Coenzyme specificity in fungal 17β-hydroxysteroid dehydrogenase

Molecular and Cellular Endocrinology 241 (2005) 80–87

Coenzyme specificity in fungal 17␤-hydroxysteroid dehydrogenase Katja Kristan a , Jure Stojan a , Gabriele M¨oller b , Jerzy Adamski b , Tea Laniˇsnik Riˇzner a,∗ b

a Institute of Biochemistry, Medical Faculty, University of Ljubljana, Vrazov trg 2, 1000 Ljubljana, Slovenia GSF–National Research Center for Environment and Health, Institute of Experimental Genetics, Genome Analysis Center, Ingolst¨adter Landstraße 1, 85764 Neuherberg, Germany

Received 18 March 2005; received in revised form 20 May 2005; accepted 20 May 2005

Abstract The 17␤-hydroxysteroid dehydrogenase from the fungus Cochliobolus lunatus is an NADP(H)-dependent member of the short-chain dehydrogenase/reductase superfamily (SDR) that belongs to the cP1 classical subfamily. Here, we have created several mutants by sitedirected mutagenesis, and through these we have studied the amino acid residues that are responsible for coenzyme binding and specificity. The Thr202Val and Thr202Ile mutants were inactive, thus confirming the importance of Thr202 for the appropriate orientation of the coenzyme that enables the hydride transfer. The Ala50Arg and Asn51Arg mutants had increased rates of NADPH dissociation, and thus an enhanced substrate oxidation with NADP+ , while the Asn51Arg mutant also showed an increased rate of NADP+ dissociation, and thus an enhanced substrate reduction with NADPH. Addition of a negatively-charged amino acid residue at the first position after the second ␤-strand (Tyr49Asp) affected the coenzyme specificity and turned the enzyme into an NAD+ -dependent oxidase resembling the cD1d subfamily members. © 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: 17␤-Hydroxysteroid dehydrogenase; Short-chain dehydrogenase/reductase; Site-directed mutagenesis; Enzyme kinetic; Coenzyme specificity

1. Introduction The protein superfamily of short-chain dehydrogenases/reductases (SDRs) includes a large number of prokaryotic and eukaryotic enzymes. To date, about 3000 members of this superfamily have been reported in different species (Oppermann et al., 2003). The SDR proteins function as NAD+ -dependent oxidases or NADPH-dependent reductases, and they are involved among others in steroid hormone metabolism, fatty acid oxidation and biotransformation of xenobiotics (Oppermann et al., 2003). Although the sequence identities between different SDR proteins are low, varying from 15% to 30%, all of the available three-dimensional (3D) structures display a highly similar ␣/␤ folding pattern, where the ␤ strands form a layer that is surrounded by three ␣ helices on each side (Oppermann et al., 2003). The SDRs have been divided into five families (classical, extended, intermediate, divergent and complex) with differ∗

Corresponding author. Tel.: +386 1543 7657; fax: +386 1543 7641. E-mail address: [email protected] (T.L. Riˇzner).

0303-7207/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mce.2005.05.007

ent motifs in the coenzyme binding and active-site regions, and different chain lengths (Kallberg et al., 2002). The classical and extended SDR families have been further divided into subfamilies based on their patterns of charged coenzymebinding residues (Kallberg et al., 2002). The classical family includes oxidoreductases, such as steroid dehydrogenases and carbonyl reductases. In the NADP(H)-preferring classical SDRs, the two negative charges of the 2 -phosphate group of the coenzyme are compensated for by one or two positively charged residues of the enzymes. The first of these is found in the Gly-X-X-X-Gly-X-Gly coenzyme binding motif towards the N-terminal, and is positioned before the second glycine. The second basic residue is situated in the first loop position after the second ␤-strand. The proteins with the basic residue in the first position fall into the cP1 subfamily, while those with the basic residue in the second position belong to the cP2 subfamily. The cP3 subfamily consists of proteins that have basic residues in both positions (Kallberg et al., 2002; Persson et al., 2003) (Table 1). The presence of an acidic residue at the C-terminus of the second ␤-strand in classical SDRs is a key determining

K. Kristan et al. / Molecular and Cellular Endocrinology 241 (2005) 80–87

81

Table 1 The alignment of the amino acid sequences around the coenzyme binding fold

Comparison of the sequences around the coenzyme binding fold of the SDR superfamily proteins, grouped according to their coenzyme specificities. The amino acids that determine the coenzyme specificities are boxed (CR: carbonyl reductase; HSD: hydroxysteroid dehydrogenase; TR: tropinone reductase, THNR: tetrahydroxynaphtalene reductase, Ver: versicolorin reductase, PcR: protochlorophyllide reductase, Dhpr: dihydropteridine reductase, 15-HPGD: 15hydroxyprostaglandin dehydrogenase, ADH: alcohol dehydrogenase).

