A Reactivating Factor for Coenzyme B12-dependent Diol Dehydratase*

Adenosylcobalamin-dependent diol dehydratase of Klebsiella oxytoca undergoes suicide inactivation by glycerol, a physiological substrate. The coenzyme is modified through irreversible cleavage of its cobalt-carbon bond, resulting in inactivation of the enzyme by tight binding of the modified coenzyme to the active site. Recombinant DdrA and DdrB proteins of K. oxytoca were co-purified to homogeneity from cell-free extracts of Escherichia coli overexpressing theddrAB genes. They existed as a tight complex,i.e. a putative reactivating factor, with an apparent molecular weight of 150,000. The factor consists of equimolar amounts of the two subunits with M r of 64,000 (A) and 14,000 (B), encoded by the ddrA and ddrB genes, respectively. Therefore, its subunit structure is most likely A2B2. The factor not only reactivated glycerol-inactivated and O2-inactivated holoenzymes but also activated enzyme-cyanocobalamin complex in the presence of free adenosylcobalamin, ATP, and Mg2+. The reactivating factor mediated ATP-dependent exchange of the enzyme-bound cyanocobalamin for free 5-adeninylpentylcobalamin in the presence of ATP and Mg2+, but the reverse was not the case. Thus, it can be concluded that the inactivated holoenzyme becomes reactivated by exchange of the enzyme-bound, adenine-lacking cobalamins for free adenosylcobalamin, an adenine-containing cobalamin.

Diol dehydratase as well as glycerol dehydratase undergoes concomitant, irreversible inactivation by glycerol during catalysis (2, 14 -16). This inactivation is mechanism-based and involves irreversible cleavage of the Co-C bond of AdoCbl forming 5Ј-deoxyadenosine and a modified coenzyme (4,14). Irreversible inactivation of the enzyme results from the tight binding of the modified, inactive cobalamin (4,14,17). Such suicide inactivation seemed enigmatic because glycerol is a growth substrate for K. oxytoca and Klebsiella pneumoniae. This apparent inconsistency was solved by our finding that the glycerol-inactivated enzymes in permeabilized cells (in situ) of K. oxytoca and K. pneumoniae undergo rapid reactivation in the presence of free AdoCbl, ATP, and Mg 2ϩ (or Mn 2ϩ ) (13,18). Because the reactivation was detectable only in situ but not in vitro, it remained unclear whether the reactivation is caused by a specific proteinous factor. Recently, we have identified the two open reading frames located in the 3Ј-flanking of the K. oxytoca diol dehydratase genes as the genes encoding a reactivating factor for glycerol-inactivated diol dehydratase and designated them as ddrA and ddrB genes (19,20).
This article reports biochemical demonstration that purified DdrA and DdrB proteins form a tight complex that serves as a reactivating factor for glycerol-inactivated holoenzyme as well as O 2 -inactivated holoenzyme in vitro in the presence of free AdoCbl, ATP, and Mg 2ϩ . Evidence for the reactivating factormediated exchange of an enzyme-bound, adenine-lacking cobalamins for a free, adenine-containing cobalamin is also described here.
Cultivation of Overexpressing E. coli Strain-Recombinant DdrA and DdrB proteins of K. oxytoca were purified to homogeneity from E. coli JM109 harboring expression plasmid pUSI2ENd(6/5b) that was aero-bically grown at 30°C in LB medium containing ampicillin (50 g/ml). Isopropyl-1-thio-␤-D-galactopyranoside was added to a concentration of 1 mM for induction, and cells were harvested in the late logarithmic phase.
Protein Assays-During purification of diol dehydratase and the ddrA and ddrB gene products, protein concentrations were routinely estimated by the method of Lowry et al. (25) with crystalline bovine serum albumin as a standard. The concentration of purified diol dehydratase was determined by measuring the absorbance at 280 nm. The molar absorption coefficient at 280 nm (⑀ M, 280 ) calculated by the method of Gill and von Hippel (26) for diol dehydratase from its deduced amino acid compositions (23) and subunit structure (22) was 120,500 M Ϫ1 cm Ϫ1 . Based on the molecular weight predicted, ⑀ 1%, 280 was calculated to be 5.81 for diol dehydratase.
