Molecular basis for specificities of reactivating factors for adenosylcobalamin-dependent diol and glycerol dehydratases Hideki Kajiura 1 , Koichi Mori 1 , Naoki Shibata 2 and Tetsuo Toraya 1 1 Department of Bioscience and Biotechnology, Graduate School of Natural Science and Technology, Okayama University, Japan 2 Graduate School of Life Science, University of Hyogo, Japan Diol dehydratase (1,2-propanediol hydro-lyase, EC 4.2.1.28) and glycerol dehydratase (glycerol hydro-lyase, EC 4.2.1.30) are isofunctional enzymes that catalyze adenosylcobalamin (AdoCbl) (coenzyme B 12 )-depen- dent conversion of 1,2-propanediol, 1,2-ethanediol, and glycerol to the corresponding aldehydes [1–5]. These enzymes encoded in the pdu (propanediol utilization) operon [6–8] and the dha (dihydroxyace- tone) regulon [9–12], respectively, are involved in pro- ducing the electron acceptors propionaldehyde and Keywords adenosylcobalamin; coenzyme B 12 ; diol dehydratase; glycerol dehydratase; reactivating factors Correspondence T. Toraya, Department of Bioscience and Biotechnology, Faculty of Engineering, Okayama University, Tsushima-naka, Okayama 700–8530, Japan Fax: +81 86 2518264 Tel: +81 86 2518194 E-mail: toraya@cc.okayama-u.ac.jp (Received 26 May 2007, revised 17 August 2007, accepted 29 August 2007) doi:10.1111/j.1742-4658.2007.06074.x Adenosylcobalamin-dependent diol and glycerol dehydratases are isofunc- tional enzymes and undergo mechanism-based inactivation by a physiologi- cal substrate glycerol during catalysis. Inactivated holoenzymes are reactivated by their own reactivating factors that mediate the ATP-depen- dent exchange of an enzyme-bound, damaged cofactor for free adenosylco- balamin through intermediary formation of apoenzyme. The reactivation takes place in two steps: (a) ADP-dependent cobalamin release and (b) ATP-dependent dissociation of the resulting apoenzyme–reactivating factor complexes. The in vitro experiments with purified proteins indicated that diol dehydratase-reactivating factor (DDR) cross-reactivates the inacti- vated glycerol dehydratase, whereas glycerol dehydratase-reactivating factor (GDR) did not cross-reactivate the inactivated diol dehydratase. We inves- tigated the molecular basis of their specificities in vitro by using purified preparations of cognate and noncognate enzymes and reactivating factors. DDR mediated the exchange of glycerol dehydratase-bound cyanocobala- min for free adeninylpentylcobalamin, whereas GDR cannot mediate the exchange of diol dehydratase-bound cyanocobalamin for free ade- ninylpentylcobalamin. As judged by denaturing PAGE, the glycerol dehydra- tase–DDR complex was cross-formed, although the diol dehydratase–GDR complex was not formed. There were no specificities of reactivating factors in the ATP-dependent dissociation of enzyme–reactivating factor complexes. Thus, it is very likely that the specificities of reactivating factors are determined by the capability of reactivating factors to form complexes with apoenzymes. A modeling study based on the crystal structures of enzymes and reactivating factors also suggested why DDR cross-forms a complex with glycerol dehydratase, and why GDR does not cross-form a complex with diol dehydratase. Abbreviations AdePeCbl, adeninylpentylcobalamin; AdoCbl, adenosylcobalamin or coenzyme B 12 ; CN-Cbl, cyanocobalamin; DDR, diol dehydratase- reactivating factor; GDR, glycerol dehydratase-reactivating factor; MBTH, 3-methyl-2-benzothiazolinone hydrazone. 5556 FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS b-hydroxypropionaldehyde. They are essential for the fermentation of 1,2-propanediol and glycerol, respec- tively [4,13–17] because these carbon sources are more reduced substrates than the corresponding carbohy- drates, and oxidation and reduction must be balanced for the bacterial growth under anaerobic conditions. In some bacteria, glycerol dehydratase can be substituted by the isofunctional diol dehydratase, which is induced at a low level by glycerol [9,14,18]. The mechanism of action of diol dehydratase has been extensively studied [4,5,19–22]. Diol and glycerol dehydratases form an adenosyl radical, a catalytic radi- cal in the active site, by homolysis of the coenzyme Co-C bond and catalyze the reactions by utilizing the high reactivity of the radical. The catalytic and inter- mediate radicals are protected by proteins from un- desirable side reactions during catalysis (so-called ‘negative catalysis’ [23]). These enzymes tend to undergo mechanism-based inactivation (suicide inacti- vation) by certain substrates or coenzyme analogs (for reviews see [4,5]). Interestingly, both of them are rap- idly inactivated by a physiological substrate glycerol during catalysis [2,24,25] or by O 2 in the absence of substrate [26,27]. The glycerol inactivation of holodiol dehydratase is a sort of mechanism-based inactivation, resulting in the disappearance of organic radical inter- mediate(s) by side reactions and leaving 5¢-deoxyade- nosine and hitherto unidentified cobalamin at the active site [24]. The O 2 inactivation in the absence of substrate results in the formation of hydroxocobalamin [26] and might be caused by the reaction of adenosyl radical with oxygen, although the inactivation prod- ucts derived from it have not yet been identified. These inactivations are accompanied by the irreversible cleav- age of the Co-C bond of the coenzyme. The resulting damaged cofactors remain tightly bound to apoen- zyme, which brings about the inactivation of enzymes. The inactivation by glycerol is enigmatic because glyc- erol is a growth substrate for the bacteria that produce these enzymes. We found that the glycerol-inactivated holoenzymes in permeabilized cells of Klebsiella pneumoniae and Klebsiella oxytoca are rapidly reacti- vated in situ in the presence of ATP and Mg 2+ (or Mn 2+ ) [28,29]. The inactive complex between enzyme and cyanocobalamin (CN-Cbl) is also activated in situ under the same conditions. We identified two ORFs in the 