- Email: [email protected]
Structural basis for chiral substrate recognition by two 2,3-butanediol dehydrogenases
FEBS Letters 584 (2010) 219–223
journal homepage: www.FEBSLetters.org
Structural basis for chiral substrate recognition by two 2,3-butanediol dehydrogenases Masato Otagiri a,1, Sadaharu Ui b,*, Yuhsuke Takusagawa b, Takashi Ohtsuki b, Genji Kurisu c,1, Masami Kusunoki b a b c
Laboratory of Environmental Molecular Biology, RIKEN, 1-7-29 Suehiro-cho, Tsurumi-ward, Kanagawa 230-0045, Japan Graduate School of Medicine and Engineering, University of Yamanashi, Takeda, Kofu 400-8511, Japan Division of Structural Biology, Institute for Protein Research, Osaka University, Yamadaoka, Suita 565-0871, Japan
a r t i c l e
i n f o
Article history: Received 1 September 2009 Revised 19 November 2009 Accepted 19 November 2009 Available online 24 November 2009 Edited by Hans Eklund Keywords: Butanediol dehydrogenase Chiral recognition Stereoisomer Short-chain dehydrogenase/reductase family X-ray crystallography
a b s t r a c t 2,3-Butanediol dehydrogenase (BDH) catalyzes the NAD-dependent redox reaction between acetoin and 2,3-butanediol. There are three types of homologous BDH, each stereospecific for both substrate and product. To establish how these homologous enzymes possess differential stereospecificities, we determined the crystal structure of L-BDH with a bound inhibitor at 2.0 Å. Comparison with the inhibitor binding mode of meso-BDH highlights the role of a hydrogen-bond from a conserved Trp residue192. Site-directed mutagenesis of three active site residues of meso-BDH, including Trp190, which corresponds to Trp192 of L-BDH, converted its stereospecificity to that of L-BDH. This result confirms the importance of conserved residues in modifying the stereospecificity of homologous enzymes. Ó 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
1. Introduction Enzymatic reactions are often stereospecific and sometimes stereoselective. Engineering the stereospecifities of enzymes, although very difficult to control, has been a significant challenge. As in the well-known structural independence between L-lactose dehydrogenase and D-lactose dehydrogenase, the artificial design of enantiospecificity possessed by a native enzyme is not so simple. 2,3-Butanediol (BD) is known as a by-product of sugar metabolism in microorganisms. There are three types of stereoisomer of BD (D-, L-, and meso-forms). Since the ratio of BD stereoisomers depends on the organisms and the culture conditions of the fermentation process, various 2,3-butanediol dehydrogenases (BDHs) catalyzing the redox reaction between acetoin (AC) and BD were identified from many microorganisms. Of these BDHs, meso-BDH from Klebsiella pneumoniae IAM 1063, specific for meso-BD and D-AC, and L-BDH from Brevibacterium saccharolyticum C-1012, spe-
Abbreviations: AC, acetoin; BD, 2,3-butanediol; BDH, butanediol dehydrogenase; SDR, short-chain dehydrogenase/reductase family * Corresponding author. Fax: +81 55 220 8538. E-mail address: [email protected] (S. Ui). 1 Authors having equally contributed to this work.
cific for L-BD and L-AC, are both biochemically and enzymatically well characterized [1]. These two enzymes share a 50% homology of their amino acid sequence [2,3]. Therefore, it is very interesting to ascertain exactly how the homologous BDHs distinguish diastereomers. L- and meso-BDH belong to the short-chain dehydrogenase/ reductase (SDR) family, based on their amino acid sequences [3]. The substrates of SDR family enzymes vary widely in molecular size and properties, as they include alcohols, glucose and steroids [4]. Many crystal structures of SDR enzymes have been reported [5], and all have core structure similar to the Rossmann fold [6]. The variable loop region located in the C-terminal side of molecule was considered to be the reason this conserved structural frame could recognize a wide variety of substrates [7]. There is thought to be a common ‘‘Asn-Ser-Tyr-Lys catalytic tetrad” in SDR family enzymes, based upon biochemical and crystallographic studies [8]. We previously determined the crystal structure of meso-BDH from Klebsiella pneumoniae IAM 1063 at 1.7 Å resolution [9], and here we describe X-ray crystallographic analysis of L-BDH from B. saccharolyticum C-1012 at 2.0 Å resolution. It is now possible to compare the chiral recognition mechanisms of the two BDHs. We discuss the structural basis for stereospecificity on the basis of crystallographic and site-directed mutagenesis experiments.
0014-5793/$36.00 Ó 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2009.11.068
220
M. Otagiri et al. / FEBS Letters 584 (2010) 219–223
Table 1 Summary of refinement statistics. L-BDH
Space group Cell parameters Resolution (Å) Rfactor/Rfreea
P1 a = 60.8 Å, b = 69.2 Å, c = 127.4 Å, a = 96.1°, b = 100.2°, c = 109.6° 40.0–2.0 19.3/24.0
No. of non-hydrogen atoms Protein 15,208 Coenzyme 352 Mercaptoethanol 32 4 Mg+ Solvent 839 Rms deviations from ideal values Bond length (Å) 0.006 Bond angles () 1.2 Ramachandran plot analysis (%) Most favored 91.1 Allowed 8.9 a
Five percent of the reflections were set aside for an Rfree calculation.