PAGE-PAGE was performed under non-denaturing conditions essentially as described by Davis (27), except that 5 mM dithiothreitol was added in the gel, or under denaturing conditions as described by Laemmli (28). Protein was stained with Coomassie Brilliant Blue R-250. Densitometric analysis of gels was performed with a Printgraph AE-6911CX system (ATTO, Tokyo, Japan) and the NIH-Image program, Version 1.61 (National Institutes of Health).
Edman Sequencing of the Subunits-A purified preparation of the complex of the DdrA and DdrB proteins was separated into the subunits (A and B polypeptides, respectively) by SDS-PAGE (15.0%) and electrophoretically transferred to a polyvinylidene difluoride membrane (Ap-plied Biosystems). Protein bands were visualized with Coomassie Brilliant Blue R-250, excised, and analyzed for NH 2 -terminal amino acid sequences on an Applied Biosystems 491 protein sequencer.
Molecular Weight Determination by Gel Filtration-The molecular weight of the complex (putative reactivating factor) was determined by gel filtration on Superose 6 column (HR10/30) using a FPLC system (Amersham Pharmacia Biotech). The purified factor and molecular weight marker proteins were applied to the column, which was equilibrated with 50 mM potassium phosphate buffer (pH 8.0) containing 0.1 M KCl and developed with the same buffer at a flow rate of 0.4 ml/min. The elution of proteins was monitored by A 280 .
Enzyme Assays-The amount of aldehydic products formed by diol dehydratase reaction was determined by the 3-methyl-2-benzothiazolinone hydrazone method (29). One unit of diol dehydratase is defined as the amount of enzyme activity that catalyzes the formation of 1 mol of propionaldehyde/min at 37°C. Reactivation of the inactivated holoenzymes and activation of the enzyme⅐CN-Cbl complex by the reactivating factor was determined using 1,2-propanediol or glycerol as a substrate in the presence of 21 M AdoCbl, 24 mM ATP, and 24 mM MgCl 2 .
The capability of reactivating factor to reactivate glycerol-inactivated or O 2 -inactivated holodiol dehydratase was assayed in vitro using 1,2propanediol as substrate. The capability of it to activate the inactive enzyme⅐CN-Cbl complex was also assayed in vitro because these two capabilities in situ were shown to be well correlated (18).

FIG. 1. Purification of the DdrA and DdrB proteins.
Fractions at each purification step were subjected to SDS-PAGE (denaturing) on 12.5% polyacrylamide gel (A) and PAGE (non-denaturing) on 6.5% polyacrylamide gel containing 5 mM dithiothreitol (B). Protein staining of the resulting gels were carried out as described under "Experimental Procedures." Molecular weight markers were SDS-7 (Sigma) (A). Positions of the A and B polypeptide are indicated with arrowheads to the right of the gel (A).

FIG. 2. Reactivation of glycerol-inactivated (A) and O 2 -inactivated (B) holodiol dehydratases by the reactivating factor.
Glycerolinactivated holoenzyme was prepared by incubation of apoenzyme (331 units) with 50 M AdoCbl at 37°C for 30 min in 2.5 ml of 0.05 M potassium phosphate buffer (pH 8) containing 30% glycerol, followed by dialysis at 4°C for 43 h against 800 volumes of 0.05 M potassium phosphate buffer (pH 8) containing 2% 1,2-propanediol. O 2 -inactivated holoenzyme was prepared by incubation of substrate-free apoenzyme (12.6 units) with 49 M AdoCbl at 37°C for 30 min in 0.18 ml of 0.05 M potassium phosphate buffer (pH 8). Glycerol-inactivated holoenzyme (1.5 units) or O 2 -inactivated holoenzyme (1.5 units) was incubated at 37°C for the indicated time periods with (q, E) and without (OE, ƒ) 47 g of the reactivating factor in 0.02 M potassium phosphate buffer (pH 8) containing 21 M AdoCbl and 1.2 M 1,2-propanediol in the presence (q, OE) and absence (E, ƒ) of 24 mM ATP/24 mM MgCl 2 , in a total volume of 50 l. Time course of 1,2-propanediol dehydration with non-inactivated enzyme was measured as a control (छ). The reaction was terminated by adding 50 l of 0.1 M potassium citrate buffer (pH 3.6). After removal of precipitate by centrifugation, the reaction mixture was appropriately diluted for determining the amount of propionaldehyde formed by the 3-methyl-2-benzothiazolinone hydrazone method (18).

