Steroid Signaling

Steroids 82 (2014) 14–22

Contents lists available at ScienceDirect

Steroids

journal homepage: www.elsevier.com/locate/steroids

Review Steroid signaling: Ligand-binding promiscuity, molecular symmetry, and the need for gating ⇑ Richard Lathe a,b,c, , Yuri Kotelevtsev a,b,d,e a State University of Pushchino, Prospekt Nauki, Pushchino 142290, Moscow Region, Russia b Pushchino Branch of the Institute of Bio-Organic Chemistry, Russian Academy of Sciences, Pushchino 142290, Moscow Region, Russia c Pieta Research, PO Box 27069, Edinburgh EH10 5YW, UK d Biomedical Centre for Research Education and Innovation (CREI), Skolkovo Institute of Science and Technology, 143025 Skolkovo, Russia e Queens Medical Research Institute, University of Edinburgh, Little France, Edinburgh EH16 4TJ, UK article info abstract

Article history: Steroid/sterol-binding receptors and enzymes are remarkably promiscuous in the range of ligands they Received 14 June 2013 can bind to and, in the case of enzymes, modify – raising the question of how specific receptor activa- Received in revised form 3 December 2013 tion is achieved in vivo. Estrogen receptors (ER) are modulated by 27-hydroxycholesterol and 5a- Accepted 6 January 2014 androstane-3b,17b-diol (Adiol), in addition to estradiol (E2), and respond to diverse small molecules Available online 21 January 2014 such as bisphenol A. Steroid-modifying enzymes are also highly promiscuous in ligand binding and metabolism. The specificity problem is compounded by the fact that the steroid core (hydrogenated Keywords: cyclopentophenanthrene ring system) has several planes of symmetry. Ligand binding can be in sym- Symmetry metrical East–West (rotation) and North–South (inversion) orientations. Hydroxysteroid dehydrogen- Promiscuity Ligand-binding ases (HSDs) can modify symmetrical 7 and 11, also 3 and 17/20, positions, exemplified here by yeast 3a,20b-HSD and mammalian 11b-HSD and 17b-HSD enzymes. Faced with promiscuity and symmetry, other strategies are clearly necessary to promote signaling selectivity in vivo. Gating regulates hormone access via enzymes that preferentially inactivate (or activate) a subclass of ligands, thereby governing which ligands gain receptor access – exemplified by 11b-HSD gating cortisol access to the mineralocorticoid receptor, and P450 CYP7B1 gating Adiol access to ER. Counter-intuitively, the specificity of steroid/sterol action is achieved not by intrinsic binding selectivity but by the combination of local metabolism and binding affinity. Ó 2014 Elsevier Inc. All rights reserved.

Contents

1. The paradigm: specificity of steroid–protein interactions ...... 15 2. What does a steroid-binding polypeptide recognize? ...... 15 2.1. Promiscuity of receptor binding ...... 15 2.2. What is the primary ER ligand? ...... 16 2.3. A further problem – steroid symmetry ...... 16 3. Promiscuity and symmetry: steroid-metabolizing enzymes ...... 17 3.1. 11b-HSD (HSD11B) ...... 17 3.2. HSD17B10 ...... 17 3.3. ACAT1 – one enzyme, two binding sites, multiple ligands...... 17 3.4. Other examples ...... 18 4. Lack of specificity – the need for gating ...... 18 4.1. HSD11B and the mineralocorticoid receptor (MR) ...... 18 4.2. CYP7B1 and estrogen receptors ...... 18 4.3. Reverse (‘positive’) gating ...... 19 4.4. Different ligands, different effects ...... 19 4.5. Targeted transport and delivery ...... 20

⇑ Corresponding author at: Pieta Research, PO Box 27069, Edinburgh EH10 5YW, UK. Tel.: +44 131 478 0684. E-mail addresses: [email protected] (R. Lathe), [email protected] (Y. Kotelevtsev).

5. Concluding remarks ...... 20 Acknowledgments ...... 20 References ...... 20

1. The paradigm: speciﬁcity of steroid–protein interactions 2. What does a steroid-binding polypeptide recognize?

The current paradigm is that steroids (and related molecules The structure of the steroid nucleus does not easily lend itself to including sterols) induce their biological effects by binding selec- highly specific molecular contacts. As shown in Fig. 1, there are few tively to cognate receptor targets (the ‘lock and key’ model). How- molecular groups in a molecule such as estradiol that permit ste- ever, there is increasing realization that the stereochemistry of reospecific interactions, raising the question of what specific con- ligand binding to nuclear receptors and metabolizing enzymes is tacts steroid receptors make with the target ligand. X-ray studies far more flexible than previously thought. This conclusion is borne have shown that, following receptor binding, steroids are largely out by evolutionary and functional studies. buried within the folded polypeptide, but the assumption that Studies on an ancient metazoon, the sponge Amphimedon queen- primary interactions are spread uniformly across the molecule slandica, revealed two nuclear hormone receptors that more closely does not appear to be correct. Instead, initial binding is likely to resemble the modern retinoid-X receptor (RXR) than the group of be driven by a small number of contacts within restricted domains true steroid receptors; it was argued that the ancestral ligand, whose of the molecule, with weak stabilizing contacts across the binding arose through a process of ‘molecular tinkering’, was possi- molecular surface – that only in a second step lead to complex bly a long-chain fatty acid [1]. True steroid-type receptors arose refolding [7]. more recently. Genome sequence comparisons have revealed that This conclusion is perhaps inevitable, regarding steroid-binding the first nuclear steroid-type receptor most closely resembled the receptors and modifying enzymes, for two reasons. First, if the estrogen receptor (ER), and studies in ancient fishes argue that gen- binding polypeptide made extensive high-affinity contacts across ome duplication then subsequently led to the evolution of a second the whole ligand then the molecule would no longer dissociate receptor that most closely resembled the progesterone receptor (PR) once a local hormone surge had abated, or following reaction com- [2,3]. Reconstruction of the putative ancestral ER-like receptor at the pletion. Second, ligand-binding specificity for modern steroids is foot of the evolutionary tree allowed demonstration that this poly- often governed by a small number of molecular differences, some- peptide responded best to E2 [3], although alternative ligands (such times a single group. If the whole molecule were to be recognized as 3-hydroxysterols) were not tested. by the binding polypeptide then the impact of a single group However, there is no relationship in the position in the evolu- would be diminished, and binding specificity would be lost. tionary tree and the chemical nature of the current ligand, leading Clearly, some compromises must be made, but it appears that ste- to the speculation that ligand-binding dependency emerged sev- roid/sterol receptors make initial high-affinity contacts with only a eral times independently [4,5] and that specific ligand binding ar- fraction of the ligand, and therefore interact with a spectrum of ose relatively late. The most likely conclusion is that early related molecules. receptors responded to a range of molecules. For example, the emergence of E2 as a selective ligand for ER must have arisen late 2.1. Promiscuity of receptor binding because there is no easy synthetic route for the generation of E2 from likely precursors. In another example, the reconstructed We define here ‘promiscuity’ as the capacity of an enzyme or ancestor of the mineralocorticoid and glucocorticoid receptors receptor to bind to (and in the case of enzymes, modify) a range (MR and GR, respectively) in jawless fishes responds best to aldo- of different substrates, often in different configurations and at dif- sterone, despite evidence that the molecule was absent at this ferent positions. The term ‘promiscuity’ has been widely used in evolutionary stage, arguing that the ancestor receptor bound to this context (e.g., [8,9]) and the concept is akin to that of ‘biological a range of related molecules that pre-dated aldosterone [6].As messiness’ [10]. we will see below, the capacity to bind to a range of related The mammalian ER affords a good example. It was proposed that molecules is conserved through to the present day, with impor- binding is driven by the steroidal A ring alone, and that the activity of tant implications for the selectivity of hormone signaling the molecule in terms of receptor activation is driven by the D ring: mechanisms. Duax’s 1978 ‘A-ring binding/D-ring acting’ model for estrogens and