factor for NAD(H) preference. This residue participates in hydrogen bonding with the 2 -and 3 -hydroxy groups of the adenine ribose moiety of NAD(H), and it repels NADP(H) electrostatically. The members that bind NAD(H) and have an aspartic or glutamic acid at the end of the second ␤-strand form the cD1d and cD1e subfamilies, respectively. The proteins that instead have a negativelycharged residue in the first or the second positions after the second ␤-strand belong to the cD2 and cD3 subfamilies, respectively (Kallberg et al., 2002; Persson et al., 2003) (Table 1). The amino acid residues that determine the coenzyme specificities have been already studied in the following SDR members: NADP(H)-dependent mouse lung carbonyl reductase (cP3 subfamily) (Nakanishi et al., 1996; Nakanishi et al., 1997); type 1 (cP2) (Huang et al., 2001) and type 3 (cP3) (McKeever et al., 2002) human 17␤-hydroxysteroid dehydrogenases (HSDs); human carbonyl reductase (cP3) (Sciotti and Wermuth, 2001); rainbow trout carbonyl reductase-like 20␤-HSD (cP3) (Guan et al., 2000); as well as in NAD(H)dependent 15-hydroxyprostaglandin dehydrogenase (cD1d) (Cho et al., 2003); type 2 11␤-HSD (cD3) (Arnold et al., 2003); and type 1 and type 2 human 3␤-HSD/isomerase (extended SDR) (Thomas et al., 2004).

We have been studying 17␤-HSD from the fungus Cochliobolus lunatus (17␤-HSDcl), which is the only fungal 17␤-HSD that has been purified and cloned to date (Laniˇsnik Riˇzner et al., 1996, 1999). Although the physiological function of the fungal 17␤-HSD has not yet been confirmed, the sequence similarity to human 17␤-HSD type 4 and 8 (30% and 29% identity, respectively), the endogenous biosynthesis of androgens (Kastelic-Suhadolc et al., 1994) and the presence of androgen binding proteins in C. lunatus (KastelicSuhadolc and Lenasi, 1993) indicate 17␤-HSDcl might be involved in steroid signalling. Considering homology to fungal reductases 17␤-HSDcl might also be involved in the biosynthesis of mycotoxins or it might be a part of fungal detoxification mechanism (Laniˇsnik Riˇzner et al., 2001). Further studies uncovering the physiological role of this enzyme are underway. This 17␤-HSD is an oxidoreductase with a molecular mass of 28 kDa that preferentially catalyses reversible oxidoreduction of estrogens and androgens with 4-estrene-3,17-dione and 4-estrene-17␤-ol-3-one as the two most preferred substrates. The enzyme has conserved classical SDR motifs, i.e. the Gly25 -X-X-X-Gly29 -X-Gly31 coenzyme binding site in the N-terminal region and the Tyr167 -X-X-X-Lys171 catalytic site in the central region. It is an NADP(H)-preferring enzyme

82

K. Kristan et al. / Molecular and Cellular Endocrinology 241 (2005) 80–87

and it has only one basic residue, Arg28, situated in the Glymotif; this therefore falls into the cP1 subfamily of the classical SDR family. The high percentage of amino acid identity between 17␤-HSDcl and 1,3,8-trihydroxynaphthalene reductase (3HNR) from the fungus Magnaporthe grisea (60%) has enabled the preparation of a homology-built structural model of 17␤-HSDcl using the known 3D structure of 3HNR as a template (Laniˇsnik Riˇzner et al., 2000). This model structure has facilitated further studies of the structure–function relationships of 17␤-HSDcl. We have previously described site-directed mutagenesis of the amino acid residues that are potentially important for catalysis in this fungal 17␤-HSD (Kristan et al., 2003). Here, we have investigated further the residues responsible for coenzyme binding and specificity of this cP1 subfamily member. Although similar studies have been reported previously for other subfamily members, this is the first such description of a cP1 subfamily member. We have thus investigated the importance of Thr202 for the appropriate binding of the coenzyme (Thr202Val, Thr202Ile), and the effects of an additional basic residue (Ala50Arg, Asn51Arg) and of an introduced Asp (Tyr49Asp, Ala50Asp) on the NAD+ /NADH and NADP+ /NADPH specificities.

