Converging on a mechanism for choline degradation

In PNAS, Craciun and Balskus (1) combine biochemical intuition with modern genome mining techniques to discover the elusive enzyme responsible for catalyzing the degradation of choline (1, Scheme 1). Choline is an essential nutrient in higher organisms that is required for the biosynthesis of the neurotransmitter acetylcholine and the head group of several phospholipids, and that serves as a source of the methyl groups for methionine and S-adenosylmethionine (2). In humans and other mammals, choline is catabolized to trimethylamine (TMA; 2, Scheme 1) by symbiotic gut microbes, and irregularities in choline and TMA metabolism have been linked to liver and cardiovascular diseases, atherosclerosis, and deficiencies in fetal brain development (3–5). TMA derived from choline is also a substrate for methanogenesis by marine microorganisms and, as such, contributes to global production of the greenhouse gas methane.

Scheme 1.

Mechanistic similarities in the catabolism of choline, ethanolamine, and glycerol; in each case, the bonds highlighted in red are cleaved by enzyme-mediated radical chemistry that results in a 1,2-migration.

To identify the unknown choline degrading enzyme, which has eluded researchers for more than a century, the authors postulated that the conversion of choline to TMA and acetaldehyde (3, Scheme 1) may share mechanistic similarities to the well-studied catabolism of ethanolamine (4, Scheme 1) to ammonia and acetaldehyde by bacterial enzymes encoded in the ethanolamine utilization (eut) gene cluster (6–8). In ethanolamine degradation, C-N bond cleavage is carried out by the adenosylcobalamin (AdoCbl)-dependent enzyme, ethanolamine ammonia lyase (EAL), encoded by the eutBC genes (6). After the EAL-catalyzed cleavage of the C-N bond of ethanolamine, the acetaldehyde product is converted to ethanol (5, Scheme 1) and acetyl-CoA (6, Scheme 1) by alcohol dehydrogenase (EutG) and aldehyde oxidoreductase (EutE) enzymes, respectively. In a clever approach to identify the gene cluster responsible for choline degradation in the anaerobic organism Desulfovibrio desulfuricans (an organism that metabolizes choline to TMA but whose genome does not encode EutBC homologs), Craciun and Balskus mined the genome for the presence of eut genes involved in acetaldehyde processing.

This approach led to the identification of the choline utilization (cut) gene cluster, which contains several tightly clustered homologs of eutE, eutG, and other eut genes. Interestingly, within this gene cluster are also encoded a predicted glycyl radical enzyme (cutC) and a glycyl radical activating protein (cutD), enzymes not known to catalyze C-N bond cleavage. Subsequent genetic disruption of the cutC gene in the anaerobic choline user Desulfovibrio alaskensis G20, and heterologous expression of the cutC and cutD genes from D. alaskensis G20 in Escherichia coli confirmed the essential role of these genes in the conversion of choline to TMA. In addition, whole-cell electron paramagnetic resonance studies showed that D. desulfuricans cells grown on media containing choline possess significant quantities of a paramagnetic species whose spectroscopic properties were consistent with a protein-derived, Cα-centered glycyl radical. Furthermore, a homology model of CutC suggested similarities in its active site architecture with another glycine radical enzyme, the B₁₂-independent glycerol dehydratase (9), which catalyzes formation of 3-hydroxypropanal (7, Scheme 1) from glycerol (8, Scheme 1). Bioinformatic analyses demonstrated that CutC homologs form a distinct clade of glycyl radical enzymes and that they are encoded in the genomes of organisms that are known to ferment choline as well as in human commensural bacteria and marine microorganisms. On the basis of these findings, the authors postulate that CutC is a choline TMA lyase enzyme that catalyzes conversion of choline to TMA using radical chemistry in a manner analogous to the EutBC-mediated C-N bond cleavage of ethanolamine.

