Phytoestrogens As Inhibitors of Fungal 17Я-Hydroxysteroid Dehydrogenase

Steroids 70 (2005) 694–703

Phytoestrogens as inhibitors of fungal 17␤-hydroxysteroid dehydrogenase

Katja Kristan a, Katja Krajnc b, Janez Konc b, Stanislav Gobec b, Jure Stojan a, Tea Lanisnikˇ Riznerˇ a,∗

a Institute of Biochemistry, Medical Faculty, University of Ljubljana, Vrazov trg 2, 1000 Ljubljana, Slovenia b Faculty of Pharmacy, University of Ljubljana, Aˇskerˇceva 7, 1000 Ljubljana, Slovenia Received 3 November 2004; received in revised form 25 February 2005; accepted 28 February 2005 Available online 4 June 2005

Abstract

Different phytoestrogens were tested as inhibitors of 17␤-hydroxysteroid dehydrogenase from the fungus Cochliobolus lunatus (17␤- HSDcl), a member of the short-chain dehydrogenase/reductase superfamily. Phytoestrogens inhibited the oxidation of 100 ␮M17␤- hydroxyestra-4-en-3-one and the reduction of 100 ␮M estra-4-en-3,17-dione, the best substrate pair known. The best inhibitors of oxidation, with IC50 below 1 ␮M, were flavones hydroxylated at positions 3, 5 and 7: 3-hydroxyflavone, 3,7-dihydroxyflavone, 5,7-dihydroxyflavone (chrysin) and 5-hydroxyflavone, together with 5-methoxyflavone. The best inhibitors of reduction were less potent; 3-hydroxyflavone, 5- methoxyflavone, coumestrol, 3,5,7,4 -tetrahydroxyflavone (kaempferol) and 5-hydroxyflavone all had IC50 values between 1 and 5 ␮M. Docking the representative inhibitors chrysin and kaempferol into the active site of 17␤-HSDcl revealed the possible binding mode, in which they are sandwiched between the nicotinamide moiety and Tyr212. The structural features of phytoestrogens, inhibitors of both oxidation and reduction catalyzed by the fungal 17␤-HSD, are similar to the reported structural features of phytoestrogen inhibitors of human 17␤-HSD types 1 and 2. © 2005 Elsevier Inc. All rights reserved.

Keywords: Phytoestrogens; Flavonoids; Hydroxysteroid dehydrogenase; Short-chain dehydrogenase/reductase superfamily; Fungi; Cochliobolus lunatus

1. Introduction ductases (4HNR) from Magnaporthe grisea, Ophiostoma ﬂoccosum and other fungi involved in melanin biosynthesis 17␤-Hydroxysteroid dehydrogenase from the ﬁlamentous [2]. Sequence similarity was found also to bacterial 7␣-HSD fungus Cochliobolus lunatus (17␤-HSDcl) is an NADPH- and even to human 17␤-HSD types 4 and 8 [2]. dependent enzyme that catalyses the oxidoreduction prefer- We are studying 17␤-HSDcl as a model enzyme of the entially of estrogens and androgens, with estra-4-en-3,17- SDR superfamily. This superfamily includes about 3000 en- dione and 17␤-hydroxyestra-4-en-3-one as the best pair of zymes involved in steroid hormone metabolism, fatty acid substrates [1]. The enzyme belongs to the short-chain de- oxidation and biotransformation of xenobiotics [3]. HSDs hydrogenase/reductase superfamily (SDR) [2]. It possesses a from the SDR superfamily are implicated in the develop- high percentage of sequence identity to fungal ketoreductases ment of polycystic kidney disease [4], regulation of blood – the versicolorin reductases from Aspergillus parasiticus and pressure [5], Alzheimer’s disease [6], obesity [7] and steroid- Emericella nidulans (Ver-1 and Ver-A) involved in myco- dependent forms of cancer [8]. Of the hormone-dependent toxin biosynthesis – and to the 1,3,8-trihydroxynaphthalene forms of cancer, huge geographical differences in incidence reductases (3HNR) and 1,3,6,8-tetrahydroxynaphthalene re- were reported for breast, endometrial and prostate cancer [9]. The incidence of these diseases is much higher in Northern ∗ Corresponding author. Tel.: +386 1543 7657; fax: +386 1543 7641. Europe and USA than in Asia [9,10]. These differences were E-mail address: [email protected] (T.L. Rizner).ˇ correlated with environmental and dietary factors, especially