3¢-flanking region of the diol dehydratase genes [30] of K. oxytoca as the genes encoding a reactivating factor for diol dehydra- tase and named them ddrAB (diol dehydratase-reacti- vating factor) genes [31]. These genes correspond to pduGH [32]. We then identified two ORFs in the proximity of the glycerol dehydratase genes [33] of K. pneumoniae as the genes encoding a reactivating factor for glycerol dehydratase and named them gdrAB (glycerol dehydratase-reactivating factor) genes [34]. Recombinant DdrA and DdrB proteins as well as GdrA and GdrB form a tight a 2 b 2 complex and actu- ally function as their reactivating factor – that is, they reactivate the glycerol-inactivated and O 2 -inactivated holoenzymes and activate the inactive enzyme–CN-Cbl complexes in vitro in the presence of AdoCbl, ATP, and Mg 2+ [35–37]. They (re)activate the complexes by mediating the ATP-dependent exchange of the enzyme- bound, adenine-lacking cobalamins for free adenosyl- cobalamin, an adenine-containing cobalamin through intermediary formation of apoenzyme. The function of reactivating factors is to release a tightly bound ade- nine-lacking cobalamin from the enzymes by a molecu- lar chaperone-like mechanism of action. It was established that the reactivation of the inactivated holoenzyme by the factors takes place in two steps: (a) ADP-dependent cobalamin release and (b) ATP- dependent dissociation of the resulting apoenzyme– reactivating factor complexes. ATP plays a dual role – as a precursor of ADP for the first step and as an effector that causes conformational change of factors into low-affinity forms of enzymes to be reactivated. The DhaF and DhaG of Citrobacter freundii were also confirmed to be involved in the reactivation of glycerol dehydratase [38]. Recently, the crystal structure of glycerol dehydratase-reactivating factor (GDR) was reported [39]. We also solved the crystal structure of diol dehydratase-reactivating factor (DDR) [40]. Based on their structures and a modeling study, the molecu- lar mechanism of the release of a damaged cofactor from inactivated holoenzymes has been proposed [40]. The transient complexes between factors and enzymes, suggested from biochemical experiments [36,37], have been postulated to be formed by subunit swapping or subunit displacement. Both DDR and GDR have dimeric (ab) 2 structures – a large subunit (a) with Mr of 64 kDa (DdrA, GdrA, DhaF) and a small subunit (b) with Mr of 14 kDa (DdrB) or 12 kDa (GdrB, DhaG) [35–38]. The identi- ties of amino acid sequences of large and small sub- units between them are 61% and 30%, respectively [31,34]. The latter value is considerably lower than those of the subunits between diol and glycerol dehy- dratases (more than 50%) [30,33]. The experiments with permeabilized cells by toluene treatment (so-called in situ system) indicated that DDR cross-reactivates the inactivated glycerol dehydratase effectively, whereas the reverse was not the case – that is, GDR did not cross-reactivate the inactivated diol dehydratase [41]. We concluded that GDR is much H. Kajiura et al. B 12 -dependent dehydratases and reactivating factors FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS 5557 more specific for the dehydratase partner than DDR, and that a large subunit of the reactivating factors principally determines the specificity for a dehydratase [41]; the reason for this, however, remained unclear. Seifert et al. reported that C. freundii GDR (DhaF– DhaG complex) cross-reactivates the glycerol-inacti- vated glycerol dehydratase of K. pneumoniae, but nei- ther glycerol dehydratase of Clostridium pasteurianum nor diol dehydratases of K. oxytoca and Salmonella typhimurium [38]. In this study, we investigated the specificities of reac- tivating factors for enzymes in vitro using purified preparations of proteins and attempted to solve why DDR cross-reactivates the inactivated glycerol dehy- dratase, and why GDR does not cross-reactivate the inactivated diol dehydratase. We reached the conclu- sion that the specificities of reactivating factors are determined by the capability of reactivating factors to form complexes with apoenzymes. A modeling study based on the crystal structures of enzymes and reacti- vating factors also supported this conclusion. These results are described here. Results Specificities of reactivating factors in the reactivation of inactivated holoenzymes The specificities of reactivating factors in the in vitro reactivation of inactivated holoenzymes were investi- gated using purified preparations of reactivating factors and glycerol-inactivated holodiol dehydratase or holo- glycerol dehydratase. Reactivation of the glycerol- inactivated holoenzymes was monitored by the recovery of their 1,2-propanediol-dehydrating activity. As shown in Fig. 1A, glycerol-inactivated holodiol dehydratase was rapidly reactivated by DDR in the presence of AdoCbl, ATP, and Mg 2+ , but was not reactivated by these either in the presence of GDR or in the absence of reactivating factors. In contrast, glycerol-inactivated hologlycerol dehydratase underwent rapid reactivation by GDR, relatively slow reactivation by DDR, and no reactivation in the absence of reactivating factors in the presence of AdoCbl, ATP, and Mg 2+ (Fig. 1B). That DDR has a broader specificity and that GDR is highly specific toward a cognate dehydratase are consistent with previous studies [38,41]. Specificities of reactivating factors in the activation of inactive enzyme–CN-Cbl complex The specificities of reactivating factors in the in vitro activation of inactive enzyme–CN-Cbl complexes were studied similarly using purified preparations of reacti- vating factors and enzyme–CN-Cbl complexes. The enzyme–CN-Cbl complexes can be considered models of inactivated holoenzymes [35–37] because CN-Cbl is an adenine-lacking cobalamin that binds tightly to the active site of the enzymes. Figure 2A shows that the diol dehydratase–CN-Cbl complex was rapidly acti- vated by DDR in the presence of AdoCbl, ATP, and Mg 