Portal de Periódicos da CAPES

Structural Rationalization for the Lack of Stereospecificity in Coenzyme B12-dependent Diol Dehydratase

2003; Elsevier BV; Volume: 278; Issue: 25 Linguagem: Inglês

10.1074/jbc.m301513200

1083-351X

Naoki Shibata, Yuka Nakanishi, Masaki Fukuoka, Mamoru Yamanishi, Noritake Yasuoka, Tetsuo Toraya,

Folate and B Vitamins Research

Adenosylcobalamin-dependent diol dehydratase of Klebsiella oxytoca is apparently not stereospecific and catalyzes the conversion of both (R)- and (S)-1,2-propanediol to propionaldehyde. To explain this unusual property of the enzyme, we analyzed the crystal structures of diol dehydratase in complexes with cyanocobalamin and (R)- or (S)-1,2-propanediol. (R)- and (S)-isomers are bound in a symmetrical manner, although the hydrogen-bonding interactions between the substrate and the active-site residues are the same. From the position of the adenosyl radical in the modeled "distal" conformation, it is reasonable for the radical to abstract the pro-R and pro-S hydrogens from (R)- and (S)-isomers, respectively. The hydroxyl groups in the substrate radicals would migrates from C(2) to C(1) by a suprafacial shift, resulting in the stereochemical inversion at C(1). This causes 60° clockwise and 70° counterclockwise rotations of the C(1)–C(2) bond of the (R)- and (S)-isomers, respectively, if viewed from K+. A modeling study of 1,1-gem-diol intermediates indicated that new radical center C(2) becomes close to the methyl group of 5′-deoxyadenosine. Thus, the hydrogen back-abstraction (recombination) from 5′-deoxyadenosine by the product radical is structurally feasible. It was also predictable that the substitution of the migrating hydroxyl group by a hydrogen atom from 5′-deoxyadenosine takes place with the inversion of the configuration at C(2) of the substrate. Stereospecific dehydration of the 1,1-gem-diol intermediates can also be rationalized by assuming that Asp-α335 and Glu-α170 function as base catalysts in the dehydration of the (R)- and (S)-isomers, respectively. The structure-based mechanism and stereochemical courses of the reaction are proposed. Adenosylcobalamin-dependent diol dehydratase of Klebsiella oxytoca is apparently not stereospecific and catalyzes the conversion of both (R)- and (S)-1,2-propanediol to propionaldehyde. To explain this unusual property of the enzyme, we analyzed the crystal structures of diol dehydratase in complexes with cyanocobalamin and (R)- or (S)-1,2-propanediol. (R)- and (S)-isomers are bound in a symmetrical manner, although the hydrogen-bonding interactions between the substrate and the active-site residues are the same. From the position of the adenosyl radical in the modeled "distal" conformation, it is reasonable for the radical to abstract the pro-R and pro-S hydrogens from (R)- and (S)-isomers, respectively. The hydroxyl groups in the substrate radicals would migrates from C(2) to C(1) by a suprafacial shift, resulting in the stereochemical inversion at C(1). This causes 60° clockwise and 70° counterclockwise rotations of the C(1)–C(2) bond of the (R)- and (S)-isomers, respectively, if viewed from K+. A modeling study of 1,1-gem-diol intermediates indicated that new radical center C(2) becomes close to the methyl group of 5′-deoxyadenosine. Thus, the hydrogen back-abstraction (recombination) from 5′-deoxyadenosine by the product radical is structurally feasible. It was also predictable that the substitution of the migrating hydroxyl group by a hydrogen atom from 5′-deoxyadenosine takes place with the inversion of the configuration at C(2) of the substrate. Stereospecific dehydration of the 1,1-gem-diol intermediates can also be rationalized by assuming that Asp-α335 and Glu-α170 