Mechanism of Anaerobic Ether Cleavage

2-Phenoxyethanol is converted into phenol and acetate by a strictly anaerobic Gram-positive bacterium,Acetobacterium strain LuPhet1. Acetate results from oxidation of acetaldehyde that is the early product of the biodegradation process (Frings, J., and Schink, B. (1994) Arch. Microbiol. 162, 199–204). Feeding experiments with resting cell suspensions and 2-phenoxyethanol bearing two deuterium atoms at either carbon of the glycolic moiety as substrate demonstrated that the carbonyl group of the acetate derives from the alcoholic function and the methyl group derives from the adjacent carbon. A concomitant migration of a deuterium atom from C-1 to C-2 was observed. These findings were confirmed by NMR analysis of the acetate obtained by fermentation of 2-phenoxy-[2-13C,1-2H2]ethanol, 2-phenoxy-[1-13C,1-2H2]ethanol, and 2-phenoxy-[1,2-13C2,1-2H2]ethanol. During the course of the biotransformation process, the molecular integrity of the glycolic unit was completely retained, no loss of the migrating deuterium occurred by exchange with the medium, and the 1,2-deuterium shift was intramolecular. A diol dehydratase-like mechanism could explain the enzymatic cleavage of the ether bond of 2-phenoxyethanol, provided that an intramolecular H/OC6H5 exchange is assumed, giving rise to the hemiacetal precursor of acetaldehyde. However, an alternative mechanism is proposed that is supported by the well recognized propensity of α-hydroxyradical and of its conjugate base (ketyl anion) to eliminate a β-positioned leaving group.

environment at high quantities, as lubricants, solubility mediators, hydrophilic moiety of nonionic surfactants and household detergents, or as a constituent of cosmetics and pharmaceutical preparations (2). PEGs were found to be degraded by various bacteria, both in the presence and the absence of molecular oxygen (aerobically, Refs. 2-6; anaerobically, Refs. [7][8][9][10][11][12]. Different reaction mechanisms are involved in PEG degradation, and it is generally accepted that they all involve the formation of a labile intermediary hemiacetal structure (1). In the presence of oxygen, such a hemiacetal can be formed through a monooxygenase-catalyzed hydroxylation of one of the methylene carbon atoms. In the absence of molecular oxygen, generation of such a hemiacetal can be achieved only with substrates containing a free hydroxyl group adjacent to the ether carbon through a hydroxyl shift reaction. Such hydroxyl shift reactions are catalyzed by diol dehydratase (EC 4.2.1.28) and glycerol dehydratase (EC 4.2.1.30) enzymes, with the substrates EG, 1,2-propanediol, or glycerol. The reaction mechanisms of these enzymes have been studied in great detail (13)(14)(15). They typically depend on adenosylcobalamin as cofactor, which provides a reversible radical source. Based on these well studied model systems, it was assumed that anaerobic PEG degradation to acetaldehyde as the first identifiable intermediate may be adenosylcobalamin-dependent as well and may proceed in a way analogous to diol dehydratase, provided that at least one terminal hydroxyl group is free for the required shift reaction (7,10,11,16,17).
The anaerobic homoacetogenic bacterium Acetobacterium strain LuPhet 1 can grow with low molecular weight PEGs as the sole source of carbon and energy but can also use EG or 2-phenoxyethanol as the sole substrate; the latter is fermented to phenol plus acetate (12) as schematized in Fig. 1. In cell-free extracts of this strain, two separate enzyme activities were detected, the one reacting with EG and the other one reacting with phenoxyethanol. Both reactions yield acetaldehyde as the first product. The authors found that the EG-degrading activity was stimulated 3.5-fold by added adenosylcobalamin and was strongly inhibited by cyano-or hydroxocobalamin or by light; the latter effect could be alleviated by adenosylcobalamin addition (12). With this, the EG-degrading enzyme behaved identically to the known diol dehydratases (18). Cleavage of 2-phenoxyethanol, on the other hand, was influenced neither by various corrinoids, including adenosylcobalamin, nor by light (12), indicating that the two enzymes are definitively different proteins and perhaps operate by different reaction mechanisms.