2. Materials and methods 2.1. Enzyme stereospecific, preparation, crystallization and X-ray data collection The stereospecific type of BDH, L-BDH or meso-BDH, was used based on its stronger substrate isomer, L-BD or meso-BD, for keeping consistency of the usage until now. Over-expression and purification of recombinant L-BDH were performed according to a previously reported method [3]. L-BDH was co-crystallized with NAD+ and mercaptoethanol (ME) by the hanging drop vapor diffusion method [10]. Details of crystallization and data collection were described [11]. 2.2. Structure determination A crystal structure of L-BDH was solved using the molecular replacement method with the program AMORE [12] using mesoBDH (PDB code 1GEG) [9] as a search model. Crystallographic refinements were carried out with the program CNS version 0.9 [13]. The molecular model was built using the graphics program O [14] interpreting 2|Fo| |Fc| and |Fo| |Fc| electron density maps, which were improved through several cycles of model building and refinement with CNS. The crystallographic statistics are summarized in Table 1. Atomic coordinates and structure factors have been deposited in the Protein Data Bank. The PDB ID is 3A28. 2.3. Site-directed mutagenesis, enzyme assay and kinetic analysis Site-directed mutagenesis was performed using the Quikchange site-directed mutagenesis kit (Stratagene). The activity of BDH was assayed by recording the absorption of NADH at 340 nm. The standard reaction mixture consisted of 0.2 M sodium phosphate, pH 8.0, 2.5 lmol NAD+, 50 lmol BD and enzyme, in a total volume of 3.0 mL. One unit of enzyme activity was defined as the amount of enzyme that forms 1 lmol of NADH per minute at 37 °C. The Ki value of ME against BDH was determined by Dixon plots [15]. The all values obtained were averages of three trials. 3. Results and discussion 3.1. Overall structure of L-BDH The crystal structure of L-BDH complexed with NAD+ and ME was determined, identifying two identical homotetramers per
asymmetric unit. The eight subunits (A–H) in the asymmetric unit can be superimposed with an averaged, pairwise root-meansquare deviation (rmsd) of 0.15 Å for 258 Ca atoms. Therefore, the structure of subunit A will be taken hereafter as representative of the eight subunits. The crystal structure of the L-BDH protomer is shown in Fig. 1A. The L-BDH subunit is composed of seven b strands, six a helices and two short a helices (aFG1 and aFG2), which form a small lobe on top of the core domain. The conserved active site residues Asn112, Ser141, Tyr154 and Lys158 are located near NAD+ molecule. The roles of catalytic residues in SDR enzymes have been intensively studied. The Tyr residue acts as a general base for proton transfer, while the Lys residue is important in cofactor-binding and lowering the pKa value of the hydroxyl group of the Tyr residue. The Ser residue is suggested to be involved in the binding of substrate, the reaction intermediate and the product and/or the hydroxyl group of the conserved Tyr residue. The Asn residue plays a role in maintaining the active site configuration, allowing the formation of a proton relay system involving Lys, Tyr, the nicotinamide ribose moiety of the coenzyme and the substrate. A NAD+ molecule is bound to the enzyme in an extended conformation at the carboxyl ends of the central parallel b-sheet. The overall conformation of NAD+ and its binding mode to the Rossmann fold is similar to that found in other SDR enzymes [16]. A close-up view of the binding site is shown in Fig. 1B. The adenine ring of NAD+ binds in a hydrophobic pocket formed by the side chains of Ile91, Val111, Asp61, Leu34 and Ala89. Hydrogen bonds are formed by the N1 atom of adenine and the side chain of Asp61. The adenine ribose and pyrophosphate moieties are situated on the turn formed by Gly9 and Gly15. The turn between the first b-strand and the second a-helix has the sequence GXXXGX, which is a characteristic fingerprint of the nucleotide-binding motif of the SDR enzymes. The 20 - and 30 -hydroxyl groups of the ribose form hydrogen bonds with the side chain of Asp33, which is common to SDR enzymes that selectively bind NAD+. The pyrophosphate moiety interacts with the enzyme through hydrogen bonds with the side chains of Gln12 and Thr189 and the main-chain nitrogen of Ile14. The nicotinamide ring moiety makes polar interactions with the main-chain of Val187 and the side chain of Thr189. These residues are located in the N-terminal part of aFG1. The ribose moiety forms hydrogen bonds with the side chains of Tyr154 and Lys158 and with the main-chain oxygen of Asn88. Other interactions with NAD+ are similar to those seen in structures of other SDR enzymes. 3.2. Comparison with other SDRs A structural comparison of L-BDH with other SDR enzymes demonstrates that L-BDH exhibits a high degree of overall structural similarity to enzymes such as meso-BDH [9], glucose dehydrogenase [18] and 3a, 20b-hydroxysteroid dehydrogenase [6]. Structural comparison of two BDHs reveals that L-BDH has a very similar overall structure to meso-BDH, except for some conformational differences on the molecular surface and in the terminal regions. The rmsd of the superimposed L-BDH and meso-BDH structures is 0.5 Å. Furthermore, the manner of tetramer formation is also identical in L-BDH and meso-BDH. Comparison of the NAD+ binding site in L-BDH and the corresponding site in meso-BDH reveals that the residues responsible for NAD+ binding and their spatial organization are highly conserved between L-BDH and mesoBDH. This result indicates that L-BDH binds the cofactor in the same way as meso-BDH. At present, the ‘Asn-Ser-Tyr-Lys catalytic tetrad’ is considered important in SDR catalysis [5]. The catalytic tetrad residues (Asn112, Ser141, Tyr154 and Lys158) are situated in identical positions in meso-BDH (Asn110, Ser139, Tyr152 and Lys156). The position of the
M. Otagiri et al. / FEBS Letters 584 (2010) 219–223
221
Fig. 1. Structure of L-BDH. (A) Ribbon drawing of the L-BDH subunit with NAD+ and an inhibitor, mercaptoethanol (BME). (B) Active site structure of L-BDH. A 2|Fo| |Fc| omit map is superimposed on NAD+ and mercaptoethanol at the 1.0 r level. The catalytic tetrad is colored in orange. All figures showing structures in this paper were produced with the program PYMOL [17].