2. Materials and methods 2.1. Site-directed mutagenesis The mutant proteins were generated using the Quick Change Site-Directed Mutagenesis Kit (Stratagene) according to the manufacturer instructions, the pGex-17␤-HSDcl expression vector, and the following primers (only forward 5 –3 primers are shown; the introduced mutations are underlined): Thr202Val: CCAGGCGGTACCGTGGTCGATATGTTCCACGAG Thr202Ile: CCAGGCGGTACCGTGATCGATATGTTCCACGAGGTC Ala50Arg: CAAAGTCGTTGTCAACTACCGCAACTCCACCAAAGACGCTG Asn51Arg: GTCGTTGTCAACTACGCCCGCTCCACCAAAGACGCTGAG Tyr49Asp: GCCAAAGTCGTTGTCAACGACGCCAACTCCACCAAAGAC Ala50Asp: CAAAGTCGTTGTCAACTACGACAACTCCACCAAAGACGCTG The fidelity of the constructs was confirmed by dideoxy sequencing.

cells. The purification was by affinity chromatography on glutathione Sepharose. The 17␤-HSDcl was cleaved from the GST with thrombin, as described previously (Laniˇsnik Riˇzner et al., 1999). Protein concentrations were determined by the Bradford method, using bovine serum albumin as a standard (Bradford, 1976). The homogeneity of the proteins was checked by SDS-PAGE and Coomassie blue staining (Laemmli, 1970). 2.3. Kinetic analyses of the wild-type and mutant enzymes 2.3.1. Determination of binding constants for the coenzymes The changes in the intrinsic fluorescence of the enzymes upon incremental additions of the coenzymes NADPH, NADP+ , NADH and NAD+ (0.0–1.0 mM) were measured on a Cary Eclipse fluorescence spectrophotometer (Varian). To ensure that the volume of coenzyme added was not more than 3% of the total volume, three stock solutions of the coenzyme were prepared—150 ␮M, 1.5 and 15 mM. A 4 mm × 10 mm cuvette holder with a magnetic stirrer was used, and each 800 ␮l sample contained 1 ␮M protein in 100 mM phosphate buffer, pH 8.0 at 25 ◦ C. The samples were excited at 290 nm, and fluorescence emission was measured at 335 nm, with excitation and emission band-passes of 5 nm. 2.3.2. Time course of enzymatic reactions The reduction of 4-estrene-3,17-dione and the oxidation of 4-estrene-17␤-ol-3-one (50–100 ␮M) in the presence of the coenzymes NADPH, NADH, NADP+ or NAD+ (50–1000 ␮M) were followed using a Beckman DU diode array spectrophotometer (Kristan et al., 2003). The time courses of the changes in absorbance at 340 nm were measured from 7 to 1000 s. All of the reactions were carried out in 100 mM phosphate buffer, pH 8.0 at 25 ◦ C, with a protein concentration of 0.5 ␮M, except for Tyr49Asp mutant where the concentration of 1.5 ␮M was used. 2.3.3. Data analysis To determine the Kd values, the fluorescence data were plotted as fluorescence (F) at 335 nm versus NAD(P)(H) concentration. The progress curves for the interconversion of the substrates 4-estrene-3,17-dione and 4-estrene-17␤-ol-3-one and of the coenzymes NAD(P)H and NAD(P)+ by the wildtype 17␤-HSDcl and by its mutants were analyzed according to procedures described previously (Laniˇsnik Riˇzner et al., 2000). The Theorell–Chance reaction mechanism (Cleland, 1963) shown in Scheme 1 was used.

2.2. Expression and purification of the mutant proteins The wild-type 17␤-HSDcl enzyme and the mutants were expressed as GST-fusion proteins in Escherichia coli JM107

Scheme 1.

K. Kristan et al. / Molecular and Cellular Endocrinology 241 (2005) 80–87

83

Table 2 Rate constants for the 17␤-HSDcl wild-type and mutants

Wild type Thr202Val Thr202Ile Ala50Arg Asn51Arg Tyr49Asp Tyr49Aspc Ala50Asp

k+1 a

k−1 b

k+2 a

k−2 a

k+3 b

k−3 a

7.1 × 105 N.A. N.A. 1.7 × 106 1.4 × 105 1852 559 8.1 × 105

1.13 N.A. N.A. 5.9 1.53 0.74 0.25 11.9

55587 N.A. N.A. 18860 152626 2139 3150 13594

15220 N.A N.A. 12932 37087 19333 42383 5281

12.1 N.A. N.A. 1.77 18.2 0.09 1.78 0.62

6.6 × 107 N.A N.A. 1.85 × 106 1.75 × 107 64.2 674 2.71 × 105

Rate constants for the reduction and oxidation of the substrates 4-estrene-3,17-dione and 4-estrene-17␤-ol-3-one, catalyzed by the 17␤-hydroxysteroid dehydrogenases from C. lunatus, in the presence of the coenzymes NADPH and NADP+ , according to the Theorell-Chance reaction mechanism, at pH 8 and 25 ◦ C (Keq = 2.46 ± 0.11). a Second-order rate constant (M−1 s−1 ). b First-order rate constant (s−1 ). c Determined for the NAD(H) pair of coenzymes, N.A.—not active.