Fig. 1. Structure of estradiol. (A) Ball and stick model; (B) spacefill (Van der Waals surface). Modeling performed using JSmol/JMol [92]. 16 R. Lathe, Y. Kotelevtsev / Steroids 82 (2014) 14–22 progestins (Fig. 2) [11]. Although this early model is a simplification, (5a-androstane-3b,17b-diol; Adiol), androstenediol (androst-5- more extensive studies confirm that the only tight-binding pocket ene-3b,17b-diol), and 27-hydroxycholesterol (27OHC) (reviewed encompasses the phenolic A ring, and the overall scope of the li- in [19]), are active as ER ligands. As we will see later, Adiol is a phys- gand-binding pocket, that is only opened upon ligand binding, can iological ligand for ERb in the prostate, and 27OHC is a physiological accommodate bulky substituents – ERa can therefore bind to a wide modulator of both ERa and ERb. Although the affinity of the recep- diversity of molecules ([7] for discussion). tors for 27OHC is far lower than for E2, the concentrations of In support, ligand-binding promiscuity for ER is demonstrated 27OHC are markedly higher – the molecule is thus likely to compete by studies on molecules that have only limited similarity to E2. A with E2 for receptor binding. diversity of alkylphenolic compounds can bind to ER and activate From an evolutionary perspective the preference of ER for E2 is transcription [12]. The pesticide DDT (dichlorodiphenyltrichloro- perplexing because E2 is only produced via a chain of conversions ethane), one of the first known xenoestrogens [13], is known to ex- (via androgens) and is therefore unlikely to have been the ancestral ert potent biological effects on the reproductive tract [14]. Other ligand for ER. On these grounds, sterols (and 3b-hydroxysteroids) molecules with ER endocrine effects include phytoestrogens and are more likely contenders as ancestral ligands. Regarding ER and dioxins [15] and these interfere with diverse aspects of estrogenic E2, it seems likely that (as with aldosterone and MR, above) the li- signaling [16]. In a comprehensive study, Blair et al. [17] screened gand evolved after the receptor. 188 environmental chemicals. In a competition assay, 100 were Interestingly, it has been speculated (as for other receptors, see found to bind to ER, and of these 26 were classified as strong bind- Section 4.4) that there may be ligand-specific functional differ- ers. The study confirmed that an aromatic A ring is absolutely re- ences: ERb complexed to Adiol may adopt a conformation that required; binding was increased by the presence of two rings and, cruits the co-repressor CtBP (C-terminal binding protein), leading although a 3-OH group markedly increased binding, this was not to anti-inflammatory effects [20] that are absent when ERb binds absolutely essential [17]. These conclusions are reinforced by a re- E2. This raises the interesting question of whether a single receptor cent study on an ancestral ER dubbed AncSR1 [9]. The small molec- (ER) can fulfil different roles depending on the nature (and orienta- ular sizes of some of these molecules (e.g., bisphenol A and phenol tion, see below) of the ligand. red; Fig. 3) [18] support the contention that initial ligand binding to ER is governed by a very limited molecular domain. 2.3. A further problem – steroid symmetry

2.2. What is the primary ER ligand? The problem of promiscuous binding to a variety of molecules with only limited similarity to the lead ligand is accentuated by If at first contact ER only binds with high affinity to part of the the fact that the steroidal molecule itself has multiple planes of molecule, notably the hydroxylated A ring, the logical consequence symmetry (Fig. 4). This leads to the speculation that the ligand is that ER in vivo must bind to a range of molecules similar to E2, molecule can bind to its target, be it a steroid-type nuclear receptor raising the question – what is the true ER ligand? There is extensive or a metabolizing enzyme, in more than one orientation. evidence that other steroids and sterols, including androstanediol Modeling studies have revealed that androstenediol is likely to dock into human ERa in two different configurations, one in which is as for E2 (76/100 poses) and a second in which the entire molecule is ‘North–South’ (up–down) inverted (24/100 poses) [21]. This raises the intriguing possibility that binding of a single molecule in different orientations could have different biological activities. These promiscuity and symmetry issues reported for ER are likely to be only the tip of the iceberg; similar considerations no doubt apply to other receptors (discussed by Baker [22]).

Fig. 2. Schematic of estrogen binding to its receptor emphasizing the primary recognition of the phenolic A ring. From [11], with permission. For more detailed description of ER ligand binding, where initial contacts with the A-ring cleft precipitate refolding of the molecule, see [7].

Fig. 4. Symmetrical positions on the steroid nucleus via rotation or inversion, exempliﬁed by estradiol. Substituents above the plane of the steroid are described as b and are shown as a solid line ( ); those below the plane are described as a and are shown by a broken line ( ). Planes of rotational symmetry are indicated, emphasizing the similarity of the 3a and 17b positions (180° rotation around the z Fig. 3. Limited similarities between estrogen mimetics and estradiol. (a) Estradiol, axis), 11b and 7a (180° rotation around the x axis), and 11b and 7b (180° rotation (b) bisphenol A, (c) phenol red. around the y axis). R. Lathe, Y. Kotelevtsev / Steroids 82 (2014) 14–22 17

3. Promiscuity and symmetry: steroid-metabolizing enzymes

Promiscuity extends to steroid/sterol metabolizing enzymes. The view ‘one pocket, one ligand-binding conformation’ was ﬁrst challenged by work on 3a,20b-hydroxysteroid dehydrogenase (3a,20b-HSD) of the yeast Streptomyces hydrogenans. The enzyme catalyzes both 3a and 20b oxidoreduction of steroid substrates. How can a single enzyme modify its substrate at two very different positions? Two models were debated: (i) a single substrate-binding site, but with two different pockets accommodating NAD/NADP cofactors, or (ii) substrate binding in two different orientations, with a single cofactor site. X-ray studies revealed a single cofactor site [23], conﬁrming the Sweet and Samant [24] proposal that the substrate can bind in two alternative orientations, such that two distinct positions are exposed to cofactor-mediated oxidoreduction (see [25]). As discussed below, this appears to be commonplace.