0039-128X/$ – see front matter © 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.steroids.2005.02.023 K. Kristan et al. / Steroids 70 (2005) 694–703 695 with phytoestrogens that are abundant in soy products, a ma- [31]. The enzyme was homogeneous by SDS-PAGE with jor component of the Asian diet [9,10]. Coomassie blue staining. Phytoestrogens are plant-derived, non-steroidal compounds with estrogenic activity that can be divided into 2.3. Enzyme assay three structural groups: flavonoids, coumestans and lignans. Flavonoids comprise flavones, flavanones and isoflavones Phytoestrogens were tested for their inhibitory action [11]. Phytoestrogens possess antiviral, anti-inflammatory, an- against recombinant 17␤-HSDcl. 17␤-HSDcl catalyzes the timutagenic and anticarcinogenic activities [12] and have dif- oxidation of 17␤-hydroxyestra-4-en-3-one to estra-4-en- ferent mechanisms of action. They can interfere with the com- 3,17-dione in the presence of NADP+ and the reduction of plex human endocrine system in several ways: estra-4-en-3,17-dione to 17␤-hydroxyestra-4-en-3-one in the presence of coenzyme NADPH. The reaction was followed (1) they can act as agonists or antagonists of estrogen recep- spectrophotometrically by measuring NADPH concentration tors (ER␣ and ER␤), pregnane X receptor and constitu- (absorbance at 340 nm) in the absence and presence of dif- tive androstane receptor [13], ferent phytoestrogens. Assays were carried out in 0.6 ml sys- (2) they have stimulatory effects on hepatic sex hormone tems in 100 mM phosphate buffer, pH 8.0, containing 1% binding globulin (SHBG), DMSO as co-solvent, as described earlier [32]. Both the sub- (3) they inhibit tyrosine kinase, and thus, prevent growth strate and the coenzyme were 100 ␮M, the concentrations factor-mediated stimulation of proliferation, of inhibitors were from 0.1 to 100 ␮M and the enzyme was (4) they modulate activity of the key enzymes in the biosyn- 0.5 ␮M. Initial velocities were calculated and percentage in- thesis and metabolism of steroid hormones [14] and in hibition was given by 100 − [(v (with inhibitor)/v (without this manner act at the pre-receptor level. 0 0 inhibitor)) × 100]. IC50 values were determined graphically Among steroid metabolizing enzymes, the inhibitory ef- from a plot of log10 (inhibitor concentration) versus percent- fect of phytoestrogens has been studied against aromatase, age inhibition using GraphPad Prism Version4.00 (GraphPad sulfatase, sulfotransferases, 5␣-reductase, 3␤-HSD 5/4 Software Inc.). isomerase, 11␤-HSD type 1 and 17␤-HSDs [12,15–20].Phy- toestrogens were shown to inhibit human 17␤-HSD types 1, 2.4. Docking of kaempferol and chrysin into the active 2, 3 and 5 [21–28]. There are, however, no reports on phytoe- site of 17β–HSDcl and 17β-HSD type 1 strogen inhibition of other human 17␤-HSD isoforms (types 4, 7, 8, 10 or 11) [29,30]. Docking was performed using Autodock3 computer In this work, we examined the action of phytoestrogens program [33]. Homology built model of 17␤-HSDcl was at the pre-receptor level by testing their inhibitory poten- used with Kollman charges and solvent parameters added tial towards the fungal 17␤-HSD. We tested the inhibitory by Autodock. The ligands kaempferol and chrysin were action of flavones, flavanones, isoflavones, coumestans and modeled with Molden [34] and further refined by Gaussian coumarins, two plant-derived organic acids, one mycotoxin 03 in vacuum. Planar structures were obtained which were and three synthetic estrogens/antiestrogens. Of these, coume- then submitted for docking as rigid bodies. Additionally, strol and different hydroxylated flavones exerted the highest CHARMM [35] force field parameters, reproducing ab initio inhibitory action. geometry were worked out. Prior to docking of the two inhibitors, the coenzyme molecule had to be added. We used, as a template, the position of NADPH from the structure 2. Experimental of the complex between THNR, coenzyme and tricyclazole (1YBV). After