This Article Abstract Full Text (PDF) All Versions of this Article: 26/4/525    most recent Final Manuscript Author Manuscript Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Simard, J. Articles by Melner, M. H. Search for Related Content PubMed PubMed Citation Articles by Simard, J. Articles by Melner, M. H.

Molecular Biology of the 3ß-Hydroxysteroid Dehydrogenase/⁵-⁴ Isomerase Gene Family

Canada Research Chair in Oncogenetics (J.S., M.-L.R., S.G., P.S.), Oncology and Molecular Endocrinology Research Center, Laval University Medical Center, and Laval University, Quebec, Canada G1V 4G2; and Departments of Obstetrics/Gynecology and Cell Biology (F.A.F., M.H.M.), Vanderbilt University School of Medicine, Nashville, Tennessee 37232

Correspondence: Address all correspondence and requests for reprints to: Professor Jacques Simard, Cancer Genomics Laboratory, T3-57, Laval University Medical Center (CHUL) Research Center, 2705 Laurier Boulevard, Québec City, Québec, Canada G1V 4G2. E-mail: jacques.simard{at}crchul.ulaval.ca

	Abstract

Top Abstract I. Introduction II. Human Type I... III. Structure-Function... IV. Evolution of the... V. Transcriptional Regulation of... VI. Ontogeny, Localization, and... VII. Molecular Genetics of... References

 
The 3ß-hydroxysteroid dehydrogenase/⁵-⁴ isomerase (3ß-HSD) isoenzymes are responsible for the oxidation and isomerization of ⁵-3ß-hydroxysteroid precursors into ⁴-ketosteroids, thus catalyzing an essential step in the formation of all classes of active steroid hormones. In humans, expression of the type I isoenzyme accounts for the 3ß-HSD activity found in placenta and peripheral tissues, whereas the type II 3ß-HSD isoenzyme is predominantly expressed in the adrenal gland, ovary, and testis, and its deficiency is responsible for a rare form of congenital adrenal hyperplasia. Phylogeny analyses of the 3ß-HSD gene family strongly suggest that the need for different 3ß-HSD genes occurred very late in mammals, with subsequent evolution in a similar manner in other lineages. Therefore, to a large extent, the 3ß-HSD gene family should have evolved to facilitate differential patterns of tissue- and cell-specific expression and regulation involving multiple signal transduction pathways, which are activated by several growth factors, steroids, and cytokines. Recent studies indicate that HSD3B2 gene regulation involves the orphan nuclear receptors steroidogenic factor-1 and dosage-sensitive sex reversal adrenal hypoplasia congenita critical region on the X chromosome gene 1 (DAX-1). Other findings suggest a potential regulatory role for STAT5 and STAT6 in transcriptional activation of HSD3B2 promoter. It was shown that epidermal growth factor (EGF) requires intact STAT5; on the other hand IL-4 induces HSD3B1 gene expression, along with IL-13, through STAT 6 activation. However, evidence suggests that multiple signal transduction pathways are involved in IL-4 mediated HSD3B1 gene expression. Indeed, a better understanding of the transcriptional factors responsible for the fine control of 3ß-HSD gene expression may provide insight into mechanisms involved in the functional cooperation between STATs and nuclear receptors as well as their potential interaction with other signaling transduction pathways such as GATA proteins. Finally, the elucidation of the molecular basis of 3ß-HSD deficiency has highlighted the fact that mutations in the HSD3B2 gene can result in a wide spectrum of molecular repercussions, which are associated with the different phenotypic manifestations of classical 3ß-HSD deficiency and also provide valuable information concerning the structure-function relationships of the 3ß-HSD superfamily. Furthermore, several recent studies using type I and type II purified enzymes have elegantly further characterized structure-function relationships responsible for kinetic differences and coenzyme specificity.

I. Introduction

A. The role of 3ß-hydroxysteroid dehydrogenase activity in steroid formation and degradation

B. Subcellular localization

II. Human Type I and II 3ß-HSD Genes and Pseudogenes

III. Structure-Function Relationships

IV. Evolution of the 3ß-HSD Gene Family

A. The rat 3ß-HSD gene family

B. The mouse 3ß-HSD gene family

C. The hamster 3ß-HSD gene family

D. Phylogeny of the 3ß-HSD gene family

E. Enzymatic characteristics of the 3-KSRs (rat liver-specific type III, mouse types IV and V, and hamster type III)

F. 17ß-HSD activity of rat type I and IV 3ß-HSDs

V. Transcriptional Regulation of Human Type I and II 3ß-HSD

A. Gonadal/adrenal isoenzyme—type II 3ß-HSD

B. Future directions in transcriptional regulation

C. Regulation of placenta/peripheral tissue type I 3ß-HSD

D. Species similarity/divergence in mechanisms

VI. Ontogeny, Localization, and Regulation of 3ß-HSD Expression

A. Adrenal

B. Ovary

C. Testis

D. Placenta

E. Liver

F. Breast

G. Prostate

H. Skin

I. Brain

J. Other expression sites

VII. Molecular Genetics of Human 3ß-HSD Deficiency

A. Clinical features

B. Biological diagnosis

C. Molecular diagnosis

D. Genotype-phenotype relationships

E. Structure-function relationships

F. Sequence variants in the HSD3B2 gene vs. nonclassical 3ß-HSD deficiency

	I. Introduction

Top Abstract I. Introduction II. Human Type I... III. Structure-Function... IV. Evolution of the... V. Transcriptional Regulation of... VI. Ontogeny, Localization, and... VII. Molecular Genetics of... References

 
STEROID HORMONES PLAY a crucial role in the differentiation, development, growth, and physiological function of most vertebrate tissues. The major pathways of steroid hormone synthesis are well established, and the sequence of the responsible steroidogenic enzymes has been elucidated (Refs.1, 2, 3, 4, 5, 6, 7 and references therein) (Fig. 1). For example, in the human, after the conversion of cholesterol to pregnenolone (PREG) by the mitochondrial side-chain cleavage system, the adrenal cortex may direct PREG toward one of three different pathways. First, PREG may remain as a C21,17-deoxysteroid and proceed down the pathway to produce the mineralocorticoid, aldosterone. Second, it may undergo 17-hydroxylation and proceed down the C21,17-hydroxy pathway to form the principal glucocorticoid, cortisol. The third option is that, after 17-hydroxylation, it may undergo cleavage of the C17–20 bond to become a C19–17-ketosteroid, leading to the formation of androgens and estrogens. As can be seen in Fig. 1, whichever pathway is followed, the subsequent formation of all classes of steroid hormones relies upon the action of the enzyme 3ß-hydroxysteroid dehydrogenase/⁵-⁴-isomerase (3ß-HSD) (8, 9, 10).

