Background on Steroidogenesis
Steroid hormones are ubiquitous regulators of a wide variety of physiologic processes. Our laboratory studies the normal and abnormal processes by which human beings make the three classes or steroid hormones: mineralocorticoids, glucocorticoids and sex steroids. Mineralocorticoids such as aldosterone instruct the kidney to retain sodium; without mineralocorticoids blood concentrations of sodium will fall, eventually leading to cardiovascular collapse and death. Glucocorticoids, such as cortisol are named for their action to mobilize carbohydrates and elevate blood sugar, but they exert a wide variety of effects, without which mammals cannot respond to stress. Sex steroids, including estrogen and testosterone responsible for secondary sexual characteristics, some aspects of growth, and are crucial for reproduction. All steroid hormones are made from cholesterol according to the pathway shown in Fig 1. Our laboratory has cloned the cDNAs and genes for many of these enzymes, characterized their developmentally programmed, tissue-specific and hormonally regulated expression, analyzed their transcriptional regulation, and identified many of the common disease-causing mutations in these genes. (Detailed reviews of the molecular biology of steroid hormone synthesis are presented in refs 74,145, 231, 234, 253 and 302)
Current Projects
Our present efforts are largely concerned with six broad areas:
Tropic hormones, such as ACTH and gonadotropins, regulate steroidogenesis both acutely (within minutes) and chronically (hours to days). The chronic regulation of steroidogenesis is primarily at the level of transcriptional regulation of the genes for the various steroidogenic enzymes, especially P450scc, the rate-limiting enzyme (see project 2). Acute regulation is mediated by governing the movement of cholesterol into mitochondria, where P450scc resides on the inner membrane. This steroidogenic acute regulatory protein (StAR) was cloned by Doug Stocco's laboratory from mouse MA-10 cells (JBC 269:28314, 1994), as it appeared to be an important factor in hormonally-regulated cholesterol flux. Our laboratory provided genetic and functional proof that StAR serves this function (148) and we also found that mutations in StAR are responsible for the most severe form of congenital adrenal hyperplasia (CAH) termed lipoid CAH (148, 158, 171, 178, 225, 294), which is a StAR gene knockout of nature (184). Using bacterially-produced wild-type and mutant human StAR, we have characterized the protein's folding using CD and FTIR spectroscopy (194) and have determined that it functions in a so-called 'molten globule' state (209, 235, 237, 280). We have shown that StAR works exclusively on the outer mitochrondrial membrane; as it resides on the OMM only transiently, before being imported and cleaved. StAR is a rare, perhaps unique example of a protein that exerts its biological activity at a cellular location other than that to which it is targeted (249). We have used mass spectrometric analysis of StAR peptides protected from proteolysis by lipid vesicles to show that only the carboxyl-terminal helix interacts with membranes (291). Molecular Dynamics simulations under various conditions, coupled with the loss of activity and cholesterol binding by disulfide mutants, have established that lateral movement of the C-helix is the essential structural change needed for activity (305). The molecular biology of StAR's action has been reviewed (312, 315), and recent studies have shown that StAR's activity as a cholesterol transfer protein is distinct from its action to stimulate steroidogenesis (313). In collaboration with H. S. Bose, a former fellow now an Associate Professor at Mercer University, we have shown that the related protein STARD6 behaves indistinguishably from StAR (321) and that StAR’s activity requires interaction with the peripheral benzodiazepine receptor (293) and with VDAC1 and phosphate carrier protein on the mitochondrial outer membrane (323).
The chronic regulation of steroidogenesis is at the level of transcriptional regulation of the genes for each of the steroidogenic enzymes. These genes are regulated in tissue-specific, developmentally programmed, and hormonally regulated fashions. There are substantial differences in the expression of these genes among various mammals (for example, rodents do not express the gene for P450c17 in their adrenals, while humans and cattle do) so that each gene must be studied separately for each species; furthermore such transfection studies are often not meaningful unless done in cells from the same species (174). We have studied the tissue-specific, developmentally programmed and normally regulated expression of the mRNAs for the human steroidogenic enzymes (for example, refs 45, 50, 58, 70, 92, 100, 128). We have also analyzed the transcriptional regulation and promoter activity of the human gene for P450scc (97, 122, 125, 150) and the human genes for P450c17 (99) and adrenodoxin reductase (118). We identified three cis-acting elements lying within intron 35 of the C4B gene 5 kb upstream from the cap site of the P450c21gene that are required for adrenal expression of P450c21 (216). We have cloned two novel transcription factors related to HIV-inducible LBP proteins that regulate placental transcription of P450scc (217, 292). We have identified novel elements required for adrenal transcription of P450c17, shown that these bind NF-1c, Sp1, and Sp3 (241) and delineated the roles of GATA factors in its tissue-specific transcription (284). Human adrenal transcription of the gene for cytochrome b5 proceeds by a similar strategy (300). Current work concerns the tissue-specific and developmental regulation of P450 oxidoreductase (POR).