Purification of Recombinant DdrA and DdrB Proteins-
The ddrA and ddrB gene products were co-purified from extracts of overexpressing E. coli. All operations were performed at 0 -4°C. Throughout the purification steps, purity of the proteins in each fraction was analyzed by SDS-PAGE.
About 10 g of wet cells grown at 30°C were suspended in 50 ml of 50 mM potassium phosphate buffer (pH 8.0) containing 2 mM EDTA and 2 mM PMSF and disrupted by sonication for 10 min at 240 W with a Kaijo Corp. TA-5287 ultrasonic destruction system (Japan). After centrifugation at 27,200 ϫ g for 30 min, the supernatant was collected. The precipitate was washed with 60 ml of the same buffer, and the washing was combined with the supernatant (cell-free extract).
Solid ammonium sulfate was added to the cell-free extract to 15% saturation. After centrifugation, solid ammonium sulfate was added again to the supernatant to 35% saturation. The precipitate was dissolved in 20 ml of 50 mM potassium phosphate buffer (pH 8.0) containing 2 mM EDTA and 2 mM PMSF and dialyzed for 1 day against 2 liters of 5 mM potassium phosphate buffer (pH 8.0) containing 0.5 mM EDTA with one buffer change.
The dialysate was applied to a DEAE-cellulose column (bed volume, 100 ml) that was equilibrated with 10 mM potassium phosphate buffer (pH 8.0). The column was washed successively with 500 ml of the same buffer and with 600 ml of 5 mM potassium phosphate buffer (pH 8. The resulting solution was applied to a hydroxyapatite column (bed volume, 70 ml) which was equilibrated with 2 mM potassium phosphate buffer (pH 8.0). The column was washed successively with 280 ml of the same buffer and with 200 ml of 6 mM potassium phosphate buffer (pH 8.0) and then developed with 350 ml of 10 mM potassium phosphate buffer (pH 8.0). The DdrA and DdrB proteins-containing fractions were pooled and concentrated to about 1.7 ml by ultrafiltration through a Diaflo PM-10 membrane and Centriplus (Amicon).
The concentrated solution was loaded onto a Sephadex G-200 column (bed volume, 150 ml) which was equilibrated with the 50 mM potassium phosphate buffer (pH 8.0). The column was developed with the same buffer, and the fractions containing the DdrA and DdrB proteins were pooled, concentrated to about 5 ml by ultrafiltration through a Centriplus, and stored at Ϫ80°C.
Purity, Molecular Weight, and Subunit Structure of the Purified Reactivating Factor-It is evident that the two bands with M r of 64,000 and 14,000 (designated A and B polypeptides, respectively) were overexpressed in E. coli JM109 carrying pUSI2ENd(6/5b) (Fig. 1A). Both of these bands were progressively enriched upon purification. When the purified preparation was electrophoresed under non-denaturing conditions in the presence of dithiothreitol, it migrated as a single band (Fig.  1B). Therefore, it was clear that the two polypeptides were co-purified as a tight complex. Because the predicted molecular weights of the DdrA and DdrB proteins are 64,266 and 13,620, respectively (20), it was highly suggested that the A and B polypeptides are the products of these genes. The NH 2 -terminal 10-amino acid sequences of the A and B polypeptides determined by Edman sequencing were MRYIAGIDIG and MNGNHSAPAI, respectively. These sequences agreed com- pletely with those deduced from the nucleotide sequences of the ddrA and ddrB genes (20). Thus, the A and B polypeptides were undoubtedly identified as the ddrA and ddrB gene products, respectively. These results indicate that the DdrA and DdrB proteins are co-purified and exist as a tight complex under non-denaturing conditions. This complex was considered as a putative reactivating factor.
