Introduction
The presence of halogen atoms is a feature of many natural products derived from microorganisms and is thought to reflect the concentrations of halide ions present in the habitat of the producing organisms. Although less common, halogenated compounds are also present in higher organisms and include the human hormone thyroxine and the analgesic alkaloid epibatidine, produced by rainforest tree frogs. In the past decade or so, different classes of enzymes catalyzing the formation of carbon-halogen bonds have been identified; of particular chemical interest are those that catalyze the halogenation of unactivated hydrocarbons such as methyl groups. Such reactions are catalyzed by enzymes that use iron(II) as a cofactor and 2-oxoglutarate (2OG) and oxygen as co-substrates (2OG oxygenases). In this issue, Galonic´ et al.1 describe studies on the chlorinase CytC3 that reveal an iron(IV) oxo intermediate. This intermediate is closely related to those observed in 2OG oxygenases catalyzing their archetypal hydroxylation reactions2, 3, 4, demonstrating that it can enable very varied types of oxidative chemistry.
The first 2OG oxygenases identified were prolyl and lysyl hydroxylases involved in collagen biosynthesis, but this class of enzyme has subsequently been shown to catalyze oxidative reactions with a range of important functional roles in most organisms. These include transcriptional regulation, DNA repair and fatty acid metabolism. In metazoans, these biological roles have so far all been ones involving the catalysis of hydroxylation reactions. In plants and micoorganisms, however, the 2OG oxygenases catalyze a plethora of oxidative reactions, which has led to the proposal that they may be the most versatile of all oxidizing biological catalysts. Some of these reactions are chemically remarkable and indeed presently cannot be achieved through synthetic—that is, non-biological—chemistry5, 6. Oxidative reactions catalyzed by 2OG oxygenases include cyclizations, ring fragmentation, C-C bond cleavage, epimerization, desaturation and the hydroxylation of aromatic rings. The discovery that 2OG oxygenases can catalyze chlorination reactions further extends the scope of the family6.
Structural studies have revealed a common structural platform for the 2OG oxygenases, the double-stranded 
-helix or jelly-roll motif that supports iron-binding residues5. In almost all studies these residues comprise one aspartyl or glutamyl residue and two histidinyl residues that form a conserved triad (such as is found, for example, in factor inhibiting HIF (FIH); Fig. 1a). The crystal structure of the chlorinating enzyme SyrB2 revealed the anticipated jelly-roll motif; unusually, however, the carboxylate of the canonical iron-binding triad was replaced by an alanyl residue, thus sterically enabling a chloride ion to complex the iron in the place of the missing carboxylate (Fig. 1a)7. The SyrB2 structures led to a proposed mechanism for chlorination in which a key intermediate is a Cl-Fe(IV)-oxo species generated by the oxidative decarboxylation of an iron–2OG complex, such as is seen with the other 2OG oxygenases. This complex can then effect hydrogen-atom abstraction from the substrate to yield a Cl–Fe(III)–OH complex and a carbon radical. Substrate chlorination can then proceed through 'rebound' of a chloride radical, analogous to the hydroxyl radical rebound postulated for both heme and non-heme hydroxylases (Fig. 1b). Precedent for the reaction of a non-hydroxyl iron ligand with a substrate comes from work on the oxidase isopenicillin N-synthase, which is closely related to the 2OG oxygenases but does not use a 2OG substrate8.
Figure 1: The structural and mechanistic similarities between Fe- and 2OG-dependent hydroxylases and Fe- and 2OG-dependent halogenases.
(a) Comparison of the iron-binding sites for (left) FIH, a 2OG-dependent hydroxylase, and (right) SyrB2, a 2OG-dependent chlorinase. (b) Outline catalytic cycles for Fe(II)- and 2OG-dependent oxygenase–catalyzed hydroxylation (left) and chlorination (right, as for CytC3) reactions. The common ferryl-oxo intermediates can also catalyze other two-electron oxidations, such as desaturations and oxidative cyclizations.
Full size image (82 KB)CytC3, a halogenase isolated from the soil bacterium Streptomyces, chlorinates the methyl group of the amino acids L-2-aminobutyric acid or L-valine in non-ribosomal peptide biosynthesis. With this enzyme, Galonic´ et al. have used stopped-flow absorption and freeze-quench Mossbauer techniques to provide kinetic and spectroscopic evidence for an Fe(II)–2OG complex (with absorption at 520 nm) and for two interconverting Cl-Fe(IV)-oxo complexes (with absorption at 318 nm). These are hypothesized to be two distinct states in rapid equilibrium, one of which is responsible for cleaving the substrate C-H bond, allowing subsequent chlorination. The work supports the proposal of a broad common mechanism for this family of enzymes.
The Fe(IV) intermediates were observed to decay slowly in the presence of deuterated substrate, suggesting that the active site can suppress 'undesirable' side reactions of the highly reactive Fe(IV)-oxo intermediates. This is consistent with the concept of negative catalysis applying to 2OG oxygenases. Many enzymes that use highly reactive intermediates have evolved to suppress side reactions in order both to enable formation of a single product and to avoid inactivation9. For the 2OG oxygenases, the energy-rich Fe(IV)-oxo intermediate can potentially catalyze many different types of two-electron oxidation, including irreversible oxidation of the active site itself, resulting in inactivation. The negative catalysis concept proposes that the enzymes (ideally) suppress all but the 'desired' reaction—that is, chlorination in the case of SyrB2. Sometimes, however, the process is imperfect and damage can occur that results in hydroxylation or fragmentation of the enzyme. Remarkably, some 2OG oxygenases can catalyze more than one type of oxidative reaction, for example hydroxylation, oxidative cyclization and desaturation, at a single active site.
The spectroscopic and kinetic characterization of intermediates, combined with structural analyses, has opened up the possibility not only of further understanding the mechanisms of the different types of oxidation reactions catalyzed by 2OG oxygenases, but also of manipulating them to catalyze useful unnatural reactions as a means to produce valuable chemicals such as pharmaceuticals. Genomic and other studies have revealed many likely 2OG oxygenases of unknown function. Although many of these candidate oxygenases may well be hydroxylases, the available evidence suggests that a significant proportion may catalyze other oxidative reactions, possibly including as-yet-uncharted chemistry. It is even possible that some of the chemically remarkable oxidation reactions catalyzed by 2OG oxygenases involved in microbial secondary metabolism actually also occur in human metabolism.