insertion of the coenzyme, it was relaxed in a 2.1. Phytoestrogens, substrates and cofactors constrained protein by using molecular mechanics optimization routines (50 steps of steepest descent, SD, followed Phytoestrogens originally purchased from ICN Biochem- by 50 steps of Gauss–Newton, ABNR). Subsequently, the icals GmbH, Steraloids Inc. and Sigma Aldrich Chemie complex structure was refined by 30 steps of Gauss–Newton GmbH were the kind gift of Dr. Jerzy Adamski (GSF, optimization using QMMM algorithm, without any con- Neuherberg, Germany). Substrates estra-4-en-3,17-dione and strains (QM atoms: Tyr, coenzyme). After docking of 17␤-hydroxyestra-4-en-3-one and coenzymes NADP+ and inhibitors, CHARMM molecular simulation program was NADPH were from Sigma Aldrich Chemie, GmbH. employed once again: molecular mechanics relaxation of coenzyme and inhibitor, followed by QMMM optimiza- 2.2. Preparation of recombinant 17β-HSDcl tion without any constrains (QM atoms: Tyr, coenzyme, inhibitor). We repeated this procedure for both the oxidized 17␤-HSDcl was expressed as a GST-fusion protein in Es- and the reduced coenzyme forms. The charges and solvent cherichia coli BL21 cells and purified as previously described parameters for the two coenzyme forms were generated by [2]. Protein concentration of the enzyme preparation was Autodock while the charges for the inhibitors were taken determined by the Bradford method with BSA as standard from Gaussian minimization. In all quantum mechanical 696 K. Kristan et al. / Steroids 70 (2005) 694–703 calculations with Gaussian and CHARMM, we used 6–31 g 100 ␮M coenzyme NADP+ (Table 1). The most effective in- basis set. hibitors, with IC50 around 1 ␮M were: 3-hydroxyflavone, 3,7- Also, docking of kaempferol into the crystal structure of dihydroxyflavone, 5-methoxyflavone, 5,7-dihydroxyflavone ␤ human 17 -HSD type 1 (PDB code 1EQU) was performed (chrysin) and 5-hydroxyflavone. Inhibitors with IC50 be- as described above. The atomic coordinates of docked in- tween 1 and 10 ␮M were: flavone, quercetin, kaempferol, hibitors into the 17␤-HSDcl model and the crystal structure apigenin, coumestrol and biochanin A. Others, luteolin, 7- of 17␤-HSD type 1 can be downloaded from the website hydroxyflavone, flavanone, naringenin and mycoestrogen ∼ http://www2.mf.uni-lj.si/ stojan/stojan.html. zearalenone had IC50 values above 10 ␮M(Table 1). Flavones hydroxylated at positions 3, 5 and 7 or methoxy- lated at position 5 were the best inhibitors in the oxidative di- 3. Results and discussion rection. The only exception was 7-hydroxyflavone with IC50 of 15 ␮M. The absence of hydroxyl groups did not drastically We tested 23 compounds, 19 plant-derived estrogenic change the inhibitory effect of flavones—the IC50 value of compounds (flavones, flavanones, isoflavones, coumestans, flavone was still below 2 ␮M. Flavones with an additional hy- coumarin and organic acids), one estrogenic mycotoxin (zear- droxyl groups at the 4 or the 3 and 4 positions (quercetin, alenone) and three synthetic estrogens/antiestrogens (tamox- apigenin, luteolin) were not as potent, and isoflavone ifen, diethylstilbestrol and equilin) as inhibitors of fungal (biochanin A) and flavanone that lack one double bond in 17␤-HSD. These substances were tested in the oxidative (ox- ring C (naringenin), were even less effective inhibitors. idation of 17␤-hydroxyestra-4-en-3-one in the presence of + NADP ) and reductive directions (reduction of estra-4-en- 3.2. Phytoestrogens as inhibitors of reduction 3,17-dione in the presence of and NADPH). Phytoestrogens were less inhibitory on the reduction of 3.1. Phytoestrogens as inhibitors of oxidation 100 ␮M estra-4-en-3,17-dione in the presence of 100 ␮M NADPH (Table 2). 3-Hydroxyflavone, 5-methoxyflavone, The majority of phytoestrogens inhibited the oxidation coumestrol, 3,5,7,4 tetrahydroxyflavone (kaempferol) and of 100 ␮M17␤-hydroxyestra-4-en-3-one in the presence of 5-hydroxyflavone, with IC50 values below 5 ␮M, were the