View larger version (35K): [in this window] [in a new window]   FIG. 1. Schematic representation of the major mammalian steroidogenic pathways. All P450s are cytochrome enzymes. P450c18, This enzyme mediates 11ß-hydroxylation and subsequent reactions involved in the biosynthesis of aldosterone; P450c11, 11ß-hydroxylase; P450aro, P450 aromatase.

A. The role of 3ß-hydroxysteroid dehydrogenase activity in steroid formation and degradation
The nicotinamide adenine dinucleotide (NAD)⁺-dependent membrane-bound enzyme 3ß-HSD, was first described in 1951 by Samuels et al. (17). It is located in the endoplasmic reticulum (ER) and mitochondria (18, 19, 20, 21, 22, 23), and it catalyzes the sequential 3ß-hydroxysteroid dehydrogenation and ⁵ to ⁴-isomerization of the ⁵-steroid precursors PREG, 17-hydroxypregnenolone (17OH-PREG), DHEA, and androst-5-ene-3ß,17ß-diol (⁵-DIOL) into their respective ⁴-ketosteroids, namely progesterone (PROG), 17-hydroxyprogesterone (17OH-PROG), ⁴-DIONE, and T. Therefore, this bifunctional dimeric enzyme is required for the biosynthesis of all classes of steroid hormones, namely glucocorticoids, mineralocorticoids, PROG, androgens, and estrogens (Fig. 1). In addition, enzymes of the 3ß-HSD family also catalyze the formation and/or degradation of 5-androstanes and 5-pregnanes, such as dihydrotestosterone (DHT) and dihydroprogesterone (DHP) (8, 9, 10). The 3ß-HSD isoenzyme therefore controls critical steroidogenic reactions in the adrenal cortex, gonads, placenta, and a variety of peripheral target tissues (24).

Transient expression of human 3ß-HSD isoenzymes provided the first direct evidence that the 3ß-HSD and ⁵-⁴-isomerase activities reside within a single protein (25, 26, 27). However, data obtained from affinity alkylation (28) and inhibition experiments (29) that suggested separate 3ß-HSD and isomerase sites are also consistent with a bifunctional catalytic site adopting a different conformation for each activity, as suggested by tryptic peptides associated with both catalytic activities localized using affinity radiolabeled steroids (30, 31, 32). Additional studies have supported the hypothesis that reduced NAD (NADH), the coenzyme product of the rate-limiting 3ß-HSD reaction, induces a conformational change around the bound 3-oxo-⁵-steroid (the 3ß-HSD product and the isomerase substrate) to activate the isomerase step (33). Finally, as revealed by site-directed mutagenesis of the human type I (placental) enzyme, His²⁶¹ appears to be a critical amino acid residue for 3ß-HSD activity, whereas Tyr²⁵³ or Tyr²⁵⁴ participate in the isomerase activity (23).

B. Subcellular localization
Many of the enzymes of the steroidogenic pathway are localized to the smooth ER with the notable exceptions of P450scc (P450 cholesterol side-chain cleavage; CYP11A1), P450c11 (CYP11B1), and aldosterone synthase (CYP11B2). 3ß-HSD subcellular localization patterns are unique in that they show various degrees of ER and mitochondrial distribution. The relevance of dual localization is unclear, yet it can be hypothesized that substrate accessibility could be limited with higher degrees of mitochondrial expression due to reduced mitochondrial transport. This would be analogous to the inability of high-efficiency catalysis of cholesterol by P450scc in the absence of the protein controlling cholesterol shuttling, steroidogenic acute regulatory protein (StAR) (34). Smooth ER localized 3ß-HSD presumably would have a greater access to cytosolic steroid precursors, such as DHEA and ⁵-DIOL.

3ß-HSD activity was detected using histochemical techniques as early as 1965 (35, 36). Similar techniques isolated this activity to the smooth ER and mitochondrial cristae (37). Its membrane localization was not known until studies localized 3ß-HSD to the microsomal fraction of human adrenal (38) and chorion/amnion fetal membranes (39), suggesting that 3ß-HSD is a membrane-associated enzyme. With the development of antibodies against 3ß-HSD, the resolution of its localization increased, and it was verified that it is associated with the ER and mitochondria in human placenta (18, 40, 41), bovine adrenocortical cells (20), and rat adrenal tissue (21).

Submitochondrial fractionation studies show that bovine adrenal 3ß-HSD is associated with the inner membrane and with a particulate fraction characterized by contact sites between the two membranes. 3ß-HSD activity was higher in this fraction than in the inner mitochondrial membrane, suggesting that intermembrane contact sites may facilitate both the access of cholesterol to the inner membrane where P450scc is localized and the rapid conversion of PREG to PROG by 3ß-HSD (19). Elegant biochemical studies have confirmed that a significant amount of adrenocortical 3ß-HSD is present in the inner mitochondrial membrane (42). Coprecipitation studies have shown that 3ß-HSD is in a functional steroidogenic complex with P450scc in the inner mitochondrial membrane (43), which provides the enzyme with immediate substrate metabolized from cholesterol transported across the mitochondrial membrane. Other work has shown that subcellular distribution in bovine and murine adrenal tissues demonstrated a higher degree of microsomal to mitochondrial localization (44, 45). A similar subcellular distribution was also recently reported in rat ovary, as revealed by immunoelectron microscopic localization, whereas in the testis, the 3ß-HSD was restricted to the mitochondria (45).