NAD+ molecule as well as surrounding residues is conserved in LBDH and meso-BDH. This structural conservation strongly suggests that the same reaction mechanism is common to both BDH enzymes. 3.3. Inhibitor-binding sites Initially, we attempted to prepare a complex of L-BDH with substrate, but these efforts failed. As seen in the meso-BDH structure, the electron density corresponding to an ME, which was included in the crystallization solution, was observed. Kinetic analysis in this study showed that ME was a competitive inhibitor of L-BDH (Ki = 4.3 mM at pH 8.0). One of the most outstanding characteristics of this L-BDH structure is the differential binding manner of ME (Fig. 2). The ME molecule is located in the catalytic cleft formed by residues Ser141, Ileu142, Ala143, Phe148, Tyr154, Pro184, Gly185, Ileu186, Met191, Trp192 and Ile195. These residues may also be involved in substrate recognition and in stereospecific redox reactions. The location of the bound ME in L-BDH is roughly the same as that in meso-BDH. The relative position of the thiol group is almost same in both BDHs with the same hydrogen bond from the conserved catalytic Tyr residue (Fig. 2). The catalytic Ser residue is also located very close to the thiol group, being 3.8 Å in L-BDH and 4.4 Å in meso-BDH. On the other hand, the hydroxyl group of ME is hydrogen bonded to Gly185, similar to meso-BDH (Gly183), and to Trp192 in L-BDH, but inconsistent with Gln140 in mesoBDH. The distinct inhibitor binding orientation in two BDHs implies that stereospecific substrate-binding is due to subtle amino acid differences. 3.4. Substrate-binding and stereospecificity of BDHs A structural comparison of the two types of BDHs has indicated that both enzymes share common overall structure, cofactor-binding mode and catalytic mechanism. Therefore, L-BDH and mesoBDH must bind their substrate in different manners. ME is a competitive inhibitor with a smaller molecular size, and the number of substituent groups is same as that of L-BD. Therefore, ME is expected to mimic the binding mode of the real substrate. Substrate binding in the BDHs was postulated based on the product configuration, the inhibitor binding mode and the stereochemical arrange-
ment toward the pro-S hydride of the bound cofactor and the catalytic residues of Tyr and Ser. The C2 and C3 chiral carbon atoms of L-BD are both in the S configuration, whereas those of meso-BD are in the S and R configuration, respectively. Thus, the only difference between L-BD and meso-BD is the configuration at the C3 chiral carbon. Superposition of bound ME within L-BDH and mesoBDH implies that the binding environments are similar at the thiol group side but they differ at the hydroxyl group side. We present here a working hypothesis for the binding mode of L-BD and meso-BD, referring to the binding mode of ME. The orientation of BD is with the C2 chiral carbon at the thiol side and the C3 carbon at the hydroxyl side. There is enough space for additional C1 and C4 methyl groups of the substrate, in both cases of L-BD and meso-BD. The structures of predicted substrate-binding sites in LBDH and meso-BDH are shown in Fig. 3. The hydroxyl groups being catalyzed are located within hydrogen bonding distances from the catalytic Tyr and/or Ser, while the other hydroxyl groups with different stereochemistry in two BDs are hydrogen bonded distinctively as seen in ME bindings. Of the 11 residues that line the active site in L-BDH, only two are different from those in meso-BDH. The different residues are Ile142 (Gly140 in meso-BDH) and Phe148 (Asn146 in meso-BDH). In the meso-BDH structure, the carbonyl group of Gly183 and the side chain NH atom of Gln140 would make hydrogen bonds with the hydroxyl group of meso-BD. In the L-BDH structure, the carbonyl group of Gly185 and the side chain NH atom of Trp192 would form hydrogen bonds with the hydroxyl group at the C3 carbon of LBD. Therefore, we propose that switching two kinds of hydrogen bonds, one from a Trp residue and the other from a Gln residue, is crucial to differentiate the stereoisomers stereospecifically. In order to verify this hypothesis, we prepared mutant enzymes and measured their stereoselectivities. 3.5. Stereospecificities of the mutant enzymes In order to deduce the mechanism of chiral recognition of BDHs, based on the shape and physiochemical properties of the substrate-binding site, site-directed mutagenesis was performed. Single and double mutants of L-BDH and meso-BDH were made, in which the two residues that differ in the substrate-binding site between two BDHs were exchanged. As shown in Table 2, stereospec-
222
M. Otagiri et al. / FEBS Letters 584 (2010) 219–223
Fig. 2. Close-up views of the mercaptoethanol-binding site in L-BDH (A) and in meso-BDH (B). Hydrogen bond distances and the distances from catalytic residues are depicted as yellow broken lines.
ificity of the mutant enzymes was analyzed with respect to their activity towards L-BD and meso-BD. Although the Q140I or N146F mutant of meso-BDH show some activity (0.1 U/mg or 19.4 U/mg, respectively) with meso-BD as substrate, no activity was detected in the I142Q or F148N mutant of L-BDH. In both L-BDH and mesoBDH, neither single nor double mutations were sufficient to change the stereospecificity of BDHs. As already discussed, 3D comparison has confirmed that a conserved tryptophan side chain provides a hydrogen bond to a substrate only in the L-BDH structure (Trp192). Since it is difficult to change the side chain conformation by site-directed mutagenesis, a non-complementary triple mutant of meso-BDH (Q140I, N146F and W190H) was prepared in order to introduce an additional hydrogen-bond donating site at Trp190 of meso-BDH, and the stereoselectivities were measured. The stereospecificity of the triple mutant of meso-BDH was shifted from the meso-type to the L-type. This change might be occurred by a more pronounced decrease for the meso-substrate. However, the mutant showed an increase in Km value such that it was higher than wild-type L-BDH. Therefore, an additional substitution is necessary to attain reversed stereospecificity. The results presented here demonstrate that while BDH stereoselectivity is controlled, at least in part, by tuning the orientation
Table 2 Specific activity of WT enzymes and their mutants. Enzymes
Specific activity (U/mg)a meso-BD
L-BD
Wild type I142Q F148N I142Q/F148N
0.1 Tb T T
4.3 (Km: 0.1 mM) T T T
meso-BDH Wild type Q140I N146F Q140I/N146F Q140I/N146F/W190H
177 (Km: 2.0 mM) 0.1 19.4 3.1 T
5.5 0.1 1.1 4.5 2.9 (Km: 70 mM)
L-BDH
a b
Value is the average obtained from three trials within ±3.2% dispersion range. T means trace activities under 0.1.