Table 1 Inhibitors of oxidation with IC50 values below 100 ␮M

Structural formulas and corresponding IC50 values are shown. K. Kristan et al. / Steroids 70 (2005) 694–703 697

Table 2 Inhibitors of reduction with IC50 values below 100 ␮M

Structural formulas and corresponding IC50 values are shown. most effective inhibitors, while 3,7-dihydroxyflavone, 5,7- weak inhibitors only in one direction. While genistein effec- dihydroxyflavone, quercetin and flavone had IC50 values be- tively inhibited the reduction (IC50 56.6 ␮M), zearalenone tween 5 and 10 ␮M. Among other substances, biochanin A, was a potent inhibitor only in the oxidative direction (IC50 luteolin, naringenin, 7-hydroxyflavone, apigenin and genis- 12 ␮M). We could not detect any inhibitory action in either tein showed IC50 values still below 100 ␮M(Table 2). direction for coumarin and the synthetic antiestrogen, equilin 3-Hydroxy and 5-methoxyflavone were the best inhibitors (Table 3). in both the reductive and oxidative directions. Other potent The weakly inhibitory isoflavone, daidzein, differs from inhibitors were compounds with similar structural features to genistein (IC50 56.6 ␮M) only in the absence of an addi- the substrates, with hydroxyl groups at positions 3 and 9 of tional hydroxyl group at position 5. It differs from another rings A and D (coumestrol) or a hydroxyl group at 4 posi- isoflavone, biochanin A, a potent inhibitor in the oxidative tion of the flavonic B ring (kaempferol). The slightly lower and reductive directions (IC50 7.3 and 11.5 ␮M, respectively), inhibitory action of phytoestrogens in the reductive direction in the absence of a 5-hydroxyl group and a methoxyl group may be explained by their weaker binding to the active site at the 4 position. It appears the hydroxylation at position 5 in the presence of NADPH. Although the difference might leads to some inhibitory action (genistein), but only methyla- be insignificant, one would expect tighter packing of the lig- tion of the 4-hydroxyl group enables appropriate accommo- ands in the presence of NADP+ with aromatic nicotinamide dation of isoflavones (biochanin A), thus further increasing moiety. their inhibitory effect. Flavanone, another weak inhibitor, differs from the po- 3.3. Structural features of phytoestrogens with little or tent flavones only in the saturation of ring C. A non-planar no inhibitory effect on fungal 17β-HSD ring C may prevent favorable orientation of ring B, which could, in turn, prevent proper binding and inhibitory action. Some of the compounds like daidzein, flavanone, plant- Zearalenone inhibited 17␤-HSD only in the oxidative direc- derived glycyrrhetinic and abietic acid, but also synthetic es- tion. Since it is oxidized with a very low turnover by the trogens/antiestrogen tamoxifen and diethylstilbestrol, were fungal 17␤-HSD (see below), it may compete effectively weak inhibitors with IC50 values above 100 ␮M in both direc- only with the substrate 17␤-hydroxyestra-4-en-3-one and in tions (Table 3). Others, like genistein and zearalenone, were this manner inhibit just the oxidation. A steroid equilin, 3- 698 K. Kristan et al. / Steroids 70 (2005) 694–703

Table 3 Weak inhibitors and non-inhibitors

∗ Structural formulas of substances with little (IC50 above 100 ␮M) or no inhibitory effect on fungal 17␤-HSD. IC50 above 100 ␮M only in the oxidative ** *** direction. IC50 above 100 ␮M only in the reductive direction. Non-inhibitor. hydroxyestra-1,3,5,7-tetraen-17on-estron, with an additional is a facultative pathogen for monocotyledonous crops, double bond in ring B, a potent inhibitor of 17␤-HSD type causing leaf blight on Sorghum, leaf spot on Cannabis sativa 1, failed to inhibit the fungal enzyme. The model structure and seedling blight on Saccharum spp. [37]. Reports of of 17␤-HSDcl suggested that Val161, speciﬁc for the fungal human disease caused by C. lunatus are rare but include enzyme, might sterically hinder the binding of equilin into endocarditis, brain abscess, skin infections, onychomycosis, the active site. Finally, coumarin appears to be too small to keratitis, pneumonia, disseminated disease, mycetoma, ﬁrmly accommodate into the active site of fungal 17␤-HSD. allergic bronchopulmonary disease and sinusitis [38–40]. Being a facultative plant pathogen and occasionally even 3.4. Phytoestrogens as putative substrates of 17β-HSD human pathogen, the fungus C. lunatus and its 17␤-HSD from the fungus Cochliobolus lunatus might encounter phytoestrogens produced by plants or even human steroid hormones. Since the physiological role of C. lunatus (anamorph Curvularia lunata) is a saprobic this fungal enzyme is not yet known, one could expect that dematiaceous mold that resides primarily in soil [36].It in addition to steroid hormones, phytoestrogens might be

Table 4 Distances between active site residues or NADP(H) and inhibitors in the model of 17␤-HSDcl O␩Tyrl67- O␥Serl53- NC4-C5/ C5/C4 (A)˚ C5/C4 (A)˚ C4 (A)˚

4.23 4.45 3.41 4.55 3.99 3.33

4.42 4.81 3.34 3.61 4.67 3.37 K. Kristan et al. / Steroids 70 (2005) 694–703 699 recognized not only as inhibitors, but also as substrates of this We examined all these compounds as putative substrates of enzyme. fungal 17␤-HSD. It was found that only the mycotoxin zear- Examination of the structural features of phytoestrogens alenone was oxidized, but with approximately 10-fold lower included in this study showed that flavones, isoflavones and rate than 17␤-hydroxyestra-4-en-3-one, the best steroidal flavanones are potential candidates for reduction at the corre- substrate of fungal 17␤-HSD. It seems, therefore, that zear- sponding 4-keto groups. Coumestrol, cumarin, plant-derived alenone is most probably not the physiological substrate of organic acids, zearalenone and equilin also possess keto this enzyme. groups that may be reduced. On the other hand, phytoestrogens and organic acids cannot be oxidized; only the es- 3.5. Docking phytoestrogens into the active site of trogenic mycotoxin zearalenone possesses one oxidizable 17β-HSDcl group. The synthetic estrogens/antiestrogens tamoxifen and diethylstilbestrol have no functional groups that can undergo Inhibitors of reduction and oxidation, chrysin and kaem- reduction or oxidation. pferol, were docked into the homology model of 17␤-HSDcl