Although the functional significance of differential 3ß-HSD subcellular localization is unknown, studies have been performed to determine whether the dynamics of 3ß-HSD subcellular localization can be altered by regulation. Because Ca²⁺ flux mediates K⁺ and A-II increases in aldosterone production by zona glomerulosa (ZG) (46), it is possible that Ca²⁺ could affect the mitochondrial to ER ratio of 3ß-HSD. However, bovine ZG cells showed that neither Ca²⁺ nor A-II had any effect on the subcellular distribution of 3ß-HSD and P450scc, but it did affect StAR localization (22). Another report showed that microsomal 3ß-HSD activity in the ovary was unchanged during mouse estrous, yet mitochondrial 3ß-HSD activity increased and doubled during diestrous in the mouse (47). These results suggest that 3ß-HSD activity could be preferentially distributed to the mitochondria under certain physiological conditions, but this may not be a general phenomenon.

	II. Human Type I and II 3ß-HSD Genes and Pseudogenes

Top Abstract I. Introduction II. Human Type I... III. Structure-Function... IV. Evolution of the... V. Transcriptional Regulation of... VI. Ontogeny, Localization, and... VII. Molecular Genetics of... References

 
During the past decade, the structure of the isoenzymes of the 3ß-HSD family has been characterized in the human and several other vertebrate species (Fig. 2). Human type I 3ß-HSD cDNA was isolated and characterized by Luu-The et al. (18, 48, 49) after purification of the 3ß-HSD enzyme from human placenta, and this sequence was later confirmed by other workers (25, 50). The second human 3ß-HSD isoenzyme, chronologically designated as type II, was isolated from a human adrenal cDNA library (27). The type I 3ß-HSD gene (HSD3B1) encodes an enzyme of 372 amino acids predominantly expressed in the placenta and peripheral tissues, such as the skin (principally in sebaceous glands), mammary gland, prostate, and several other normal and tumor tissues (27, 51, 52, 53, 54). The purified enzyme has a Michaelis constant (K_m) of 3.7 µM and maximal velocity (V_max) of 43 nmol/min·mg for 3ß-HSD substrate (DHEA) and a K_m of 28 µM and V_max of 598 nmol/min·mg for the isomerase substrate (5-androstene-3,17-dione) (54). In comparison, the type II gene (HSD3B2), which encodes a protein of 371 amino acids, shares 93.5% identity with the type I and is almost exclusively expressed in the adrenals, the ovary, and testis (27, 54, 55). The purified enzyme has a K_m of 47 µM and V_max of 82 nmol/min·mg for 3ß-HSD substrate (DHEA) and a K_m of 88 µM and V_max of 970 nmol/min·mg for the isomerase substrate (5-androstene-3,17-dione). The higher affinity of type I 3ß-HSD could facilitate steroid formation from relatively low concentrations of substrates usually present in peripheral tissues. Based on their differential tissue-specific expression pattern, it is not surprising that classical 3ß-HSD deficiency, which will be discussed further in Section VII, results from mutations in the HSD3B2 gene, whereas the HSD3B1 gene is normal in affected individuals (56, 57, 58, 59).

View larger version (45K): [in this window] [in a new window]   FIG. 2. Comparison of the amino acid sequences of members of the 3ß-HSD gene family: human types I and II; macaque; bovine; rat types I, II, III, and IV; mouse types I, II, III, IV, V, and VI; hamster types I, II, and III; horse; pig; chicken; rainbow trout; catfish; and eel types I and II. Residues common to the human type I 3ß-HSD are represented by a dot. The members of the mammalian 3ß-HSD family have been chronologically designated according to their order of elucidation in each species. The numbers indicated above refer to the human type II sequence. The missense mutations associated with 3ß-HSD deficiency are shown by an arrow indicating their position in the human type II sequence. [Adapted from Ref.576 .]

View larger version (17K): [in this window] [in a new window]   FIG. 3. Chromosomal localization showing the two expressed genes HSD3B1 and HSD3B2, and five pseudogenes, 1–5. The orientation of four genes is shown by the arrow that points toward the stop codon or its homolog. Clones of yeast artificial chromosomes (alphanumeric identification) are shown as a contig. The information regarding the order of the markers was obtained from the Whitehead Institute/MIT Center for Genome Research, Cambridge, Massachusetts (http://www-genome.wi.mit.edu/). Structure of human type I and type II 3ß-HSD genes, mRNA species, and the corresponding proteins. Exons are represented by boxes in which hatched lines demarcate the coding regions, whereas open boxes represent the noncoding regions. Introns are represented by black bold lines. [Adapted from Ref.576 .]

	III. Structure-Function Relationships

Top Abstract I. Introduction II. Human Type I... III. Structure-Function... IV. Evolution of the... V. Transcriptional Regulation of... VI. Ontogeny, Localization, and... VII. Molecular Genetics of... References

The 3ß-HSD isoenzymes belong to the short-chain alcohol dehydrogenase superfamily, mainly determined by the nucleotide-binding site sequence located at the amino terminus. It consists of a ß-strand, -helix, ß-strand in a fold that provides a hydrophobic pocket for the AMP part of the nucleotide factor. The turn between the first ß-strand and the -helix is a glycine-rich segment, Gly-X-X-Gly-X-X-Gly, similar to the common Rossmann fold sequence Gly-X-Gly-X-X-Gly conserved among most NAD(H)-binding enzymes (68). This well-conserved glycine-rich fragment forms a hydrophobic pocket that allows close association of the AMP part of the cofactor. A preliminary study of rat type III enzyme has targeted Asp (36) as the amino acid that may be responsible for the strict NAD⁺ specificity of the enzyme (8). More recent mutagenesis studies in human type I enzyme demonstrate that the D36A/K37R mutant shifts cofactor preferences of both 3ß-HSD and isomerase activities from NAD(H) to NADP(H), thus showing that the two activities utilize a common coenzyme domain (69).

Affinity labeling of purified human type I identified two tryptic peptides, comprising amino acids Asn¹⁷⁶ to Arg¹⁸⁶ and Gly²⁵¹ to Lys²⁷⁴ that contain residues involved in the putative substrate-binding domain (30). These studies have shown that the Gly²⁵¹ to Lys²⁷⁴ peptide was associated with the site of isomerase activity, whereas Tyr²⁵³ appears to function as the general proton donor in the isomerase reaction (24). His²⁶¹ also appears to be a critical residue for the 3ß-HSD activity (23). Additional kinetic analyses of D257L and D258L mutants suggest that this region is part of the isomerase substrate domain (69).

