The β-oxidation pathway for the catabolism of fatty acids is encoded by most genomes. A key step in the pathway is the hydration of enoyl CoA esters, which is catalyzed by enoyl CoA hydratase (also called crotonase). The reaction catalyzed by crotonase has been subjected to mechanistic scrutiny [20
]; several high-resolution structures are available for the rat mitochondrial enzyme [23
]. As a result of these studies, the mechanism is known to involve two glutamate residues that function as acid/base catalysts: in the direction of hydration of enoyl CoAs, Glu144 activates a water molecule for nucleophilic attack and Glu164 delivers a proton to the α carbon. Although never observed directly, the mechanism probably involves the transient formation of a thioester enolase anion intermediate that is stabilized by enhanced hydrogen-bonding interactions with an 'oxyanion hole' [25
]. As judged by the structures of complexes with substrate analogs/products, the 'oxyanion hole' is formed by two peptidic NH groups that bind the thioester carbonyl group.
Database searches with the sequence of the structurally characterized crotonase yield many 'hits', as judged by conserved consensus sequences for both 'halves' of the oxyanion hole [26
]. The reactions catalyzed by some of these hits have been biochemically characterized, leading to the discovery that crotonase is a member of a mechanistically diverse superfamily: the reactions can be 'explained' by a mechanism that involves the transient formation and stabilization of a thioester enolate anion intermediate.
In addition to crotonase, structurally characterized members of the superfamily include (Figure ): 4-chlorobenzoyl CoA dehalogenase, which catalyzes a nucleophilic aromatic substitution [28
-dienoyl CoA isomerase, which catalyzes a 1,5-proton transfer reaction [29
]; and methylmalonyl CoA decarboxylase, which catalyzes a decarboxylation reaction [30
]. In the case of the dehalogenase, homologs of Glu144 and Glu164 in crotonase are missing, but the functional groups involved have been identified and subjected to mechanistic scrutiny assisted by site-directed mutagenesis. In the case of the isomerase and decarboxylase, the sequences of both contain a homolog for either Glu144 or Glu164 but not both; the mechanisms of the reactions are not yet certain.
Different reactions catalyzed by members of a mechanistically diverse superfamily: reactions catalyzed by members of the crotonase superfamily.
Biochemically characterized members of the superfamily for which no structure is yet available include (Figure ): 1,4-dihydroxynaphthoyl CoA synthase, which catalyzes formation of a carbon-carbon bond in a Dieckman reaction [31
]; 2-ketocyclohexyl CoA hydrolases, which catalyzes cleavage of a carbon-carbon bond in a retro Dieckman reaction [32
-enoyl CoA isomerase, which catalyzes a 1,3-proton transfer reaction [33
]; feruloyl CoA hydratase/lyase, which catalyzes successive hydration and carbon-carbon bond cleavage reactions [34
]; and carnitinyl CoA epimerase, which inverts the configuration of an unactivated carbon by an unknown mechanism [35
]. The hydratase/lyase and epimerase contain homologs of both Glu144 and Glu164 in crotonase; as a result, the mechanisms of both may involve hydration/dehydration partial reactions involving a thioester enolase anion intermediate. Mechanisms involving carbon-carbon bond cleavage reactions via retroaldol mechanisms are also possible, however. Irrespective of the precise reaction coordinates that describe the reactions catalyzed by these enzymes, all probably involve the formation and transient stabilization of a thioester enolate anion intermediate.
The database searches also yield members of the crotonase superfamily that catalyze reactions that probably do not involve formation of a thioester enolate anion but, instead, an anionic tetrahedral intermediate in thioester or peptide bond hydrolysis that would be stabilized by the conserved oxyanion hole (Figure ): 3-OH isobutyryl CoA hydrolase [36
] and the ClpP protease [37
]. The hydrolase occurs in the catabolic pathway for valine; the preceding enzyme in the pathway, also a member of the crotonase superfamily, catalyzes the hydration of α-methacryl CoA using homologs of Glu144 and Glu164 in rat mitochondrial crotonase. Although structure/function relationships have not yet been characterized in detail for the protease, the structure allows the suggestion that the functional groups are configured in a catalytic triad (Ser-His-Glu) structurally analogous to those found in other examples of analogous serine proteases [38
]. That the hydrolase and protease are members of the crotonase superfamily provides compelling evidence that the underlying chemical strategy supported by the α/β fold is stabilization of oxyanion intermediates.
We therefore conclude that structure/function relationships in the crotonase superfamily cannot be extrapolated simply to newly discovered members from the sequence and available structures. Database annotations indicate, however, that newly discovered members of this superfamily are 'probable enoyl CoA hydratases/isomerases'. Table compares the number of members of the crotonase superfamily encoded by several microbial genomes, as assessed by the presence of consensus sequences for the oxyanion hole, with the number of proteins that could catalyze an enoyl CoA hydratase reaction, as assessed by the presence of homologs of both Glu144 and Glu164 in rat mitochondrial crotonase. Despite the differences, the databases invariably include the members of the superfamily that do not contain homologs of Glu144 and Glu164 in rat mitochondrial crotonase as 'likely' or 'probable' 'enoyl CoA hydratases/isomerases'. We believe that, in many cases, these annotations are incorrect or misleading. Although we recognize this problem, given our expertise in structure/function relationships for several members of this superfamily, most users of the sequence databases will not. The likely consequences are that annotations of genome sequences completed in the future will use these incorrect annotations and continue to misrepresent the functions of orthologous ORFs that are identified, and that novel metabolic pathways involving members of the crotonase superfamily will be ignored or assigned incorrect functions. Both types of incorrect assignment will undermine the considerable potential of genome sequencing projects in defining new metabolic capabilities.
Members of the crotonase superfamily encoded by microbial genomes, and the likely enoyl CoA hydratases