Fig. 1. Docking phytoestrogens into the active site of fungal 17␤-HSD. Stereo view of the phytoestrogen binding site of 17␤-HSDcl showing the docked coenzyme NADP(H), amino acid residues within contact distance of 4.5 A˚ and two orientations of (A) chrysin (red and yellow) and (B) kaempferol (blue and orange) (For interpretation of the references to colour in this ﬁgure legend, the reader is referred to the web version of the article.). 700 K. Kristan et al. / Steroids 70 (2005) 694–703 using Autodock. Docking revealed that inhibitors of oxida- within the contact distance of 4.5 A:˚ Ala228, Ala231, Ala270, tion and reduction accommodate into the substrate-binding Asn154, Thr155, His164, Val107, Pro197, Gly198, Gly199, region of the active site in similar manner as previously sug- Thr202, Met204, Phe205 and Val208 (Fig. 1; Table 4). The gested for androstenedione [41]. bond between rings A/C and ring B allows for a certain degree Fig. 1 shows the position of the coenzyme and of the im- of rotation, but the planar conformation of kaempferol in the portant amino acid residues in the active site of the fungal crystal structure of quercetin 2,3-dioxygenase (PDB code: 17␤-HSD, together with the docked inhibitors. It can be seen 1H1M) strongly suggests the same situation in our enzyme that both kaempferol and chrysin may adopt two different (Fig. 1). positions where (1) the hydroxylated rings A and C or (2) Different charges, assigned for two coenzyme forms ring B with (kaempferol) or without (chrysin) a 4 -OH group (NADP+/NADPH), did not affect the docking predicted by are oriented towards the catalytic amino acid residues Tyr167 Autodock. Also, presence of an additional hydroxyl group and Ser153. Both chrysin and kaempferol are sandwiched be- at the B ring of kaempferol did not inﬂuence the positions tween the nicotinamide moiety of the coenzyme and Tyr212, substantially (Fig. 1; Table 4). On the basis of docking at- and are surrounded by the following amino acid residues tempts by Autodock, we could not clearly explain the weaker

Fig. 2. Sequence alignment of 17␤-HSDcl, 3HNR, 17␤-HSD types 1 and 2. Alignment of amino acid sequences of four members of the SDR superfamily: 17␤- HSD from C. lunatus (17␤-HSDcl), trihydroxynaphthalene reductase from M. grisea (3HNR) and human 17␤-HSD types 1 (17␤-HSD1) and 2 (17␤-HSD2). Asterisks indicate the identical amino acids; amino acids of the steroid/phytoestrogen binding site are in bold. K. Kristan et al. / Steroids 70 (2005) 694–703 701 binding of phytoestrogens to the active site in the presence When we compared the structural features of phytoestro- of NADPH, but this issue might be resolved by further crys- gens, inhibitors of reduction and oxidation catalyzed by the tallization studies that are underway. fungal 17␤-HSD, with those of phytoestrogenic inhibitors of human 17␤-HSD isoenzymes, we saw striking similarities. 3.6. Comparison of phytoestrogens as inhibitors of The best inhibitors of oxidation were flavones hydroxylated fungal and human 17β-HSDs at positions 3, 5 and 7, similar to those reported for the oxidative isoform 17␤-HSD type 2. The best inhibitors of reduc- Our results on phytoestrogen inhibition were compared tion were coumestrol, kaempferol and flavones hydroxylated with those previously reported on the inhibition of human at positions 3, 5, 7 and/or 4, resembling inhibitors of both 17␤-HSDs types 1, 2, 3 and 5. The most effective inhibitor of types 1 and 2 17␤-HSD. There was less similarity with in- the reductive isoform 17␤-HSD type 1 was coumestrol. Of hibitors of 17␤-HSD types 3 and 5. the flavones and isoflavones, those hydroxylated at positions Fungal 17␤-HSD possesses only 21.2 and 18.1% identity 4, 5 and 7, with or without a hydroxyl group at position 3 to the human 17␤-HSD types 1 and 2, respectively (Fig. 2). of ring C (kaempferol, apigenin, naringenin, genistein), were In order to understand the structural similarity between in- also very potent [12,18,22]. The structural demands for the inhibitors of fungal and human enzymes, we compared the 17␤- hibition of the oxidative isoform 17␤-HSD type 2 were differ- HSDcl model structure with the crystal structures of 17␤- ent. Flavones hydroxylated at positions 3, 5 and 7 of rings A HSD type 1 in complex with different ligands (PDB codes and C were the most effective, while a hydroxyl or methoxyl 1EQU, 1A27). As expected, the active sites differ in amino group at position 4 of ring B had no effect on the inhibitory acid composition (except for both catalytic Tyr155/Tyr167 properties [19]. In contrast to the best inhibitors of 17␤-HSD and Ser142/Ser153 as well as Tyr218/Tyr212). In both type 1, the hydroxyl group at position 3 appeared to be criti- enzymes, steroidal substrate binds in a hydrophobic tun- cal for inhibitors of type 2 [19]. The reductive 17␤-HSD type nel containing some H-bond forming residues. Docking of 3 was strongly inhibited by 7-hydroxyflavone, biochanin A kaempferol into the crystal structure of human 17␤-HSD and 7-hydroxycoumarins [15,16], while the reductive 17␤- type 1 (PDB code 1EQU) revealed that kaempferol occupies HSD type 5, the only 17␤-HSD that belongs to the aldo–keto the equilin-binding cavity (Fig. 3). Among different bind- reductase superfamily, was inhibited by zearalenone, coume- ing modes, the position where 7-OH group points towards strol, the flavone quercetin and the isoflavone biochanin A the catalytical Tyr155 and Ser142, 5-OH and 4-keto groups [20,21]. Thus, changes in the number and location of hy- face Tyr218 and Ser222, and the 4-OH group lays close to droxyl groups discriminate between inhibitors of different Glu282 and His221 explains why flavones with 4-OH group isozymes. act as the best flavonic inhibitors of this enzyme. Glu282

