The use of 3-[2-(N,N-diethyl-N-methylammonium)ethyl]-7-methoxy-4-methylcoumarin (AMMC) as a specific CYP2D6 probe in human liver microsomes.

Recently, a novel nonfluorescent probe 3-[2-(N,N-diethyl-N-methylammonium)-ethyl]-7-methoxy-4-methylcoumarin (AMMC), which produces a fluorescent metabolite AMHC (3-[2-(N,N-diethyl-N-methylammonium)ethyl]-7-hydroxy-4-methylcoumarin) was used with microsomes containing recombinant enzymes (rCYP) to monitor CYP2D6 inhibition in a microtiter plate assay. This article describes the studies that were performed in human liver microsomes (HLM) to establish the selectivity of AMMC toward CYP2D6. Metabolism studies in HLM showed that AMMC was converted to one metabolite identified by mass spectrometry as AMHC. Kinetic studies indicated an apparent K(m) of 3 microM with a V(max) of 20 pmol/min. mg of protein for the O-demethylation reaction. The O-demethylation of AMMC in HLM was inhibited significantly in the presence of a CYP2D6 inhibitory antibody. Using a panel of various HLM preparations (n = 12), a good correlation (r(2) = 0.95) was obtained between AMMC O-demethylation and bufuralol metabolism, a known CYP2D6 substrate, but not with probes for the other major xenobiotic metabolizing CYPs. Finally, only rCYP2D6 showed detectable metabolism in experiments conducted with rCYPs using AMMC at a concentration of 1.5 microM (near K(m)). However, at a concentration of 25 microM AMMC, rCYP1A also contributed significantly to the formation of AMHC. Knowing the experimental conditions under which AMMC was selective for CYP2D6, a microtiter assay was developed to study the inhibition of various compounds in HLM using the fluorescence of AMHC as an indication of CYP2D6 activity. The inhibition potential of various chemicals was found to be comparable to those determined using the standard CYP2D6 probe, bufuralol, which requires high-performance liquid chromatography separation for the analysis of its CYP2D6-mediated 1'-hydoxylated metabolite.

Fluorometric high-throughput screening for inhibitors of cytochrome P450.

Rapid screening for cytochrome P450 inhibitors is part of the current paradigm for avoiding development of drugs likely to give clinical pharmacokinetic drug-drug interactions and associated toxicities. We have developed microtiter plate-based, direct, fluorometric assays for the activities of the principal human drug-metabolizing enzymes, CYP1A2, CYP2C8, CYP2C9, CYP2C19, CYP2D6, and CYP3A4, as well as for CYP2A6, which is an important enzyme in environmental toxicology. These assays are rapid and compatible with existing high-throughput assay instrumentation. For CYP1A2, CYP2C8, CYP2C9, CYP2C19, and CYP2D6, the potency of enzyme inhibition (IC50) is consistent regardless of the probe substrate or assay method employed. In contrast, CYP3A4 inhibition for an individual inhibitor shows significant differences in potency (>300-fold) depending on the probe substrate being used. We have investigated these differences through the use of several structurally distinct fluorescent substrates for CYP3A4 and several classical substrate probes (e.g., testosterone, nifedipine, and midazolam), with a panel of known, clinically significant, CYP3A4 inhibitors. The use of multiple probe substrates appears to be needed to characterize the inhibition potential of xenobiotics for CYP3A4.

Substrate-dependent modulation of CYP3A4 catalytic activity: analysis of 27 test compounds with four fluorometric substrates.