In contrast to other short chain dehydrogenases with a single catalytic Y-X-X-X-K motif (5, 70, 71), there are two potential catalytic motifs (Y¹⁵⁴-X-X-X-K¹⁵⁸ and Y²⁶⁹-X-X-X-K²⁷³) in the primary structure of all 3ß-HSDs. Human type I and type II only differ at position 156 in this motif, type I having a tyrosine whereas type II has a histidine residue. The H156Y mutant form of the type I enzyme shifts the substrate kinetics for DHEA and PREG to the same K_m and V_max values exhibited by the type II enzyme; thus, H¹⁵⁶ in the type I vs. Y¹⁵⁶ in type II 3ß-HSD accounts for the substantially higher affinity of the type I 3ß-HSD activity for these substrates and inhibitor epostane relative to the type II enzyme (72).

Two membrane-binding domains lying between residues 72 and 89 in the NH₂-terminal region and between residues 283 and 310 in the COOH-terminal region were identified. Indeed, deletion of the 283–310 region causes the enzyme to localize in the cytosol without affecting its activities (73). The region is therefore a critical membrane domain of 3ß-HSD that can be deleted without compromising enzyme function (54, 73). Deletion of residues 72–89 in the NH₂-terminal region produces a mutant protein that is distributed among the microsomes, mitochondria, and cytosol (73). Because 28% of the 3ß-HSD and isomerase activities remain in the membranes of microsomes and mitochondria, the presence of the 283–310 domain in this mutant allows the protein to retain significant hydrophobicity. However, a majority (72%) of the protein is shifted into the cytosol, so the 72–89 region does contribute to membrane association. The 8-fold loss of both 3ß-HSD and isomerase activity that results from the 72–89 deletion underscores the importance of this region to enzyme function (73). The data obtained by Thomas’ group (73) with the human type I enzyme are consistent with one of our previous studies in which the increased polarity of the domain between residues 75 and 91 in the rat type II 3ß-HSD/isomerase was responsible for its having much lower activity than the rat type I enzyme (74). Thus, the presence of this highly conserved hydrophobic domain may be crucial to activity in the entire 3ß-HSD gene family. The expression of an active soluble 283–310 deletion mutant of the type I enzyme in a baculovirus expression system provides a valuable tool for crystallographic studies that may ultimately determine the tertiary/quaternary structure of the enzyme (73). A three-dimensional ribbon model has been constructed by Thomas’ group (69), based on the homology data for human type I 3ß-HSD and UDP-galactose-4-epimerase. This also represents a useful tool for interpreting biochemical data and designing inhibitors (Fig. 4) (69).

View larger version (67K): [in this window] [in a new window]   FIG. 4. Ribbon structure of human type I 3ß-HSD/isomerase based on homology modeling with key amino acids identified. The primary sequences of 3ß-HSD/isomerase (green) and UDP-galactose-4-epimerase (yellow) were aligned using ClustalX. The NAD and DHEA structures are included. The key Asp36 residue is shown hydrogen binding (gray dotted lines) to the 2', 3'-hydroxyl groups of the adenosyl ribose group of NAD. The catalytic Tyr154 and Lys158 residues for human type I 3ß-HSD activity, the catalytic Tyr253 and Asp257 residues for isomerase activity, and the Asp241 residue that bridges the upper isomerase domain with the lower coenzyme domain are also shown. This ribbon model represents the 3ß-HSD/isomerase structure in the 3ß-HSD conformation. The oxygen atoms are red, nitrogen atoms are blue, carbon atoms are gray, and phosphorus atoms are pink. [Reproduced from J.L. Thomas et al.: J Biol Chem 278:35483–35490, 2003 (69 ), copyright 2003, with permission from The American Society for Biochemistry and Molecular Biology.]

	IV. Evolution of the 3ß-HSD Gene Family

Top Abstract I. Introduction II. Human Type I... III. Structure-Function... IV. Evolution of the... V. Transcriptional Regulation of... VI. Ontogeny, Localization, and... VII. Molecular Genetics of... References

View larger version (24K): [in this window] [in a new window]   FIG. 5. Enzymatic 17ß-HSD and 3ß-HSD activities of rat types I and IV 3ß-HSD expressed in intact cells. Reaction 1 corresponds to the androgenic 17ß-HSD activity measured in cells expressing the rat type I and IV 3ß-HSD isoenzymes. Reaction 2 corresponds to the 3ß-HSD activity present in cells expressing rat type I or IV 3ß-HSD enzymes. Reaction 3 corresponds to 3-HSD present in some cells. Reaction 4 corresponds to the 3-KSR activity present in liver cells expressing the rat 3-KRS (type III) enzyme. The hatched arrows indicate predominant reactions expected to use primarily NAD+ as cofactor, whereas the black arrows indicate the predominant reactions expected to use primarily NADPH as cofactor. [Adapted from Ref.84 .]

It is hypothesized that mouse 3ß-HSD type VI cDNA is orthologous to human 3ß-HSD type I cDNA, which has been shown to be the only isoenzyme expressed in the placenta and the skin. The demonstration that the type VI isoenzyme in the mouse functions as a 3ß-HSD and is the predominant isoenzyme expressed during the first half of pregnancy in uterine tissue and embryonic cells, suggests that this isoenzyme may be involved in the local production of PROG, which is required for the successful implantation of the blastocyst and/or maintenance of pregnancy (92).

C. The hamster 3ß-HSD gene family
The hamster is a rodent species, but in contrast to the rat and mouse, in which the principal corticosteroid is corticosterone, the principal corticosteroid in the hamster is cortisol. A study on the regulation of adrenal steroidogenic enzymes suggested that the hamster could be a good model for studying human steroidogenesis. With this in mind, three isoenzymes of 3ß-HSD were characterized in the hamster (96, 97). The type I isoenzyme was isolated from an adrenal cDNA library and was identified as being a low K_m 3ß-HSD (K_m: PREG, 5.5 µM; DHEA, 2.4 µM). A separate isozyme, designated type II was isolated from the kidney and was also found to be a low K_m 3ß-HSD (K_m: PREG, 8.8 µM; DHEA, 2.9 µM). Two cDNAs were isolated from the liver, one which was identical to the type II sequence found in the kidney, and a distinct cDNA encoding an isoform designated as type III, which does not possess any steroid dehydrogenase activity but functions as a 3-KSR. There is sexual dimorphic expression of this liver-specific type III 3ß-HSD in the hamster, as seen for the rat liver-specific type III KSR. As is the case for both the rat and mouse, a high affinity 3ß-HSD is expressed in the adrenal and gonad of the hamster, consistent with the steroidogenic role of these tissues (96).