Fig. 3. Comparison of phytoestrogen binding sites. Schematic view of the interactions between kaempferol and amino acid residues within 4.5 A˚ in (A) fungal 17␤-HSDcl and (B) human 17␤-HSD type 1. Distances between non-hydrogen atoms are in A.˚ 702 K. Kristan et al. / Steroids 70 (2005) 694–703

(and His221) may additionally stabilize 4 -hydroxyflavones [10] Tham DM, Christopher D, Gardner D, Haskell WL. Potential health (as well as coumestrol) similarly, as reported for 3-OH group benefits of dietary phytoestrogens: a review of the clinical, epi- of estrogens [42–44]. demiological, and mechanistic evidence. J Clin Endocrinol Metab 1998;83:2223–35. Comparison of the active sites of fungal and human en- [11] Ososki AL, Kennelly EJ. Phytoestrogens: a review of the present zymes thus revealed phytoestrogens bind in different ori- state of research. Phytother Res 2003;17:845–69. entations, both occupying the steroid binding pocket. The [12] Le Bail J-C, Champavier Y, Chulia A-J, Habrioux G. Effects of major difference we noticed is the presence of Glu282 (and phytoestrogens on aromatase, 3␤ and 17␤-hydroxysteroid dehydro- His221) in human enzyme, which seems to determine enzyme genase activities and human breast cancer. Life Sci 2000;66:1281– ␤ 91. selectivity for 4 -hydroxyflavones. Fungal 17 -HSD has no [13] Jacobs MN, Lewis DF. Steroid hormone receptors and dietary lig- Glu282/His221 corresponding residues; therefore, the stabi- ands: a selective review. Proc Nutr Soc 2002;61:105–22. lization of phytoestrogens in this manner is not expected. [14] Wuttke W, Jarry H, Westphalen S, Christoffel V, Seidlova-Wuttke D. Phytoestrogens with (kaempferol) or without (chrysin) a 4- Phytoestrogens for hormone replacement therapy? J Steroid Biochem OH group may thus bind in two modes where those without Mol Biol 2003;83:133–47. [15] Kirk CJ, Harris RM, Wood DM, Waring RH, Hughes PJ. Do dietary 4 -OH group show lower IC50 values. phytoestrogens influence susceptibility to hormone-dependent cancer Finally, it should be emphasized that the lowest IC50 val- by disrupting the metabolism of endogenous estrogens? Biochem Soc ues determined for phytoestrogens were in the low ␮M range. Trans 2001;29:209–15. This is about 100-fold lower than the optimal substrate con- [16] Ohno S, Matsumoto N, Watanabe M, Nakajin S. Flavonoid inhibition ␤ centration for our spectrophotometric assay. Additionally, the of overexpressed human 3 -hydroxysteroid dehydrogenase type II. J Steroid Biochem Mol Biol 2004;88:175–82. relatively high enzyme concentration used shows that most [17] Ohno S, Shinoda S, Toyoshima S, Nakazawa H, Makino T, Nakajin of the compounds tested are, in fact, tight binding inhibitors S. Effects of flavonoid phytochemicals on cortisol production and on acting in equimolar concentration ratios. activities of steroidogenic enzymes in human adrenocortical H295R cells. J Steroid Biochem Mol Biol 2002;80:355–63. [18] Lephart ED, Lund TD, Horvath TL. Brain androgen and progesterone metabolizing enzymes: biosynthesis, distribution and function. Brain Acknowledgments Res Rev 2001;37:25–37. [19] Weber KS, Jacobson NA, Setchell KD, Lephart ED. Brain aro- This work was supported by the Ministry of Education, matase and 5alpha-reductase, regulatory behaviors and testosterone Science and Sport of Slovenia. The authors thank Dr. Jerzy levels in adult rats on phytoestrogen diets. Proc Soc Exp Biol Med 1999;221:131–5. Adamski (GSF—National Research Centre for Health and [20] Schweizer RA, Atanasov AG, Frey BM, Odermatt A. A rapid screen- Environment, Institute of Experimental