of substrate in the binding site, it still remains unclear which factors underlie the affinity for substrate. Further structural analysis of the crystallized mutants will help to clarify the subtle structural differences at the substrate-binding site, and this will deepen our understanding of these enzymes. Structural investigations of the
Fig. 3. Schematic drawings of the substrate binding model for L-BD (a) and meso-BD (b). Hydrogen bonds are indicated with dashed lines. The catalytic Tyr and Ser residues are colored in brown.
M. Otagiri et al. / FEBS Letters 584 (2010) 219–223
mutants are currently under way. This information will enable further rational engineering of the enzyme to change the topology of the substrate-binding site and adjust substrate stereoselectivity. Acknowledgements We thank Professor S. Wakatsuki and the beamline staff at beamline BL-6A of the Photon Factory, Japan. References [1] Ui, S., Matsuyama, N., Masuda, H. and Muraki, H. (1984) Mechanism for the formation of 2, 3-butanediol stereoisomers in Klebsiella pneumoniae. J. Ferment. Techol. 62, 551–559. [2] Ui, S., Okajima, Y., Mimura, A., Kanai, H., Kobayashi, T. and Kudo, T. (1997) Sequence analysis of the gene for and characterization of D-acetoin forming meso-2, 3-butanediol dehydrogenase of Klebsiella pneumoniae expressed in Escherichia coli. J. Ferment. Bioeng. 86, 290–295. [3] Ui, S., Otagiri, M., Mimura, A., Dohmae, N., Takio, K., Ohkuma, M. and Kudo, T. (1998) Cloning, expression and nucleotide sequence of the L-2, 3-butanediol dehydrogenase gene from Brevibacterium saccharolyticum C-1012. J. Ferment. Bioeng. 83, 32–37. [4] Jörnvall, H., Persson, B., Krook, M., Arian, S., Gonzalez-Duarte, R., Jeffery, J. and Ghosh, D. (1995) Short-chain dehydrogenases/reductases (SDR). Biochemistry 34, 6003–6013. [5] Filling, C., Berndt, K.D., Benach, J., Knapp, S., Prozorovski, T., Nordling, E., Ladenstein, R., Jörnvall, H. and Oppermann, U. (2002) Critical residues for structure and catalysis in short-chain dehydrogenase/reductases. J. Biol. Chem. 227, 25677–25684. [6] Gosh, D., Weeks, C.M., Grouchlski, P., Duax, W.L., Erman, M., Rimsay, R.L. and Orr, J.C. (1991) Three-dimensional structure of holo 3,20-hydroxysteroid dehydrogenase: a member of a short-chain dehydrogenase family. Proc. Natl. Acad. Sci. USA 88, 10064–10068. [7] Brenach, J., Atrian, S., Gonzalez-Duarte, R. and Ladenstein, R. (1998) The refined crystal structure of Drosophila lebanonensis alcohol dehydrogenase at 1.9 Å resolution. J. Mol. Biol. 282, 383–399.
223
[8] Tanaka, N., Nonaka, T., Nakanishi, M., Deyashiki, Y., Hara, A. and Mitsui, Y. (1996) Crystal structure of the ternary complex of mouse lung carbonyl reductase at 1.8 Å resolution: the structural origin of coenzyme specificity in the short-chain dehydrogenase/reductase family. Structure 4, 33–45. [9] Otagiri, M., Kurisu, G., Ui, S., Takusagawa, Y., Ohkuma, M., Kudo, T. and Kusunoki, M. (2001) Crystal structure of meso-2,3-butanediol dehydogenase in a complex with NAD+ and inhibitor mercaptoethanol at 1.7 Å resolution for understanding of chiral substrate recognition mechanisms. J. Biochem. (Tokyo) 129, 205–208. [10] McPherson, A. (1982) Preparation and Analysis of Protein Crystals, first ed, John Wiley, New York. pp 96–97. [11] Otagiri, M., Kurisu, G., Ui, S., Ohkuma, M., Kudo, T. and Kusunoki, M. (2001) Crystallization and preliminary X-ray studies of L-(+)-2,3-butanediol dehydrogenase from Brevibacterium saccharolyticum C-I012. Protein Peptide Lett. 8, 57–61. [12] Navza, J. (1994) AMoRe: an automated package for molecular replacement. Acta Crystallogr. Sect. A 50, 157–163. [13] Brünger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., GrosseKunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., Read, R.J., Rice, L.M., Simonson, T. and Warren, G.L. (1998) Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921. [14] Jones, T.A., Zou, J.Y., Cowan, S.W. and Kjeldgaard, M. (1991) Improved methods for binding protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A47, 110–119. [15] Dixon, M. (1953) The determination of enzyme inhibitor constants. Biochem. J. 55, 170–171. [16] Rossmnn, M.G., Liljas, A., Branden, C.I. and Banaszak, L.J. (1975) Evolutionary and structural relationships among dehydrogenases, third ed (Boyer, P.D., Ed.), The enzymes, vol. 11A, pp. 61–102, Academic Press, New York. [17] DeLano, W.L. (2002) The PyMOL molecular graphics system, DeLano Scientific, Palo Alto, CA. [18] Yamamoto, K., Kurisu, G., Kusunoki, M., Tabata, S., Urabe, I. and Osaki, S. (2001) Crystal structure of glucose dehydrogenase from Bacillus megaterium IWG3 at 1.7 Å resolution. J. Biochem. (Tokyo) 129, 303–312.