Where E represents the enzyme, A, Q, B and P are the NAD(P)H and NAD(P)+ , and the oxidized and reduced substrates, respectively. The kinetic parameters of the second or first-order rate constants (see Table 2) were determined by fitting the differential equations, characteristic for the Theorell–Chance mechanism, to the experimental curves. In the fitting, we used the ratios k−1 /k+1 and k+3 /k−3 as determined from fluorescence experiments. Taking the enzyme independent equilibrium constant (k−1 k−2 k−3 /k+1 k+2 k+3 ) as fixed, there were only three parameters to be evaluated from a given set of curves. In addition to the kinetic rate constants, the actual concentrations were also fitting parameters in the evaluation. To compare our rate constants that were obtained under non-steady-state conditions with those of the data published for other enzymes, the classical kinetic parameters of KM and kcat , and the initial rates were derived in terms of six rate constants (Cleland, 1963), as follows: kcat

k+2 [B]k+3 = k+2 [B] + k+3

and

Km =

(k−1 + k+2 [B])k+3 k+1 (k+2 [B] + k+3 )

kcat =

k−2 [P]k−1 k−2 [P] + k−1

Km =

(k+3 + k−2 [P])k−1 k−3 (k−2 [P] + k−1 )

v=

and for NADP+ ;

k+2 [B]k+3 [A][E]t /(k+2 [B] + k+3 [A] + (k−1 + k+2 [B])k+3 /k+1 (k+2 [B] + k+3 ) for NADPH

v=

for NADPH;

and

k−2 [P]k−1 [Q][E]t /k−2 [P] + k−1 ) [Q] + (k+3 + k−2 [P])k−1 /k−3 (k−2 [P] + k−1 ) for NADP

3. Results and discussion The aim of this study was to identify the amino acid residues that are important for the coenzyme binding and specificity of 17␤-HSDcl, a model for SDR enzymes of cP1 subfamily. We created several mutants by site-directed mutagenesis, and through these we have studied the importance of Thr202 (Thr202Val, Thr202Ile) in the correct binding of the coenzyme, the effects of an additional basic residue (Ala50Arg, Asn51Arg), and the effects of an introduced Asp (Ala50Asp, Tyr49Asp) on the NAD+ /NADH and NADP+ /NADPH specificities. Stereo view of the NADP(H) binding site in 17␤-HSDcl model structure is shown in Fig. 1. 3.1. The role of Thr202 in the binding of the coenzyme In HSDs from the SDR superfamily, the bound coenzyme lies across the Rossmann fold in an extended synconformation. This orients the B-face of the coenzyme into the active-site cleft and enables the transfer of the 4-pro(S)-hydrogen of the coenzyme to the steroid to form axial alcohols (Penning, 1997). The amino acid residues around the nicotinamide moiety contribute to the correct accommodation of the coenzyme. In the 17␤-HSDcl model structure, two threonines (Thr200 and Thr202) were in close contact with the amide group of the nicotinamide. The first one, Thr200, is not conserved among the members of the SDR superfamily since it has only been found in NAD+ -dependent Streptomyces hydrogenans 3␣,20␤-HSD (Ghosh et al., 1994) and Drosophila lebanonensis alcohol dehydrogenase (Benach et al., 1998). The crystal structures revealed that this threonine interacts with the carbonyl part of the nicotinamide moiety. When Thr200 was replaced by Val, as reported previously, substrate oxidation was more affected than reduction, indicating that this Thr200 is involved in the binding and dissociation of NADP(H) (Kristan et al., 2003).

84

K. Kristan et al. / Molecular and Cellular Endocrinology 241 (2005) 80–87

Fig. 1. The amino acid residues that are important for the coenzyme specificity in 17␤-HSDcl. Stereo view of the NADP(H) binding site from the homology build model structure of 17-HSDcl is shown. The Arg28, Tyr49, Ala50, Asn51, Thr200 and Thr202 amino acid residues are indicated.

The second threonine in 17␤-HSDcl, Thr202, is relatively conserved among the members of the SDR superfamily (Kallberg et al., 2002; Persson et al., 2003; Shi and Lin, 2004). In crystal structures of several SDR members, the respective threonine forms a bifurcated hydrogen bond with the amino group of the nicotinamide moiety and one oxygen of the pyrophosphate moiety (Shi and Lin, 2004). It has previously been shown that the homologous Thr is critical for the interaction with NAD+ in the human NAD(H)-dependent 15-hydroxyprostaglandin dehydrogenase, a member of cD1d classical subfamily (Zhou and Tai, 1999). The Thr188Ala and Thr188Tyr mutants were inactive, while Thr188Ser remained active, although with a 100-fold higher Km for NAD+ when compared to the wild type (Zhou and Tai, 1999). We studied the role of this Thr202 in the 17␤-HSDcl cP1 member using two mutants, Thr202Val and Thr202Ile. Both of these mutants were inactive, thus confirming the importance of this Thr202 for the appropriate orientation of the coenzyme and the consequent hydride transfer (Table 2). 3.2. Introduction of the second basic residue In the NADP(H)-dependent enzymes, one or two positively-charged residues are conserved that determine the