D. Phylogeny of the 3ß-HSD gene family
McBride et al. (66) indicated no evidence for the presence of other members of the human 3ß-HSD family within the physical contig of 0.5 Mb by Southern blot analysis, thus suggesting that in humans there is no comparable liver-specific 3-KSR sharing a high percentage of identity with other members of the HSD3B cluster. Such a conclusion is also well supported by phylogenetic analysis of the mammalian 3ß-HSD gene family. Unexpectedly, the phylogenetic tree strongly suggests that independent gene duplications occurred in different species (66, 91), (V. Laudet, personal communication). As illustrated in Fig. 6, our recent analysis shows a first complex of three genes from primates and suggests that an ancestral gene duplicated specifically in the primate lineage to give rise to human types I and II, whereas the macaque gene is the homolog of human type II. It is very likely that an ortholog of the human type I exists in the macaque genome, but yet remains to be identified. The second complex clusters together the single 3ß-HSD species characterized in bovine, pig, and horse. The third complex clusters together three clear classes of rodent 3ß-HSD genes; firstly, the rat type I, II, and IV as well as the mouse type I, II, III, and VI; secondly, the mouse type IV and V and rat type III, the specific 3-KSRs; and thirdly the hamster type I, II, and III. Because the hamster type III is a liver-specific 3-KSR (97), it is surprising that it is not included in the second class of rodent genes. These findings strongly suggest that the 3ß-HSD genes were independently duplicated or triplicated three times in the lineage of the rat, the mouse, and the hamster. It is difficult to understand why the duplication failed to occur earlier in mammalian evolution if there are physiological needs and/or advantages for the presence of multiple isoenzymes. These data may indicate that the need for different 3ß-HSD genes occurred very late in mammals, with subsequent evolution in a similar manner in other lineages.

View larger version (17K): [in this window] [in a new window]   FIG. 6. Unrooted phylogenetic tree constructed by the Neighbor-Joining method using 1000 bootstrap replicates. Multiple nucleotide alignments of 3ß-HSDs from different species were obtained using PILEUP (Wisconsin GCG package), and the phylogenetic analysis was performed by PAUPsearch, which provides a GCG interface to the tree-searching options in the PAUP program, version 4.0.0d55 (Phylogenetic Analysis Using Parsimony) (590 ). [Adapted from Ref.576 .]

View larger version (27K): [in this window] [in a new window]   FIG. 7. Major bile acid biosynthesis pathways. Two major bile acid biosynthesis pathways are shown. Only major enzymes and intermediates are shown.

View larger version (52K): [in this window] [in a new window]   FIG. 8. Top, Schematic representation of HSD3B7 gene, mRNA, and corresponding protein. Exons are represented by hatched boxes indicating coding region, whereas open boxes represent noncoding regions. Asterisk represents an alternative noncoding exon. Introns are represented by black bold lines. Mutations causing progressive intrahepatic cholestasis are identified on the gene. The nucleotide numbers indicating the positions of individual mutations refer to C27 3ß-HSD cDNA (GenBank accession no. AF277719). Bottom, Alignment of amino acid sequences of human type I and type VII 3ß-HSD. Residues common to both types are identified by black areas, whereas similar residues are identified by gray areas. Positions of mutations causing progressive intrahepatic cholestasis are identified by an arrow in reference to amino acid sequences on the primary structure. [Adapted from Ref.591 .]

Examination of the 3ß-HSD isoenzymes shows a typical ßß dinucleotide-binding fold with Asp (36) located in the position predicted for the acidic residue that participates in hydrogen bond formation with the 2'-hydroxyl moiety of the adenosine ribose of all known NAD-dependent dehydrogenases (Fig. 2). Using site-directed mutagenesis, it has been shown that the presence of a Tyr residue instead of an Asp residue at position 36 in the typical ßß dinucleotide-binding fold of the cofactor binding domain of rat type III is responsible for the difference in cofactor specificity of the rat 3-KSR (type III) protein, but this alteration is not sufficient to explain the low activity of the enzyme with ⁵-3ß-hydroxysteroid substrates (103). The physiological importance of this peculiar member of the rat 3ß-HSD family is well supported by the finding that mouse types IV and V and hamster type III also possess this specific 3-KSR activity (88, 91, 96). The 3-KSR activity of these three latter enzymes is most likely due to the presence of Phe (36) in place of Arg (36), as suggested by the data resulting from the study on Asp (36) in the rat type III sequence.

F. 17ß-HSD activity of rat type I and IV 3ß-HSDs
Using cell homogenate preparations, it was first noted that the rat type I 3ß-HSD protein has androgenic 17ß-HSD-like activity. It was found that the affinity for the DHT substrate by the enzyme is similar to that of the substrates for 3ß-HSD activity, although with a much lower velocity (86). However, using intact cells transfected with the rat type I, it was found that the enzyme catalyzed almost exclusively the conversion of DHT into androstanedione via the 17ß-HSD-like activity. The intrinsic 17ß-HSD activity of rat types I and IV 3ß-HSD is specific to 5-androstane-17ß-ol steroids, therefore suggesting that it plays a key role in controlling the bioavailability of the active androgen DHT.

The predominance of this secondary 17ß-HSD activity over the primary 3ß-HSD activity in intact cells most likely results from the high bioavailability of NAD⁺ in mammalian cells relative to the low levels of the intracellular pool of NADH. The apparent discrepancy between the oxidoreductase activity of the rat type I and IV proteins, as revealed by the interconversion of DHT and 3ß-DIOL using homogenate preparations from mammalian cells expressing the recombinant isoenzymes, and the lack of significant 3-KSR activity, measured in intact cells in culture expressing the same recombinant enzymes, can also be explained by the low levels of the intracellular pool of NADH relative to the high bioavailability of NAD⁺ in mammalian cells. In agreement with this explanation, in cultured HeLa, JEG-3, and SW-13 cells expressing rat types I and IV 3ß-HSDs, ⁵-hydroxysteroid precursors are efficiently converted into their corresponding ⁴-3-keto steroids, which is well known to require NAD⁺ as the allosteric cofactor. The highly efficient conversion of DHT into 3ß-DIOL in cultured HeLa cells expressing rat 3-KSR (type III) also argues in favor of the bioavailability of NADPH within the cells (84, 86). It is also of interest that 3-HSD is known to use NADPH as a cofactor, which is in agreement with the predominant metabolic pathways using NAD⁺ or NADPH as cofactors postulated to be present in transfected cells in culture. These results also emphasize the important fact that assessment of enzyme activity should always be performed in intact cells to more closely resemble the situation in vivo, an issue that will be discussed in more detail in Section VII.