Inhibition of cytochrome P450 catalytic activity is a principal mechanism for pharmacokinetic drug-drug interactions. Rapid, in vitro testing for cytochrome P450 inhibition potential is part of the current paradigm for identifying drug candidates likely to give such interactions. We have explored the extent that qualitative and quantitative inhibition parameters are dependent on the cytochrome P450 (CYP) 3A4 probe substrate. Inhibition potential (e.g., IC(50) values from 8-point inhibition curves) or activation potential for most compounds varied dramatically depending on the fluorometric probe substrates for CYP3A4 [benzyloxyresorufin (BzRes), 7-benzyloxy-4-trifluoromethylcoumarin (BFC), 7-benzyloxyquinoline (BQ), and dibenzylfluorescein (DBF)]. For 21 compounds that were primarily inhibitors, the range of IC(50) values for the four substrates varied from 2.1- to 195-fold with an average of 29-fold. While the rank order of sensitivity among the fluorometric substrates varied among the individual inhibitors, on average, BFC dealkylation was the most sensitive to inhibition, while BQ dealkylation was least sensitive. Partial inhibition was observed with BzRes and BQ but not for BFC and DBF. BzRes was more prone to activation, whereas dramatic changes in IC(50) values were observed when the BQ concentration was below the S(50). Three different correlation analyses indicated that IC(50) values with BFC, BQ, and DBF correlated well with each other, whereas the response with BzRes correlated more weakly with the other substrates. One of these correlation analyses was extended to the percent inhibition of 10 microM inhibitor with the standard CYP3A4 probe substrates testosterone, midazolam, and nifedipine. In this analysis the responses with BQ, BFC and DBF correlated well with testosterone and midazolam but more poorly with nifedipine. In the aggregate, BFC and DBF appear more suitable as an initial screen for CYP3A4 inhibition. However, the substrate-dependent effects reported here and by others indicate that all CYP3A4 inhibition data should be interpreted with caution.

CYP3A4 allelic variants with amino acid substitutions in exons 7 and 12: evidence for an allelic variant with altered catalytic activity.

OBJECTIVE: To determine the existence of mutant and variant CgammaP3A4 alleles in three racial groups and to assess functions of the variant alleles by complementary deoxyribonucleic acid (cDNA) expression. METHODS: A bacterial artificial chromosome that contains the complete CgammaP3A4 gene was isolated and the exons and surrounding introns were directly sequenced to develop primers to polymerase chain reaction (PCR) amplify and sequence the gene from lymphocyte DNA. DNA samples from Chinese, black, and white subjects were screened. Mutating the affected amino acid in the wild-type cDNA and expressing the variant enzyme with use of the baculovirus system was used to functionally evaluate the variant allele having a missense mutation. RESULTS: To investigate the existence of mutant and variant CgammaP3A4 alleles in humans, all 13 exons and the 5'-flanking region of the human CgammaP3A4 gene in three racial groups were sequenced and four alleles were identified. An A-->G point mutation in the 5'-flanking region of the human CgammaP3A4 gene, designated CgammaP3A4*1B, was found in the three different racial groups. The frequency of this allele in a white population was 4.2%, whereas it was 66.7% in black subjects. The CgammaP3A4*1B allele was not found in Chinese subjects. A second variant allele, designated CgammaP3A4*2, having a Ser222Pro change, was found at a frequency of 2.7% in the white population and was absent in the black subjects and Chinese subjects analyzed. Baculovirus-directed cDNA expression revealed that the CYP3A4*2 P450 had a lower intrinsic clearance for the CYP3A4 substrate nifedipine compared with the wild-type enzyme but was not significantly different from the wild-type enzyme for testosterone 6beta-hydroxylation. Another rare allele, designated CgammaP3A4*3, was found in a single Chinese subject who had a Met445Thr change in the conserved heme-binding region of the P450. CONCLUSIONS: These are the first examples of potential function polymorphisms resulting from missense mutations in the CgammaP3A4 gene. The CgammaP3A4*2 allele was found to encode a P450 with substrate-dependent altered kinetics compared with the wild-type P450.

The use of heterologously expressed drug metabolizing enzymes--state of the art and prospects for the future.

The first report of the functional, heterologous expression of a mammalian cytochrome P450 (CYP) enzyme occurred more than a decade ago. In the intervening years, these expression systems have been optimized with regard to the specific requirements for production of catalytically active enzymes. In this review, we discuss the strengths and limitations of heterologously expressed enzymes as they affect in vitro drug metabolism studies. Emphasis is given to new applications (screens for CYP inhibition and novel enzyme mixtures) that have been enabled by high level, functional expression of CYP enzymes.

Smoking-attributable medical care costs in the USA.