2+ , but hardly activated either in the presence of GDR or in the absence of reactivating factors. Con- versely, the glycerol dehydratase–CN-Cbl complex underwent rapid activation by GDR, relatively slow but significant activation by DDR, and essentially no activation in the absence of reactivating factors under the same conditions (Fig. 2B). Again, it was demon- strated that DDR acts on both diol and glycerol dehy- dratases, whereas GDR is more specific for glycerol dehydratase. These conclusions are in good agreement with earlier results [38,41]. Specificities of reactivating factors in the promotion of exchange of enzyme-bound CN-Cbl for free adeninylpentylcobalamin (AdePeCbl) The absolute requirement for free AdoCbl in addition to ATP and Mg 2+ in both the reactivation of glycerol- inactivated holoenzymes and the activation of the enzyme–CN-Cbl complexes strongly indicates that the Fig. 1. Specificities of reactivating factors in the reactivation of glycerol-inactivated holodiol dehydratase (A) and hologlycerol de- hydratase (B). Glycerol-inactivated holoenzymes (0.75 unit) were incubated at 37 °C for the indicated time without (open squares) and with 24 lg of DDR (open circles) or GDR (closed circles) in 0.03 M potassium phosphate buffer (pH 8) containing 21 lM AdoCbl and 1.2 M 1,2-propanediol in the presence of 24 mM ATP ⁄ 24 mM MgCl 2 , in a total volume of 25 lL. The amount of propionaldehyde formed was determined as described in the text. The extents of reactivation of diol dehydratase by DDR and of glycerol by dehydra- tase by GDR were 82% and 25% for diol and glycerol dehydrata- ses, respectively. B 12 -dependent dehydratases and reactivating factors H. Kajiura et al. 5558 FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS reactivation of inactivated holoenzymes and the activa- tion of the inactive enzyme–CN-Cbl complexes take place by exchange of the enzyme-bound damaged cofactor and CN-Cbl, respectively, for free intact co- enzyme. CN-Cbl and AdePeCbl can be considered models of the damaged cofactor (adenine-lacking cobalamin) and intact coenzyme (adenine-containing cobalamin), respectively. As shown in Fig. 3A, when the diol dehydratase–CN-Cbl complex was incubated with free AdePeCbl, ATP, and Mg 2+ in the presence of DDR, followed by dialysis to remove unbound co- balamins, the spectrum of the dialyzate indicated that the enzyme-bound CN-Cbl was displaced by Ade- PeCbl. Such an exchange did not occur in the presence of GDR (Fig. 3B) or in the absence of reactivating fac- tors (Fig. 3A) under the same conditions. In contrast, upon incubation of the glycerol dehydratase–CN-Cbl complex with free AdePeCbl, ATP, and Mg 2+ in the presence of GDR, followed by dialysis, the enzyme- bound CN-Cbl was displaced by AdePeCbl (Fig. 3F). Such an exchange occurred in the presence of DDR as well (Fig. 3E), but not in the absence of reactivating factors (Fig. 3E) under the same conditions. The enzyme-bound AdPeCbl did not undergo displacement by free CN-Cbl under any conditions (Fig. 3C,D,G,H). It is thus evident that reactivating factor mediates the exchange of the enzyme-bound, adenine-lacking cobal- amin for free, adenine-containing cobalamin toward a cognate dehydratase. In addition, DDR can mediate a similar exchange with glycerol dehydratase, a non- cognate enzyme, whereas GDR cannot mediate the Fig. 3. Specificities of reactivating factors in the promotion of exchange of enzyme-bound CN-Cbl for free AdePeCbl. Diol dehydratase–CN- Cbl (A,B), glycerol dehydratase–CN-Cbl (E,F), diol dehydratase–AdePeCbl (C,D), and glycerol dehydratase–AdePeCbl (G,H) complexes were prepared by incubation of apoenzymes (50 units) with 33 l M CN-Cbl or AdePeCbl at 37 °C for 30 min in 0.2 mL of 0.05 M potassium phos- phate buffer (pH 8) containing 0.3 M 1,2-propanediol in the dark. The enzyme–CN-Cbl (A,B,E,F) and enzyme–AdePeCbl (C,D,G,H) complexes were incubated at 37 °C for 30 min without (broken lines in A,C,E,G) and with 1.25 mg of DDR (solid lines in A,C,E,G) or GDR (solid line in B,D,F,H) in 0.04 M potassium phosphate buffer (pH 8) containing 20 lM AdePeCbl (A,B,E,F) or CN-Cbl (C,D,G,H) and 20 mM ATP ⁄ 20 mM MgCl 2 , in a total volume of 0.5 mL. The mixtures were then dialyzed at 4 °C for 48 h against 1000 volumes of 0.01 M potassium phosphate buffer (pH 8) containing 0.3 M 1,2-propanediol with a buffer change. The spectra of dialyzates were measured with a minus cobalamin con- trol as reference and corrected for dilution. Fig. 2. Specificities of reactivating factors in the activation of diol dehydratase–CN-Cbl (A) and glycerol dehydratase–CN-Cbl (B) com- plexes. The enzyme–CN-Cbl complexes (0.75 unit) were incubated at 37 °C for the indicated time without (open squares) and with 24 lg of DDR (open circles) or GDR (closed circles) in 0.03 M potas- sium phosphate buffer (pH 8) containing 21 l M AdoCbl and 1.2 M 1,2-propanediol in the presence of 24 mM ATP ⁄ 24 mM MgCl 2 ,ina total volume of 25 lL. The amount of propionaldehyde formed was determined as described in the text. The extents of activation of diol dehydratase by DDR and of glycerol dehydratase by GDR were 67% and 65%, respectively. H. Kajiura et al. B 12 -dependent dehydratases and reactivating factors FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS 5559 exchange of diol dehydratase-bound CN-Cbl for free AdePeCbl – that is, GDR is more specific for glycerol dehydratase, the cognate dehydratase. Specificities of reactivating factors in the complex formation with dehydratases DDR and GDR form a complex with apoenzymes of diol and glycerol dehydratases, respectively, and tran- sient formation of such complexes seems to result in the dissociation of a tightly bound damaged cofactor or adenine-lacking cobalamin [36,37]. To elucidate the molecular basis of the specificities of reactivating fac- tors in the reactivation of inactivated holoenzymes and the activation of inactive enzyme–cobalamin com- plexes, the possibility of cross-formation of noncognate enzyme–reactivating