3.1. 11b-HSD (HSD11B)

The property of bifunctionality is shared by the enzymes that catalyze the interconversion of ‘inert’ (11-keto; cortisone in humans and 11-dehydrocorticosterone in mice and rats) and ‘active’ 11b-hydroxylated glucocorticoids (cortisol in humans, corticoste- Fig. 5. Molecular modeling of HSD11B1 interactions with 7b-hydroxy DHEA and corticosterone (11b,21-dihydroxyprogesterone) showing conservation of distances rone in mice and rats), thereby regulating the biological activity between the target hydroxyl and both C4 of NADP and the catalytic tyrosine of glucocorticoids [26]. However, the 7a and 11b positions are (Tyr183) residue. From [32] with permission. symmetrically positioned ([27] and Fig. 4), and some HSD11B enzymes can also modify the 7-position of the steroid backbone [28–31] – although this may not be true of all HDS11B enzymes be- 2-methyl-3-hydroxybutyryl CoA dehydrogenase implicated in iso- cause there are suggestions of marked species-specificity in the leucine catabolism [42]; such substrates are presumed to fold to alternative orientations that can be accommodated. adopt a steroid-like configuration in part of the molecule. In exploration of the mechanism, Nashev et al. [32] expressed Importantly, HSD1710 is capable of modifying both the 3a and mouse HSD11B1 together with hexose-6-phosphate dehydroge- 17b positions of steroid substrates. Several studies have addressed nase to drive the production of NADPH cofactor. In addition to cat- the substrate specificity of the enzyme [40,43,44]. Modifications alyzing 11b oxidoreduction, HSD11B1 was found to catalyze catalyzed include 3a and 17b oxidoreduction, but there is also efficiently the interconversion of 7-keto and 7-hydroxylated DHEA, significant activity at the 20, 21, and 7 (both a and b) positions. at both the 7a and 7b positions. Furthermore, HSD11B1 was found Modification at 3a is of considerable interest because 3a-hydrox- to catalyze hepatic reduction of 7-oxolithocholic acid [33]. Molec- ysteroids (and potentially sterols) have potent anesthetic activity ular modeling of HSD11B1 has revealed that substrate can bind in and directly target GABA and intracellular (mitochondrial) two different orientations and, for example, the 11b and 7b posi- receptors that control neuronal activation (reviewed in [45]). tions adopt an almost identical binding configuration in relation- Generally, different HSD17 enzymes are reported, in addition to ship to NADP cofactor and the catalytic tyrosine residue in the catalyzing interconversions at the 17 position, to catalyze enzyme (Fig. 5) [34], despite East–West rotation of the molecule. conversions at the 3 and 7 positions, and also at the 20 (in Activity at 7a may reflect North–South inversion of the molecule 20-hydroxyprogesterone), 21 (in 11-dehydrocorticosterone) and (Fig. 4). 24 (in 3-ketostearoyl-CoA) positions [46]. Interestingly, the enzymes HSD11B1 and 2 are not the only enzymes capable of catalyzing 11b oxidoreduction: in mice knocked- out for both enzymes significant HSD11B activity remains (Y.K., 3.3. ACAT1 – one enzyme, two binding sites, multiple ligands unpublished data). Identification of the other enzymes with HSD11B activity will be of interest. ACAT, also known as sterol O-acyltransferase, affords another example. ACAT1 is responsible for converting cholesterol to cho- 3.2. HSD17B10 lesteryl esters, thereby promoting cholesterol sequestration in fatty droplets, and has been implicated in foam-cell formation A further case is the Alzheimer disease Ab-binding dehydroge- and atherogenesis. In addition to modifying cholesterol (predomi- nase. Yeast two-hybrid screening using Ab as bait identified a nantly at the 3b position), substrates include lathosterol, cholesta- new binding partner dubbed ‘ERAB’ [35] or ABAD (Ab-binding alco- nol, 7-dehydrocholesterol, 7a-hydroxycholesterol, and 25- hol dehydrogenase). The protein has now been shown to be iden- hydroxycholesterol, as well as a range of dietary plant sterols tical to type 10 17b-HSD (HSD17B10) [36]. HSD17B enzymes [47–50]. ACAT can also metabolize the steroid pregnenolone [51]. typically catalyze the interconversion of estradiol (17b-hydroxy) ACAT1 contains not just one, but two, binding sites for sterols. and estrone (17-oxo), and regulate the bioavailability of active hor- One represents the enzyme active site; the second site governs mone in tissues such as ovary [37]. However, this specific isoform allosteric activation of the enzyme. Allosteric activation is medi- is present at highest levels in hippocampus [38] and, moreover, un- ated largely by side-chain oxidized cholesterols, and not by close like other HSDs, the enzyme is membrane-bound [39] and is prin- analogs such as 7-ketocholesterol [48]. However, once cholesterol cipally confined to mitochondria [40,41]. The true in vivo substrate is bound to the allosteric site, the enzyme becomes highly promis- for the enzyme has not been established. HSD17B10 has other cuous towards different substrate sterols [48]. In fact, ACAT1 activ- substrates in addition to steroids/sterols, and also acts as a ity on pregnenolone (above) is crucially dependent on cholesterol. 18 R. Lathe, Y. Kotelevtsev / Steroids 82 (2014) 14–22

In the absence of cholesterol, pregnenolone is a poor substrate but, 4. Lack of specificity – the need for gating in the presence of cholesterol, becomes an excellent substrate [51]. ACAT is also modulated by, and may metabolize, estrogens. In The lack of selectivity, at first sight, would appear to be incom- addition to binding the anti-estrogen tamoxifen, ACAT activity is patible with accurate hormone signaling. Therefore, other mecha- inhibited by E2 itself, but only at concentrations in the 10 lM nisms must be exploited to boost selectivity. We discuss ‘gating’, range [52] (E2 generally circulates in the nM range). Conversely, where combinations of two (or more) ligand-binding polypeptides the prototypical ACAT inhibitor Sah 58-035 is a potent ER agonist are used to achieve the required in vivo selectivity in target tissues. [53]. We define gating as a process in which local enzyme action in a Steroid symmetry considerations would predict that ACAT is target tissue dictates whether a systemically delivered molecule likely to esterify E2 at the 17b position (and potentially at the 3 po- can, or cannot, gain access to a target binding site. We distinguish sition). 17b esters of E2 are known to accumulate in association between gating and detoxification (oxidative and other metabo- with low-density lipoprotein [54] and have been specifically impli- lisms principally mediated by the liver); although, given the poten- cated in mammary tumorigenesis [55]. The related enzyme leci- tial to regulate hormone availability, it would be surprising if thin/cholesterol acyltransferase (LCAT) has been reported to hepatic metabolism does not in some instances play a specific reg- catalyze preferential esterification of E2 at the 17b position, in ulatory role (not reviewed here). In the following we focus on two addition to modifying Adiol at both 3b and 17b [56]. It is therefore examples where gating has been demonstrated. likely that ACAT also modifies E2 at 17b but this remains to be demonstrated. 4.1. HSD11B and the mineralocorticoid receptor (MR)