Cytochrome P450c17, found in the endoplasmic reticulum, catalyzes both 17α-hydroxylase activity (reactions I and III in Fig 1) and 17,20 lyase activity (reactions II and IV in Fig 1). Both 17α-hydroxylationand scission of the 17,20 carbon-carbon bond occur on the single active site of P450c17. There is only one gene encoding one species of human P450c17 mRNA (49, 60), yet these two activities are regulated separately in distinct developmental-specific and hormonally-induced fashions (109, 130). We have shown that serine/threonine phosphorylation of P450c17 by a cAMP-dependent kinase is required for the enzyme to acquire 17,20 lyase activity (154), thus providing the first functionally important example of post-translational regulation of a steroidogenic enzyme. The most important factor in the differential regulation of the hydroxylase and lyase activities of P450c17 appears to be electron donation by redox partners (130, 173). We have built a computer graphic model of human P450c17 (210) and have shown that mutations in the redox-partner binding site of this enzyme causes isolated 17,20 lyase deficiency while sparing 17α-hydroxylase activity (182, 206). Cytochrome b5, which has been regarded as an alternative electron donor for P450c17, facilitates 17,20 lyase activity allosterically but without electron donation (191). Most recently, we have shown that PP2A is the physiologically relevant phosphatase that counter balances the kinase activity, and that the action of PP2A on P450c17 is regulated by phosphoprotein SET (268). Both cytochrome b5 and P450c17 phosphorylation can elicit a maximal 17,20 lyase reaction (297). Current efforts are directed in two areas. First we are determining which of the 54 Ser and Thr residues in the enzyme are phosphorylated and which ones influence 17,20 lyase activity. Second, we are seeking to identify and, if necessary, clone the responsible kinase. Preliminary data involve a pathway that includes ROCK1 (324).
The CYP21 gene encoding P450c21, the human adrenal steroid 21-hydroxylase, lies in the midst of the class III region of the human major histocompatibility locus on chromosome 6p21.3 (Fig 2). The CYP21 gene and the immediately upstream C4 gene for the fourth component of serum complement are duplicated in the human genome thus: 5' C4A,21A,C4B,21B 3'. The CYP21 gene is of considerable medical genetic interest because disorders of P450c21 cause the common form of congenital adrenal hyperplasia that affects about 1 in 15,000 people. Our initial efforts concerned cloning the bovine gene (44) and human cDNA (57) and characterizing the gene in affected patients. This led to two unexpected findings. First, we found that the predominant mechanism by which the humanCYP21 gene is disrupted is by gene conversion, rather than by point mutation or gene deletion (57, 83, 93, 111); second, we found a previously unidentified gene encoded on the opposite strand of DNA that physically overlapped the CYP21 gene (89). The discovery of this opposite-strand gene led to a thorough characterization of the mechanism by which the C4/CYP21 gene locus was duplicated (117) and to a complete characterization of the large (~80 kb) XB gene on the opposite strand, which we found to encode an extracellular matrix protein that we named Tenascin-X (136, 168). This work also led to the discovery of several other duplicated transcription units in this locus, including the genes we term Y (131) and Z (155), and the discovery of the 'XB-short' gene within Tenascin-X (152). Recent work in collaboration with Dr. James Bristow has shown that mutations in Tenascin-X cause a distinct autosomal recessive form of the Ehlers-Danlos syndrome (179, 243), and we have now identified the key regulatory regions in the principal TN-X promoter (260). Current efforts concern the safety, efficacy and ethics of prenatal treatment of 21-hydroxylase deficiency with dexamethasone (175, 256) and the enzymatic basis of extra-adrenal 21-hydroxylation (334).
Vitamin D is a seco-steroid hormone that can be synthesized in skin; or ingested in the diet. Vitamin D requires bio-activation by sequential 25-hydroxylation in the liver and 1α-hydroxylation by the kidney to produce 1,25(OH)2D. These activities are catalyzed by separate mitochondrial P450 enzymes. The 1 α-hydroxylation is the rate-limiting, hormonally regulated step, but the very low abundance of this enzyme precluded its cloning until recently. Human skin keratinocytes acquire robust 1α-hydroxylase activity when grown in serum-free low-calcium medium. In collaboration with Dr. A. A. Portale, we cloned human 1 α-hydroxylase (P450c1α) cDNA from keratinocytes, proved it produced 1,25(OH)2D by GC/MS of the biosynthetic steroid, showed that mutations in this gene cause vitamin D-dependent rickets type I (VDDRI), sequenced the entire gene, devised rapid PCR tactics to study other patients and analyzed the evolutionary relationships of all the mitochondrial P450 enzymes (185, 186). Analysis of 19 patients in 17 families with VDDR-I showed that all had 1α-hydroxylase mutations and identified the mutation that commonly affects French Canadian patients (201). Analysis of individuals with clinically mild forms of VDDR-I identified mutations with residual activity (250). We have shown that P450c1α gene transcription is regulated by dietary phosphorous (244) and a recent study identified common mutations in the Korean population (317).
6) P450
oxidoreductase (POR)
All microsomal (type 2) cytochrome P450 enzymes must receive electrons from NADPH via the intermediary of P450 oxidoreductase (POR) to achieve catalysis. POR receives electrons from NADPH via FAD moiety, transfers the electrons to its FMN moiety, and then to the P450 (299). Others have determined the structure of rat POR crystallographically (PNAS 94: 8411, 1997) and shown that knockout of the POR gene causes early embryonic lethality in mice (J Biol. Chem. 277: 6536, 2002; Mol. Cell Biol. 23: 6103, 2003). We recently described the first cases of human POR deficiency, causing a spectrum of disease from simple infertility, to substantially disordered steroidogenesis associated with Antley-Bixler skeletal malformation syndrome (279, 297, 333). We have shown that human POR transcription is initiated at an untranslated exon ~38 kb upstream from the first coding exon (314). Sequencing of the POR gene in 842 normal persons showed that 28% of human alleles carry the variant A503V (319). This variant has decreased capacity to support the catalytic activities of P450c17 (297, 319) but normal ability to support the activities of P450c21 (332) and the hepatic drug-metabolizing enzymes CYP1A2 and CYP2C19 (331). Current work concerns the roles of A503V with CYP2D6 and CYP3A4, which metabolize most clinically used drugs, and the regulation of POR gene transcription.
[Home] [Walter L. Miller] [Background and Current Projects]
[Lab Members]
[Selected Publications] [Lab Alumni] [Job Opportunities] [How to Contact Us]