In relation to an enzyme having dual activity, such secondary activity could be explained by the binding of the steroid in the inverted substrate orientation, in this case from the C-17 rather than C-3 extremity, to the same active site responsible for the primary activity of the enzyme. The physiological relevance of this secondary 17ß-HSD activity is also supported by the observation that the purified bovine adrenal 3ß-HSD enzyme (20) also possesses the 17ß-HSD-like activity (84, 86). Moreover, it was recently shown that mouse types I and VI 3ß-HSD isoforms display significant 17ß-HSD-like activity (617).

Another enzyme known to catalyze the oxidation of both 3ß- and 17ß-hydroxy groups of certain hydroxysteroids is the NAD⁺-dependent 3ß-17ß-hydroxysteroid dehydrogenase from Pseudomonas testosteroni (104). The elucidation of the cDNA sequence of this enzyme revealed that it is a member of the short-chain dehydrogenase/reductase superfamily, however, sharing more homology with the 17ß-HSD than with the 3ß-HSD enzymes (105). It is also known that some members of the 17ß-HSD family, such as human type 2 (106) and human type 7 (107) 17ß-HSDs, possess dual 3ß/17ß activity.

	V. Transcriptional Regulation of Human Type I and II 3ß-HSD

Top Abstract I. Introduction II. Human Type I... III. Structure-Function... IV. Evolution of the... V. Transcriptional Regulation of... VI. Ontogeny, Localization, and... VII. Molecular Genetics of... References

 
The human HSD3B2 gene is the major form expressed in the adrenal cortex, ovary, and testicular Leydig cells. It is most homologous to the type I gene expressed in mice, rats, and other species. The different gene isoforms are so named because of their chronological isolation. The current state of knowledge concerning the transcriptional control human HSD3B2 gene will be discussed first.

A. Gonadal/adrenal isoenzyme—type II 3ß-HSD
1. Steroidogenic factor-1 (SF-1).
Initial studies investigating the transcriptional regulation of the human HSD3B2 gene primarily focused on the trophic hormones known to regulate expression of other genes, including ACTH in the adrenal cortex, LH/human chorionic gonadotropin (hCG) in theca cells and corpus luteum (CL), and LH in testis Leydig cells. Until fairly recently, data regarding the transcriptional regulation of 3ß-HSD was highly limited. After the isolation and sequencing of genomic clones for human type II 3ß-HSD, several studies analyzed the promoter and 5'-flanking regions looking for regulatory elements essential for expression and regulation in steroidogenic cells. Initial examinations of the sequence were not very revealing. Because cAMP was a known intracellular mediator of trophic hormone stimulation of 3ß-HSD expression, it was interesting that no identifiable cAMP response elements were observed in the proximal 1.3 kb of the 5'-flanking sequence. There were two putative activator protein-1 (AP-1) elements at –576 and –977 that matched functional AP-1 elements in other genes (e.g., the sequence at –977 is an exact match of an AP-1 element in the simian virus 40 promoter). This limited information formed the basis of the initial functional studies of the 3ß-HSD promoter by transfection into steroidogenic cells.

The H295R human adrenocortical carcinoma cell line was chosen as a steroidogenic cell model due to its relatively high level of differentiated functions, including responsiveness to adrenal trophic hormones such as ACTH and synthesis of cortisol. Interestingly, these cells have characteristics of all three zones of the normal adrenal gland, the zona reticularis (ZR), zona fasciculata (ZF), and ZG (108). As such, these cells may represent a pluripotent adrenal cortical stem cell that transformed before terminal differentiation. The H295R cells can be cultured under different specific conditions that promote the differentiated characteristics of cells from the individual zones (108). In these studies, the cells were treated with activators of the protein kinase A (PKA) pathway, such as forskolin and dibutyryl cAMP, to increase the expression of enzymes P450 17-hydroxylase/17,20-lyase (P450c17), 3ß-HSD, and P450c11 as well as cortisol production, thereby promoting the characteristics of ZF cells. Promoter-reporter constructs were prepared that used a series of 5'-deletions of the human HSD3B2 5'-flanking sequence and promoter, fused to the chloramphenicol acetyltransferase (CAT) gene. Transfections were performed, and the cells were treated in the presence and absence of phorbol ester stimulation. The surprising finding of these initial studies was that the putative AP-1 elements were nonfunctional with respect to either cAMP or phorbol ester stimulation. Both putative AP-1 elements could be removed without affecting the cAMP- or phorbol ester-stimulated promoter activity, with 5'-deletion down to –100 bp. The promoter-reporter constructs did not lose responsiveness to cAMP or phorbol ester stimulation until a further deletion was made to –52 (109).

The sequence between –100 and –52 contained an element at –64 to –56 that is an 8/9 match for the consensus regulatory element that binds the orphan nuclear receptor SF-1, also referred to as adrenal 4 binding protein. New nomenclature for members of the nuclear receptor superfamily designates this family as NR5A and SF-1 as NR5A1a (110). The essential nature of the SF-1 element was tested further using two approaches. First, promoter-reporter constructs were transfected into a nonsteroidogenic cell type that does not express SF-1 (HeLa cells) in the presence and absence of cotransfection with an SF-1 expression vector. Second, the importance of SF-1 was tested with the use of a 2-bp point mutation in the element (from TCAAGGTAA to TCAATTTAA), whereas all other sequences in the –100-bp HSD3B2 gene promoter remained as in the wild type. The 2-bp point mutation in the SF-1 element dramatically abrogated the cAMP and phorbol ester response of the promoter (109). These studies indicated that SF-1 was essential for cAMP or phorbol ester-stimulated steroidogenesis and suggested that trophic hormone stimulation of type II 3ß-HSD expression involved SF-1 activation.