Since 2-phenoxyethanol is a monosubstituted ethylene glycol, it allows us to study the assumed shift reaction in greater detail because theoretically, either the free hydroxyl group or the phenoxy residue can be shifted to form a hemiacetal as an intermediate. We therefore tried to distinguish between those two possible pathways by application of specifically deuterated and/or 13 C-labeled 2-phenoxyethanol preparations to resting cell suspensions of Acetobacterium strain LuPhet 1 and subsequent analysis of the produced acetate.

EXPERIMENTAL PROCEDURES
General Methods-TLC was performed on Silica Gel F 254 -precoated aluminum sheets (0.2-mm layer, Merck, Darmstadt, Germany); components were detected by spraying a ceric sulfate ammonium molybdate solution followed by heating to ϳ150°C. Silica gel (Merck, 40 -63 m) was used for FC. GC analyses were carried out on a DANI 3800 gas chromatograph (DANI, Monza, Italy) using a homemade glass column (2 m ϫ 2 mm inner diameter) packed with 20% Carbowax 20M on Chromosorb W (60 -80 mesh). GC parameters were as follows: injector, 220°C; detector (flame ionization detection), 220°C; carrier, N 2 (30 ml/min); oven, from 60 to 200°C at 10°C/min. 1 H and 13 C NMR spectra were acquired at 400.132 and 100.613 MHz on a Bruker AVANCE 400 Spectrometer using an Xwin-nmr software package and at 200.133 and 50.330 MHz on a Bruker AC 200 (Bruker, Karlsruhe, Germany) equipped with an ASPECT 2000 data system. Chemical shifts (␦) are given in parts per million and were referenced to the signals of CDCl 3 (␦ H 7.25 and ␦ C 77.00 ppm) or to 3-(trimethylsilyl)propionic-2,2,3,3-d 4 acid sodium salt (␦ Me 0 ppm) in the case of D 2 O/NaOD (pH Ͼ10) solutions. 13 C NMR signal multiplicities were based on attached proton test spectra. 13 C NMR spectra for quantitative analyses were obtained by the inverse gated decoupling pulse sequence and a relaxation delay of 300 s (19). EIMS spectra were run on a VG 7070 EQ mass spectrometer (VG Instruments, Manchester, UK) operating at 70 eV. All reagents were of commercial quality or purified prior to use by standard methods. Ethyl bromo-[2-13 C]acetate, bromo-[1-13 C]acetate, and bromo-[1,2-13 C]acetate were from Aldrich.
Medium and Growth Conditions-Acetobacterium strain LuPhet 1 (DSM 9077) was grown at 28°C in the dark in bicarbonate-buffered (30 mM, pH 7.2), sulfide-reduced (1 mM) freshwater mineral medium (20) with 10 mM 2-phenoxyethanol as sole organic carbon substrate under a N 2 /CO 2 atmosphere (80:20 v/v) as described previously (12). 2-Phenoxyethanol was added from anoxic filter-sterilized stock solutions. Besides other vitamins, the medium contained about 40 nM cyanocobalamin. The addition of a few crystals of dithionite shortened the lag phases. Cells were grown as batch cultures of 0.5-or 1-liter volume in infusion bottles sealed with butyl rubber septa. Growth was followed by measuring turbidity at 578 nm.