3.4. Other examples As noted earlier, HSD11B enzymes catalyze the interconversion of ‘inert’ (11-keto; cortisone/11-dehydrocorticosterone) and ‘ac- These cases are not unusual: the capacity to bind to and/or tive’ 11b-hydroxylated glucocorticoids (cortisol/corticosterone). modify steroid substrates at different positions appears to be wide- The ability of the enzyme to gate the action of glucocorticoids spread. As a further illustration, some 3b-HSD enzymes can modify and mineralocorticoids emerged from studies of patients with the 17b position (reviewed in [57]) and a small number of modifi- apparent mineralocorticoid excess as a result of deficiency of cations can convert an enzyme catalyzing 3a oxidoreduction to HSD11B2 [65]. Briefly, kidney MRs are activated by both aldoste- one catalyzing 17b or 20a conversions [58]. Indeed, HSD11B ap- rone and cortisol, molecules that differ centrally in oxidation of pears to have evolved from an ancestral HSD17B2-type enzyme aldosterone at the 18 position, three carbons away from the 11 po- [59]. Furthermore, some enzymes are able to modify the choles- sition. For this reason aldosterone is not a substrate for HSD11B2, terol side-chain (positions 20–27), suggesting that these carbons whereas cortisol is an excellent substrate. HSD11B2 thus inacti- might possibly fold back to adopt a compact configuration resem- vates cortisol, but not aldosterone. bling the 3D structure of the steroid nucleus. Normally, the action of local HSD11B2 prevents cortisol action Together, these studies confirm that steroid-binding enzymes at MR, and the receptor therefore responds only to aldosterone – make initial high-affinity contacts with only a small fraction of but, in patients lacking this enzyme, cortisol is able to activate the molecule and, depending on the orientation in which the sub- renal MR, leading to apparent mineralocorticoid excess ([65], strate inserts into the binding site, catalyze modifications at differ- reviewed in [26,66,67]). Selective activation of MR by aldosterone ent positions on the steroid nucleus. Notably, these molecules can – and not by cortisol – is achieved not by the intrinsic ligand-bind- adopt alternative configurations, both via rotation (East–West) and ing selectivity of the receptor but by the substrate specificity of inversion (North–South or up–down; Fig. 4). Of these different locally expressed HSD11B enzymes (Fig. 6). positions, the 3a and 17b (or perhaps 20b) appear to have the clos- est resemblances, as do the 7a/b and 11b positions (Fig. 4). 4.2. CYP7B1 and estrogen receptors This raises several interesting questions. For example, does EBI2, a receptor for 7a-hydroxylated sterols [60,61], respond to One of the two major ERs in mammals, ERb, binds to both E2 11b-hydroxylated molecules? Do LXRs, that bind to 22, 24, 25 and the widely circulating steroid Adiol. However, ERb responds and/or 27-oxidized cholesterols [62–64], also bind to 17b/3a- poorly to 7a-hydroxylated Adiol, whose formation by hydroxyl- hydroxylated molecules? ation of Adiol is catalyzed by the oxysterol 7a-hydroxylase CYP7B1

Fig. 6. Gating of cortisol access to the mineralocorticoid receptor (MR) by 11b-HSD. Insert panel adapted from Kotelevtsev et al. [93]. Note that the exact hormonal ligands differ between human and rodent, but gating takes place in both cases. R. Lathe, Y. Kotelevtsev / Steroids 82 (2014) 14–22 19

(noting, importantly, that CYP7B1 is only poorly active against E2 receptor is required for immune cell migration. The most potent itself). In the developing rodent prostate, ERb activation is neces- activators are 7a-hydroxylated 25-hydroxycholesterols [60,61]. sary to control cellular proliferation [68]. Inactivation of this CYP In this case the combined activities of two enzymes, CYP7B1 and led to hyperactivation of ERb by Adiol and markedly reduced cellu- cholesterol 25-hydroxylase (CH25H), are required to generate lar proliferation [68]. Similar gating takes place in females, and fe- EBI2 ligand, and failure of immune cell migration was observed male Cyp7b1À/À mice display early uterine and mammary gland in mice deficient in CH25H [60]. proliferation attributed to hyperactivation of ERb by Adiol [69].A LXR (liver X receptor) positive gating has also been demon- parallel CNS hyperproliferation phenotype was seen in Cyp7b1À/À strated. LXRa and LXRb are members of the nuclear family of tran- mice owing to overactivation of brain ERb by Adiol [70]. Thus, scription factors that regulate cellular cholesterol efflux. LXR the specificity of ERb activation by E2, and not by Adiol (or indeed activation promotes export of excess cholesterol from peripheral by androstenediol or 27OHC), is ensured by the substrate specific- tissues to the liver and bile for excretion. The best ligands are cho- ity of locally expressed CYP7B1, in conjunction with the intrinsic lesterols oxidized at the 22, 24, and/or 25 positions [62], but these activity of the receptor (Fig. 7). sterols are present only at low concentrations, and 27OHC is more Gating extends to other ER ligands. 27OHC is now recognized to likely to be the endogenous ligand [63,64]. The conclusion that the be a selective estrogen receptor modulator (SERM) at both ERa and activity of cholesterol 27-hydroxylase, CYP27, is required for posi- ERb [71,72]. Although it has a lower binding affinity for ERs than tive gating of LXR was borne out by inspection of fibroblasts cul- does E2, circulating concentrations are a little under the lM range, tured from a patient with CYP27 deficiency (cerebrotendinous and local tissue levels (such as in atherosclerotic plaques) can even xanthomatosis, CTX) where upregulation of a LXR target gene in re- reach mM concentrations. In vascular endothelial cells 27OHC is sponse to cholesterol loading was markedly impaired [63]. Inter- now known to antagonize transcriptional activation by ERs, and estingly, Janowski et al. [62] reported that 22(S)-hydroxy

27OHC binds to recombinant ERa and ERb with Ki values of 1.2 cholesterol binds to both LXR subtypes in vitro, but does not mod- and 0.42 lM respectively [72,73]. However, CYP7B1-mediated ulate LXR activity in cells; they speculated that selective ACAT- metabolism of 27OHC [74–76] gates access to ERs. Following vas- mediated metabolism of 22(S)-hydroxycholesterol may gate cular injury, E2 promotes repair of endothelial denudation in wild- receptor access [62]. type mice, but repair was significantly reduced in Cyp7b1À/À mice, Gating appears to be a generic feature of steroid/sterol biology. arguing that lack of CYP7B1 allows 27OHC to antagonize ER activ- To give one invertebrate example, in Caenorhabditis elegans the ste- ity [73]. rol dafachronic acids (DAs) regulate lifespan via the nuclear recep- Gating seems to be the rule rather than the exception. To pick tor DAF-12 (a homolog of LXR and FXR). DAF-12 activation requires one further example, progesterone binds more tightly to MR than a short but pivotal pathway of metabolic conversions from choles- does aldosterone, but has only low activation activity, leading to terol to DAs, and the activity of this pathway is governed by potent in vitro anti-mineralocorticoid activity [77,78]. The same environmental stress – thus attuning development/diapause, is not seen in vivo, suggesting that PROG action at MR is locally reproduction, and longevity to environmental signals (e.g. [82]). gated by metabolizing enzymes [79]. Although the key enzymes It remains an open question whether environmental stresses also involved have not been identified, evidence is emerging that operate via steroid/sterol pathways to regulate growth and aging CYP27 also participates in gating PROG access to MR [80]. Some in mammals. progestins are also ligands of AR [81], unlike natural PROG itself (which does not bind significantly to AR), but it is not known if gating processes also operate at this target. 4.4. Different ligands, different effects