The mechanisms by which cAMP and phorbol esters stimulate SF-1-mediated transcription of the HSD3B2 gene are not yet clear. In the GnRH promoter, another SF-1 responsive gene, the essential response of SF-1 involves an interaction between SF-1 and another transcription factor, early growth response protein (111, 112). However, the HSD3B2 promoter differs from the GnRH promoter in that it lacks proximal early growth response recognition sequences. However, it is possible that the interaction of SF-1 with another yet undescribed transcription factor is a necessary component to the cAMP/phorbol ester stimulation.

2. Stat5.
The involvement of transcription factors other than SF-1 in the control of the human HSD3B2 gene has become evident following the discovery of a regulatory element, which interacts with Stat5. The Stat family of proteins is named for the acronym (signal transducers and activators of transcription). The Stats comprise a family of cytoplasmic transcription factors that are activated by tyrosine kinases followed by nuclear translocation and binding to specific regulatory elements (reviewed in Ref.113). The Stats are activated by a number of extracellular protein ligands including cytokines, growth factors, and prolactin (PRL)/GH. The activation involves either tyrosine kinase activity that is intrinsic to the receptors for these ligands or tyrosine kinases that associate with the receptors such as the Janus kinases. Seven different Stat genes have been identified (Stats 1, 2, 3, 4, 5a, 5b, and 6). Targeted gene disruptions have been performed for each of these seven Stat genes in mice, resulting in different phenotypes, which include immune deficiencies in Stat 1 knockouts, embryonic lethals in knockouts of Stats 2 and 3, and deficient breast development and lactation in Stat5 knockouts. Interestingly, the Stat5a Stat5b double knockout displays luteal failure (114), which involves one of the key tissues expressing 3ß-HSD.

Stat5 was independently identified and cloned in studies concerning a transcription factor important for PRL activation of the ß-casein promoter in mammary epithelial cells (115, 116). The potential involvement of PRL and Stat5 in the human 3ß-HSD promoter is an intriguing concept. Prolactin and placental lactogens are known to be important luteotropic hormones in many species including bovine, porcine, and rodent (117). The human placenta expresses human placental lactogen, which was previously referred to as human chorionic sommatomammotropin. The involvement of PRL as a potential luteotropic hormone in humans has not been examined in detail, although conditions of hypo- and hyperprolactinemia are known indicators of female infertility (118, 119).

Two independent findings suggested the possibility that Stat5 could be mediating regulation of the 3ß-HSD promoter. First, analysis by Western blot showed that Stat5 was induced on d 5 (after hCG) in pseudopregnant rat ovaries approximately 10-fold over controls (120). This was a specific effect because Stat3 levels remained unchanged. These data suggested that Stat5 was being specifically induced during luteinization and that it could be involved in mediating the regulation of luteal function. Because this represents an important site of Stat5-mediated activity, it suggested a role for Stat5 in luteal function. Stat5 is expressed in a wide variety of tissues, and its expression is fairly constitutive. However, it has been shown to be induced in mammary epithelial cells during differentiation of the mammary gland into the active lactation state (121). A second finding was the discovery of a 9/9 match for a Stat5 consensus regulatory element in the human HSD3B2 gene promoter during the examination of SF-1 action. The sequence TTCTGAGAA at –118 to –110 is a 9/9 match to the consensus regulatory element TTCNNNGAA for Stat5 (Fig. 9). These findings suggested a potential regulatory role for Stat5.

View larger version (42K): [in this window] [in a new window]   FIG. 9. Schematic representation of type I and II 3ß-HSD promoter regions (A) and alignment of the promoter sequences (B). [Reproduced from S. Gingras et al.: J Steroid Biochem Mol Biol 76:213–225, 2001 (196 ), copyright 2001, with permission from Elsevier.]

The functional Stat5 element in the HSD3B2 promoter may also integrate signaling from other ligand-stimulated pathways. For example, angiotensin II has been shown to stimulate Stat5 in cardiac myocytes (123). A similar signaling pathway in adrenal ZG cells would allow angiotensin II stimulation of 3ß-HSD expression in the synthesis of mineralocorticoids. Another ligand that could utilize this pathway is EGF. Stat5 has been shown to be the major Stat protein activated in EGF stimulation of the mouse liver (124). Recent studies have demonstrated that EGF stimulates cortisol synthesis in H295 adrenocortical cells as well as stimulating 3ß-HSD mRNA levels and promoter activity (125). Furthermore, Stat5 and a functional Stat5 regulatory element in the HSD3B2 promoter are required for this stimulation (Fig. 10) (125).

View larger version (14K): [in this window] [in a new window]   FIG. 10. EGF activates type II 3ß-HSD reporter activity through a Stat5-dependent mechanism. A, Increasing concentrations of EGF result in increased type II 3ß-HSD reporter activity. HeLa cells were cotransfected with a –301+45 fragment of the type II 3ß-HSD promotor fused to a CAT reporter gene (–301 CAT; 5 µg), ovine Stat5 (5 µg), ß-galactosidase (0.5 µg), and control DNA for a total of 15.5 µg using the calcium phosphate precipitation method, followed by treatment for 24 h with increasing doses of EGF (0100 ng/ml). B, hStat5A and hStat5B isoforms in activation of 3ß-HSD reporter activity. HeLa cells were cotransfected with a –301+45 fragment of the type II 3ß-HSD promotor fused to a CAT reporter gene (–301 CAT; 5 µg), hStat5A, or hStat5B (5 µg), ß-galactosidase (0.5 µg), and control DNA for a total of 15.5 µg using the calcium phosphate precipitation method, followed by treatment for 24 h with EGF (25 ng/ml). The number above the bar is fold activation compared with the identically transfected group minus EGF. C, A Stat5 regulatory element confers EGF responsiveness to the human type II 3ß-HSD promotor region. HeLa cells were cotransfected with –301 (wild-type) or –301 (mutant) CAT reporter contructs (5 µg), ovine Stat5 (5 µg), ß-galactosidase (0.5 µg), and control DNA for a total of 15.5 µg using the calcium phosphate precipitation method, followed by treatment for 24 h with EGF (25 ng/ml). Data represent the mean ± SE of triplicate cultures after correction for transfection efficiency from a representative experiment of two performed. [Reproduced from F.A. Feltus et al.: Endocrinology 144:1847–1853, 2003 (125 ), copyright 2003, The Endocrine Society.]