function as base catalysts in the dehydration of the (R)- and (S)-isomers, respectively. The structure-based mechanism and stereochemical courses of the reaction are proposed. Enzymes are generally stereospecific for their substrates. Adenosylcobalamin (AdoCbl) 1The abbreviations used are: AdoCbl, adenosylcobalamin or coenzyme B12; CN-Cbl, cyanocobalamin; LDAO, lauryldimethylamine oxide; PEG, polyethylene glycol. (coenzyme B12)-dependent diol dehydratase (1Lee Jr., H.A. Abeles R.H. J. Biol. Chem. 1963; 238: 2367-2373Abstract Full Text PDF PubMed Google Scholar, 2Toraya T. Shirakashi T. Kosuga T. Fukui S. Biochem. Biophys. Res. Commun. 1976; 69: 475-481Crossref PubMed Scopus (125) Google Scholar) is, however, not stereospecific and catalyzes the conversion of both (R)- and (S)-1,2-propanediols to propionaldehyde (shown below in Reaction 1, where R = CH3, H, HOCH2). R-CHCH2OH→R-CH2CHO+H2O |OH Reaction 1 Kinetic measurements provided some clues to solve the enigma of apparent lack of stereospecificity of diol dehydratase. When each enantiomer of 1,2-propanediol is run independently, the rate with the (R)-isomer is 1.7–1.8 times higher than that with the (S)-isomer (3Zagalak B. Frey P.A. Karabatsos G.L. Abeles R.H. J. Biol. Chem. 1966; 24: 3028-3035Google Scholar, 4Yamane T. Kato T. Shimizu S. Fukui S. Arch. Biochem. Biophys. 1966; 113: 362-366Crossref PubMed Scopus (10) Google Scholar, 5Bachovchin W.W. Eagar Jr., R.G. Moore K.W. Richards J.H. Biochemistry. 1977; 16: 1082-1092Crossref PubMed Scopus (99) Google Scholar). However, when racemic 1,2-propanediol is used as substrate, the (S)-isomer reacts at a faster rate than the (R)-isomer (6Jensen F.R. Neese R.A. Biochem. Biophys. Res. Commun. 1975; 62: 816-821Crossref PubMed Scopus (9) Google Scholar). This reversal is because of the ratio of Km values (Km(R)/Km(S) = 3.1–3.2). These lines of evidence suggested that the (R)- and (S)-isomers are bound to the enzyme in two different modes with different catalytic efficiency and binding affinity. In other words, the enzyme recognizes the enantiomers as "different" substrates. The stereochemistry of the diol dehydratase reaction established by the labeling experiments is summarized as follows (Fig. 1A). It was shown by Rétey et al. (7Rétey J. Umani-Ronchi A. Seibl J. Arigoni D. Experientia. 1966; 22: 502-503Crossref PubMed Scopus (112) Google Scholar) that [18O]- and unlabeled propionaldehydes are formed from [1-18O]-(S)- and [1-18O]-(R)-1,2-propanediols, respectively. This indicated that 1,1-gem-diol is formed as an intermediate. The initial migration of an OH group from C(2) to C(1) is stereoselective, and the dehydration of the resulting gem-diol undergoes steric control by the enzyme, with only one of the two OH groups on the prochiral center being eliminated. However, the absolute configurations of the gem-diol intermediates and why and how a specific hydroxyl group undergoes elimination remained to be elucidated. The hydrogen atom moves to the adjacent carbon atom without exchange with solvent protons (8Brownstein A.M. Abeles R.H. J. Biol. Chem. 1961; 236: 1199-1200Abstract Full Text PDF Google Scholar). Abeles and co-workers (3Zagalak B. Frey P.A. Karabatsos G.L. Abeles R.H. J. Biol. Chem. 1966; 24: 3028-3035Google Scholar) demonstrated that the pro-S and pro-R hydrogen atoms on C(1) of (S)- and (R)-1,2-propanediols, respectively, migrate to C(2). The migrating OH group is replaced by the hydrogen atom with an accompanying inversion of the configuration at C(2) (3Zagalak B. Frey P.A. Karabatsos G.L. Abeles R.H. J. Biol. Chem. 1966; 24: 3028-3035Google Scholar, 9Rétey J. Umani-Ronchi A. Arigoni D. Experientia. 