Cell Suspension Experiments with Labeled 2-Phenoxyethanol-Cell suspensions were prepared under strictly anoxic conditions in an anoxic chamber (Coy Laboratory Products, Ann Arbor, MI) with an atmosphere of 5% H 2 in N 2 . Bacteria were harvested in the late exponential growth phase (A 578 ϭ 0.1) by centrifugation at 11,000 ϫ g and 4°C for 30 min. Polypropylene centrifuge beakers were preincubated in the chamber for 2-3 days. Cells were washed once with degassed potassium phosphate buffer (50 mM, pH 7.0) prereduced with 2.5 mM titanium(I-II)citrate and then resuspended in freshwater mineral medium without substrate (bicarbonate-buffered, 30 mM, pH 7.2, and sulfide-reduced, 1 mM) and transferred into a serum bottle sealed with a butyl rubber stopper. The headspace in the bottle was exchanged to N 2 /CO 2 (80:20 v/v), and the cell suspension was incubated at 28°C under protection from light. The reaction was started by the addition of labeled 2-phenoxyethanol to about 10 mM concentration. Aliquots (50 l) were taken at regular intervals with a gas-tight syringe and injected into 200 l of H 3 PO 4 (100 mM) to stop all enzymatic reactions. 2-Phenoxyethanol, phenol, and acetate were analyzed with a high performance liquid chromatography system (System Gold, Beckman Instruments) equipped with an AQ-ODS column (4.6 by 250 mm) from YMC Europe (Schermbeck, Germany) with an eluent composed of ammonium phosphate buffer (100 mM, pH 2.6) and methanol. The three compounds were measured simultaneously using a gradient from 5% methanol increasing to 60% methanol and detection at a 206-nm wavelength. Concentrations were calculated via external standards. The protein content in the cell suspension varied between 0.09 and 0.4 mg/ml. The reaction was stopped after substrate depletion or after a maximum of 28 h by centrifugation at 11,000 ϫ g for 30 min and at 4°C. The supernatant was filtered through a cellulose acetate membrane filter with a pore size of 0.2 m and stored at 4°C. From the supernatant the acetate was isolated by the procedure described below.
Isolation of Acetate from the Reaction Mixture-The neutral or slightly alkaline aqueous phase was extracted three times with chloroform to reduce the phenol and 2-phenoxyethanol content before acidification to pH 1-2 by the addition of concentrated HCl. Then some NaCl was added to the aqueous phase, and the acetic acid was extracted with diethyl ether at least twice with a 5:1 ether-to-water volume ratio. The ether phase was concentrated to few milliliters in vacuo, and the acetic acid was dissociated by the addition of a sufficient amount of sodium hydroxide ( (5 ml) was refluxed under N 2 , monitoring the reaction progress by TLC (petroleum ether/diethyl ether, 8:2) and GC. After 4 h, the reaction mixture was diluted with water (10 ml), acidified with 2 N HCl, and extracted with diethyl ether. The organic phase was washed with saturated NaHCO 3 , washed with water, and dried over Na 2 SO 4 . Solvent removal under reduced pressure followed by FC of the residue (eluent as above) afforded ethyl phenoxyacetate (2) (230 mg, 71% yield); pure by TLC (R f 0.61) and GC (t R 6.8 min), 1 H NMR and EIMS as in Ref. 23; 13 C NMR as in Ref. 24. Compound 2 (200 mg, 1.1 mmol) in dry diethyl ether (2 ml) was added dropwise to a cold (0°C) suspension of LiAlD 4 (92.4 mg, 2.2 mmol) in dry diethyl ether (4 ml), and the reaction mixture was refluxed with stirring under N 2 for 5 h (GC control). After cooling to room temperature, a saturated solution of Na 2 SO 4 was carefully added. The white salts were removed by filtration and then washed with diethyl ether. The filtrate was washed with water, dried (Na 2 SO 4 ), and evaporated under reduced pressure to give the title compound 3 (142 mg, 92% yield) pure by GC (t R 8.  (7)-Benzyloxyacetyl chloride (4.7 g, 25.5 mmol) was added via a syringe over 15 min to an ice-cooled solution of pyridine (5 ml) and ethyl alcohol (4 ml) in dry dichloromethane (15 ml) under N 2 with stirring. The reaction mixture was allowed to warm to room temperature, and stirring was continued for 30 min followed by quenching with 1 N HCl (20 ml). The two phases were separated, and the organic one was washed with water (2 ϫ 20 ml), dried over Na 2 SO 4 , and concentrated under reduced pressure to give a viscous oil. After purification by FC (petroleum ether/ethyl acetate, 5:2), ethyl benzyloxyacetate (4) (4.7 g, 95% yield) was obtained, pure by TLC (R f 0.46, eluent as above), 1 H, and 13 C NMR (25). To an ice-cooled solution of the ester 4 (4.7 g, 24.2 mmol) in dry diethyl ether (40 ml) was added LiAlD 4 (1.2 g, 47.6 mmol) in several portions. After further addition of diethyl ether (10 ml), the reaction mixture was refluxed for 3 h. Workup as described above for compound 3 afforded 2-benzyloxy-[1,1-2 H 2 ]ethanol (5) (3.3 g, 89% yield), pure by TLC (R f 0.15, eluent as above) and GC (t R 8.7 min), which was used for the next step without further purification; 1  A stirred solution of PPh 3 (2.0 g, 7.8 mmol) and diisopropyl azodicarboxylate (1.5 ml, 7.8 mmol) in tetrahydrofuran (80 ml) at 0°C was treated, sequentially, with a solution of freshly distilled phenol (1.1 g, 11.7 mmol) in tetrahydrofuran (5 ml) and then with a solution of 2-benzyloxy-[1,1-2 H 2 ]ethanol (5) (1.0 g, 6.5 mmol) over a period of 15 min. The reaction mixture was allowed to warm to room temperature, stirred for an additional 1 h (TLC control), and quenched by the addition of water (5 ml) and a few drops of concentrated HCl. The solvent was removed under reduced pressure, and the residue was taken up with diethyl ether (60 ml). Insoluble materials were removed by filtration, and the filtrate was washed with 2 N NaOH and with water, dried (Na 2 SO 4 ), and concentrated to approximately a half-volume under re-  (Fig. 2A). After the complete fermentation of 3 by Acetobacterium under a N 2 /CO 2 atmosphere, sodium acetate was isolated from the culture supernatant and examined by 1 H and 13 C NMR spectroscopy. It is understood that throughout this study, spectra of sodium acetate (proton-decoupled in the case of 13 C) were recorded using NaOD/D 2 O at pH Ͼ 10. The methyl regions of these spectra exhibited peaks assignable to a mixture of monoand non-deuterated acetate molecules only (Fig. 3, A and B). Monodeuterated molecules are revealed by the typical patterns of 1 H and 13 C NMR signals due to the CH 2 D and 13 CH 2 D groups. In both cases, this pattern consists of a 1:1:1 triplet (27) ( 2 J HD ϭ 2.09 Hz, J CD ϭ 19.5 Hz) (28,29), which is upfield with respect to the non-deuterated methyl group ( 2 ⌬H(D) ϭ 13.5 ppb, ⌬C(D) ϭ 0.254 ppm) (29,30). The presence of non-deuterated molecules besides the monodeuterated ones in the fermentation acetate (ϳ35% as calculated from the integrated peak areas in the 1 H NMR spectrum, taking into account the number of protons of the two species) can be explained by considering additional acetate synthesis from CO 2 by this acetogenic bacterium (12) (Fig. 1). In addition, a partial loss of both deuterium atoms during the conversion of 2-phenoxyethanol into phenol and acetate could not be excluded.

2-Phenoxy-[2-2 H 2 ]ethanol
When Acetobacterium cells were fed with 2-phenoxy-[2-2 H 2 ]ethanol (7) prepared as shown in Fig. 2B, the resulting acetate was found to be a mixture of dideuterated and nondeuterated molecules in the ratio ϳ2.5:1. In fact, in the 1 H and 13 C NMR spectra of this acetate, an upfield quintet (1:2:3:2:1) (27) was present beside the singlets due to the non-deuterated methyl group ( 2 ⌬H(D 2 ) ϭ 27.0 ppb, ⌬C(D 2 ) ϭ 0.467 ppm) (29,30), thus indicating the occurrence of CHD 2 and 13 CHD 2 groups (Fig. 3, C and D). The complete absence of CHD 2 -CO 2 Ϫ and CH 2 D-CO 2 Ϫ species in the product from the former and the latter experiment, respectively, clearly resulted from a comparison of the corresponding NMR spectra. The results of the experiments carried out with 2-phenoxyethanol bearing the dideuterated methylene group at either position of the glycol unit were in agreement with each other and consistent with the conversion of carbon-1 into the carboxylic group of the acetate and of carbon-2 into the methyl group. The most striking feature of this biotransformation appeared to be the shift of a deuterium (hydrogen) atom from carbon-1 to carbon-2 (Reaction 1).