cofactor specificity. These positively-charged residues compensate for the two negative charges of the 2 -phosphate group of the coenzyme. The model structure of 17␤-HSDcl revealed Arg28 as the only candidate that could interact with the 2 -phosphate. We have previously confirmed the importance of this Arg28 with two mutants, Arg28Ala and Arg28Glu (Kristan et al., 2003). Arg28Ala remained active, although significantly less so than the wild-type enzyme, while we observed a very low activity for the Arg28Glu mutant, and only after a 1-h incubation. Since this Arg28 might also interact with the negatively-charged oxygens from the pyrophosphate moiety of the coenzymes, as has been suggested for the homologous Arg39 from 3HNR of M. grisea (Andersson et al., 1996), this might be the reason for the low activity of the Arg28Glu mutant with all four of the coenzymes. To investigate the effects of an additional basic amino acid on the NADPH specificity, we introduced the second basic amino acid at two different positions, thus generating the mutants Ala50Arg and Asn51Arg. The titration of the Ala50Arg mutant with the coenzymes NADPH and NADP+ revealed higher values for their dissociation constants (Kd ), when compared with the wild-type enzyme. The Kd for NADPH and NADP+ increased 2.2and 4.5-fold, respectively. The Kd for NADPH is the ratio

K. Kristan et al. / Molecular and Cellular Endocrinology 241 (2005) 80–87

85

Table 3 Kinetic constants for the 17␤-HSDcl wild-type and mutants

Wild type Ala50Arg Asn51Arg Tyr49Asp Tyr49Aspb Ala50Asp

Wild type Ala50Arg Asn51Arg Tyr49Asp Tyr49Aspb Ala50Asp

KdNADPH a

NADP+ a KM

NADPH KM

NADP+ KM

NADPH kcat

NADP+ kcat

1.6 3.5 10.6 399 447 14.7

0.2 0.9 1.0 1448 2641 2.3

6.5 2.2 9.2 156 859 5.2

0.09 1.36 1.74 8737 497 4.1

3.8 0.9 8.3 0.06 0.3 0.4

0.7 1.1 1.1 0.5 0.2 0.5

NADPH kcat /Km

NADP+ kcat /Km

vNADPH 0

vNADP+ 0

5.9 × 105 4.1 × 105 9.0 × 105 415 311 8.3 × 104

7.3 × 106 7.8 × 105 6.3 × 106 61 475 1.2 × 105

1.79 0.45 2.54 0.012 0.014 0.20

0.32 0.52 0.54 0.003 0.02 0.24

Kinetic constants Kd (␮M), KM (␮M), kcat (s−1 ), kcat /Km (s−1 M−1 ), v0 (␮M/s)) for the reduction and oxidation of the substrates 4-estrene-3,17-dione and 4-estrene-17␤-ol-3-one (100 ␮M), catalyzed by the 17␤-hydroxysteroid dehydrogenases from C. lunatus, in the presence of the coenzymes NADPH and NADP+ (100 ␮M), at pH 8 and 25 ◦ C. a Determined by fluorescence titrations. b Determined for the NAD(H) pair of coenzymes.

k−1 /k+1 , and a more extensive data analysis revealed that the rate constants k−1 (the rate of NADPH dissociation) and k+1 increased. The Kd for NADP+ is the ratio k+3 /k−3 and both of these constants, k+3 (the rate of NADP+ dissociation) and k−3 , decreased. The dissociation rates of NADPH and NADP+ also affected the kcat values: the kcat for the substrate reduction with NADPH decreased four-fold, while the kcat for the oxidation of the substrate with NADP+ increased 1.6-fold. The Ala50Arg substitution thus leads to a higher kcat and initial rates only in the oxidative direction (Tables 2 and 3). This mutation, on the other hand, does not affect the enzyme activity in the presence of the coenzymes NAD+ and NADH. Also with Asn51Arg, the Kd values for NADPH and NADP+ increased 6.6- and 5.0-fold, respectively. However, in this case the rates of NADPH and NADP+ dissociation (k−1 and k+3 ) increased and the rates of NADPH and NADP+ binding (k+1 and k−3 ) decreased. Both kcat values for the reduction and oxidation increased 2.2- and 1.7-fold, respectively. Thus, according to the higher kcat values and initial rates, the introduced Arg at position 51 increased enzyme activity in both the reductive and oxidative directions with the coenzymes NADPH and NADP+ (Tables 2 and 3). Again, the mutation did not affect the enzyme activity in the presence of the coenzymes NAD+ and NADH. The introduction of a second basic residue has also been reported for a cP2 member, human type1 17␤-HSD, which acts in vitro as an NADH/NADPH-dependent reductase or as an NAD+ /NADP+ -dependent oxidase. In this case, the introduction of a positively-charged Lys at position 12 within the Gly-X-X-X-Gly-X-Gly motif increased the relative specificity of the enzyme for NADP(H) and significantly decreased the enzyme affinity for NAD(H) (Huang et al., 2001). The apparent activity of Ser12Lys measured in cultured Sf9 cells was about 175% that of the wild-type enzyme for estrone