Medical care costs attributable to cigarette smoking are estimated using an econometric model of annual individual expenditures for four types of medical services: ambulatory, hospital, prescription drug, and other (which includes home health and durable medical equipment and excludes dental and mental health). The model follows the two-part specification of Duan et al. (1983). Estimation is carried out using the 1987 National Medical Expenditure Survey. Fitted values are used to calculate smoking-attributable fractions (SAFs) of expense by type of service and by age and gender category. The overall weighted average SAF is 6.54%. SAFs are generally largest for ambulatory and smallest for hospital expenses. They are larger for males and for the older age categories. The model is analyzed for heteroscedasticity and goodness of fit. Additional analysis using the National Health Interview Survey is conducted to test for the possible effect of not being able to include alcohol consumption in the primary model. A balanced repeated replication analysis is conducted to evaluate the variance of the SAFs. Variances are found to be acceptably small. An extension of the model to support evaluation of smoking-attributable costs for special populations such as individual states, and special insurance pools such as Medicaid recipients, is described. Results for the fifty states are presented. Conclusions and subjects for further research are discussed.

Mechanism-based inactivation of human cytochrome P450 3A4 by grapefruit juice and red wine.

Grapefruit juice is well documented to cause clinically significant increases in the plasma concentrations of many therapeutic agents. These interactions are believed to be mediated via inhibition of intestinal cytochrome P-450 3A4 (CYP3A4) by flavonoids and/or other chemicals in grapefruit juice, although the mechanism of that inhibition has not been fully characterized. Like grapefruit juice, red wine contains large amounts of flavonoids and other xenobiotics which could also mediate CYP3A4 inhibition. In this study, we investigated the mechanism of inhibition of CYP3A4 by grapefruit juice and also examined the ability of red wine to inhibit this enzyme. Both red wine and grapefruit juice potently inhibited CYP3A4 activity in a concentration-dependent manner. At 8% of natural strength, enzyme activity was inhibited almost 90 and 84%, respectively, by grapefruit juice and red wine. In contrast, white wine did not appreciably inhibit CYP3A4 activity. Grapefruit juice irreversibly inactivated CYP3A4 in a time- and NADPH-dependent manner. The rate of inactivation mediated by grapefruit juice was similar to that mediated by troleandomycin, a potent mechanism-based inhibitor of CYP3A4. Red wine also inactivated CYP3A4 but at a rate approximately 16% that of grapefruit juice. Inhibition of CYP3A4 by red wine is primarily reversible in nature. The clinical implications of this research are discussed.

The R144C change in the CYP2C9*2 allele alters interaction of the cytochrome P450 with NADPH:cytochrome P450 oxidoreductase.

We have examined the kinetics of substrate metabolism by cDNA-expressed human CYP2C9 and the R144C variant. Both enzymes exhibited similar apparent K(m) values for (S)-warfarin 7-hydroxylation, diclofenac 4'-hydroxylation and lauric acid 11-hydroxylation. In contrast, the R144C variant (relative to CYP2C9) had slower rates of metabolism for all three substrates. The difference was most pronounced for (S)-warfarin. Surprisingly, the magnitude of the difference was found to be dependent on the cytochrome P450 to NADPH-cytochrome P450 reductase (OR) ratio in the system (the difference being more pronounced at higher OR to P450 ratios) implying that the R144C change affects interaction of the P450 with OR. The rates of (S)-warfarin 7-hydroxylation by CYP2C9 and the R144C variant also exhibited differential dependence on salt concentration which further supported a difference in interaction with OR. When OR was bypassed and the hydroxylation was supported by cumene hydroperoxide, no difference in the rates of diclofenac 4'-hydroxylation was observed for CYP2C9 and the R144C variant regardless of OR to P450 ratio. However, for (S)-warfarin 7-hydroxylation, some OR-dependence was maintained even when the reaction was supported by cumene hydroperoxide. Finally, we compared CYP2C9 activity and CYP2C9 protein levels for human lymphoblast expressed (high OR to P450 ratio) to human liver microsomes using immunoblotting and enzyme selective substrates. Human liver microsomal CYP2C9 and human lymphoblast-expressed CYP2C9 showed comparable amounts of activity per unit enzyme. This final observation indicates that the high OR to P450 ratio is the preferred model and predicts that the R144C change in human liver microsomal CYP2C9 should markedly reduce the rates of substrate metabolism. The implications of these observations for the interpretation of results with cDNA-expressed enzymes is discussed.

Horseradish peroxidase Phe172-->Tyr mutant. Sequential formation of compound I with a porphyrin radical cation and a protein radical.