1966; 22: 72-73Crossref PubMed Scopus (84) Google Scholar). However, when the enantiomeric pair of ethylene glycols labeled stereospecifically with deuterium and tritium is used as substrates, racemic products are obtained (10Arigoni D. Zagalak B. Friedrich W. Vitamin B12. Walter de Gruyter & Co., Berlin1979: 389-410Google Scholar). This indicated that rapid internal rotation in the intermediate occurs before the hydrogen recombination. Abeles and co-workers (11Frey P.A. Essenberg M.K. Abeles R.H. J. Biol. Chem. 1967; 242: 5369-5377Abstract Full Text PDF PubMed Google Scholar, 12Frey P.A. Abeles R.H. J. Biol. Chem. 1966; 241: 2732-2777Abstract Full Text PDF PubMed Google Scholar, 13Frey P.A. Kerwar S.S. Abeles R.H. Biochem. Biophys. Res. Commun. 1967; 29: 873-879Crossref PubMed Scopus (25) Google Scholar, 14Frey P.A. Essenberg M.K. Abeles R.H. Kerwar S.S. J. Am. Chem. Soc. 1970; 92: 4488-4489Crossref PubMed Scopus (20) Google Scholar) demonstrated that the enzyme-bound AdoCbl serves as an intermediate hydrogen carrier, first accepting a hydrogen atom from C(1) of the substrate to C(5′) of the coenzyme and then, in a subsequent step, giving a hydrogen back to C(2) of the product. 5′-Deoxyadenosine was postulated to be an intermediate (11Frey P.A. Essenberg M.K. Abeles R.H. J. Biol. Chem. 1967; 242: 5369-5377Abstract Full Text PDF PubMed Google Scholar, 15Essenberg M.K. Frey P.A. Abeles R.H. J. Am. Chem. Soc. 1971; 93: 1242-1251Crossref PubMed Scopus (99) Google Scholar, 16Abeles R.H. Zagalak B. J. Biol. Chem. 1966; 241: 1245-1246Abstract Full Text PDF PubMed Google Scholar). The formation of cob(II)alamin and an organic radical intermediate during catalysis was demonstrated by optical and electron paramagnetic resonance spectroscopies (17Abeles R.H. Lee Jr., H.A. Ann. N. Y. Acad. Sci. 1964; 112: 695-702Crossref PubMed Scopus (21) Google Scholar, 18Finlay T.H. Valinsky J. Mildvan A.S. Abeles R.H. J. Biol. Chem. 1973; 248: 1285-1290Abstract Full Text PDF PubMed Google Scholar, 19Valinsky J.E. Abeles R.H. Fee J.A. J. Am. Chem. Soc. 1974; 96: 4709-4710Crossref PubMed Scopus (58) Google Scholar, 20Toraya T. Ushio K. Fukui S. Hogenkamp H.P.C. J. Biol. Chem. 1977; 252: 963-970Abstract Full Text PDF PubMed Google Scholar, 21Toraya T. Krodel E. Mildvan A.S. Abeles R.H. Biochemistry. 1979; 18: 417-426Crossref PubMed Scopus (118) Google Scholar, 22Cockle S.A. Hill H.A.O. Williams R.J.P. Davies S.P. Foster M.A. J. Am. Chem. Soc. 1972; 94: 275-277Crossref PubMed Scopus (88) Google Scholar). From these lines of evidence, the minimal mechanism for diol dehydratase was established (Fig. 1, B and C) (21Toraya T. Krodel E. Mildvan A.S. Abeles R.H. Biochemistry. 1979; 18: 417-426Crossref PubMed Scopus (118) Google Scholar, 23Abeles R.H. Dolphin D. Acc. Chem. Res. 1976; 9: 114-120Crossref Scopus (238) Google Scholar, 24Abeles R.H. Zagalak B. Friedrich W. Vitamin B12. Walter de Gruyter & Co., Berlin1979: 373-388Google Scholar, 25Toraya T. Fukui S. Dolphin D. B12. John Wiley & Sons, Inc., New York1982: 233-262Google Scholar). In the diol dehydratase reaction, X and H are the OH group on C(2) and a hydrogen atom on C(1), respectively, and a water molecule is subsequently eliminated from a gem-diol intermediate. This has now been accepted as a general mechanism for all the AdoCbl-dependent intramolecular group-transfer reactions (26Dolphin D. B12. 