factor complexes were examined. The complex formation was analyzed by nondenatur- ing PAGE. When apodiol dehydratase was incubated with reactivating factors either in the presence of ADP or in the absence of nucleotides, it formed a new major band with DDR but not with GDR (Fig. 4A,B, lanes 1 and 2). In the presence of ATP, the new band did not appear (Fig. 4C, lanes 1 and 2). When apogly- erol dehydratase was incubated with reactivating fac- tors, it formed new major bands with either GDR or DDR in the presence of ADP or in the absence of nucleotides (Fig. 4A,B, lanes 3 and 4). In the presence of ATP, the new major band with DDR did not appear, whereas a part of the new band with GDR remained (Fig. 4C, lanes 3 and 4). To identify the new major band formed between glycerol dehydratase and DDR, the band was analyzed by two-dimensional PAGE (nondenaturing PAGE in the first dimension and SDS⁄ PAGE in the second). The analysis provided all of the a, b, and c subunits of glycerol dehydratase and the a and b subunits of DDR (Fig. 5C), indicating that this new band corresponds to a complex between them. The protein band observed above the band of DDR a subunit seems to be not an impurity band but a band due to the insufficient denaturation of glycerol dehydratase upon SDS ⁄ PAGE in the second dimen- sion, because the same band was observed when this enzyme was subjected to SDS ⁄ PAGE without heat treatment in the sample buffer (data not shown). Such a band was not observed when diol dehydratase, DDR or GDR in the sample buffer was applied even without heat treatment. The bands observed in combinations ABC DEF Fig. 4. Specificities of reactivating factors in the complex formation with dehydratases. Apoenzymes (ApoE) (A–C) or enzyme–CN-Cbl com- plexes (EÆCN-Cbl) (D–F) (0.35 unit) were incubated at 37 °C for 10 min without and with 15 lg of DDR or GDR in a volume of 6.5 lL, and the mixtures were further incubated at 37 °C for 10 min in the absence (A,D) and presence of 21 m M ADP ⁄ 21 mM MgCl 2 (B,E) or 21 mM ATP ⁄ 21 mM MgCl 2 (C,F) in 35 mM potassium phosphate buffer (pH 8), in a total volume of 7.5 lL. The mixtures were then subjected to non- denaturing PAGE (5% gel) in the absence (A,D) and presence of 1 m M ADP ⁄ 1mM MgCl 2 (B,E) or 1 mM ATP ⁄ 1mM MgCl 2 (C,F). Positions of diol dehydratase (D), glycerol dehydratase (G), DDR (DR), and GDR (GR) are indicated with arrowheads to the right of the gels, and their complexes 1–4 to the left. Bands 1 and 2 correspond to 1 : 2 and 1 : 1 DD–DDR complexes, respectively, and bands 3 and 4 to GD–GDR and GD–DDR complexes, respectively. BPB, Bromophenol blue. B 12 -dependent dehydratases and reactivating factors H. Kajiura et al. 5560 FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS of cognate dehydratases and reactivating factors were confirmed to be complexes between them (Fig. 5A,B), in accordance with our previous results [36,37]. When similar experiments were carried out with the enzyme– CN-Cbl complexes, essentially the same results as with apoenzymes were obtained in the presence of ADP (Fig. 4E), but almost no enzyme–reactivating factor complexes were formed in the presence of ATP or in the absence of nucleotides (Fig. 4D,F). These results are also consistent with published results [36,37]. In the cases of both DDR and GDR, tightly bound CN-Cbl is released upon the binding of reactivating factors to the enzyme–CN-Cbl complexes. Thus, it is very likely that the specificities of reactivating factors in the reactivation of inactivated holoenzymes and the activation of inactive enzymeÆcobalamin complexes are determined by the capability of reactivating factors to form complexes with apoenzymes. Specificities of reactivating factors in the inhibition of apoenzymes and the reversal by ATP The specificities of reactivating factors in the complex formation with dehydratases were further investigated by inhibition experiments. Apoenzymes of diol and glycerol dehydratases were strongly inhibited in a time- dependent manner by pre-incubation with DDR and GDR, respectively, either in the presence of ADP or in the absence of nucleotides, in accordance with previous results [36,37]. ADP alone did not inhibit the enzymatic activity (data not shown). This inhibition is due to the complex formation between enzymes and reactivating factors [36,37]. In contrast, diol dehydratase was not inhibited at all by pre-incubation with GDR, a noncog- nate reactivating factor, either in the presence of ADP or in the absence of nucleotides (Fig. 6B,C), whereas glycerol dehydratase was strongly inhibited by pre-incu- bation with DDR either in the presence of ADP or in the absence of nucleotides (Fig. 6F,G). Again, these results indicate that diol dehydratase–GDR complex was not formed, although glycerol dehydratase–DDR complex was cross-formed. In all cases, the inhibition by reactivating factors were completely reversed when assayed in the presence of ATP. This is because the enzymeÆreactivating factor complexes dissociate into apoenzymes and reactivating factors in the presence of ATP [36,37]. No inhibition was observed by the pre- incubation of apoenzymes and reactivating factors in the presence of ATP (Fig. 6D,H). Therefore, it can be concluded that there are no specificities of reactivating factors in the ATP-dependent dissociation of enzyme– reactivating factor complexes. Buried surface areas between the a subunits of reactivating factors and the b subunits of dehydratases To estimate the strengths of interactions between the a subunits of reactivating factors and the b subunits of dehydratases, modeling studies were carried out on the AB C Fig. 5. Identification of a new band as a noncognate enzymeÆreactivating factor complex by two-dimensional PAGE. Apodiol dehydratase (A) and apoglycerol dehydratase (B,C) (0.35 unit) were incubated at 37 °C for 10 min with 15 lg of DDR (A,C) or GDR (B) in a volume of 6.5 lL. The mixtures were further incubated at 37 °C for 10 min in the presence of 21 m M ADP ⁄ 21 