A gene coding for the F172Y mutant of horseradish peroxidase isozyme C (HRP) has been constructed and expressed in both Spodoptera frugiperda (SF-9) and Trichoplusia ni egg cell homogenate (HighFive) cells. Homology modeling with respect to three peroxidases for which crystal structures are available places Phe172 on the proximal side of the heme in the vicinity of porphyrin pyrrole ring C. The pH optimum and spectroscopic properties of the F172Y mutant are essentially identical to those of wild type HRP. Vmax values show that the mutant protein retains most of the guaiacol oxidizing activity. Stopped flow studies indicate that Compound I is formed with H2O2 at the same rate (kappa 1 = 1.6 x 10(7) M-1 s-1) at both pH 6.0 and 8.0 as it is with the wild type enzyme. This Compound I species decays rapidly at a rate kappa 2 = 1.01 s-1, pH 7.0, to a second two-electron oxidized species that retains the ferryl (FeIV = O) absorption. EPR studies establish that a ferryl porphyrin radical cation is present in the initial Compound I, but electron transfer from the protein results in formation of a second Compound I species with an unpaired electron on the protein (presumably on Tyr172). The presence or absence of oxidizable amino acids adjacent to the heme is thus a key determinant of whether the second oxidation equivalent in Compound I is found as a porphyrin or protein radical cation.

Chloroperoxidase-catalyzed benzylic hydroxylation.

Chloroperoxidase oxidizes p-methylanisole and p-ethylanisole to 4-methoxybenzyl alcohol and 1-(4'-methoxyphenyl)ethanol, respectively. It ineffectively oxidizes toluene to benzyl alcohol but does not appear to oxidize toluene substituted with strong electron-withdrawing groups. O-Demethylation is also observed. The enzyme is sensitive to substituents at other than the para position and does not detectably catalyze benzylic hydroxylation of p-methylanisole if it bears additional methyl or methoxy groups. An exception is 1,2-(methylenedioxy)-4-methylbenzene, which is oxidized to both 3,4-(methylenedioxy)benzyl alcohol and 2-hydroxy-4-methylphenol. Studies with H2(18)O2 indicate that all the oxygen incorporated into the product in the oxidation of p-methylanisole to 4-methoxybenzyl alcohol derives from the peroxide. The mono- and dideuterated methyl analogues of p-methylanisole are oxidized with apparent intramolecular isotope effects of 3.51 and 3.34, respectively. Abstraction of a hydrogen from a carbon bearing a hydroxyl group competes effectively with benzylic oxidation because 2-[1,1-2H2]phenylethanol is oxidized to 2-[1-2H]- rather than 2-[1,2-2H2]phenylacetaldehyde. Aldehyde formation therefore involves abstraction of the carbinol hydrogen rather than hydrogen migration to a benzylic carbocation intermediate. Chloroperoxidase resembles cytochrome P450 in that it catalyzes benzylic hydroxylation reactions but it has a more limited substrate specificity.

CDP-6-deoxy-delta 3,4-glucoseen reductase from Yersinia pseudotuberculosis: enzyme purification and characterization of the cloned gene.

The 3,6-dideoxyhexoses, usually confined to the cell wall lipopolysaccharide of gram-negative bacteria, are essential to serological specificity and are formed via a complex biosynthetic pathway beginning with CDP-D-hexoses. In particular, the biosynthesis of CDP-ascarylose, one of the naturally occurring 3,6-dideoxyhexoses, consists of five enzymatic steps, with CDP-6-deoxy-delta 3,4-glucoseen reductase (E3) participating as the key enzyme in this catalysis. This enzyme has been previously purified from Yersinia pseudotuberculosis by an unusual procedure (protocol I) including a trypsin digestion step (O. Han, V.P. Miller, and H.-W. Liu, J. Biol. Chem. 265:8033-8041, 1990). However, the cloned gene showed disparity with the expected gene characteristics, and upon expression, the resulting gene product exhibited no E3 activity. These findings strongly suggested that the protein isolated by protocol I may have been misidentified as E3. A reinvestigation of the purification protocol produced a new and improved procedure (protocol II) consisting of DEAE-Sephacel, phenyl-Sepharose, Cibacron blue A, and Sephadex G-100 chromatography, which efficiently yielded a new homogeneous enzyme composed of a single polypeptide with a molecular weight of 39,000. This highly purified protein had a specific activity nearly 8,000-fold higher than that of cell lysates, and more importantly, the corresponding gene (ascD) was found to be part of the ascarylose biosynthetic cluster. Presented are the identification and confirmation of the E3 gene through cloning and overexpression and the culminating purification and unambiguous assignment of homogeneous E3. The nucleotide and translated amino acid sequences of the genuine E3 are also presented.