reduction, whereas for estradiol oxidation it was only about 27% that of the wild type (Huang et al., 2001). The cP1 member used in the present study cannot be directly compared with human type 1 17␤-HSD since it acts in vitro as an NADPH-dependent reductase or as an NADP+ -dependent oxidase, while in the presence of NAD(H), the reaction proceeds so slowly that the rate constants cannot be determined. When Arg was introduced at position 50, only the NADP+ -dependent oxidase activity increased (162%), while the NADPH-dependent reductase activity decreased (25%). When Arg was introduced at position 51, the enzyme activity increased in both directions, reductive and oxidative (142% and 171%, respectively) (Table 3). Neither of these substitutions affected the enzyme affinity for the coenzyme pair NAD(H). The small differences in the activities of both of these mutants compared to the wild-type enzyme in the reductive direction indicate that as 17␤-HSDcl possesses only one basic amino acid residue, it is already adapted to NADP(H) specific oxidoreductions, and that the introduction of another basic residue does not improve the activity significantly. 3.3. Introduction of an acidic residue The presence of an acidic residue that electrostatically repels NADP(H) is a key determining factor for the NAD(H) preference in the NAD(H)-preferring enzymes. There are still some NAD(H)-preferring enzymes that also have a conserved basic amino acid residue in addition to the acidic residue, which is typical for an NADP(H)-preference, such as with 3␣,20␤-HSD from S. hydrogenans (Arg16, the first basic amino acid; Protein Database (PDB) code: 1HDC) and alcohol dehydrogenase from D. lebanonensis (Arg39, the second basic amino acid; PDB code: 1A4U) (Table 1).

86

K. Kristan et al. / Molecular and Cellular Endocrinology 241 (2005) 80–87

Here, we introduced an aspartic acid into 17␤-HSDcl at two positions, 49 and 50. The kinetic analysis showed that the introduction of the aspartic acid at the Tyr49 position (Tyr49Asp) switched the enzyme cofactor preference from NADP(H) to NAD(H), and changed the catalytic designation of the enzyme from a reductase to an oxidase. This newly designed enzyme, with the Asp residue at the end of the second ␤-strand, resembles members of the cD1d subfamily. The titration of this mutant with the coenzymes NADPH and NADP+ revealed that the Kd values for both of the coenzymes drastically increased. The Kd for NADPH increased 250-fold, and the Kd for NADP+ increased 7900-fold, when compared with the wild-type enzyme. The kcat values for NADPH and NADP+ decreased and the Km values increased, resulting in drastically decreased kcat /Km values (1900- and 1.2 × 105 fold, respectively) (Tables 2 and 3). The corresponding values for the coenzymes NADH and NAD+ could not be determined for the wild-type enzyme, but they could be determined for the Tyr49Asp mutant. We then compared the kinetic parameters for the NAD(H) pair of coenzymes with those for NADP(H). The Kd values for the coenzymes NADH and NAD+ were similar to those determined for NADPH and NADP. The kcat /Km values for NAD+ and NADH were higher (7.8-fold) and almost the same (0.75fold) when compared to the kcat /Km values for NADP+ and NADPH, respectively (Tables 2 and 3). Our results thus demonstrate that the mutation of Tyr49Asp increases the enzyme oxidative activity in the presence of NAD+ . Addition of a negatively-charged amino acid residue at the end of the second ␤-strand (Tyr49Asp) thus affected the coenzyme specificity and turned this cP1 member into an oxidase that resembled the cD1d members. Less drastic effects on enzyme activity were observed when we substituted the amino acid at the first position after the second ␤-strand, Ala50, with Asp, to mimic the cD2 SDR members. The kinetic parameters for the Ala50Asp mutant revealed 9.2- and 12.4-fold higher Kd values for NADPH and NADP+ , respectively. The kcat values for NADPH and NADP+ decreased, leading to lower kcat /Km values (Tables 2 and 3). These results thus demonstrate that the introduction of an Asp into the Ala50 position decreases the enzyme activity in the presence of NADPH and NADP+ , but does not improve the enzyme activity with the NADH/NAD+ pair of coenzymes. A switch of the coenzyme specificity was first reported for mouse lung carbonyl reductase, a cP3 SDR member. The described mutation resulted in a 200-fold increase of the Km value for NADPH and a 16-fold decrease in the Km for NAD+ (Nakanishi et al., 1996). Also in the Tyr49Asp 17␤-HSDcl mutant, the Km values for NADP(H) drastically increased, by 24-fold for NADPH and by 105 -fold for NADP+ . In type 1 17␤-HSD, a cP2 member, the coenzyme specificity was switched when a negatively-charged amino acid residue, Asp, was introduced at Leu36, the position typical for cD1d members (Huang et al., 2001). In human enzyme this mutation switched the enzyme coenzyme preference from NADP+