mM MgCl 2 in 35 mM potassium phosphate buf- fer (pH 8), in a total volume of 7.5 lL. The mixtures were then subjected to nondenaturing PAGE (6% gel) in the presence of 1 m M ADP ⁄ 1mM MgCl 2 (first dimension, from left to right), followed by SDS ⁄ PAGE (6 and 14% gels) (second dimension, from top to bottom). Positions of the subunits of dehydratases and reactivating factors are indicated with arrowheads to the right of the gels. BPB, Bromophenol blue. Subscripts D, G, DR, and GR are the same as those in the legend to Fig. 4. H. Kajiura et al. B 12 -dependent dehydratases and reactivating factors FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS 5561 bases of three-dimensional structures of dehydratases and reactivating factors. When the b subunits of diol and glycerol dehydratases are superimposed on the b subunits of DDR and GDR, respectively, the buried surface areas between the DDR a and diol dehydratase b subunits and between the GDR a and glycerol dehy- dratase b subunits are quite similar (710 and 708 A ˚ 2 , respectively). In the model of the glycerol dehydratase– DDR cross-formed complex, the buried surface area is 742 A ˚ 2 , significantly higher than those of the cog- nate complexes. On the other hand, in the diol dehydratase–GDR model, this value is decreased to 698 A ˚ 2 . Discussion DDR (re)activates both diol and glycerol dehydratases, but GDR acts only on glycerol dehydratase, a cognate enzyme, and the a subunit of reactivating factors prin- cipally determines the specificity for a dehydratase [41]. The molecular basis of these findings has remained obscure until recently. It was established with DDR and GDR that the reactivation of inactivated holoen- zymes and the activation of inactive enzyme–cobala- min complexes by reactivating factors take place in two steps: (a) ADP-dependent cobalamin release, and (b) ATP-dependent dissociation of the resulting apoen- zyme–reactivating factor complexes [36,37]. ATP serves as a precursor of ADP in the first step and as an effec- tor in the second step. ATP and ADP thus function as a nucleotide switch that modulates the affinity of reac- tivating factors for the enzymes to be (re)activated. In this context, one possibility may be that the specifici- ties of reactivating factors in the reactivation of inacti- vated holoenzymes and the activation of inactive enzyme–cobalamin complexes are determined by their capability to form a complex with apoenzymes of diol or glycerol dehydratase. The transient formation of such complexes results in the dissociation of a tightly bound damaged cofactor or adenine-lacking cobalamin [36,37]. The modeling study based on the crystal struc- tures of diol dehydratase and DDR suggested that the binding of diol dehydratase b subunit to the DDR a subunit induces steric repulsion between the a sub- units of enzyme and DDR, leading to the release of a damaged cofactor from inactivated holoenzymes [40]. Another possibility may be that enzyme–reactivating factor complexes are formed even with the noncognate dehydratase, but the specificities of reactivating fac- tors might be determined by their effectiveness of Fig. 6. Specificities of reactivating factors in the inhibition of apoenzymes and the reversal by ATP. Apodiol dehydratase (A–D) and apoglycer- ol dehydratase (E–H) (0.74 unit) were preliminarily incubated without (broken lines in A,E) and with (solid lines) 23 lg of GDR (B–D) and DDR (F–H), respectively, in the absence (A,B,E,F) and presence of 16 m M ADP ⁄ 16 mM MgCl 2 (C,G) or 10 mM ATP ⁄ 10 mM MgCl 2 (D,H) in 38 mM potassium phosphate buffer (pH 8) containing 0.9 M 1,2-propanediol, in a total volume of 15 lL. The mixtures were incubated at 37 °C for the indicated time periods, and then diluted 600-fold with 50 m M potassium phosphate buffer (pH 8) containing 2% 1,2-propanediol. AdoCbl (15 l M), 1,2-propanediol (0.1 M) and KCl (50 mM) were added to 0.2 mL of the diluted mixtures without (open circles) and with (closed cir- cles) additional ATP and MgCl 2 (10 mM each) to a total volume of 1 mL. After incubation at 37 °C for 10 min, the amount of propionaldehyde formed was determined as described in the text. B 12 -dependent dehydratases and reactivating factors H. Kajiura et al. 5562 FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS nucleotide switch – that is, some complexes are disso- ciable by the binding of ATP, but others are not. All the data reported in this paper indicated that the for- mer possibility is likely, but not the latter. Thus, we concluded that specificities of reactivating factors are determined by their capability to form a complex with apodehydratases. The structure-based subunit swap- ping or displacement models of DDR and GDR [39,40] as well as biochemical data with DDR indicates that the b subunit of DDR is released upon complex formation between diol dehydratase and DDR. Thus, the reason why the a subunit of reactivating factors principally determines the specificity for a dehydratase also became clear in this study. Why does DDR cross-form a complex with glycerol dehydratase, and why does GDR not cross-form a complex with diol dehydratase? To solve this enigma, modeling studies were carried out on the bases of three-dimensional structures of dehydratases and reac- tivating factors. The following speculations were pos- sible. As described above, the buried surface areas between the a subunits of reactivating factors and the b subunits of dehydratases decrease in the following order: glycerol dehydratase–DDR > diol dehydra- tase–DDR ¼ glycerol dehydratase–GDR > diol dehy- dratase–GDR. This order suggests the order of relative strengths of interactions between the b sub- units of enzymes and the a subunits of reactivating factors and thus seems to reflect on the stability of complexes. Another factor possibly determining the specificity is the differences of the surface charge dis- tribution on the area. However, the surface charge distributions of DDR and GDR on the area are quite similar (data not