Cofactor characterization and mechanistic studies of CDP-6-deoxy-delta 3,4-glucoseen reductase: exploration into a novel enzymatic C-O bond cleavage event.

The CDP-6-deoxy-delta 3,4-glucoseen reductase (E3) is a NADH-dependent enzyme which catalyzes the key reduction of the C-3 deoxygenation step during the formation of CDP-ascarylose, a 3,6-dideoxyhexose found in the lipopolysaccharide of Yersinia pseudotuberculosis. This highly purified enzyme is also a NADH oxidase capable of mediating the direct electron transfer from NADH to O2, forming H2O2. While previous work showed that E3 contains no common cofactor, one FAD and one plant ferredoxin type [2Fe-2S] center were found in this study to be associated with each molecule of E3. The iron-sulfur center is essential for E3 activity since bleaching of the [2Fe-2S] center leads to inactive enzyme. These results suggest that E3 employs a short electron-transport chain composed of both FAD and the iron-sulfur center to shuttle electrons from NADH to its acceptor. The order of electron flow, as indicated by EPR measurement with partially reduced E3, starts with hydride reduction of FAD by NADH. The iron-sulfur cluster, receiving electrons one at a time from the reduced flavin, relays the reducing equivalents via another iron-sulfur center in the active site of E1 to its final acceptor, the E1-bound PMP-glucoseen adduct. The participation of a one-electron-carrying iron-sulfur center in this reduction is advantageous since both electrons are dispatched from the same redox state of the prosthetic group, allowing electrons of equal energy to be delivered to the final acceptor.(ABSTRACT TRUNCATED AT 250 WORDS)

Cytochrome P450cam-catalyzed oxidation of a hypersensitive radical probe.

trans-1-Phenyl-2-vinylcyclopropane, a hypersensitive radical probe, is oxidized by cytochrome P450cam (CYP101) to a diastereomeric mixture of the corresponding epoxide (81%), (trans-2-phenylcyclopropyl)acetaldehyde (6%), and trans-5-phenyl-2-penten-1,5-diol (13%). trans-5-Phenyl-2-penten-1-ol and (trans-2-phenylcyclopropyl)ethane-1,2-diol are not detectably formed. Authentic standards of all the products have been synthesized and used to establish the identities (or the absence) of the metabolites. Studies with [18O]H2O demonstrate that the oxygens at positions 1 and 5 in the rearranged diol derive from molecular oxygen and water, respectively. Catalytic turnover of the enzyme is required for product formation from the olefin, but incubation of the epoxide metabolite with the enzyme, or with buffer alone, yields both the aldehyde and the rearranged diol products. The absence of trans-5-phenyl-2-penten-1-ol implies that the lifetime of the putative radical intermediate is so short that its existence as a discrete entity is questionable. A cationic intermediate is unlikely but cannot be excluded because the same metabolites are formed in a secondary reaction, even at pH 8.0, from the epoxide. The results provide no evidence for the involvement of radicals or cations in the epoxidation reaction, in agreement with results on the oxidation of olefins in organic solvents by metalloporphyrin catalysts.

Monooxygenase activity of cytochrome c peroxidase.

