weddington, S.C., McLean, W.S., Nayfeh, S.N., and French, F.S. (1974). Androgen binding protein (ABP) in rabbit testis and epididymis. Steroids 24, 123-134.

Wenn, R.V., Kamberi, I.A., Keyvanjah, M., and Johannes, A. (1977). Distribution of testosterone-estradiol binding globulin (TeBG) in the higher vertebrates. Endokrinologie 69, 151-156.

Wiley, S.R., Kraus, R.J., and Mertz, J.E. (1992). Functional binding of the “TATA” box binding component of transcription factor TFIID to the -30 region of TATA-less promoters. Proceedings of the National Academy of Sciences of the United States of America 89, 5814-5818.

Wood, I.C., Roopra, A., and Buckley, N.J. (1996). Neural specific expression of the m4 muscarinic acetylcholine receptor gene is mediated by a RE1/NRSE-type silencing element. Journal of Biological Chemistry 271, 14221- 14225.

Wu, C.H., Motohashi, T., Abdel-Rahman, H.A., Flickinger, G.L., and Mikhail, G. (1976). Free and protein-bound plasma estradiol-17 beta during the menstrual cycle. Journal of Clinical Endocrinology & Metabolism 43, 436-445.

Xie, M., Shao, G., Buyse, I.M., and Huang, S. (1997). Transcriptional repression mediated by the PR domain zinc finger gene RIZ. Journal of Biological Chemistry 272, 26360-26366.

Yamagata, K., Furuta, H., Oda, N., Kaisaki, P.J., Menzel, S., Cox, NJ, Fajans, S.S., Signorini, S., Stoffel, M., and Bell, G.I. (1996). Mutations in the hepatocyte nuclear factor-4alpha gene in maturity-onset diabetes of the young (MODY1) Nature 384, 458-460.