2. John Wiley & Sons, Inc., New York1982Google Scholar, 27Banerjee R. Chemistry and Biochemistry of B12. John Wiley & Sons, Inc., New York1999Google Scholar), in which a hydrogen atom migrates from one carbon atom of the substrate to an adjacent carbon atom in exchange for group X that moves in the opposite direction. That is, the enzyme-coenzyme interaction leads to the activation of the Co–C bond of the coenzyme. Substrate binding triggers the homolysis of the Co–C bond, forming the adenosyl radical and cob(II)alamin. The adenosyl radical abstracts a hydrogen atom from the substrate, producing a substrate-derived radical and 5′-deoxyadenosine. The substrate radical rearranges to the product radical, which then abstracts a hydrogen atom back from 5′-deoxyadenosine. This leads to the formation of the product and regeneration of the coenzyme. The three-dimensional structures of AdoCbl-dependent enzymes have been solved to understand their fine mechanisms of action (28Mancia F. Keep N.H. Nakagawa A. Leadlay P.F. McSweeney S. Rasmussen B. Bösecke P. Diat O. Evans P.R. Structure. 1996; 4: 339-350Abstract Full Text Full Text PDF PubMed Scopus (429) Google Scholar, 29Shibata N. Masuda J. Tobimatsu T. Toraya T. Suto K. Morimoto Y. Yasuoka N. Structure. 1999; 7: 997-1008Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar, 30Reitzer R. Gruber K. Jogl G. Wagner U.G. Bothe H. Buckel W. Kratky C. Structure. 1999; 7: 891-902Abstract Full Text Full Text PDF PubMed Scopus (195) Google Scholar, 31Sintchak M.D. Arjara G. Kellogg B.A. Stubbe J. Drennan C.L. Nat. Struct. Biol. 2002; 9: 293-300Crossref PubMed Scopus (169) Google Scholar, 32Yamanishi M. Yunoki M. Tobimatsu T. Sato H. Matsui J. Dokiya A. Iuchi Y. Oe K. Suto K. Shibata N. Morimoto Y. Yasuoka N. Toraya T. Eur. J. Biochem. 2002; 269: 4484-4494Crossref PubMed Scopus (96) Google Scholar). We have determined the crystal structures of diol dehydratase in complexes with CN-Cbl (29Shibata N. Masuda J. Tobimatsu T. Toraya T. Suto K. Morimoto Y. Yasuoka N. Structure. 1999; 7: 997-1008Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar, 33Shibata N. Masuda J. Morimoto Y. Yasuoka N. Toraya T. Biochemistry. 2002; 41: 12607-12617Crossref PubMed Scopus (54) Google Scholar) or adeninylpentylcobalamin (34Masuda J. Shibata N. Morimoto Y. Toraya T. Yasuoka N. Structure. 2000; 8: 775-788Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar) with (29Shibata N. Masuda J. Tobimatsu T. Toraya T. Suto K. Morimoto Y. Yasuoka N. Structure. 1999; 7: 997-1008Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar, 34Masuda J. Shibata N. Morimoto Y. Toraya T. Yasuoka N. Structure. 2000; 8: 775-788Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar) and without (33Shibata N. Masuda J. Morimoto Y. Yasuoka N. Toraya T. Biochemistry. 2002; 41: 12607-12617Crossref PubMed Scopus (54) Google Scholar) substrate 1,2-propanediol. The O(1) and O(2) of the substrate are hydrogen-bonded to the respective pair of amino acid residues, namely (Glu-α170 and Gln-α296) and (Asp-α335 and His-α143), respectively (29Shibata N. Masuda J. Tobimatsu T. Toraya T. Suto K. Morimoto Y. Yasuoka N. Structure. 1999; 7: 997-1008Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar). The 1,2-propanediol bound at the active site was assigned to the (S)-isomer, although the crystals were grown in the presence of racemic substrate. The structure-based mechanism of reaction with the (S)-isomer has been proposed (29Shibata N. Masuda J. Tobimatsu T. Toraya T. Suto K. Morimoto Y. Yasuoka N. Structure. 1999; 7: 997-1008Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar, 33Shibata N. Masuda J. Morimoto Y. Yasuoka N. Toraya T. Biochemistry. 2002; 41: 12607-12617Crossref PubMed Scopus (54) Google Scholar, 34Masuda J. Shibata N. Morimoto Y. Toraya T. Yasuoka N. Structure. 