Recombinant cytochrome c peroxidase (CcP) and a W51A mutant of CcP, in contrast to other classical peroxidases, react with phenylhydrazine to give sigma-bonded phenyl-iron complexes. The conclusion that the heme iron is accessible to substrates is supported by the observation that CcP and W51A CcP oxidize thioanisole to the racemic sulfoxide with quantitative incorporation of oxygen from H2O2. Definitive evidence for an open active site is provided by stereoselective epoxidation by both enzymes of styrene, cis-beta-methylstyrene, and trans-beta-methylstyrene. trans-beta-methylstyrene yields exclusively the trans-epoxide, but styrene yields the epoxide and phenylacetaldehyde, and cis-beta-methylstyrene yields both the cis- and trans-epoxides and 1-phenyl-2-propanone. The sulfoxide, stereoretentive epoxides, and 1-phenyl-2-propanone are formed by ferryl oxygen transfer mechanisms because their oxygen atom derives from H2O2. In contrast, the oxygen in the trans-epoxide from the cis-olefin derives primarily from molecular oxygen and is probably introduced by a protein cooxidation mechanism. cis-[1,2-2H]-1-Phenyl-1-propene is oxidized to [1,1-2H]-1-phenyl-2-propanone without a detectable isotope effect on the epoxide:ketone product ratio. The phenyl-iron complex is not formed and substrate oxidation is not observed when the prosthetic group is replaced by delta-meso-ethylheme. CcP thus has a sufficiently open active site to form a phenyl-iron complex, to oxidize thioanisole to the sulfoxide, and to epoxidize styrene and beta-methylstyrene. The results indicate that a ferryl (Fe(IV) = O)/protein radical pair can be coupled to achieve two-electron oxidations. The unique ability of CcP to catalyze monooxygenation reactions does not conflict with its peroxidase function because cytochrome c is oxidized at a distinct surface site (DePillis, G. D., Sishta, B. P., Mauk, A. G., and Ortiz de Montellano, P. R. (1991) J. Biol. Chem. 266, 19334-19341).

Mechanistic and stereochemical studies of a unique dehydration catalyzed by CDP-4-keto-6-deoxy-D-glucose-3-dehydrase: a pyridoxamine 5'-phosphate dependent enzyme isolated from Yersinia pseudotuberculosis.

CDP-4-keto-6-deoxy-D-glucose-3-dehydrase (E1) purified from Yersinia pseudotuberculosis is a pyridoxamine 5'-phosphate (PMP) dependent enzyme which catalyzes the C-O bond cleavage at C-3 of a CDP-4-keto-6-deoxy-D-glucose substrate, a key step in the formation of 3,6-dideoxyhexoses. Since enzyme E1 utilizes the PMP cofactor in a unique manner, it is essential to establish its role in E1 catalysis. When an incubation was conducted in [18O]H2O, incorporation of 18O into positions C-3 and C-4 of the recovered substrate was observed. This result not only provided the evidence necessary to reveal the reversibility of E1 catalysis but also lent credence to the formation of a delta 3,4-glucoseen intermediate. In view of E1 catalysis being initiated by a C-4' deprotonation of the PMP-substrate complex, the stereochemical course of this step was examined using chemically synthesized (4'S)- and (4'R)-[4'-3H]PMP as probes. Our results clearly demonstrated that the stereochemistry of this deprotonation is pro-S specific, which is in agreement with the stereochemical consistency found with other vitamin B6 phosphate dependent enzymes. The fact that reprotonation at C-4' of the PMP-delta 3,4-glucoseen complex in the reverse direction of E1 catalysis was also found to be pro-S stereospecific strongly suggested that enzyme E1, like most of its counterparts, has the si face of its cofactor-substrate complex exposed to solvent and accessible to active-site catalytic groups as well.(ABSTRACT TRUNCATED AT 250 WORDS)

Mechanistic studies of the biosynthesis of 3,6-dideoxyhexoses in Yersinia pseudotuberculosis. Purification and characterization of CDP-6-deoxy-delta 3,4-glucoseen reductase based on its NADH:dichlorophenolindolphenol oxidoreductase activity.

CDP-6-deoxy-delta 3,4-glucoseen reductase, the key enzyme catalyzing the biosynthetic formation of CDP-ascarylose (CDP-3,6-dideoxy-L-arabino-hexose), was purified from Yersinia pseudotuberculosis by monitoring its NADH:dichlorophenolindolphenol oxidoreductase activity. A protocol consisting of DEAE-cellulose, phenyl-Sepharose, and Sephadex G-100 column chromatography yielded a mixture of two proteins. The low molecular weight protein contaminant was removed by limited tryptic digestion leaving a purified enzyme consisting of a single polypeptide with a molecular weight of 41,000. A weak, featureless uv spectrum above 300 nm suggested no common chromophoric cofactor contributes to enzyme activity and no protein-associated metals were detected. The stereospecificity of nicotinamide oxidation was determined to be pro-R stereospecific. Reduction of ferricyanide during NADH oxidation and confirmation of the intermediacy of O2- in the reaction flux suggested that enzyme-catalyzed H2O2 formation is not a direct two-electron reduction of molecular oxygen, but is rather the consequence of an enzymatic 2e-/1e- switch. The sugar deoxygenation reaction may therefore proceed through a radical mechanism.