To gain further insight into the process schematized in Reaction 1, samples of 2-phenoxyethanol enriched with 13 C at 1and/or 2-position and dideuterated at the alcoholic function, i.e. 8, 9, and 10 ( Fig. 4), were prepared from the proper ethyl [ 13 C]bromoacetate. The quantitative determination of differently labeled species (isotopomers) in the acetate recovered from feeding experiments performed with these samples was based on peak area measurements in 1 H and 13 C NMR spectra, provided that the latter were obtained by the inverse gated decoupling method (19). The identification of signals due to isotopomeric molecules was made possible by exploiting deuterium effects on the shielding of 1 H and 13 C nuclei as well as spin-spin coupling constants.
After fermentation of sample 8, the 1 H NMR spectrum of the resulting acetate showed signals of CH 2 D (triplet) and CH 3 at a ratio from which ϳ45% dilution of the biotransformation product with de novo synthesized acetate could be calculated (neglecting satellite peaks due to 13  complete retention of the migrating deuterium atom (within the limits of the experimental error) was indicated by the 13 CH 3 /CH 3 peak area ratio approximating the value of 13 C natural abundance (ϳ1.1%). In accordance with this assumption, the peak intensities measured in the 13 C NMR spectrum appeared in the expected proportions, i.e. ϳ1:16:3 for the 13 CH 3 group (singlet at ␦ 26.27, acetate coming from the acetogenic activity of the microorganism), for the 13 CH 2 D group (triplet upfield shifted, acetate coming from the 2-phenoxyethanol supplied), and for the 13 CO 2 Ϫ (singlet at ␦ 184.39, corresponding to the 13 C natural abundance level of the whole acetate recovered from the fermentation experiment).
Only the peak due to the [ 13 C]carboxylate group was detectable in the 13 C NMR spectrum of the acetate arising from the bioconversion of 9. The 1 H NMR spectrum displayed singlet at ␦ 2.031 and a 1:1:1:1:1:1 system centered at ␦ 2.017, really a doublet ( 2 J HC ϭ 5.9 Hz) (31) of triplet ( 2 J HD ) in agreement with the presence of two species only, i.e. CH 3 -CO 2 Ϫ and CH 2 D-13 CO 2 Ϫ , in the ratio of ϳ1:2. The absence of the isotopomer CH 2 D-CO 2 Ϫ allows the exclusion of an exchange with the medium of the carbonyl group (at the level of acetyl-CoA) (Fig. 1). Such an exchange has been reported to occur by the action of carbon monoxide dehydrogenase (32), an enzyme that is present in our strain of Acetobacterium (12).