To determine the subunit composition of the factor, the complex purified to homogeneity was separated into subunits by SDS-PAGE in a 15% gel and stained with Coomassie Brilliant Blue R-250. Densitometric analysis of the bands, together with molecular weights of the subunits predicted from the amino acid composition, indicated that the molar ratio of the A and B polypeptides in the complex was approximately 1:1.1. The apparent molecular weight of the complex determined by FPLC with a calibrated Superose 6 column (HR 10/30) was approximately 150,000 (data not shown). By taking the predicted molecular weights of the subunits into consideration, it can be concluded that the subunit structure of the putative reactivating factor is most likely A 2 B 2 .
The molar absorption coefficient at 280 nm (⑀ M, 280 ) for the complex (reactivating factor), calculated by the method of Gill and von Hippel (26) from its amino acid compositions and subunit structure, was 58,140 M Ϫ1 cm Ϫ1 . ⑀ 1%, 280 for the reactivating factor was calculated to be 3.73 on the basis of the predicted molecular weight.
In Vitro Reactivation of Inactivated Holoenzymes by the Reactivating Factor-The capability of the putative reactivating factor to reactivate the glycerol-inactivated holodiol dehydratase was examined in vitro using 1,2-propanediol as substrate. As illustrated in Fig. 2A, in vitro reactivation of the glycerol-inactivated holoenzyme by the purified factor in the presence of AdoCbl, ATP, and Mg 2ϩ was observed for the first time. The reactivation did not take place at all without the factor or with the factor but in the absence of ATP/Mg 2ϩ . Free AdoCbl was also absolutely required for the reactivation (data not shown). Thus, it is evident that the factor actually functions as a reactivating factor for glycerol-inactivated diol dehydratase. The product formed increased exponentially at the initial stage of reaction and then almost linearly with time of incubation. By comparison of the maximum slope with the control, the extent of reactivation under the conditions employed was estimated to be approximately 64%.
Diol dehydratase holoenzyme is known to undergo inactivation by O 2 in the absence of substrate (30). This inactivation is considered because of reaction of the activated Co-C bond of the enzyme-bound coenzyme with O 2 . Fig. 2B shows that the O 2 -inactivated holoenzyme also undergoes reactivation by the factor in the presence of AdoCbl, ATP, and Mg 2ϩ . The extent of reactivation increased with time of incubation and reached at least 71% at 20 min. Again, the reactivation was strictly dependent on the factor and on ATP/Mg 2ϩ and free AdoCbl.
In Vitro Activation of the Enzyme⅐CN-Cbl Complex by the Reactivating Factor-The capability of the reactivating factor to activate the inactive complex of diol dehydratase with CN-Cbl was also examined in vitro because the inactive enzyme⅐CN-Cbl complex can be considered as a model of the inactivated holoenzyme. Fig. 3A indicates that the enzyme⅐CN-Cbl complex is rapidly activated by the factor in the presence of AdoCbl, ATP, and Mg 2ϩ . This activation by the factor also required ATP/Mg 2ϩ (Fig. 3A) in addition to free AdoCbl (data not shown). Approximately 76% of the enzyme⅐CN-Cbl complex underwent activation by 20 min of incubation under the conditions. As shown in Fig. 3B, the extent of reactivation was dependent on a molar ratio of the factor to the enzyme⅐CN-Cbl complex. From the double-reciprocal plot, the concentration of the factor giving half-maximal activation of 1.2 M enzyme⅐CN-Cbl complex was calculated to be 3.5 M.