2000; 8: 775-788Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). To confirm this assignment and to solve the mechanism of reaction with the (R)-isomer, we crystallized the diol dehydratase·CN-Cbl complexes in the presence of (R)- and (S)-enantiomers and analyzed their three-dimensional structures. The conformations of the enzyme-bound 1,1-gem-diol intermediates from enantiomeric substrates were modeled. Based on the crystal structures and the models, we elucidated the stereochemical courses of the diol dehydratase reaction with the (R)- and (S)-isomers. We propose here the reason why diol dehydratase is apparently not stereospecific for the enantiomers, although hydrogen abstraction and recombination, as well as OH group migration, proceed in stereospecific manners. Materials—(R)-1,2-Propanediol was purchased from TCI, Tokyo, Japan. (S)-1,2-Propanediol was synthesized by reduction of l-lactide (TCI, Tokyo, Japan) with LiAlH4 and purified by distillation under reduced pressure. Crystalline AdoCbl was a gift from Eizai, Tokyo, Japan. All other chemicals were reagent grade commercial products and used without further purification. Purification of Enzyme and Substitution of Racemic 1,2-Propanediol by Enantiomeric Substrates—Diol dehydratase of Klebsiella oxytoca ATCC 8724 was overexpressed in Escherichia coli and purified by the following modification of the method described previously (35Tobimatsu T. Hara T. Sakaguchi M. Kishimoto Y. Wada Y. Isoda M. Sakai T. Toraya T. J. Biol. Chem. 1995; 270: 7142-7148Abstract Full Text Full Text PDF PubMed Scopus (65) Google Scholar, 36Tobimatsu T. Sakai T. Hashida Y. Mizoguchi N. Miyoshi S. Toraya T. Arch. Biochem. Biophys. 1997; 347: 132-140Crossref PubMed Scopus (50) Google Scholar). The enzyme extracted from the crude membrane fractions with 1% Brij35 was loaded onto a DEAE-cellulose column, and the column was washed with 0.01 m potassium phosphate buffer (pH 8.0) containing 2% racemic 1,2-propanediol and 1% phenylmethanesulfonyl fluoride, as described previously. The column was washed again with 0.01 m potassium phosphate buffer (pH 8.0) containing 2% (R)- or (S)-1,2-propanediol and then eluted with the same buffer containing 0.15 m KCl, 2% (R)- or (S)-1,2-propanediol, and 20 mm sucrose monocaprate. After addition of LDAO to a concentration of 0.1%, the apoenzyme was concentrated and incubated with 4 mm CN-Cbl. The resulting enzyme·CN-Cbl complex was dialyzed against 0.01 m potassium phosphate buffer (pH 8.0) containing 2% (R)- or (S)-1,2-propanediol, 0.1% LDAO, and 20 μm CN-Cbl to remove excess unbound CN-Cbl. Crystallization—The final solution contained 50–60 mg/ml of each enzyme·CN-Cbl complex. Crystals of (R)- or (S)-1,2-propanediol-bound enzyme·CN-Cbl complex were grown by the sandwich-drop vapor diffusion method at 4 °C against a reservoir containing 15% (w/v) PEG6000, 0.23 m ammonium sulfate, 0.02 m Tris·HCl buffer (pH 8.0), and 0.2% LDAO (29Shibata N. Masuda J. Tobimatsu T. Toraya T. Suto K. Morimoto Y. Yasuoka N. Structure. 1999; 7: 997-1008Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar, 37Masuda J. Yamaguchi T. Tobimatsu T. Toraya T. Suto K. Shibata M. Morimoto Y. Higuchi Y. Yasuoka N. Acta Crystallogr. Sect. D Biol. Crystallogr. 1999; 55: 907-909Crossref PubMed Scopus (14) Google Scholar). Data Collection—For both complexes, the diffraction data sets were collected at the BL41XU beamline, SPring-8, Japan. The crystal of (S)-1,2-propanediol-bound form was transferred to a cryoprotectant solution containing 15% PEG20000, 17.5% ethylene glycol, and other components of the protein drop except 1,2-propanediol and PEG6000, which had been used for the crystals grown in the presence of racemic 1,2-propanediol. In the case of (R)-isomer, crystals were dissolved in several cryoprotectant solutions available, but the addition of PEG400 directly to a protein drop to a final concentration of ∼20% kept crystals intact. For both forms, the crystal was transferred into a cryo-stream with a cryo-loop and was flash-cooled at 100 K. A total of 180 ° of data were measured. All diffraction data were processed and scaled using DENZO and SCALEPACK (38Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref PubMed Scopus (38572) Google Scholar). Structure Determination—Two kinds of starting models, substrate-bound (29Shibata N. Masuda J. Tobimatsu T. Toraya T. Suto K. Morimoto Y. Yasuoka N. Structure. 