Genetics, Neuher- ing assay for inhibitors of 11beta-hydroxysteroid dehydrogenases berg, Germany) for donating phytoestrogens. (11beta-HSD): flavanone selectively inhibits 11beta-HSD1 reductase activity. Mol Cell Endocrinol 2003;212:41–9. [21] Le Lain R, Nicholls PJ, Smith HJ, Mahrlouie FH. Inhibitors of human and rat testes microsomal 17␤-hydroxysteroid dehydrogenase References (17␤-HSD) as potent agents for prostatic cancer. J Enzyme Inhib 2001;16:35–45. [1] Lanisnikˇ Riznerˇ T, Stojan J, Adamski J. 17␤-Hydroxysteroid de- [22] Le Lain R, Barrell KJ, Saeed GS, Nicholls PJ, Simons C, Kirby A, et hydrogenase from the fungus Cochliobolus lunatus: structural and al. Some coumarins and triphenylethene derivatives as inhibitors of functional aspects. Chem Biol Interact 2001;130–132:793–803. human testes microsomal 17␤-hydroxysteroid dehydrogenase (17␤- [2] Lanisnikˇ Riznerˇ T, Moeller G, Thole TH, Zakelj-Mavriˇ cM,ˇ HSD type 3): further studies with tamoxifen on the rat testes micro- Adamski J. A novel 17␤-hydroxysteroid dehydrogenase in the fun- somal enzyme. J Enzyme Inhib 2002;17:93–100. gus Cochliobolus lunatus: new insights into the evolution of steroid- [23] Le Bail J-C, Pouget C, Fagnere C, Basly J-P, Chulia A-J, Habri- hormone signalling. Biochem J 1999;337:425–31. oux G. Chalcones are potent inhibitors of aromatase and 17␤- [3] Oppermann UCT, Filling C, Jornvall H. Forms and functions of hydroxysteroid dehydrogenase activities. Life Sci 2001;68:751–61. human SDR enzymes. Chem Biol Interact 2001;130–132:699–705. [24] Makela S, Poutanen M, Lehtimaki J, Kostian M-L, Santii R, [4] Fomitcheva J, Baker ME, Anderson E, Lee GY, Aziz N. Charac- Vihko R. Estrogen-specific 17␤-hydroxysteroid oxidoreductase type terization of Ke6, a new 17␤-hydroxysteroid dehydrogenase, and its 1 (E.C.1.1.1.62) as a possible target for the action of phytoestrogens. expression in gonadal tissues. J Biol Chem 1998;273:22664–71. Proc Soc Exp Biol Med 1995;208:51–9. [5] Tedde R, Krozowski ZS, Pala A, Li KX, Shackleton CH, Mantero [25] Makela S, Poutanen M, Kostian ML, Lehtimaki N, Strauss L, San- F, et al. Molecular basis for hypertension in the “type II variant” of tii R, et al. Inhibiton of 17␤-hydroxysteroid oxidoreductase by apparent mineralocorticoid excess. Am J Hum Genet 1998;63:370–9. flavonoids in breast and prostate cancer cells. Proc Soc Exp Biol [6] He XZ, Merz G, Mehta P, Schultz H, Yang SY. Human brain short- Med 1998;217:306–10. chain l-3-hydroxyacyl coenzyme A dehydrogenase is a single do- [26] Krazeisen A, Breitling R, Moeller G, Adamski J. Phytoestrogens main multifunctional enzyme. J Biol Chem 1999;274:15014–9. inhibit human 17␤-hydroxysteroid dehydrogenase type 5. Mol Cell [7] Masuzaki H, Peterson J, Shinyama H, Morton NM, Mullins JJ, Seckl Endocrinol 2001;171:151–62. JR, et al. A transgenic model of visceral obesity and the metabolic [27] Krazeisen A, Breitling R, Moeller G, Adamski J. Human 17␤- syndrome. Science 2001;294:2071–2. hydroxysteroid dehydrogenase type 5 is inhibited by dieatary [8] Lin SX, Han Q, Azzi A, Zhu D, Gangloff A, Campbell RL. 3D- flavonoids. Adv Exp Med Biol 2002;505:151–61. structure of human estrogenic 17␤-HSD1: binding with various [28] Poirier D. Ihibitors of 17␤-hydroxysteroid dehydrogenases. Curr steroids. J Steroid Biochem Mol Biol 2000;73:183. Med Chem 2003;10:453–77. [9] Adlercreutz H, Mazur W. Phyto-estrogens and western disease. Ann [29] Peltoketo H, Luu-The V, Simard J, Adamski J. 17␤-Hydroxysteroid Med 1997;29:95–120. dehydrogenase (HSD)/17-ketosteroid reductase (KSR) family; K. Kristan et al. / Steroids 70 (2005) 694–703 703