The possibility was discussed (Section 2.2, above) that binding 4.3. Reverse (‘positive’) gating of different ligands to ER (E2 vs Adiol) might exert different functions. This concept has been amply conﬁrmed for MR. Indeed, The action of HSD11B or CYP7B prevents steroid/sterol access to although mutation studies have demonstrated that speciﬁcity for nuclear receptors, respectively MR and ER, but in other cases enzy- aldosterone (versus cortisol/corticosterone) is achieved by matic conversion is required to generate ligand. For example, the HSD11B-mediated gating, this is unlikely to be the whole story. Epstein–Barr virus-induced gene 2 (EBI2) G protein-coupled (i) Cortisol/corticosterone circulates at up to 100-fold higher levels

Fig. 7. Gating of 5a-androstane-3b,17b-diol (Adiol) access to ERb by the action of CYP7B1. Insert panel adapted from Sugiyama et al. [70]. 20 R. Lathe, Y. Kotelevtsev / Steroids 82 (2014) 14–22 than aldosterone, and HSD11B2 would therefore need to be >99% – there are (i) multiple keys, and the guardians of the keys (gating effective in inactivating cortisol to prevent entirely its action at enzymes) govern whether the lock is opened and, indeed, (ii) dif- MR. (ii) Some tissues (e.g., hippocampus and macrophages) express ferent keys may open the lock in different ways. The design of high levels of MR but do not express detectable HSD11B2, but – de- pharmaceutical steroidal molecules needs to take into account spite the absence of the gating enzyme – receptor signaling is prin- both local (and systemic) metabolism and transport, as well as dif- cipally in response to aldosterone (and not, except in stress ferential ligand-binding conformations, and not only binding situations, in response to cortisol/corticosterone). How is this affinity. achieved? As recently reviewed [66], in addition to differential binding kinetics, the conformation of the ligand-bound receptor Acknowledgments (and the effects it exerts via downstream effectors) may differ materially according to the ligand. A single receptor (in this case A preliminary version of this paper was presented at the 2012 MR) can therefore produce a range of responses depending on meeting of the European Network on Oxysterol Research (ENOR) the nature of the ligand [66] and, potentially, on its binding orien- in Djion, France. This work was funded, in part, by a Government tation, further challenging the ‘lock and key’ paradigm. of Russia Grant (11.G34.31.0032/14.12.2010) to Y.K. We thank Rebecca Pruss (Marseille) for critical reading of the manuscript 4.5. Targeted transport and delivery and many suggestions, and anonymous reviewers for further helpful insights. Enzymatic gating is one mechanism by which specificity can be enhanced in vivo. Clearly tissue- and cell-specific expression pat- terns of metabolizing enzymes and receptor targets also play a piv- References otal role in ensuring that ligands act in the right place, and at the [1] Bridgham JT, Eick GN, Larroux C, Deshpande K, Harms MJ, Gauthier ME, et al. right time (not reviewed here). Selective delivery is another potent Protein evolution by molecular tinkering: diversification of the nuclear mechanism. In addition to a range of steroid/sterol binding pro- receptor superfamily from a ligand-dependent ancestor. PLoS Biol 2010;8. teins [including low density lipoproteins (LDL), high density lipo- [2] Thornton JW. Evolution of vertebrate steroid receptors from an ancestral estrogen receptor by ligand exploitation and serial genome expansions. Proc proteins (HDL), their integral apolipoproteins, as well as their Natl Acad Sci USA 2001;98:5671–6. uptake receptors] and metabolizing enzymes (e.g., sulfotransferas- [3] Thornton JW, Need E, Crews D. Resurrecting the ancestral steroid receptor: es, methyltransferases, hydroxylases, conjugating enzymes), many ancient origin of estrogen signaling. Science 2003;301:1714–7. [4] Escriva H, Safi R, Hanni C, Langlois MC, Saumitou-Laprade P, Stehelin D, et al. of which act systemically via the liver, a further mechanism for Ligand binding was acquired during evolution of nuclear receptors. Proc Natl focusing activity involves selective trans-cellular transport (e.g., Acad Sci USA 1997;94:6803–8. [83]) and intracellular targeting. The lipid-transfer sterol carrier [5] Escriva H, Delaunay F, Laudet V. Ligand binding and nuclear receptor evolution. BioEssays 2000;22:717–27. proteins (SCPs) and oxysterol binding proteins (OSBPs) are a case [6] Bridgham JT, Carroll SM, Thornton JW. Evolution of hormone-receptor in point. OSBPs contain two membrane-binding surfaces, and func- complexity by molecular exploitation. Science 2006;312:97–101. tion to transfer sterols from one intracellular membrane to another [7] Katzenellenbogen JA. The 2010 Philip S. Portoghese Medicinal Chemistry [84–86]. In human there is a family of 12 OSBPs [87,88]. Emerging Lectureship: addressing the ‘‘core issue’’ in the design of estrogen receptor ligands. J Med Chem 2011;54:5271–82. evidence indicates that each has a distinct profile of sterol binding [8] Jones SA, Moore LB, Shenk JL, Wisely GB, Hamilton GA, McKee DD, et al. The and, significantly, subcellular localization [89]. This means that pregnane X receptor: a promiscuous xenobiotic receptor that has diverged each binding molecule is differentially targeted within the cell. Evi- during evolution. Mol Endocrinol 2000;14:27–39. [9] Eick GN, Colucci JK, Harms MJ, Ortlund EA, Thornton JW. Evolution of minimal dence for selective intracellular transport and delivery of steroids specificity and promiscuity in steroid hormone receptors. PLoS Genet is less well documented, although SCP2 (also known as nonspecific 2012;8:e1003072. lipid-transfer protein) is highly promiscuous in the lipid moieties it [10] Tawfik DS. Messy biology and the origins of evolutionary innovations. Nat Chem Biol 2010;6:692–6. transports [90]. [11] Duax WL, Ghosh D, Pletnev V, Griffin JF. Three-dimensional structures of The evolutionary impact of such transport proteins is high- steroids and their protein targets. Pure Appl Chem 1996;68:1297–302. lighted by the finding that HSD17B4 is a chimeric protein that con- [12] White R, Jobling S, Hoare SA, Sumpter JP, Parker MG. Environmentally persistent alkylphenolic compounds are estrogenic. Endocrinology tains both an HSD domain and a C-terminal region with high 1994;135:175–82. similarity to SCP2 [91]. It is possible, but so far unproven, that [13] McLachlan JA, Arnold SF. Environmental estrogens. Am Sci 1996;84:452–61. OSBPs and other transport proteins pre-select particular substrates [14] Tiemann U. In vivo and in vitro effects of the organochlorine pesticides DDT, TCPM, methoxychlor, and lindane on the female reproductive tract of and orientations for presentation to receptors and metabolizing mammals: a review. Reprod Toxicol 2008;25:316–26. enzymes, conferring a further layer of molecular selectivity. [15] Lorand T, Vigh E, Garai J. Hormonal action of plant derived and anthropogenic non-steroidal estrogenic compounds: phytoestrogens and xenoestrogens. Curr Med Chem 2010;17:3542–74. 5. Concluding remarks [16] Tilghman SL, Nierth-Simpson EN, Wallace R, Burow ME, McLachlan JA. Environmental hormones: multiple pathways for response may lead to multiple disease outcomes. Steroids 2010;75:520–3. Polycyclic molecules such as steroids and sterols would appear [17] Blair RM, Fang H, Branham WS, Hass BS, Dial SL, Moland CL, et al. The estrogen to be well adapted for signaling purposes. They are derived from receptor relative binding affinities of 188 natural and xenochemicals: structural diversity of ligands. Toxicol Sci 2000;54:138–53. one of the most abundant cellular components, cholesterol, and li- [18] Randerath K, Moorthy B, Mabon N, Sriram P. Tamoxifen: evidence by pid solubility and structural rigidity lend themselves to signaling – 32P-postlabeling and use of metabolic inhibitors for two distinct pathways perhaps explaining the ubiquity of this type of molecule in a sys- leading to mouse hepatic DNA adduct formation and identification of temic signaling role. However, there are few molecular groups 4-hydroxytamoxifen as a proximate metabolite. Carcinogenesis 1994;15: 2087–94. available for highly specific interactions. In consequence, binding [19] Baker ME. What are physiological estrogens? Steroids 2013;78:337–40. polypeptides are remarkably promiscuous and interact with multi- [20] Saijo K, Collier JG, Li AC, Katzenellenbogen JA, Glass CK. An ADIOL-ERbeta-CtBP ple ligands, with docking taking place in alternative orientations. It transrepression pathway negatively regulates microglia-mediated inflammation. Cell 2011;145:584–95. is this adaptability of steroids which makes them natural ligands of [21] Baker ME, Uh KY, Chandsawangbhuwana C. 3D models of human ERalpha and multiple receptors. However, the selectivity of hormone action ERbeta complexed with 5-androsten-3beta,17beta-diol. Steroids seen today is only achieved through tissue-specific ligand metabo- 2012;77:1192–7. [22] Baker ME. Insights from the structure of estrogen receptor into the evolution lism, or ‘gating’ (combined with other mechanisms including of estrogens: implications for endocrine disruption. Biochem Pharmacol selective delivery). In short, the ‘lock and key’ model is incomplete 2011;82:1–8. R. Lathe, Y. Kotelevtsev / Steroids 82 (2014) 14–22 21