Direct Evidence for Cobalamin Exchange-The reactivating factor, free AdoCbl, ATP, and Mg 2ϩ were absolutely required for both reactivation of the glycerol-inactivated holoenzyme and activation of the enzyme⅐CN-Cbl complex. ADP was not able to replace ATP. From the absolute requirement for free AdoCbl, it was strongly suggested that the reactivation of the glycerol-inactivated holoenzyme and activation of the enzyme⅐CN-Cbl complex occurs by exchange of the enzymebound, modified coenzyme and CN-Cbl, respectively, for free intact AdoCbl. We have previously reported that the in situ reactivation of the glycerol-inactivated holoenzyme or the in situ activation of inactive cobalamin-enzyme complexes takes place by exchange of enzyme-bound cobalamins for AdoCbl (13,18).
To examine whether the reactivating factor mediates such exchange, the enzyme⅐CN-Cbl complex was subjected to incubation with and without the reactivating factor in the presence of AdePeCbl, ATP, and Mg 2ϩ , followed by dialysis to remove unbound cobalamins. AdePeCbl, an inactive analog of AdoCbl containing the adenine ring in the upper axial ligand, was used instead of AdoCbl itself because the complex of diol dehydratase with AdoCbl (regular holoenzyme) is catalytically active and rather susceptible to inactivation even in the presence of substrate (31). As depicted in Fig. 4A, the spectrum of the dialysate indicates that the enzyme-bound CN-Cbl was replaced by AdePeCbl with the reactivating factor. This exchange never occurred without the factor. Thus, it is evident that the reactivating factor mediates the exchange of the enzyme-bound CN-Cbl for free AdePeCbl. In contrast, the reverse was not the case. That is, the replacement of the enzyme-bound AdePeCbl by free CN-Cbl did not occur at all even with the factor in the presence of ATP and Mg 2ϩ (Fig. 4B). Thus, it is quite likely that the reactivating factor mediates the exchange of enzymebound, adenine-lacking cobalamins for free, adenine-containing cobalamins. Because the coenzyme loses the adenine moiety by irreversible cleavage of the Co-C bond in the inactivation of holoenzymes by glycerol or O 2 , it can therefore be concluded that the reactivating factor reactivates the inactivated holoenzymes or activates the enzyme⅐CN-Cbl complex by mediating the ATP-dependent exchange of the enzymebound, modified coenzyme or CN-Cbl, i.e. adenine-lacking cobalamins, for free intact AdoCbl, i.e. an adenine-containing cobalamin.
Numbers of Reactivation per Diol Dehydratase and per Reactivating Factor-The data shown in Figs. 2 and 3 were obtained with 1,2-propanediol as substrate for measuring enzyme activity restored. Because 1,2-propanediol does not cause suicide inactivation at a significant rate, there remains a possibility that the reactivating factor may mediate only a single exchange. To test this possibility, the time course of glycerol dehydration was determined with and without the reactivating factor. As shown in Fig. 5A, this substrate brought about complete inactivation of the enzyme within 3 min. However, when the reactivating factor and ATP/Mg 2ϩ were supplemented to the reaction mixture in addition to AdoCbl, an initial rapid phase of glycerol dehydration was followed by a slower but almost constant rate of the dehydration. Furthermore, when the reactivating factor was added to the completely inactivated enzyme together with ATP/Mg 2ϩ in the presence of free AdoCbl, the inactivated enzyme underwent rapid reactivation. The product formed increased linearly with time of incubation, although the rate of product formation with glycerol was much slower than that with 1,2-propanediol. As shown in Fig. 5B, the amount of the product formed by the reactivated enzyme in 4 h was dependent on a molar ratio of the factor to diol dehydratase. On the assumption that the probability of the mechanism-based inactivation by glycerol is not affected by the presence of the factor and ATP/Mg 2ϩ , the numbers of reactivation per diol dehydratase and per reactivating factor were calculated to reach 6 and 2, respectively. These numbers were 5.5 and 1.5, respectively, when the reaction was terminated at 100 min of incubation. It is therefore evident that diol dehydratase undergoes multiple reactivation during dehydration of