1999; 7: 997-1008Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar) at 100 K and substrate-free (33Shibata N. Masuda J. Morimoto Y. Yasuoka N. Toraya T. Biochemistry. 2002; 41: 12607-12617Crossref PubMed Scopus (54) Google Scholar) forms of the enzyme·CN-Cbl complex, for structure determination were used for the first refinement step to cross-check influence of model bias of (S)-1,2-propanediol molecule that had been assigned to the substrate-bound form so far. Coordinates of (S)-1,2-propanediol were deleted from the original substrate-bound structure. Refinement and model building were carried out with CNS (39Brünger A.T. Warren G.L. Acta Crystallogr. Sect. D Biol. Crystallogr. 1998; 54: 905-921Crossref PubMed Scopus (16967) Google Scholar) and XFIT (40McRee D.E. Practical Protein Crystallography. Academic Press, San Diego, CA1993Google Scholar). After rigid-body refinement, 2mFo–DFc electron density maps of (R)-1,2-propanediol from both starting models clearly indicated that C(3) atom of (R)-1,2-propanediol shares the similar position to that of (S)-1,2-propanediol. As expected, the (S)-1,2-propanediol-bound form clearly displayed the same binding mode as observed in the structures crystallized in the presence of racemic 1,2-propanediol (29Shibata N. Masuda J. Tobimatsu T. Toraya T. Suto K. Morimoto Y. Yasuoka N. Structure. 1999; 7: 997-1008Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar, 34Masuda J. Shibata N. Morimoto Y. Toraya T. Yasuoka N. Structure. 2000; 8: 775-788Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). Thus we could assign the initial coordinates of substrate molecule for both forms. Each model was built with rounds of manual rebuilding, positional refinement, and individual B-factor refinement, using all data between 30 and 2.3 Å for the (R)-1,2-propanediol-bound form and between 30 and 1.8 Å for the (S)-1,2-propanediol-bound form. Non-crystallographic symmetry restraint was applied for the (R)-1,2-propanediol-bound form. Refinement statistics are shown in Table I.Table ICrystallographic statistics(R)-Isomer-bound(S)-Isomer-boundData collectionWavelength (Å)0.7080.708Resolution range (Å)aThe values in parentheses are for the highest resolution shell.30-2.30 (2.38-2.30)30-1.80 (1.86-1.80)Space groupP212121P212121Unit cell (Å)a74.6473.83b122.27121.95c207.43208.24Measured reflections1,033,4831,991,323Unique reflections79,408152,278RmergeaThe values in parentheses are for the highest resolution shell.0.095 (0.413)0.055 (0.119)Completeness (%)aThe values in parentheses are for the highest resolution shell.93.1 (80.8)87.3 (67.1)RefinementResolution range (Å)30-2.3030-1.80Rwork/Rfree0.208/0.2550.195/0.230Total number of atoms14,22514,908a The values in parentheses are for the highest resolution shell. Open table in a new tab Geometry Calculation and Modeling Study—The first αβγ unit of the (αβγ)2 heterohexamer was used for geometry calculation and modeling study of the active-site vicinity of the crystal structures published so far, because the average B-factor of the first unit is much lower than that of the second unit. In this paper, that is the case for the (S)-1,2-propanediol-bound form, 23.1 Å2 for the first unit and 36.0 Å2 for the second unit. In the case of (R)-isomer-bound form, however, the average B-factor of the (R)-1,2-propanediol molecule of the first unit (25.8 Å2) is slightly higher than that of the second unit (23.5 Å2), and the overall B-factor of the first αβγ unit (37.6 Å2) is slightly lower than that of the second unit (39.4 Å2), but they are comparable with each other. The part of the final 2mFo–DFc electron density map covering the second substrate is clearer than that of the first molecule. The second unit, therefore, is used for geometry calculation and modeling study in this paper. The coordinates of the adenosyl group in the "distal" conformation are derived from the superimposed model of the adenosyl group on the adenine ring of the enzyme·adeninylpentylcobalamin complex (34Masuda J. Shibata N. Morimoto Y. Toraya T. Yasuoka N. Structure. 2000; 8: 775-788Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar). Libraries for bond length and bond angle are based on the original topology and parameter files of CNS. Modeling of 1,1-gem-diol intermediates was performed with XFIT, followed by hydrogen generation and energy minimization with CNS. As a result of the energy minimization, the distance between O(1) and O(2) of the 1,1-gem-diol intermediate molecule is shorter than that of the 1,2-propanediol molecule for both (R)- and (S)-isomers. Fitting of the 1,1-gem-diol intermediate molecule on the 1,2-propanediol molecule was performed as follows: (i) overlap the O(1)–O(2) line of 1,1-gem-diol intermediate molecule on that of 1,2-propanediol molecule so that the center of the lines of both molecules share the same position; (ii) minimize the distance between C(5′) of the adenosyl group and C(2) of the 1,1-gem-diol intermediate molecule by rotations of the ribose moiety around the glycosidic linkage and of the 1,1-gem-diol molecule around the O(1)–O(2) line. Fig. 5, A and B and Fig. 6, A–D are generated from these modeled structures. Unless otherwise indicated, structural figures were created with MOLSCRIPT (41Kraulis J. J. Appl. Crystallogr. 1991; 24: 946Crossref Google Scholar) and RASTER3D (42Merritt E.A. Murphy M.E.P. Acta Crystallogr. Sect. D Biol. Crystallogr. 1994; 50: 869-873Crossref PubMed Scopus (2857) Google Scholar).Fig. 6Stereospecific elimination of the OH group in the dehydration of 1,1-gem-diol intermediates. A, the (R)-isomer-derived 1,1-gem-diol intermediate viewed from O(1) along the O(1)–C(1) bond. B, the (S)-isomer-derived 1,1-gem-diol intermediate viewed from O(1) along the O(1)–C(1) bond. C, the (R)-isomer-derived 1,1-gem-diol intermediate viewed from O(2) along the O(2)–C(1) bond. D, the (S)-isomer-derived 1,1-gem-diol intermediate viewed from O(2) along the O(2)–C(1) bond. Schematic drawings are given to the right of each figure.View Large Image Figure ViewerDownload (PPT) Modes of Binding of (R)- and (S)-1,2-Propanediols to Diol Dehydratase—The x-ray structures of diol dehydratase·CN-Cbl complexes crystallized in the presence of (R)- and (S)-1,2-propanediols were determined at 2.3- and 1.8-Å resolutions, respectively (Table I). Overall structures of both complexes were essentially identical to each other except for the substrates bound at the active site. As shown in Fig. 2B, the configuration and the conformation of the (S)-1,2-propanediol fitted to the electron density map were the same as those in the enzyme·CN-Cbl complex crystallized in the presence of racemic 1,2-propanediol (29Shibata N. Masuda J. Tobimatsu T. Toraya T. Suto K. Morimoto Y. Yasuoka N. Structure. 1999; 7: 997-1008Abstract Full Text Full Text PDF PubMed Scopus (217) Google Scholar). This provided a definite proof for our previous assignment of the (S)-isomer to the electron density map. In contrast, the electron density contoured for the (R)-enantiomer was quite different from that for the (S)-isomer (Fig. 2A). It rather resembles that corresponding to the 1,2-propanediol that is bound to the active site of glycerol dehydratase (32Yamanishi M. Yunoki M. Tobimatsu T. Sato H. Matsui J. Dokiya A. Iuchi Y. Oe K. Suto K. Shibata N. Morimoto Y. Yasuoka N. Toraya T. Eur. J. Biochem. 