[23] Ghosh D, Weeks CM, Grochulski P, Duax WL, Erman M, Rimsay RL, et al. Three- [50] An S, Cho KH, Lee WS, Lee JO, Paik YK, Jeong TS. A critical role for the histidine dimensional structure of holo 3 alpha,20 beta-hydroxysteroid dehydrogenase: residues in the catalytic function of acyl-CoA:cholesterol acyltransferase a member of a short-chain dehydrogenase family. Proc Natl Acad Sci USA catalysis: evidence for catalytic difference between ACAT1 and ACAT2. FEBS 1991;88:10064–8. Lett 2006;580:2741–9. [24] Sweet F, Samant BR. Bifunctional enzyme activity at the same active site: study [51] Rogers MA, Liu J, Kushnir MM, Bryleva E, Rockwood AL, Meikle AW, et al. of 3 alpha and 20 beta activity by affinity alkylation of 3 alpha, 20 beta- Cellular pregnenolone esterification by acyl-CoA:cholesterol acyltransferase. J hydroxysteroid dehydrogenase with 17-(bromoacetoxy)steroids. Biochemistry Biol Chem 2012;287:17483–92. 1980;19:978–86. [52] de Medina P, Payre BL, Bernad J, Bosser I, Pipy B, Silvente-Poirot S, et al. [25] Duax WL, Thomas J, Pletnev V, Addlagatta A, Huether R, Habegger L, et al. Tamoxifen is a potent inhibitor of cholesterol esterification and prevents the Determining structure and function of steroid dehydrogenase enzymes by formation of foam cells. J Pharmacol Exp Ther 2004;308:1165–73. sequence analysis, homology modeling, and rational mutational analysis. Ann [53] de Medina P, Boubekeur N, Balaguer P, Favre G, Silvente-Poirot S, Poirot M. The N Y Acad Sci 2005;1061:135–48. prototypical inhibitor of cholesterol esterification, Sah 58-035 [3- [26] Seckl JR, Walker BR. Minireview: 11beta-hydroxysteroid dehydrogenase type [decyldimethylsilyl]-n-[2-(4-methylphenyl)-1-phenylethyl]propanamide], is 1 – a tissue-specific amplifier of glucocorticoid action. Endocrinology an agonist of estrogen receptors. J Pharmacol Exp Ther 2006;319:139–49. 2001;142:1371–6. [54] Helisten H, Hockerstedt A, Wahala K, Tiitinen A, Adlercreutz H, Jauhiainen M, [27] Lathe R. Steroid and sterol 7-hydroxylation: ancient pathways. Steroids et al. Accumulation of high-density lipoprotein-derived estradiol-17beta fatty 2002;67:967–77. acid esters in low-density lipoprotein particles. J Clin Endocrinol Metab [28] Song W, Chen J, Dean WL, Redinger RN, Prough RA. Purification and 2001;86:1294–300. characterization of hamster liver microsomal 7alpha-hydroxycholesterol [55] Mills LH, Yu J, Xu XM, Lee AJ, Zhu BT. Naturally-occurring estradiol-17beta- dehydrogenase. Similarity to type I 11beta-hydroxysteroid dehydrogenase. J fatty acid esters, but not estradiol-17beta, preferentially induce mammary Biol Chem 1998;273:16223–8. tumorigenesis in female rats: implications for an important role in human [29] Robinzon B, Michael KK, Ripp SL, Winters SJ, Prough RA. Glucocorticoids breast cancer. Toxicol Appl Pharmacol 2008;229:332–41. inhibit interconversion of 7-hydroxy and 7-oxo metabolites of [56] Kanji SS, Kuohung W, Labaree DC, Hochberg RB. Regiospecific esterification of dehydroepiandrosterone: a role for 11beta-hydroxysteroid dehydrogenases? estrogens by lecithin:cholesterol acyltransferase. J Clin Endocrinol Metab Arch Biochem Biophys 2003;412:251–8. 1999;84:2481–8. [30] Robinzon B, Prough RA. Interactions between dehydroepiandrosterone and [57] Simard J, Ricketts ML, Gingras S, Soucy P, Feltus FA, Melner MH. Molecular glucocorticoid metabolism in pig kidney: nuclear and microsomal 11beta- biology of the 3beta-hydroxysteroid dehydrogenase/delta5-delta4 isomerase hydroxysteroid dehydrogenases. Arch Biochem Biophys 2005;442:33–40. gene family. Endocr Rev 2005;26:525–82. [31] Muller C, Pompon D, Urban P, Morfin R. Inter-conversion of 7alpha- and 7beta- [58] Ma H, Penning TM. Conversion of mammalian 3alpha-hydroxysteroid hydroxy-dehydroepiandrosterone by the human 11beta-hydroxysteroid dehydrogenase to 20alpha-hydroxysteroid dehydrogenase using loop dehydrogenase type 1. J Steroid Biochem Mol Biol 2006;99:215–22. chimeras: changing specificity from androgens to progestins. Proc Natl Acad [32] Nashev LG, Chandsawangbhuwana C, Balazs Z, Atanasov AG, Dick B, Frey FJ, Sci USA 1999;96:11161–6. et al. Hexose-6-phosphate dehydrogenase modulates 11beta-hydroxysteroid [59] Baker ME. 11Beta-hydroxysteroid dehydrogenase-type 2 evolved from an dehydrogenase type 1-dependent metabolism of 7-keto- and 7beta-hydroxy- ancestral 17beta-hydroxysteroid dehydrogenase-type 2. Biochem Biophys Res neurosteroids. PLoS ONE 2007;2:e561. Commun 2010;399:215–20. [33] Odermatt A, Da Cunha T, Penno CA, Chandsawangbhuwana C, Reichert C, Wolf A, [60] Liu C, Yang XV, Wu J, Kuei C, Mani NS, Zhang L, et al. Oxysterols direct B-cell etal.Hepaticreductionofthe secondarybileacid7-oxolithocholic acid ismediated migration through EBI2. Nature 2011;475:519–23. by 11beta-hydroxysteroid dehydrogenase 1. Biochem J 2011;436:621–9. [61] Hannedouche S, Zhang J, Yi T, Shen W, Nguyen D, Pereira JP, et al. Oxysterols [34] Balazs Z, Nashev LG, Chandsawangbhuwana C, Baker ME, Odermatt A. Hexose- direct immune cell migration via EBI2. Nature 2011;475:524–7. 