to NAD+ , with a 220-fold change in specificity (Huang et al., 2001). With 17␤-HSDcl here, we cannot compare the NAD(H)-dependent reactions catalyzed by the Tyr49Asp mutant with the reaction catalyzed by the wild-type enzyme, but the comparison between the kcat /Km values for NAD+ and for NADP+ revealed a 7.8-fold difference. 4. Conclusions We have here examined the coenzyme specificities of 17␤HSDcl mutants. Loss of activity after replacement of the Thr202 (Thr202Val/Ile) that is situated near the nicotinamide and pyrophosphate moiety point to its pivotal role in the correct accomodation of the coenzyme and the consequent hydride transfer. The introduction of the second basic residue at position 50 (Ala50Arg) increases the rate of NADPH dissociation and thus facilitates the reaction in an oxidative direction, while the introduction of an Arg at position 51 (Asn51Arg) increases the rates of NADP+ and NADPH dissociation and thus affects the enzyme activity in both directions. The small differences in the activities of both of these mutants in the reductive direction indicate that as 17␤-HSDcl possesses only one positively-charged amino acid residue, it is already adapted to NADP(H)-specific oxidoreductions and so the introduction of additional basic residue does not improve the activity significantly. The introduction of an acidic residue at position 49 (Tyr49Asp) changes the coenzyme specificities from an NADPH-dependent reductase to an NAD+ -dependent oxidase, while the introduction of this acidic residue at position 50 (Ala50Asp) does not affect the coenzyme specificity. Acknowledgements The work was supported by Ministry of Education, Science and Sport of Slovenia, by a NATO grant to J.A. and T.L.R. and by a WFS Scholarship to K.K. We thank Dr. Karin Pritsch for use of the fluorescence spectrophotometer (GSF, Institute of Soil Ecology, Neuherberg, Germany). References Andersson, S., Jordan, D., Schneider, G., Lindqvist, Y., 1996. Crystal structure of the ternary complex of 1,3 8-trihydroxynaphthalene reductase from Magnaporthe grisea with NADPH and an active-site inhibitor. Structure 4, 1161–1170. Arnold, P., Tam, S., Yan, L, Baker, M.E., Frey, F.J., Odermatt, A., 2003. Glutamate-115 renders specificity of human 11beta-hydroxysteroid dehydrogenase type 2 for the cofactor NAD+ . Mol. Cell. Endocrinol. 201, 177–187. Benach, J., Atrian, S., Gonzales-Duarte, R., Ladenstein, R., 1998. The refined crystal structure of Drosophila lebanonensis alcohol dehydrogenase at 1.9 A resolution. J. Mol. Biol. 282, 383–399. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Anal. Biochem. 72, 248–254.

K. Kristan et al. / Molecular and Cellular Endocrinology 241 (2005) 80–87 Cho, H., Oliveira, M.A., Tai, H.H., 2003. Critical residues for the coenzyme specificity of NAD+ -dependent 15-hydroxyprostaglandin dehydrogenase. Arch. Biochem. Biophys. 419, 139–146. Cleland, W.W., 1963. The kinetics of enzyme-catalyzed reactions with two or more substrates or products. Biochim. Biophys. Acta 67, 104–137. Ghosh, D., Wawrzak, Z., Weeks, C.M., Duax, W.L., Erman, M., 1994. The refined three-dimensional structure of 3␣, 20␤-hydroxysteroid dehydrogenase and possible roles of the residues conserved in shortchain dehydrogenases. Structure 2, 629–640. Guan, G., Todo, T., Tanaka, M., Young, G., Nagahama, Y., 2000. Isoleucine-15 of rainbow trout carbonyl reductase-like 20␤hydroxysteroid dehydrogenase is critical for coenzyme (NADPH) binding. Proc. Natl. Acad. Sci. U.S.A. 97, 3079–3083. Huang, Y.W., Pineau, I., Chang, H., Azzi, A., Bellemare, V., Laberge, S., Lin, S.X., 2001. Critical residues for the specificity of cofactors and substrates in human estrogenic 17␤-hydroxysteroid dehydrogenase 1: variants designed from the three-dimensional structure of the enzyme. Mol. Endocrinol. 15, 2010–2020. Kallberg, Y., Oppermann, U., J¨ornvall, H., Persson, B., 2002. Shortchain dehydrogenases/reductases (SDRs). Eur. J. Biochem. 269, 4409– 4417. Kastelic-Suhadolc, T., Lenasi, H., 1993. Androgen binding proteins in Cochliobolus lunatus. FEMS Microbiol. Lett. 108, 111– 126. ˇ Kastelic-Suhadolc, T., Plemenitaˇs, A., Zigon, D., 1994. Isolation and identification of testosterone and androstenedione in the fungus Cochliobolus lunatus. Steroids 59, 357–361. Kristan, K., Laniˇsnik Riˇzner, T., Stojan, J., Gerber, J.K., Kremmer, E., Adamski, J., 2003. Significance of individual amino acid residues for coenzyme and substrate specificity of 17␤-hydroxysteroid dehydrogenase from the fungus Cochliobolus lunatus. Chem. Biol. Interact. 143–144, 493–501. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680– 685. ˇ Laniˇsnik Riˇzner, T., Zakelj-Mavriˇ c, M., Plemenitaˇs, A., Zorko, M., 1996. Purification and characterization of 17␤-hydroxysteroid dehydrogenase from the filamentous fungus Cochliobolus lunatus. J. Steroid Biochem. Mol. Biol. 59, 205–214. ˇ Laniˇsnik Riˇzner, T., M¨oller, G., Thole, H.H., Zakelj-Mavriˇ c, M., Adamski, J., 1999. A novel 17␤-hydroxysteroid dehydrogenase in the fungus Cochliobolus lunatus: new insights into the evolution of steroidhormone signaling. Biochem. J. 337, 425–431.