journal homepage: www.FEBSLetters.org
Structural basis for chiral substrate recognition by two 2,3-butanediol dehydrogenases Masato Otagiri a,1, Sadaharu Ui b,*, Yuhsuke Takusagawa b, Takashi Ohtsuki b, Genji Kurisu c,1, Masami Kusunoki b a b c
Laboratory of Environmental Molecular Biology, RIKEN, 1-7-29 Suehiro-cho, Tsurumi-ward, Kanagawa 230-0045, Japan Graduate School of Medicine and Engineering, University of Yamanashi, Takeda, Kofu 400-8511, Japan Division of Structural Biology, Institute for Protein Research, Osaka University, Yamadaoka, Suita 565-0871, Japan
a r t i c l e
i n f o
Article history: Received 1 September 2009 Revised 19 November 2009 Accepted 19 November 2009 Available online 24 November 2009 Edited by Hans Eklund Keywords: Butanediol dehydrogenase Chiral recognition Stereoisomer Short-chain dehydrogenase/reductase family X-ray crystallography
a b s t r a c t 2,3-Butanediol dehydrogenase (BDH) catalyzes the NAD-dependent redox reaction between acetoin and 2,3-butanediol. There are three types of homologous BDH, each stereospecific for both substrate and product. To establish how these homologous enzymes possess differential stereospecificities, we determined the crystal structure of L-BDH with a bound inhibitor at 2.0 Å. Comparison with the inhibitor binding mode of meso-BDH highlights the role of a hydrogen-bond from a conserved Trp residue192. Site-directed mutagenesis of three active site residues of meso-BDH, including Trp190, which corresponds to Trp192 of L-BDH, converted its stereospecificity to that of L-BDH. This result confirms the importance of conserved residues in modifying the stereospecificity of homologous enzymes. Ó 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
1. Introduction Enzymatic reactions are often stereospecific and sometimes stereoselective. Engineering the stereospecifities of enzymes, although very difficult to control, has been a significant challenge. As in the well-known structural independence between L-lactose dehydrogenase and D-lactose dehydrogenase, the artificial design of enantiospecificity possessed by a native enzyme is not so simple. 2,3-Butanediol (BD) is known as a by-product of sugar metabolism in microorganisms. There are three types of stereoisomer of BD (D-, L-, and meso-forms). Since the ratio of BD stereoisomers depends on the organisms and the culture conditions of the fermentation process, various 2,3-butanediol dehydrogenases (BDHs) catalyzing the redox reaction between acetoin (AC) and BD were identified from many microorganisms. Of these BDHs, meso-BDH from Klebsiella pneumoniae IAM 1063, specific for meso-BD and D-AC, and L-BDH from Brevibacterium saccharolyticum C-1012, spe-
Abbreviations: AC, acetoin; BD, 2,3-butanediol; BDH, butanediol dehydrogenase; SDR, short-chain dehydrogenase/reductase family * Corresponding author. Fax: +81 55 220 8538. E-mail address: [email protected] (S. Ui). 1 Authors having equally contributed to this work.
cific for L-BD and L-AC, are both biochemically and enzymatically well characterized [1]. These two enzymes share a 50% homology of their amino acid sequence [2,3]. Therefore, it is very interesting to ascertain exactly how the homologous BDHs distinguish diastereomers. L- and meso-BDH belong to the short-chain dehydrogenase/ reductase (SDR) family, based on their amino acid sequences [3]. The substrates of SDR family enzymes vary widely in molecular size and properties, as they include alcohols, glucose and steroids [4]. Many crystal structures of SDR enzymes have been reported [5], and all have core structure similar to the Rossmann fold [6]. The variable loop region located in the C-terminal side of molecule was considered to be the reason this conserved structural frame could recognize a wide variety of substrates [7]. There is thought to be a common ‘‘Asn-Ser-Tyr-Lys catalytic tetrad” in SDR family enzymes, based upon biochemical and crystallographic studies [8]. We previously determined the crystal structure of meso-BDH from Klebsiella pneumoniae IAM 1063 at 1.7 Å resolution [9], and here we describe X-ray crystallographic analysis of L-BDH from B. saccharolyticum C-1012 at 2.0 Å resolution. It is now possible to compare the chiral recognition mechanisms of the two BDHs. We discuss the structural basis for stereospecificity on the basis of crystallographic and site-directed mutagenesis experiments.
0014-5793/$36.00 Ó 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2009.11.068
220
M. Otagiri et al. / FEBS Letters 584 (2010) 219–223
Table 1 Summary of refinement statistics. L-BDH
Space group Cell parameters Resolution (Å) Rfactor/Rfreea
P1 a = 60.8 Å, b = 69.2 Å, c = 127.4 Å, a = 96.1°, b = 100.2°, c = 109.6° 40.0–2.0 19.3/24.0
No. of non-hydrogen atoms Protein 15,208 Coenzyme 352 Mercaptoethanol 32 4 Mg+ Solvent 839 Rms deviations from ideal values Bond length (Å) 0.006 Bond angles () 1.2 Ramachandran plot analysis (%) Most favored 91.1 Allowed 8.9 a
Five percent of the reflections were set aside for an Rfree calculation.