6-phosphate dehydrogenase modulates the effect of inhibitors and alternative [62] Janowski BA, Grogan MJ, Jones SA, Wisely GB, Kliewer SA, Corey EJ, et al. substrates of 11beta-hydroxysteroid dehydrogenase 1. Mol Cell Endocrinol Structural requirements of ligands for the oxysterol liver X receptors LXRalpha 2009;301:117–22. and LXRbeta. Proc Natl Acad Sci USA 1999;96:266–71. [35] Yan SD, Fu J, Soto C, Chen X, Zhu H, AlMohanna F, et al. An intracellular protein [63] Fu X, Menke JG, Chen Y, Zhou G, MacNaul KL, Wright SD, et al. 27- that binds amyloid-beta peptide and mediates neurotoxicity in Alzheimer’s Hydroxycholesterol is an endogenous ligand for liver X receptor in disease. Nature 1997;389:689–95. cholesterol-loaded cells. J Biol Chem 2001;276:38378–87. [36] He XY, Schulz H, Yang SY. A human brain L-3-hydroxyacyl-coenzyme A [64] Nelson ER, DuSell CD, Wang X, Howe MK, Evans G, Michalek RD, et al. The dehydrogenase is identical to an amyloid beta-peptide-binding protein oxysterol, 27-hydroxycholesterol, links cholesterol metabolism to bone involved in Alzheimer’s disease. J Biol Chem 1998;273:10741–6. homeostasis through its actions on the estrogen and liver X receptors. [37] Vihko P, Isomaa V, Ghosh D. Structure and function of 17beta-hydroxysteroid Endocrinology 2011;152:4691–705. dehydrogenase type 1 and type 2. Mol Cell Endocrinol 2001;171:71–6. [65] Benediktsson R, Edwards CR. Apparent mineralocorticoid excess. J Hum [38] He XY, Wegiel J, Yang SY. Intracellular oxidation of allopregnanolone by Hypertens 1994;8:371–5. human brain type 10 17beta-hydroxysteroid dehydrogenase. Brain Res [66] Fuller PJ, Yao Y, Yang J, Young MJ. Mechanisms of ligand specificity of the 2005;1040:29–35. mineralocorticoid receptor. J Endocrinol 2012;213:15–24. [39] Sambamurti K, Lahiri DK. ERAB contains a putative noncleavable signal [67] Odermatt A, Kratschmar DV. Tissue-specific modulation of mineralocorticoid peptide. Biochem Biophys Res Commun 1998;249:546–9. receptor function by 11beta-hydroxysteroid dehydrogenases: an overview. [40] He XY, Merz G, Yang YZ, Mehta P, Schulz H, Yang SY. Characterization and Mol Cell Endocrinol 2012;350:168–86. localization of human type10 17beta-hydroxysteroid dehydrogenase. Eur J [68] Weihua Z, Lathe R, Warner M, Gustafsson JA. An endocrine pathway in the Biochem 2001;268:4899–907. prostate, ERbeta, AR, 5alpha-androstane-3beta,17beta-diol, and CYP7B1, [41] He XY, Wen GY, Merz G, Lin D, Yang YZ, Mehta P, et al. Abundant type 10 17 regulates prostate growth. Proc Natl Acad Sci USA 2002;99:13589–94. beta-hydroxysteroid dehydrogenase in the hippocampus of mouse [69] Omoto Y,Lathe R,Warner M, GustafssonJA. Early onset ofpuberty and early ovarian Alzheimer’s disease model. Brain Res Mol Brain Res 2002;99:46–53. failure in CYP7B1 knockout mice. Proc Natl Acad Sci USA 2005;102:2814–9. [42] Yan SD, Shi Y, Zhu A, Fu J, Zhu H, Zhu Y, et al. Role of ERAB/L-3-hydroxyacyl- [70] Sugiyama N, Andersson S, Lathe R, Fan X, Alonso-Magdalena P, Schwend T, coenzyme A dehydrogenase type II activity in Abeta-induced cytotoxicity. J et al. Spatiotemporal dynamics of the expression of estrogen receptors in the Biol Chem 1999;274:2145–56. postnatal mouse brain. Mol Psychiatry 2009;14(223–32):117. [43] He XY, Merz G, Mehta P, Schulz H, Yang SY. Human brain short chain L-3- [71] DuSell CD, Umetani M, Shaul PW, Mangelsdorf DJ, McDonnell DP. 27- hydroxyacyl coenzyme A dehydrogenase is a single-domain multifunctional Hydroxycholesterol is an endogenous selective estrogen receptor modulator. enzyme. Characterization of a novel 17beta-hydroxysteroid dehydrogenase. J Mol Endocrinol 2008;22:65–77. Biol Chem 1999;274:15014–9. [72] Umetani M, Shaul PW. 27-Hydroxycholesterol: the first identified endogenous [44] Shafqat N, Marschall HU, Filling C, Nordling E, Wu XQ, Bjork L, et al. Expanded SERM. Trends Endocrinol Metab 2011;22:130–5. substrate screenings of human and Drosophila type 10 17beta-hydroxysteroid [73] Umetani M, Domoto H, Gormley AK, Yuhanna IS, Cummins CL, Javitt NB, et al. dehydrogenases (HSDs) reveal multiple specificities in bile acid and steroid 27-Hydroxycholesterol is an endogenous SERM that inhibits the hormone metabolism: characterization of multifunctional 3alpha/7alpha/ cardiovascular effects of estrogen. Nat Med 2007;13:1185–92. 7beta/17beta/20beta/21-HSD. Biochem J 2003;376:49–60. [74] Rose KA, Stapleton G, Dott K, Kieny MP, Best R, Schwarz M, et al. Cyp7b, a novel [45] Lathe R, Seckl JR. Neurosteroids and brain sterols. In: Mason JI, editor. Genetics brain cytochrome P450, catalyzes the synthesis of neurosteroids 7alpha- of steroid biosynthesis and function. Amsterdam: Harwood Academic; 2002. p. hydroxy dehydroepiandrosterone and 7alpha- hydroxy pregnenolone. Proc 405–72. Natl Acad Sci USA 1997;94:4925–30. [46] Mindnich R, Moller G, Adamski J. The role of 17 beta-hydroxysteroid [75] Martin KO, Reiss AB, Lathe R, Javitt NB. 7 Alpha-hydroxylation of 27- dehydrogenases. Mol Cell Endocrinol 2004;218:7–20. hydroxycholesterol: biologic role in the regulation of cholesterol synthesis 7 [47] Tavani DM, Nes WR, Billheimer JT. The sterol substrate specificity of acyl CoA: alpha-hydroxylation of 27-hydroxycholesterol: biologic role in the regulation cholesterol acyltransferase from rat liver. J Lipid Res 1982;23:774–81. of cholesterol synthesis. J Lipid Res 1997;38:1053–8. [48] Zhang Y, Yu C, Liu J, Spencer TA, Chang CC, Chang TY. Cholesterol is superior to [76] Setchell KD, Schwarz M, O’Connell NC, Lund EG, Davis DL, Lathe R, et al. 7-ketocholesterol or 7 alpha-hydroxycholesterol as an allosteric activator for Identification of a new inborn error in bile acid synthesis: mutation of the acyl-coenzyme A: cholesterol acyltransferase 1. J Biol Chem oxysterol 7alpha-hydroxylase gene causes severe neonatal liver disease. J Clin 2003;278:11642–7. Invest 1998;102:1690–703. [49] Lin DS, Steiner RD, Merkens LS, Pappu AS, Connor WE. The effects of sterol [77] Quinkler M, Diederich S. Difference of in vivo and in vitro antiminer structure upon sterol esterification. Atherosclerosis 2010;208:155–60. alocorticoid potency of progesterone. Endocr Res 2002;28:465–70. 22 R. Lathe, Y. Kotelevtsev / Steroids 82 (2014) 14–22