Assuming the participation of the enzyme/coenzyme system as a hydrogen carrier in the hydrogen 1,2-shift during the biodegradation of 2-phenoxyethanol, two possibilities could be envisaged: (i) the hydrogen (deuterium) atom is abstracted from a substrate molecule, temporarily retained by the enzyme,   and then transferred to another molecule (intermolecular transfer) or (ii) the migrating hydrogen (deuterium) is returned to the same glycolic unit from which it had been abstracted (enzyme-mediated intramolecular transfer). To estimate the relative extent of the two events, compound 10 was administered to a cell suspension of Acetobacterium after dilution (18 to 100) with unlabeled 2-phenoxyethanol. When the acetate isolated at the end of this fermentation was examined by 1 H NMR (Fig. 5A), no signals assignable to the CH 2 D group were ob-served, i.e. no signals of a triplet 13.5 ppb upfield shifted from the singlet due to the CH 3 group (Fig. 3A). This result was consistent with a complete intramolecularity of the C-1 hydrogen migration. In addition, well resolved systems of satellite peaks were present in the proton NMR spectrum (Fig. 5A) due to isotopomers containing 13  Signal assignments are as follows: a, 13  expected deuterium isotope shift), was indicative of a strong prevalence of 13 CH 2 D-13 CO 2 Ϫ species among the [ 13 C]methyl isotopomeric mixture (J HC ϭ 126.7 Hz) (31). The presence of a very minor concentration of 13 CH 3 CO 2 Ϫ isotopomer was recognizable by the slightly higher intensity of the downfield peak of each sextet and could be explained in terms of natural 13 C abundance in the acetate molecule accompanying the doubly 13 C-labeled ones.
These findings were corroborated by the following considerations: (i) by inspection of methyl and carboxyl regions of the 13 C NMR spectrum (Fig. 5, B and C) the composition of the 13 C isotopomeric mixture was estimated to be: 13  Ϫ (e) Յ1% (d at ␦ 26.29 and d at ␦ 184.40, J CC ); (ii) the ratio of acetate resulting from acetogenic activity to that produced by transformation of 2-phenoxyethanol was found to be 1:2.6. This value was calculated from the ratio between CH 3 -CO 2 Ϫ molecules (measured as the area of the singlet at ␦ H 1.90) and 13 CH 2 D-13 CO 2 molecules (measured as the total area of the satellite signals, decreased by 1.1% of the CH 3 area and then corrected for the number of hydrogen atoms in the monodeuterated methyl group), taking into account the concentration (18%) of the labeled substrate in the sample fermented; (iii) the percentage of isotopomers b and c in the 13 C isotopomeric mixture was found to be very close to the expected one (Ϯ5%) for 13 C-labeled species present at the natural abundance level in the portion of acetate (87% of the total) arising in part (28%) from de novo synthesis and in part (59%) from 2-phenoxyethanol used to dilute the doubly labeled substrate. As regards the isotopomers d and e, which are present in trace amount in the acetate examined, their formation might depend on hydrogen and CO exchange reactions (32) occurring to a very small extent. Thus, the conversion of 2-phenoxyethanol into acetate appears to be an essentially straightforward process, as shown in Reaction 1, involving an intramolecular hydrogen migration in the first step. DISCUSSION In the light of a previous report (12) and in light of the results obtained by feeding experiments performed using 1 H-and 13 Clabeled substrates and resting cell suspensions of Acetobacterium strain LuPhet1, the conversion of 2-phenoxyethanol into acetate and phenol can be summarized as follows. Acetate originates from the glycolic moiety of 2-phenoxyethanol through elimination of phenol with formation of acetaldehyde, which is then oxidized in subsequent steps with retention of its molecular integrity (Fig. 1). In the first reaction, the alcoholic function of the substrate becomes a formyl group, whereas the adjacent methylene group is transformed into a methyl group with concomitant 1,2-hydrogen shift (Reaction 1). These features are strongly reminiscent of the dioldehydratase-catalyzed reactions for which a generally accepted mechanism is schematized in Fig. 6A for 1,2-ethanediol (11, R ϭ H) (18,33). The whole process encompasses a double H/OH interchange giving rise to the gem-diol (14, R ϭ H) (15,34,35) that rapidly collapses to the aldehyde 15. Its radical nature has largely been proven (33,36,37) and appears to be consistent with the transfer of a hydrogen atom from the C-1 of the substrate to a transient radical (X ⅐ ) and then back to C-2 of the productrelated radical (13, R ϭ H), generated in turn by a hydroxyl 1,2-shift.
If an analogous rearrangement occurs in the anaerobic degradation of 2-phenoxyethanol (11, R ϭ C 6 H 5 ) with formation of the labile hemiacetal (14, R ϭ C 6 H 5 ), the migration of the phenoxyl group should be assumed given the metabolic corre-lation between each carbon atom of the glycolic unit of 2-phenoxyethanol and those of the acetate molecule (Reaction 1). Thus, the opposite pathway, i.e. the 1,2-hydroxyl shift suggested previously (11,12), has to be ruled out.