nomenclature and main characteristics of the 17HSD/KSR enzymes. lunata and a mixture of the two fungi. Crop Protect 2003;22:623– J Mol Endocrinol 1999;23:1–11. 8. [30] Midnich R, Moeller G, Adamski J. The role of 17beta-hydro- [38] Fernandez M, Noyola DE, Rossmann SN, Edwards MS. Cuta- xysteroid dehydrogenases. Mol Cell Endorinol 2004;218:7–20. neous phaeohyphomycosis caused by Curvularia lunata and a re- [31] Bradford MM. A rapid and sensitive method for the quantitation of view of Curvularia infections in pediatrics. Pediatr Infect Dis J microgram quantities of protein utilizing the principle of protein-dye 1999;18:727–31. binding. Anal Biochem 1976;72:248–54. [39] Kaushik S, Ram J, Chakrabarty A, Dogra MR, Brar GS, Gupta A. [32] Gobec S, Sova M, Kristan K, Lanisnikˇ Riznerˇ T. Cinnamic acid Curvularia lunata endophthalmitis with secondary keratitis. Am J esters as potent inhibitors of fungal 17␤-hydroxysteroid dehydroge- Ophthalmol 2001;131:140–2. nase – a model enzyme of the short-chain dehydrogenase/reductase [40] Janaki C, Sentamilselvi G, Janaki VR, Devesh S, Ajithados K. superfamily. Bioorg Med Chem Lett 2004;14:3933–6. Case report. Eumycetoma due to Curvularia lunata. Mycoses [33] Morris GM, Goodsell DS, Halliday RS, Huey R, Hart WE, Belew 1999;42:345–6. RK, et al. Automated docking using a lamarckian genetic algo- [41] Lanisnikˇ Riznerˇ T, Adamski J, Stojan J. 17␤-Hydroxysteroid dehy- rithm and empirical binding free energy function. J Comput Chem drogenase from Cochliobolus lunatus: model structure and substrate 1998;19:1639–62. speciﬁcity. Arch Biochem Biophys 2000;384:255–62. [34] Schaftenaar G, Noordik JH. Molden: a pre- and post-processing pro- [42] Azzi A, Rehse PH, Zhu D-W, Campbell RL, Labrie F, Lin S-X. gram for molecular and electronic structures. J Comput-Aided Mol Crystal structure of human estrogenic 17␤-hydroxysteroid dehydro- Design 2000;14:123–34. genase complexed with 17␤-estradiol. Nat Struct Biol 1996;3:665–8. [35] Brooks BR, Bruccoleri RE, Olafson BD, States DJ, Swaminathan S, [43] Sawicki MW, Erman M, Puranen T, Vihko P, Ghosh D. Structure Karplus M. CHARMM: a program for macromolecular energy, mini- of the ternary complex of human 17␤-hydroxysteroid dehydroge- mization, and dynamics calculations. J Comp Chem 1983;4:187–217. nase type 1 with 3-hydroxyestra-1,3,5,7-tetraen-17-one (equilin) and [36] Rinaldi MG, Phillips P, Schwartz JG, Winn RE, Holt GR, Shagets NADP+. Proc Natl Acad Sci USA 1999;96:840–5. FW, et al. Human Curvularia infections. Report of ﬁve cases and [44] Shi R, Lin S-X. Cofactor hydrogen bonding onto the protein review of the literature. Diagn Microbiol Infect Dis 1987;6:27–39. main chain is conserved in the short-chain dehydrogenase/reductase [37] Prom L, Waniska R, Kollo A, Rooney W. Response of eight family and contributes to nicotinamide orientation. J Biol Chem sorghum cultivars inoculated with Fusarium thapsinum, Curvularia 2004;279:16778–85.