[78] Geller DS, Farhi A, Pinkerton N, Fradley M, Moritz M, Spitzer A, et al. Activating [86] Raychaudhuri S, Prinz WA. The diverse functions of oxysterol-binding mineralocorticoid receptor mutation in hypertension exacerbated by proteins. Annu Rev Cell Dev Biol 2010;26:157–77. pregnancy. Science 2000;289:119–23. [87] Lehto M, Laitinen S, Chinetti G, Johansson M, Ehnholm C, Staels B, et al. The [79] Quinkler M, Meyer B, Oelkers W, Diederich S. Renal inactivation, OSBP-related protein family in humans. J Lipid Res 2001;42:1203–13. mineralocorticoid generation, and 11beta-hydroxysteroid dehydrogenase [88] Fairn GD, McMaster CR. Emerging roles of the oxysterol-binding protein family inhibition ameliorate the antimineralocorticoid effect of progesterone in metabolism, transport, and signaling. Cell Mol Life Sci 2008;65:228–36. in vivo. J Clin Endocrinol Metab 2003;88:3767–72. [89] Weber-Boyat M, Olkkonen VM. Visualization of ORP–VAP complexes at [80] Escher G, Vogeli I, Escher R, Tuckey RC, Erickson S, Krozowski Z, et al. Role of putative membrane contact sites: novel clues to oxysterol binding-protein CYP27A1 in progesterone metabolism in vitro and in vivo. Am J Physiol function. Biochem Pharmacol 2013;86:89–95. Endocrinol Metab 2009;297:E949–55. [90] Schroeder F, Atshaves BP, McIntosh AL, Gallegos AM, Storey SM, Parr RD, et al. [81] Darney PD. The androgenicity of progestins. Am J Med 1995;98:104S–10S. Sterol carrier protein-2: new roles in regulating lipid rafts and signaling. [82] Wollam J, Magner DB, Magomedova L, Rass E, Shen Y, Rottiers V, et al. A novel Biochim Biophys Acta 2007;1771:700–18. 3-hydroxysteroid dehydrogenase that regulates reproductive development [91] de Launoit Y, Adamski J. Unique multifunctional HSD17B4 gene product: and longevity. PLoS Biol 2012;10:e1001305. 17beta-hydroxysteroid dehydrogenase 4 and D-3-hydroxyacyl-coenzyme A [83] Oram JF, Heinecke JW. ATP-binding cassette transporter A1: a cell cholesterol dehydrogenase/hydratase involved in Zellweger syndrome. J Mol Endocrinol exporter that protects against cardiovascular disease. Physiol Rev 1999;22:227–40. 2005;85:1343–72. [92] Hanson RM. Jmol – a paradigm shift in crystallographic visualization. J Appl [84] Schulz TA, Choi MG, Raychaudhuri S, Mears JA, Ghirlando R, Hinshaw JE, et al. Crystallogr 2010:1250–60. Lipid-regulated sterol transfer between closely apposed membranes by [93] Kotelevtsev Y, Brown RW, Fleming S, Kenyon C, Edwards CR, Seckl JR, et al. oxysterol-binding protein homologues. J Cell Biol 2009;187:889–903. Hypertension in mice lacking 11beta-hydroxysteroid dehydrogenase type 2. J [85] Lev S. Non-vesicular lipid transport by lipid-transfer proteins and beyond. Nat Clin Invest 1999;103:683–9. Rev Mol Cell Biol 2010;11:739–50.