87

Laniˇsnik Riˇzner, T., Adamski, J., Stojan, J., 2000. 17␤-hydroxysteroid dehydrogenase from Cochliobolus lunatus: model structure and substrate specificity. Archiv. Biochem. Biophys. 384, 255–262. Laniˇsnik Riˇzner, T., Stojan, J., Adamski, J., 2001. Searching for the physiological function of 17␤-hydroxysteroid dehydrogenase from the fungus Cochliobolus lunatus: studies of substrate specificity and expression analysis. Mol. Cell. Endocrinol. 171, 193–198. McKeever, B.M., Hawkins, B.K., Geissler, W.M., Wu, L., Sheridan, R.P., Mosley, R.T., Andersson, S., 2002. Amino acid substitution of arginine 80 in 17␤-hydroxysteroid dehydrogenase type 3 and its effect on NADPH cofactor binding and oxidation/reduction kinetics. Biochim. Biophys. Acta 1601, 29–37. Nakanishi, M., Kakumoto, M., Matsuura, K., Deyashiki, Y., Tanaka, N., Nonaka, T., Mitsui, Y., Hara, A., 1996. Involvement of two basic residues (Lys-17 and Arg-39) of mouse lung carbonyl reductase in NADP(H)-binding and fatty acid activation: site-directed mutagenesis and kinetic analyses. J. Biochem. 120, 257–263. Nakanishi, M., Matsuura, K., Kaibe, H., Tanaka, N., Nonaka, T., Mitsui, Y., Hara, A., 1997. Switch of coenzyme specificity of mouse lung carbonyl reductase by substitution of threonine 38 with aspartic acid. J. Biol. Chem. 272, 2218–2222. Oppermann, U., Filling, C., Hult, M., Shafgat, N., Wu, X., Lindh, M., Shafqat, J., Nordling, E., Kallberg, Y., Persson, B., J¨ornvall, H., 2003. Short-chain dehydrogenases/reductases (SDR): the 2002 update. Chem. Biol. Interact. 143–144, 247–253. Penning, T.M., 1997. Molecular endocrinology of hydroxysteroid dehydrogenases. Endocrine Rev. 18, 281–305. Persson, B., Kallberg, Y., Oppermann, U., Jornvall, H., 2003. Coenzyme-based functional assignments of short-chain dehydrogenases/reductases (SDRs). Chem. Biol. Interact. 143–144, 271–278. Sciotti, M., Wermuth, B., 2001. Coenzyme specificity of human monomeric carbonyl reductase: contribution of Lys-15 Ala-37 and Arg-38. Chem. Biol. Interact. 130–132, 871–878. Shi, R., Lin, S.-X., 2004. Cofactor hydrogen bonding onto the protein main chain is conserved in the short chain dehydrogenase/reductase family and contributes to nicotinamide orientation. J. Biol. Chem. 279, 16778–16785. Thomas, J.L., Duax, W.L., Addlagatta, A., Kacsoh, B., Brandt, S.E., Norris, W.B., 2004. Structure/function aspects of human 3␤hydroxysteroid dehydrogenase. Mol. Cell. Endocrinol. 215, 73–82. Zhou, H., Tai, H., 1999. Threonine 188 is critical for interaction with NAD+ in human NAD+ -dependent 15-hydroxyprostaglandin dehydrogenase. Biochem. Biophys. Res. Commun. 257, 414–417.