3. DAX-1.
DAX-1 (dosage-sensitive sex reversal adrenal hypoplasia congenita critical region on the X chromosome gene 1) was originally isolated by positional cloning from patients presenting with adrenal congenita hypoplasia associated with hypogonadotropic hypogonadism (127, 128). Mutations in DAX-1 have been found to be the cause of adrenal hypoplasia congenita and hypogonadotropic hypogonadism. The possible role of an atypical member of the nuclear hormone receptor superfamily on 3ß-HSD promoter activity has come from studies examining the effects of DAX-1 overexpression on adrenal cell steroidogenesis and steroidogenic enzyme gene expression. The structure of DAX-1 indicates that it lacks the most highly conserved DNA-binding domain typical of members of the nuclear hormone receptor superfamily (127). This has raised questions as to whether this transcription factor binds to DNA. Data have been presented suggesting that DAX-1 has unique mechanisms of DNA binding through putative stem-loop structures (129). An alternative hypothesis is suggested by the sequencing of DAX-1 from another species, which apparently lacks the analogous domain necessary for DNA binding (130). This implies that possible protein-protein interactions between DAX-1 and other transcription factors may be critical to the function of this orphan nuclear receptor. Additional work in this field is needed to determine the mechanisms of DAX-1 actions.

The overexpression of DAX-1 in Y-1 adrenal cells inhibits steroidogenesis (131). Associated with this inhibition are strong inhibitory effects on the expression of mRNAs for StAR, P450scc, and 3ß-HSD (Fig. 11) (131). The exact mechanisms by which DAX-1 overexpression affects 3ß-HSD expression remain unclear. Additional studies of these mechanisms are needed to elucidate the factors involved.

View larger version (39K): [in this window] [in a new window]   FIG. 11. Effect of DAX-1 on the expression of StAR, P450scc, and 3ß-HSD in Y-1 cells. A, Northern blot showing expression of StAR, P450scc, and 3ß-HSD RNA transcripts in Y-1/neo, Y-1/hDAX-1 and Y-1/RIAB cells. Total RNA was extracted from each cell line growing in basal conditions (lanes 1, 4, 7, 10, and 12) or after 6 h (lanes 2, 5, and 8) and 24 h (lanes 3, 6, 9, 11, and 13) forskolin (10 µg/ml) stimulation. RNA was transferred on a nylon membrane and hybridized with StAR, P450scc, 3ß-HSD, and GAPDH probes. B, Western blot showing expression of StAR, P450scc, and 3ß-HSD proteins in Y-1/neo, Y-1/hDAX-1, and Y-1/RIAB cells. Mitochondrial extracts were prepared from cell lines in basal conditions (lanes 1, 3, and 5) and after 16 h forskolin (10 µg/ml) stimulation (lanes 2, 4, and 6). Western blots were sequentially probed with specific antibodies directed to StAR, P450scc, and 3ß-HSD, respectively. [Reproduced from E. Lalli et al.: Endocrinology 139:4237–4243, 1998 (592 ), copyright 1998, The Endocrine Society.]

It has been recently demonstrated that glucocorticoids stimulate type II 3ß-HSD mRNA levels in H295R cells and stimulate HSD3B2 gene promoter activity (125). The mechanisms by which this stimulation occurs are beginning to be characterized and are linked to Stat5. Inactivating point mutations in the Stat5 regulatory element of the HSD3B2 gene promoter abolishes glucocorticoid regulation (125). Additionally, point mutations in Stat5, which convert the critical tyrosine phosphorylation residue to a phenylalanine and abolish Stat5 activation, also abolish its regulation by glucocorticoids. These data indicate that Stat5 is critical to the mechanism of HSD3B2 gene promoter activation by glucocorticoids. The exact mechanisms involved in this action are not yet characterized but could involve protein-protein interactions between glucocorticoid receptor and Stat5, as suggested for the ß-casein promoter in mammary glands (136). These data are intriguing because they point to new ways in which nuclear receptors function in transcriptional activation. They also raise a number of questions concerning the established mechanisms of nuclear receptor action. For example, the domains of the nuclear receptor proteins, which are critical for different aspects of their function, have been mapped out in relation to the traditional DNA binding-dependent mechanism of action. With these new nontraditional mechanisms, which domains of the nuclear receptors are critical to their functions? In addition to structure-function questions, what is the influence of known agonists and antagonists on the efficacy of activation? What is the effect of other nonsteroid factors, which are known to activate other intracellular signaling pathways on steroid-regulated transcription? With these new mechanisms, each of these questions will need to be explored to characterize the essential processes.

5. IL-4.
Immune cell populations in the ovary change during the reproductive cycle and cytokines from these immune cells have been shown to affect steroidogenesis (140). Recent data indicate that IL-4 stimulates 3ß-HSD mRNA levels in primary cultures of human granulosa-lutein cells (141). IL-4 has been shown to primarily activate Stat6 through Stat6 regulatory elements, one of which is present in the human HSD3B2 promoter at –160 to –151 (Fig. 9). However, the activation of Stat protein signaling often involves multiple Stat proteins with some overlap in function (113). Although IL-4 stimulation is associated with Stat6 activation, IL-4 stimulation of HSD3B2 gene promoter activity requires both an intact Stat5 and an intact Stat6 regulatory element (141).

6. GATA proteins.
The GATA proteins are a family of zinc finger transcription factors that bind to GATA regulatory motifs (A/TGATAA/G) in the promoter regions of numerous target genes. Although they were originally identified as crucial regulators of hematopoietic cell differentiation and heart development, the expression of GATA factors is not limited to these two tissues. Interestingly, transcription factors belonging to the GATA family are emerging as novel regulators of steroidogenesis. Indeed, members belonging to this family, namely GATA-4 and GATA-6, are strongly expressed in steroidogenic cells of both the fetal and adult adrenals and gonads (142). In these tissues, several target genes for GATA factors have been identified, such as StAR (143, 144, 145,