2. Materials and methods 2.1. Enzyme stereospecific, preparation, crystallization and X-ray data collection The stereospecific type of BDH, L-BDH or meso-BDH, was used based on its stronger substrate isomer, L-BD or meso-BD, for keeping consistency of the usage until now. Over-expression and purification of recombinant L-BDH were performed according to a previously reported method [3]. L-BDH was co-crystallized with NAD+ and mercaptoethanol (ME) by the hanging drop vapor diffusion method [10]. Details of crystallization and data collection were described [11]. 2.2. Structure determination A crystal structure of L-BDH was solved using the molecular replacement method with the program AMORE [12] using mesoBDH (PDB code 1GEG) [9] as a search model. Crystallographic refinements were carried out with the program CNS version 0.9 [13]. The molecular model was built using the graphics program O [14] interpreting 2|Fo| |Fc| and |Fo| |Fc| electron density maps, which were improved through several cycles of model building and refinement with CNS. The crystallographic statistics are summarized in Table 1. Atomic coordinates and structure factors have been deposited in the Protein Data Bank. The PDB ID is 3A28. 2.3. Site-directed mutagenesis, enzyme assay and kinetic analysis Site-directed mutagenesis was performed using the Quikchange site-directed mutagenesis kit (Stratagene). The activity of BDH was assayed by recording the absorption of NADH at 340 nm. The standard reaction mixture consisted of 0.2 M sodium phosphate, pH 8.0, 2.5 lmol NAD+, 50 lmol BD and enzyme, in a total volume of 3.0 mL. One unit of enzyme activity was defined as the amount of enzyme that forms 1 lmol of NADH per minute at 37 °C. The Ki value of ME against BDH was determined by Dixon plots [15]. The all values obtained were averages of three trials. 3. Results and discussion 3.1. Overall structure of L-BDH The crystal structure of L-BDH complexed with NAD+ and ME was determined, identifying two identical homotetramers per
asymmetric unit. The eight subunits (A–H) in the asymmetric unit can be superimposed with an averaged, pairwise root-meansquare deviation (rmsd) of 0.15 Å for 258 Ca atoms. Therefore, the structure of subunit A will be taken hereafter as representative of the eight subunits. The crystal structure of the L-BDH protomer is shown in Fig. 1A. The L-BDH subunit is composed of seven b strands, six a helices and two short a helices (aFG1 and aFG2), which form a small lobe on top of the core domain. The conserved active site residues Asn112, Ser141, Tyr154 and Lys158 are located near NAD+ molecule. The roles of catalytic residues in SDR enzymes have been intensively studied. The Tyr residue acts as a general base for proton transfer, while the Lys residue is important in cofactor-binding and lowering the pKa value of the hydroxyl group of the Tyr residue. The Ser residue is suggested to be involved in the binding of substrate, the reaction intermediate and the product and/or the hydroxyl group of the conserved Tyr residue. The Asn residue plays a role in maintaining the active site configuration, allowing the formation of a proton relay system involving Lys, Tyr, the nicotinamide ribose moiety of the coenzyme and the substrate. A NAD+ molecule is bound to the enzyme in an extended conformation at the carboxyl ends of the central parallel b-sheet. The overall conformation of NAD+ and its binding mode to the Rossmann fold is similar to that found in other SDR enzymes [16]. A close-up view of the binding site is shown in Fig. 1B. The adenine ring of NAD+ binds in a hydrophobic pocket formed by the side chains of Ile91, Val111, Asp61, Leu34 and Ala89. Hydrogen bonds are formed by the N1 atom of adenine and the side chain of Asp61. The adenine ribose and pyrophosphate moieties are situated on the turn formed by Gly9 and Gly15. The turn between the first b-strand and the second a-helix has the sequence GXXXGX, which is a characteristic fingerprint of the nucleotide-binding motif of the SDR enzymes. The 20 - and 30 -hydroxyl groups of the ribose form hydrogen bonds with the side chain of Asp33, which is common to SDR enzymes that selectively bind NAD+. The pyrophosphate moiety interacts with the enzyme through hydrogen bonds with the side chains of Gln12 and Thr189 and the main-chain nitrogen of Ile14. The nicotinamide ring moiety makes polar interactions with the main-chain of Val187 and the side chain of Thr189. These residues are located in the N-terminal part of aFG1. The ribose moiety forms hydrogen bonds with the side chains of Tyr154 and Lys158 and with the main-chain oxygen of Asn88. Other interactions with NAD+ are similar to those seen in structures of other SDR enzymes. 3.2. Comparison with other SDRs A structural comparison of L-BDH with other SDR enzymes demonstrates that L-BDH exhibits a high degree of overall structural similarity to enzymes such as meso-BDH [9], glucose dehydrogenase [18] and 3a, 20b-hydroxysteroid dehydrogenase [6]. Structural comparison of two BDHs reveals that L-BDH has a very similar overall structure to meso-BDH, except for some conformational differences on the molecular surface and in the terminal regions. The rmsd of the superimposed L-BDH and meso-BDH structures is 0.5 Å. Furthermore, the manner of tetramer formation is also identical in L-BDH and meso-BDH. Comparison of the NAD+ binding site in L-BDH and the corresponding site in meso-BDH reveals that the residues responsible for NAD+ binding and their spatial organization are highly conserved between L-BDH and mesoBDH. This result indicates that L-BDH binds the cofactor in the same way as meso-BDH. At present, the ‘Asn-Ser-Tyr-Lys catalytic tetrad’ is considered important in SDR catalysis [5]. The catalytic tetrad residues (Asn112, Ser141, Tyr154 and Lys158) are situated in identical positions in meso-BDH (Asn110, Ser139, Tyr152 and Lys156). The position of the
M. Otagiri et al. / FEBS Letters 584 (2010) 219–223
221
Fig. 1. Structure of L-BDH. (A) Ribbon drawing of the L-BDH subunit with NAD+ and an inhibitor, mercaptoethanol (BME). (B) Active site structure of L-BDH. A 2|Fo| |Fc| omit map is superimposed on NAD+ and mercaptoethanol at the 1.0 r level. The catalytic tetrad is colored in orange. All figures showing structures in this paper were produced with the program PYMOL [17].