shown). Thus these modeling studies suggest that the differences of the buried surface areas between the cognate and cross-formed complexes pro- vide the most plausible explanation for the above- mentioned enigma. Reactivating factors, such as DDR and GDR, may be not special but rather general for radical B 12 enzymes, because, in general, their holoenzymes tend to undergo inactivation during catalysis or by oxygen in the absence of substrate. As suggested from genetic evidence [42] as well as from fragmentary similarity with DdrA and GdrA [4], EutA has been identified as a reactivating factor for ethanolamine ammonia-lyase, although the details of its mechanism of action have not yet been reported [43]. Recently, MeaB, a bacte- rial homolog of MMAA or CblA [44], has been sug- gested to function in the GTP-dependent assembly of holomethylmalonyl-CoA mutase and subsequent pro- tection of radical intermediates during catalysis [45,46]. Experimental procedures Materials Crystalline AdoCbl was a gift from Eisai Co., Ltd. (Tokyo, Japan). CN-Cbl was obtained from GlaxoSmithKline, London, UK. AdePeCbl was prepared as described before [47]. All other chemicals were analytical grade reagents and used without further purification. K. oxytoca diol dehydratase and DDR were purified to homogeneity from overexpressing Escherichia coli JM109 harbouring expression plasmids pUSI2E(DD) [30] and pUSI2ENd(6 ⁄ 5b) [31], respectively, as reported previously [35,48]. K. pneumoniae glycerol dehydratase and GDR were purified to homogeneity from overexpressing E. coli JM109 harbouring expression plasmids pUSI2E(GD) [33] and E. coli BL21(DE3) harbouring expression plasmids pET (gdrB-gdrA) [37], respectively, as reported previously [37,49]. Enzyme and protein assays Activities of diol and glycerol dehydratases were assayed in the dark by the 3-methyl-2-benzothiazolinone hydrazone (MBTH) method [50]. The standard reaction mixture con- taining an appropriate amount of apoenzyme, 15 lm Ado- Cbl, 0.1 m 1,2-propanediol, 50 mm KCl, and 35 mm potassium phosphate buffer (pH 8.0), in a total volume of 1.0 mL, was incubated at 37 °C for 10 min. After the reac- tion was terminated by adding 1 mL of 0.1 m potassium citrate buffer (pH 3.6), MBTHÆ HCl was added to a final concentration of 0.9 mm, and the mixture was incubated again at 37 °C for 15 min. The amount of propionaldehyde formed was determined by measuring the absorbance at 305 nm. One unit is defined as the amount of enzyme activ- ity that catalyzes the formation of 1 lmol propionalde- hydeÆmin )1 at 37 °C under the standard assay conditions. Protein concentrations of purified enzymes and reactivat- ing factors were determined by measuring the absorbance at 280 nm. The molar absorption coefficients at 280 nm, calcu- lated by the method of Gill & von Hippel [51] from their deduced amino acid composition and subunit structure, were 120 500 m )1 Æcm )1 for diol dehydratase [35], 112 100 m )1 Æcm )1 for glycerol dehydratase [49], 58 140 m )1 Æcm )1 for DDR [35], and 86 500 m )1 Æcm )1 for GDR [37], respectively. Assays of reactivating factors DDR and GDR activities were assayed by their capability of reactivating the glycerol-inactivated holoenzymes and activating the inactive enzyme–CN-Cbl complexes of diol and glycerol dehydratases, respectively [35–37]. Glycerol- inactivated holoenzymes were prepared by incubation of substrate-free apoenzymes (15 units) with 15 lm AdoCbl at 37 °C for 30 min in 50 lL of 0.03 m potassium phosphate buffer (pH 8) containing 0.3 m glycerol and 0.05 m KCl, H. Kajiura et al. B 12 -dependent dehydratases and reactivating factors FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS 5563 followed by dialysis at 4 °C for 48 h against 1000 volumes of 0.05 m potassium phosphate buffer (pH 8) containing 0.3 m 1,2-propanediol with a buffer change. Complexes of enzymes with CN-Cbl were prepared by incubation of apoenzymes (15 units) with 11 lm CN-Cbl at 37 °C for 30 min in 0.2 mL of 0.05 m potassium phosphate buffer (pH 8) containing 0.3 m 1,2-propanediol. In the standard assays, glycerol-inactivated holoenzymes or inactive enzyme–CN-Cbl complexes (0.75 unit) was incubated at 37 °C with appropriate amounts of DDR or GDR in 0.03 m potassium phosphate buffer (pH 8) containing 21 lm AdoCbl and 1.2 m 1,2-propanediol in the presence of 24 mm ATP ⁄ 24 mm MgCl 2 , in a total volume of 25 lL. The reaction was terminated by adding 25 lL of 0.1 m potassium citrate buffer (pH 3.6). After removal of precipi- tate by centrifugation, the reaction mixture was diluted appropriately to determine the amount of propionaldehyde using the MBTH method [50]. PAGE PAGE was performed under nondenaturing conditions as described by Davis [52] or under denaturing conditions as described by Laemmli [53]. Protein was stained with Coo- massie Brilliant Blue G-250. Nondenaturing PAGE of diol and glycerol dehydratases was performed in the presence of 0.1 m 1,2-propanediol to prevent their subunits from disso- ciation [54]. In some experiments, ATP or ADP was also added with MgCl 2 (1 mm each) and KCl (2 mm) to gels and electrode buffer. Complex formation was analyzed by nondenaturing PAGE. Enzyme–CN-Cbl complexes were prepared by incubation of apoenzyme (2.25 units) with 50 lm CN-Cbl at 37 °C for 30 min in 22.5 lLof33mm potassium phosphate buffer (pH 8). Modeling studies The following coordinates were used for the modeling stud- ies: nucleotide-free DDR [40], 2D0P; nucleotide-free GDR [39], 1NBW; CN-Cbl-bound diol dehydratase [55], 1EGM; CN-Cbl-bound glycerol dehydratase [49], 1IWP. Superim- positions of the diol dehydratase and glycerol dehydratase b subunits on the DDR and GDR b subunits were per- formed with the EBI SSM [56] under the pairwise three- dimensional alignment mode. Buried molecular surface areas were calculated with the program CNS [57], for which the probe radius was 1.4 A ˚ , equivalent to the size of a water molecule. The averaged values between the A–B and C–D interfaces are used for discussion. Acknowledgements This work was supported in part by Grants-in-Aid for Scientific Research [(B) 13480195 and 17370038 and Priority Areas 513 to (TT)] from the Japan Society for Promotion of Science and the Ministry of Education, Culture, Sports, Science and Technology, Japan, and the Grant of Natural Sciences Research Assistance from the Asahi Glass Foundation, Tokyo, Japan. References 1 Lee HA Jr & Abeles RH (1963) Purification and prop- erties of dioldehydrase, an enzyme requiring a cobamide coenzyme. J Biol Chem 238, 2367–2373. 