glycerol. However, by taking the dimeric subunit structure of the factor into consideration, it was suggested that the reactivating factor failed in catalyzing multiple exchanges of tightly bound, modified coenzyme for exogenous AdoCbl on diol dehydratase under the experimental conditions employed. DISCUSSION AdoCbl-dependent enzymatic reactions are initiated by homolysis of the Co-C bond of the enzyme-bound coenzyme and proceed via radical mechanisms (1,3,4,(32)(33)(34)(35)(36). Such reactions are considered to need the assistance of high reactivity of a free radical. Highly reactive radical intermediates must sustain their reactivity at the active sites and become extinct in the only way destined for the reaction. Once a radical intermediate becomes extinct or stabilized by side reactions, regeneration of AdoCbl becomes impossible, resulting in irreversible modification of the coenzyme (3). This leads to inactivation of enzymes because the modified coenzymes remain tightly bound to enzymes and not exchangeable with free AdoCbl. Are such inactivated enzymes reactivated? This would be important for cellular economics of energy. The data presented in this paper gave at least one answer to this question. That is, the purified complex of the DdrA and DdrB proteins reactivates glycerolinactivated and O 2 -inactivated holoenzymes. This result was as expected because the ddrAB genes were identified as the genes of K. oxytoca encoding a putative reactivating factor for inactivated diol dehydratase (20). This is the first biochemical demonstration of such reactivation by these gene products.
We have previously demonstrated with toluene-treated cells of K. oxytoca and K. pneumoniae (18) and recombinant E. coli harboring plasmid pUCDD11 (19) that the complexes of diol dehydratase or glycerol dehydratase with CN-Cbl, aquacobalamin, and pentylcobalamin undergo in situ activation, whereas the complexes with adeninylbutylcobalamin and AdePeCbl do not. These facts suggest that the affinity of the enzyme for cobalamins lacking the adenine moiety in the upper axial ligand is selectively lowered in situ in the presence of ATP and Mg 2ϩ , resulting in replacement of these enzyme-bound cobalamins by free AdoCbl. It seemed likely that the binding affinity for the modified coenzymes in the glycerol-inactivated and O 2inactivated holoenzymes is also lowered in situ in the presence of ATP and Mg 2ϩ because the adenosyl group is irreversibly severed from the cobalamin moiety during the inactivation processes (4,14,30). Evidence for such an exchange mechanism of reactivation was obtained here in vitro for the first time. It was demonstrated that the reactivating factor reactivates the inactivated holoenzyme or activates the enzyme⅐CN-Cbl complex by mediating the exchange of the enzyme-bound, adenine-lacking cobalamin for free AdoCbl, an adenine-containing cobalamin.
Because the adenine-lacking cobalamins are also bound by diol dehydratase so tightly, such exchange never occurs without the reactivating factor. Complex formation between the factor and diol dehydratase was observed (data not shown). Therefore, it can be postulated that the reactivating factor selectively lowers the affinity of diol dehydratase for adeninelacking cobalamins by forming a complex with the enzyme protein and transiently affecting its higher-order structures. The free energy required for such structural transitions may be provided by coupling with the hydrolysis of ATP because the reactivating factor shows low but distinct ATPase activity (data not shown). Non-hydrolyzable ATP analogs were not effective for the reactivation of the glycerol-inactivated holoenzyme and the activation of the enzyme⅐CN-Cbl complex. Such mechanism of action of the reactivating factor seems similar to those of molecular chaperones. In this sense, the diol dehydratase-reactivating factor may be considered as a new type of molecular chaperone which is involved in reactivation of inactivated enzymes. Detailed mechanism of action of the factor is under current investigation.
The proteins homologous to the Ddr proteins were assumed to serve as a reactivating factor for inactivated glycerol dehy-dratase (20). Therefore, such reactivating factor may be a factor of general importance for AdoCbl-dependent enzymes.