NAD+ molecule as well as surrounding residues is conserved in LBDH and meso-BDH. This structural conservation strongly suggests that the same reaction mechanism is common to both BDH enzymes. 3.3. Inhibitor-binding sites Initially, we attempted to prepare a complex of L-BDH with substrate, but these efforts failed. As seen in the meso-BDH structure, the electron density corresponding to an ME, which was included in the crystallization solution, was observed. Kinetic analysis in this study showed that ME was a competitive inhibitor of L-BDH (Ki = 4.3 mM at pH 8.0). One of the most outstanding characteristics of this L-BDH structure is the differential binding manner of ME (Fig. 2). The ME molecule is located in the catalytic cleft formed by residues Ser141, Ileu142, Ala143, Phe148, Tyr154, Pro184, Gly185, Ileu186, Met191, Trp192 and Ile195. These residues may also be involved in substrate recognition and in stereospecific redox reactions. The location of the bound ME in L-BDH is roughly the same as that in meso-BDH. The relative position of the thiol group is almost same in both BDHs with the same hydrogen bond from the conserved catalytic Tyr residue (Fig. 2). The catalytic Ser residue is also located very close to the thiol group, being 3.8 Å in L-BDH and 4.4 Å in meso-BDH. On the other hand, the hydroxyl group of ME is hydrogen bonded to Gly185, similar to meso-BDH (Gly183), and to Trp192 in L-BDH, but inconsistent with Gln140 in mesoBDH. The distinct inhibitor binding orientation in two BDHs implies that stereospecific substrate-binding is due to subtle amino acid differences. 3.4. Substrate-binding and stereospecificity of BDHs A structural comparison of the two types of BDHs has indicated that both enzymes share common overall structure, cofactor-binding mode and catalytic mechanism. Therefore, L-BDH and mesoBDH must bind their substrate in different manners. ME is a competitive inhibitor with a smaller molecular size, and the number of substituent groups is same as that of L-BD. Therefore, ME is expected to mimic the binding mode of the real substrate. Substrate binding in the BDHs was postulated based on the product configuration, the inhibitor binding mode and the stereochemical arrange-
ment toward the pro-S hydride of the bound cofactor and the catalytic residues of Tyr and Ser. The C2 and C3 chiral carbon atoms of L-BD are both in the S configuration, whereas those of meso-BD are in the S and R configuration, respectively. Thus, the only difference between L-BD and meso-BD is the configuration at the C3 chiral carbon. Superposition of bound ME within L-BDH and mesoBDH implies that the binding environments are similar at the thiol group side but they differ at the hydroxyl group side. We present here a working hypothesis for the binding mode of L-BD and meso-BD, referring to the binding mode of ME. The orientation of BD is with the C2 chiral carbon at the thiol side and the C3 carbon at the hydroxyl side. There is enough space for additional C1 and C4 methyl groups of the substrate, in both cases of L-BD and meso-BD. The structures of predicted substrate-binding sites in LBDH and meso-BDH are shown in Fig. 3. The hydroxyl groups being catalyzed are located within hydrogen bonding distances from the catalytic Tyr and/or Ser, while the other hydroxyl groups with different stereochemistry in two BDs are hydrogen bonded distinctively as seen in ME bindings. Of the 11 residues that line the active site in L-BDH, only two are different from those in meso-BDH. The different residues are Ile142 (Gly140 in meso-BDH) and Phe148 (Asn146 in meso-BDH). In the meso-BDH structure, the carbonyl group of Gly183 and the side chain NH atom of Gln140 would make hydrogen bonds with the hydroxyl group of meso-BD. In the L-BDH structure, the carbonyl group of Gly185 and the side chain NH atom of Trp192 would form hydrogen bonds with the hydroxyl group at the C3 carbon of LBD. Therefore, we propose that switching two kinds of hydrogen bonds, one from a Trp residue and the other from a Gln residue, is crucial to differentiate the stereoisomers stereospecifically. In order to verify this hypothesis, we prepared mutant enzymes and measured their stereoselectivities. 3.5. Stereospecificities of the mutant enzymes In order to deduce the mechanism of chiral recognition of BDHs, based on the shape and physiochemical properties of the substrate-binding site, site-directed mutagenesis was performed. Single and double mutants of L-BDH and meso-BDH were made, in which the two residues that differ in the substrate-binding site between two BDHs were exchanged. As shown in Table 2, stereospec-
222
M. Otagiri et al. / FEBS Letters 584 (2010) 219–223
Fig. 2. Close-up views of the mercaptoethanol-binding site in L-BDH (A) and in meso-BDH (B). Hydrogen bond distances and the distances from catalytic residues are depicted as yellow broken lines.
ificity of the mutant enzymes was analyzed with respect to their activity towards L-BD and meso-BD. Although the Q140I or N146F mutant of meso-BDH show some activity (0.1 U/mg or 19.4 U/mg, respectively) with meso-BD as substrate, no activity was detected in the I142Q or F148N mutant of L-BDH. In both L-BDH and mesoBDH, neither single nor double mutations were sufficient to change the stereospecificity of BDHs. As already discussed, 3D comparison has confirmed that a conserved tryptophan side chain provides a hydrogen bond to a substrate only in the L-BDH structure (Trp192). Since it is difficult to change the side chain conformation by site-directed mutagenesis, a non-complementary triple mutant of meso-BDH (Q140I, N146F and W190H) was prepared in order to introduce an additional hydrogen-bond donating site at Trp190 of meso-BDH, and the stereoselectivities were measured. The stereospecificity of the triple mutant of meso-BDH was shifted from the meso-type to the L-type. This change might be occurred by a more pronounced decrease for the meso-substrate. However, the mutant showed an increase in Km value such that it was higher than wild-type L-BDH. Therefore, an additional substitution is necessary to attain reversed stereospecificity. The results presented here demonstrate that while BDH stereoselectivity is controlled, at least in part, by tuning the orientation
Table 2 Specific activity of WT enzymes and their mutants. Enzymes
Specific activity (U/mg)a meso-BD
L-BD
Wild type I142Q F148N I142Q/F148N
0.1 Tb T T
4.3 (Km: 0.1 mM) T T T
meso-BDH Wild type Q140I N146F Q140I/N146F Q140I/N146F/W190H
177 (Km: 2.0 mM) 0.1 19.4 3.1 T
5.5 0.1 1.1 4.5 2.9 (Km: 70 mM)
L-BDH
a b
Value is the average obtained from three trials within ±3.2% dispersion range. T means trace activities under 0.1.
of substrate in the binding site, it still remains unclear which factors underlie the affinity for substrate. Further structural analysis of the crystallized mutants will help to clarify the subtle structural differences at the substrate-binding site, and this will deepen our understanding of these enzymes. Structural investigations of the
Fig. 3. Schematic drawings of the substrate binding model for L-BD (a) and meso-BD (b). Hydrogen bonds are indicated with dashed lines. The catalytic Tyr and Ser residues are colored in brown.