2 Toraya T, Shirakashi T, Kosuga T & Fukui S (1976) Substrate specificity of coenzyme B 12 -dependent diol dehydrase: glycerol as both a good substrate and a potent inactivator. Biochem Biophys Res Commun 69, 475–480. 3 Pawelkiewicz J & Zagalak B (1965) Enzymic conversion of glycerol into b-hydroxypropionaldehyde in a cell-free extract from Aerobacter aerogenes. Acta Biochim Pol 12, 207–218. 4 Toraya T (2003) Radical catalysis in coenzyme B 12 -dependent isomerization (eliminating) reactions. Chem Rev 103, 2095–2127. 5 Toraya T (2000) Radical catalysis of B 12 enzymes: struc- ture, mechanism, inactivation, and reactivation of diol and glycerol dehydratases. Cell Mol Life Sci 57, 106–127. 6 Jeter RM (1990) Cobalamin-dependent 1,2-propanediol utilization by Salmonella typhimurium. J Gen Microbiol 136, 887–896. 7 Bobik TA, Ailion M & Roth JR (1992) A single regula- tory gene integrates control of vitamin B 12 synthesis and propanediol degradation. J Bacteriol 174, 2253–2266. 8 Rondon MR & Escalante-Semerena JC (1992) The poc locus is required for 1,2-propanediol-dependent tran- scription of the cobalamin biosynthetic (cob) and pro- panediol utilization (pdu) genes of Salmonella typhimurium. J Bacteriol 174, 2267–2272. 9 Forage RG & Foster MA (1982) Glycerol fermentation in Klebsiella pneumoniae: functions of the coenzyme B 12 -dependent glycerol and diol dehydratases. J Bacte- riol 149, 413–419. 10 Forage RG & Lin ECC (1982) dha system mediating aerobic and anaerobic dissimilation of glycerol in Kle- bsiella pneumoniae NCIB 418. J Bacteriol 151, 591–599. 11 Ruch FE, Lengeler J & Lin ECC (1974) Regulation of glycerol catabolism in Klebsiella aerogenes. J Bacteriol 119, 50–56. 12 Seyfried M, Daniel R & Gottschalk G (1996) Cloning, sequencing, and overexpression of the genes encoding coenzyme B 12 -dependent glycerol dehydratase of Citro- bacter freundii. J Bacteriol 178, 5793–5796. 13 Toraya T, Honda S & Fukui S (1979) Fermentation of 1,2-propanediol and 1,2-ethanediol by some genera of B 12 -dependent dehydratases and reactivating factors H. Kajiura et al. 5564 FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS Enterobacteriaceae, involving coenzyme B 12 -dependent diol dehydratase. J Bacteriol 139, 39–47. 14 Toraya T, Kuno S & Fukui S (1980) Distribution of coenzyme B 12 -dependent diol dehydratase and glycerol dehydratase in selected genera of Enterobacteriaceae and Propionibacteriaceae. J Bacteriol 141, 1439–1442. 15 Toraya T & Fukui S (1982) Diol dehydratase. In: B 12 (Dolphin, D, ed.), Vol. 2, pp. 233–262. John Wiley & Sons, New York, NY. 16 Toraya T (1994) Diol dehydratase and glycerol dehydra- tase, coenzyme B 12 -dependent isozymes. In Metal Ions in Bioloical Systems (Sigel, H & Sigel, A, eds), Vol. 30, pp. 217–254. Dekker, New York, NY. 17 Daniel R, Bobik TA & Gottschalk G (1999) Biochemis- try of coenzyme B 12 -dependent glycerol and diol dehy- dratases and organization of the encoding genes. FEMS Microbiol Rev 22, 553–566. 18 Toraya T, Honda S, Kuno S & Fukui S (1978) Coen- zyme B 12 -dependent diol dehydratase: regulation of apo- enzyme synthesis in Klebsiella pneumoniae (Aerobacter aerogenes) ATCC 8724. J Bacteriol 135, 726–729. 19 Abeles RH & Dolphin D (1976) The vitamin B 12 coen- zyme. Acc Chem Res 9, 114–120. 20 Re ´ tey J, Umani-Ronchi A, Seibl J & Arigoni D (1966) On the mechanism of the propanediol dehydrase reac- tion. Experientia 22, 502–503. 21 Re ´ tey J, Umani-Ronchi A & Arigoni D (1966) On the stereochemistry of the propanediol dehydrase reaction. Experientia 22, 72–73. 22 Frey PA (1990) Importance of organic radicals in enzy- matic cleavage of unactivated C–H bonds. Chem Rev 90, 1343–1357. 23 Re ´ tey J (1990) Enzymic reaction selectivity by negative catalysis or how do enzymes deal with highly reactive intermediates? Angew Chem Int Ed Engl 29, 355–361. 24 Bachovchin WW, Eagar RG Jr, Moore KW & Richards JH (1977) Mechanism of action of adenosylcobalamin: glycerol and other substrate analogues as substrates and inactivators for propanediol dehydratase – kinetics, ste- reospecificity, and mechanism. Biochemistry 16 , 1082– 1092. 25 Poznanskaya AA, Yakusheva MI & Yakovlev VA (1977) Study of the mechanism of action of adenosylco- balamin-dependent glycerol dehydratase from Aerobac- ter aerogenes. II. The inactivation kinetics of glycerol dehydratase complexes with adenosylcobalamin and its analogs. Biochim Biophys Acta 484, 236–243. 26 Wagner OW, Lee HA Jr, Frey PA & Abeles RH (1966) Studies on the mechanism of action of cobamide coen- zymes. Chemical properties of the enzyme-coenzyme complex. J Biol Chem 241, 1751–1762. 27 Stroinski A, Pawelkiewicz J & Johnson BC (1974) Allo- steric interactions in glycerol dehydratase. Purification of enzyme and effects of positive and negative cooper- ativity for glycerol. Arch Biochem Biophys 162, 321–330. 28 Honda S, Toraya T & Fukui S (1980) In situ reactiva- tion of glycerol-inactivated coenzyme B 12 -dependent enzymes, glycerol dehydratase and diol dehydratase. J Bacteriol 143, 1458–1465. 29 Ushio K, Honda S, Toraya T & Fukui S (1982) The mechanism of in situ reactivation of glycerol-inacti- vated coenzyme B 12 -dependent enzymes, glycerol dehydratase and diol dehydratase. J Nutr Sci Vitaminol 28, 225–236. 30 Tobimatsu T, Hara T, Sakaguchi M, Kishimoto Y, Wada Y, Isoda M, Sakai T & Toraya T (1995) Molecular cloning, sequencing, and expression of the genes encoding adenosylcobalamin-dependent diol dehydrase of Klebsiella oxytoca. J Biol Chem 270, 7142–7148. 