2002; 269: 4484-4494Crossref PubMed Scopus (96) Google Scholar), an isofunctional AdoCbl-dependent enzyme. It should be noted that the hydrogen-bonding interactions between substrate and the active-site amino acid residues are the same in the (R)- and (S)-1,2-propanediol-bound forms of the enzyme. That is, O(1) is hydrogen-bonded to -COO– of Glu-α170 and Nϵ2 of Gln-α296 and O(2) to -COO– of Asp-α335 and Nϵ2 of His-α143. Previously, we assumed that (R)- and (S)-1,2-propanediols are bound to the active site in the opposite orientation as mirror images with their C(1)–C(2) bonds aligned and that Glu-α170 and Asp-α335 might at least partially share a common catalytic function (43Toraya T. Cell. Mol. Life Sci. 2000; 57: 106-127Crossref PubMed Scopus (145) Google Scholar). Although this pseudo-symmetry was not the case, it became evident that there is another plane of symmetry relating the (R)- and (S)-substrates bound by diol dehydratase. Stereospecificity of Hydrogen Abstraction from Substrates— Abeles and co-workers (3Zagalak B. Frey P.A. Karabatsos G.L. Abeles R.H. J. Biol. Chem. 1966; 24: 3028-3035Google Scholar) demonstrated that the pro-R and pro-S hydrogen atoms on C(1) of (R)- and (S)-1,2-propanediols, respectively, migrate to C(2). The enzyme-bound AdoCbl serves as an intermediate hydrogen carrier, first accepting a hydrogen atom from C(1) of the substrate to C(5′) of the coenzyme and then, in a subsequent step, giving a hydrogen back to C(2) of the product (11Frey P.A. Essenberg M.K. Abeles R.H. J. Biol. Chem. 1967; 242: 5369-5377Abstract Full Text PDF PubMed Google Scholar, 12Frey P.A. Abeles R.H. J. Biol. Chem. 1966; 241: 2732-2777Abstract Full Text PDF PubMed Google Scholar, 13Frey P.A. Kerwar S.S. Abeles R.H. Biochem. Biophys. Res. Commun. 1967; 29: 873-879Crossref PubMed Scopus (25) Google Scholar, 14Frey P.A. Essenberg M.K. Abeles R.H. Kerwar S.S. J. Am. Chem. Soc. 1970; 92: 4488-4489Crossref PubMed Scopus (20) Google Scholar). Fig. 3, A and B shows the structures of the active sites of the (R)- and (S)-isomer-bound forms of the enzyme·cobalamin complex, respectively. The position of the adenosyl radical in the distal conformation was obtained by the modeling study, as described previously (34Masuda J. Shibata N. Morimoto Y. Toraya T. Yasuoka N. Structure. 2000; 8: 775-788Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar); that is, the adenine moiety of the coenzyme adenosyl group was superimposed on that of the enzyme-bound adeninylpentylcobalamin, and the ribose moiety was rotated around the glycosidic linkage so that C(5′) could come closest to C(1) of the (R)- and (S)-isomers. Fig. 3C shows that the (R)- and (S)-enantiomers are bound to the active site of the enzyme in a symmetrical mode as the mirror image with respect to the plane including K+, O(1), and O(2). The distances from C(5′) of the adenosyl radical to C(1)-HR, C(1)-HS, and C(2)-H of (R)-1,2-propanediol are 1.45, 2.53, and 3.73 Å, respectively. The angle of C(4′)–C(5′)–C(1)HR is 137.6°. Thus, it is reasonable to predict that the adenosyl radical abstracts the pro-R hydrogen atom from C(1) of the (R)-isomer. In contrast, the distances from C(5′) of the adenosyl radical to C(1)-HR, C(1)-HS, and C(2)-H of (S)-1,2-propanediol are 2.55, 1.46, and 3.59 Å, respectively. The angle of C(4′)–C(5′)–C(1)HS is 128.5°. Thus, it is also quite reasonable for the radical to abstract the pro-S hydrogen atom from C(1) of the (S)-isomer. These predictions are just as expected from the experimental results of Zagalak et al.