M. Otagiri et al. / FEBS Letters 584 (2010) 219–223
mutants are currently under way. This information will enable further rational engineering of the enzyme to change the topology of the substrate-binding site and adjust substrate stereoselectivity. Acknowledgements We thank Professor S. Wakatsuki and the beamline staff at beamline BL-6A of the Photon Factory, Japan. References [1] Ui, S., Matsuyama, N., Masuda, H. and Muraki, H. (1984) Mechanism for the formation of 2, 3-butanediol stereoisomers in Klebsiella pneumoniae. J. Ferment. Techol. 62, 551–559. [2] Ui, S., Okajima, Y., Mimura, A., Kanai, H., Kobayashi, T. and Kudo, T. (1997) Sequence analysis of the gene for and characterization of D-acetoin forming meso-2, 3-butanediol dehydrogenase of Klebsiella pneumoniae expressed in Escherichia coli. J. Ferment. Bioeng. 86, 290–295. [3] Ui, S., Otagiri, M., Mimura, A., Dohmae, N., Takio, K., Ohkuma, M. and Kudo, T. (1998) Cloning, expression and nucleotide sequence of the L-2, 3-butanediol dehydrogenase gene from Brevibacterium saccharolyticum C-1012. J. Ferment. Bioeng. 83, 32–37. [4] Jörnvall, H., Persson, B., Krook, M., Arian, S., Gonzalez-Duarte, R., Jeffery, J. and Ghosh, D. (1995) Short-chain dehydrogenases/reductases (SDR). Biochemistry 34, 6003–6013. [5] Filling, C., Berndt, K.D., Benach, J., Knapp, S., Prozorovski, T., Nordling, E., Ladenstein, R., Jörnvall, H. and Oppermann, U. (2002) Critical residues for structure and catalysis in short-chain dehydrogenase/reductases. J. Biol. Chem. 227, 25677–25684. [6] Gosh, D., Weeks, C.M., Grouchlski, P., Duax, W.L., Erman, M., Rimsay, R.L. and Orr, J.C. (1991) Three-dimensional structure of holo 3,20-hydroxysteroid dehydrogenase: a member of a short-chain dehydrogenase family. Proc. Natl. Acad. Sci. USA 88, 10064–10068. [7] Brenach, J., Atrian, S., Gonzalez-Duarte, R. and Ladenstein, R. (1998) The refined crystal structure of Drosophila lebanonensis alcohol dehydrogenase at 1.9 Å resolution. J. Mol. Biol. 282, 383–399.
223
[8] Tanaka, N., Nonaka, T., Nakanishi, M., Deyashiki, Y., Hara, A. and Mitsui, Y. (1996) Crystal structure of the ternary complex of mouse lung carbonyl reductase at 1.8 Å resolution: the structural origin of coenzyme specificity in the short-chain dehydrogenase/reductase family. Structure 4, 33–45. [9] Otagiri, M., Kurisu, G., Ui, S., Takusagawa, Y., Ohkuma, M., Kudo, T. and Kusunoki, M. (2001) Crystal structure of meso-2,3-butanediol dehydogenase in a complex with NAD+ and inhibitor mercaptoethanol at 1.7 Å resolution for understanding of chiral substrate recognition mechanisms. J. Biochem. (Tokyo) 129, 205–208. [10] McPherson, A. (1982) Preparation and Analysis of Protein Crystals, first ed, John Wiley, New York. pp 96–97. [11] Otagiri, M., Kurisu, G., Ui, S., Ohkuma, M., Kudo, T. and Kusunoki, M. (2001) Crystallization and preliminary X-ray studies of L-(+)-2,3-butanediol dehydrogenase from Brevibacterium saccharolyticum C-I012. Protein Peptide Lett. 8, 57–61. [12] Navza, J. (1994) AMoRe: an automated package for molecular replacement. Acta Crystallogr. Sect. A 50, 157–163. [13] Brünger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., GrosseKunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., Read, R.J., Rice, L.M., Simonson, T. and Warren, G.L. (1998) Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921. [14] Jones, T.A., Zou, J.Y., Cowan, S.W. and Kjeldgaard, M. (1991) Improved methods for binding protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A47, 110–119. [15] Dixon, M. (1953) The determination of enzyme inhibitor constants. Biochem. J. 55, 170–171. [16] Rossmnn, M.G., Liljas, A., Branden, C.I. and Banaszak, L.J. (1975) Evolutionary and structural relationships among dehydrogenases, third ed (Boyer, P.D., Ed.), The enzymes, vol. 11A, pp. 61–102, Academic Press, New York. [17] DeLano, W.L. (2002) The PyMOL molecular graphics system, DeLano Scientific, Palo Alto, CA. [18] Yamamoto, K., Kurisu, G., Kusunoki, M., Tabata, S., Urabe, I. and Osaki, S. (2001) Crystal structure of glucose dehydrogenase from Bacillus megaterium IWG3 at 1.7 Å resolution. J. Biochem. (Tokyo) 129, 303–312.