31 Mori K, Tobimatsu T, Hara T & Toraya T (1997) Characterization, sequencing, and expression of the genes encoding a reactivating factor for glycerol-inacti- vated adenosylcobalamin-dependent diol dehydratase. J Biol Chem 272, 32034–32041. 32 Bobik TA, Xu Y, Jeter RM, Otto KE & Roth JR (1997) Propanediol utilization genes (pdu)ofSalmonella typhimurium: three genes for the propanediol dehydra- tase. J Bacteriol 179, 6633–6639. 33 Tobimatsu T, Azuma M, Matsubara H, Takatori H, Niida T, Nishimoto K, Satoh H, Hayashi R & Toraya T (1996) Cloning, sequencing, and high level expression of the genes encoding adenosylcobalamin-dependent glycerol dehydrase of Klebsiella pneumoniae. J Biol Chem 271, 22352–22357. 34 Tobimatsu T, Kajiura H, Yunoki M, Azuma M & Toraya T (1999) Identification and expression of the genes encoding a reactivating factor for adenosylcobala- min-dependent glycerol dehydratase. J Bacteriol 181, 4110–4113. 35 Toraya T & Mori K (1999) A reactivating factor for coenzyme B 12 -dependent diol dehydratase. J Biol Chem 274, 3372–3377. 36 Mori K & Toraya T (1999) Mechanism of reactivation of coenzyme B 12 -dependent diol dehydratase by a molecular chaperone-like reactivating factor. Biochemis- try 38, 13170–13178. 37 Kajiura H, Mori K, Tobimatsu T & Toraya T (2001) Characterization and mechanism of action of a reacti- vating factor for adenosylcobalamin-dependent glycerol dehydratase. J Biol Chem 276, 36514–36519. 38 Seifert C, Bowien S, Gottschalk G & Daniel R (2001) Identification and expression of the genes and purifica- tion and characterization of the gene products involved in reactivation of coenzyme B 12 -dependent glycerol dehydratase of Citrobacter freundii. Eur J Biochem 268, 2369–2378. 39 Liao D-I, Reiss L, Turner I Jr & Dotson G (2003) Structure of glycerol dehydratase reactivase: a new type of molecular chaperone. Structure 11 , 109–119. H. Kajiura et al. B 12 -dependent dehydratases and reactivating factors FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS 5565 [...]...B12-dependent dehydratases and reactivating factors H Kajiura et al 40 Shibata N, Mori K, Hieda N, Higuchi Y, Yamanishi M & Toraya T (2005) Release of a damaged cofactor from a coenzyme B12-dependent enzyme: X-ray structures of diol dehydratase -reactivating factor Structure 13, 1745–1754 41 Tobimatsu T, Kajiura H & Toraya T (2000) Specificities of reactivating factors for adenosylcobalamin-dependent diol dehydratase... adenosylcobalamin-dependent diol dehydratase and glycerol dehydratase Arch Microbiol 174, 81–88 42 Kofoid E, Rappleye C, Stojiljkovic I & Roth J (1999) The 17-gene ethanolamine (eut) operon of Salmonella typhimurium encodes five homologues of carboxysome shell proteins J Bacteriol 181, 5317–5329 43 Mori K, Bando R, Hieda N & Toraya T (2004) Identification of a reactivating factor for adenosylcobalamindependent ethanolamine... Rosenblatt DS & Gravel RA (2002) Identification of the gene responsible for the cblA complementation group of vitamin B12-responsive methylmalonic acidemia based on analysis of prokaryotic gene arrangements Proc Natl Acad Sci USA 99, 15554– 15559 45 Korotkova N & Lidstrom ME (2004) MeaB is a component of the methylmalonyl-CoA mutase complex required for protection of the enzyme from inactivation J Biol Chem... Assembly and protection of the radical enzyme, methylmalonyl-CoA mutase, by its chaperone Biochemistry 45, 9300–9306 47 Sando GN, Grant ME & Hogenkamp HPC (1976) The interaction of adeninylalkylcobalamins with ribonucleotide reductase Biochim Biophys Acta 428, 228–232 48 Tobimatsu T, Sakai T, Hashida Y, Mizoguchi N, Miyoshi S & Toraya T (1997) Heterologous expression, purification, and properties of diol. .. Hogenkamp HPC (1977) Studies on the mechanism of the adenosylcobalamin-dependent diol dehydrase reaction by the use of analogs of the coenzyme J Biol Chem 252, 963–970 51 Gill SC & von Hippel PH (1989) Calculation of protein extinction coefficients from amino acid sequence data Anal Biochem 182, 319–326 52 Davis BJ (1964) Disc electrophoresis II Method and application to human serum proteins Ann NY... purification, and properties of diol dehydratase, an adenosylcobalamin-dependent enzyme of Klebsiella oxytoca Arch Biochem Biophys 347, 132–140 5566 49 Yamanishi M, Yunoki M, Tobimatsu T, Sato H, Matsui J, Dokiya A, Iuchi Y, Oe K, Suto K, Shibata N, et al (2002) The crystal structure of coenzyme B12dependent glycerol dehydratase in complex with cobalamin and propane-1,2 -diol Eur J Biochem 269, 4484– 4494 50 Toraya... Cleavage of structural proteins during the assembly of the head of bacteriophage T4 Nature 227, 680–685 54 Poznanskaya AA, Tanizawa K, Soda K, Toraya T & Fukui S (1979) Coenzyme B12-dependent diol dehydrase: purification, subunit heterogeneity, and reversible association Arch Biochem Biophys 194, 379–386 55 Shibata N, Masuda J, Tobimatsu T, Toraya T, Suto K, Morimoto Y & Yasuoka N (1999) A new mode of B12... Masuda J, Tobimatsu T, Toraya T, Suto K, Morimoto Y & Yasuoka N (1999) A new mode of B12 binding and the direct participation of potassium ion in enzyme catalysis: X-ray structure of diol dehydratase Structure 7, 997–1008 56 Krissinel E & Henrick K (2004) Secondary-structure matching (SSM), a new tool for fast protein structure alignment in three dimensions Acta Cryst D60, 2256– 2268 57 Brunger AT,... D60, 2256– 2268 57 Brunger AT, Adams PD, Clore GM, DeLano WL, ¨ Gros P, Grosse-Kunstleve RW, Jiang J-S, Kuszewski J, Nilges M, Pannu NS, et al (1998) Crystallography & NMR system (CNS): a new software suite for macromolecular structure determination Acta Cryst D54, 905–921 FEBS Journal 274 (2007) 5556–5566 ª 2007 The Authors Journal compilation ª 2007 FEBS . Molecular basis for specificities of reactivating factors for adenosylcobalamin-dependent diol and glycerol dehydratases Hideki Kajiura 1 ,. the complex formation with dehydratases DDR and GDR form a complex with apoenzymes of diol and glycerol dehydratases, respectively, and tran- sient formation of