Considering that no evidence has been given so far for the formation of the hemiacetal (14, R ϭ C 6 H 5 ), an alternative mechanism can be envisaged with regard to the subsequent transformation of the radical intermediate (12, R ϭ C 6 H 5 ) (Fig.  6A). This mechanism (Fig. 6B), based on the intermediacy of the resonance stabilized (␣-carbonyl 7 enoxy) radical 18 (33), is supported by the propensity of ketyls (radical anions) (e.g. 17) to eliminate adjacent leaving groups as a result of their electron-rich character (38,39). The cleavage of the ␤-C,O-bond can also be facilitated by stereoelectronic effects in the appropriate conformation of the radical anion 17 (40). It is well known that ␣-hydroxy radicals are up to 10 5 times more acidic than the corresponding alcohols (CH 2 OH-⅐ CHOH has pK a values of ϳ10 -12) (41). In addition, a base-promoted hydrogen abstraction as schematized in formula 16 is coherent with the marked lowering of gas-phase C-H bond dissociation energy observed when going from 1-alkanols (e.g. 94 Ϯ 2 kcal mol Ϫ1 for H-CH 2 OH) (42) to alcoholate ions (e.g. 85 kcal mol Ϫ1 for H-CH 2 O Ϫ ) (43). ␣-Oxo radicals have been proposed as intermediates in a number of enzymatic reactions (36,38,39,43,44).
We have found that in the biotransformation of 2-phenoxyethanol, the exchange of the migrating hydrogen atom with the medium occurs only to a negligible extent, if at all, and that its 1,2-shift is intramolecular (even if enzyme-mediated).
FIG. 6. Hypothetical reaction mechanisms for anaerobic glycol ether cleavage. A, commonly accepted reaction mechanism of diol dehydratases (R ϭ H); X ⅐ denoting 5Ј-deoxyadenosyl radical or a protein-based radical. B, putative mechanism of the enzyme-catalyzed C-O cleavage of phenoxyethanol by Acetobacterium sp.; X ⅐ denotes a proteinbased radical. C, alternative pathway for the conversion of ␣-oxo radical 18 into acetaldehyde; NuH, nucleophile (e.g. H 2 O).
The fact that the hydrogen atom abstracted from the C-1 position of the substrate is returned quantitatively to the adjacent position of the same molecule requires that the hydrogen carrier be monoprotic (XH in Fig. 6B). It can be noted that such a facet of the phenoxyethanol acetaldehyde lyase recalls the reaction mechanism of adenosylcobalamin-dependent ribonucleotide reductase of Lactobacillus leichmannii, which involves a protein-based cysteinyl radical as a catalytically competent intermediate (43). Although the ␣-oxo radicals appear to be thermodynamically capable of hydrogen abstraction from a thiol group (XH ϭ Enz-SH in Fig. 6B), given that gas-phase bond dissociation energies of H-SR compounds are in the range 88 -92 kcal mol Ϫ1 (43) and bond dissociation energy of H-CH 2 COCH 3 was estimated at ϳ91 kcal mol Ϫ1 (45), a temporary addition of a nucleophile to the carbonyl group of the radical 18 might occur (Fig. 6C). This further step would remove the resonance stabilization in 18 (ϳ8 kcal mol Ϫ1 ) (see Table IV, footnote k in Ref. 45), thus facilitating the formation of the C-H bond by the intermediate 19 to give the labile diol 20 (bond dissociation energy for H-CH 2 R ϳ98 kcal mol Ϫ1 ) (42). A similar addition (with NuH ϭ H 2 O) has been suggested in the case of the diol dehydratase reaction mechanism (18,36,37,39). It remains to be elucidated whether this reaction mechanism also underlies anaerobic cleavage of PEG and its derivatives.