The well-studied chemical mechanisms of EAL (10) and the B₁₂-independent glycerol dehydratase (9, 11) can serve as a basis for a mechanistic model of catalysis by choline TMA lyase. In EAL (Scheme 2), substrate binding triggers homolytic cleavage of the Co-C bond of the AdoCbl coenzyme, generating cob(II)alamin and the transient, reactive 5′-deoxyadenosyl radical (5′-dAdo•). 5′-dAdo• subsequently abstracts a hydrogen atom (H•) from C1 of ethanolamine to generate the well-characterized substrate radical, 9 (Scheme 2) (12). The exact chemical mechanism from this point remains unclear, but several computational studies agree that the lowest energy reaction coordinate involves a 1,2-migration of the amine group to generate the carbinolamine product radical (10, Scheme 2) (13, 14), perhaps via a cyclic transition state such as 11 (Scheme 2). Partial protonation of the migrating amine group and partial deprotonation of the hydroxyl group at C1 by active site amino acids likely play critical, synergistic roles in lowering the energy of the transition state leading to migration (13, 15). Consistent with this hypothesis, a recent X-ray crystal structure of EAL from E. coli revealed a constellation of conserved acid/base and polar amino acid residues that seemed poised to guide the amine group along a 1,2-migration trajectory (16). After amine migration, the reactive carbinolamine radical (10, Scheme 2) abstracts a hydrogen atom from 5′-dAdo to form carbinolamine (12, Scheme 2) and regenerate the resting state of the AdoCbl coenzyme. To complete catalysis, the amino group of the carbinolamine is eliminated to give the reaction products ammonia and acetaldehyde.

Scheme 2.

Putative chemical mechanism of ethanolamine degradation by the adenosylcobalamin-dependent enzyme, ethanolamine ammonia lyase (EAL).

Extrapolating our current understanding of EAL and glycyl radical enzymes to choline TMA lyase, CutD will first generate a glycyl radical on CutC (13, Scheme 3). As in other glycyl radical enzymes, CutC is expected to use an active site thiyl radical (14, Scheme 3) for the initial H• abstraction from substrate (17–19). After formation of the substrate radical (15, Scheme 3), the trimethylamino group at C2 may undergo a 1,2-migration to C1 (path a) to give a carbinolamine radical (16, Scheme 3) analogous to the proposed product radical in the EAL-catalyzed reaction (10, Scheme 2). According to the glycerol dehydratase homology model proposed by Craciun and Balskus (1), this migration may be facilitated by Glu493-mediated partial deprotonation of the C1-OH group and perhaps by electrostatic interactions between the migrating TMA group and the conserved Asp218 residue (1, 9, 11). Regeneration of the cysteine thiyl radical and elimination of TMA would complete the CutC catalytic cycle. In line with a recent computational study of the B₁₂-independent glycerol dehydratase (11), another potential model for CutC catalysis could involve direct elimination of the TMA group from the substrate radical to give the resonance stabilized product radical (17, path b, Scheme 3). In the case of EAL, however, both experimental and computational studies have shown that the sluggish reactivity of stable product radicals similar to 17 (Scheme 3) hinder regeneration of the active form of the AdoCbl coenzyme and act instead as suicide inhibitors (13, 20). Clearly, further mechanistic, structural, and computational studies of choline TMA lyases will be required to better interrogate their chemical mechanism and reaction energetics, to assess the feasibility of the key 1,2-TMA group migration step (15 → 16, Scheme 3) and to define the roles of active site amino acids in this biologically important enzymatic transformation.

Scheme 3.

Putative chemical mechanisms of choline degradation by the glycyl radical enzyme, choline trimethylamine lyase (CutC). Amino acids are numbered according to their positions in CutC from D. desulfuricans.

Whereas many details of the EAL and choline TMA lyase-catalyzed reactions have not yet been elucidated, the C-N bond cleavage reactions catalyzed by these enzymes illustrate an interesting example of convergent evolution—whereby structurally unrelated enzymes have evolved to catalyze similar chemical transformations. Interestingly, this is not the first reported example of convergent evolution between AdoCbl- and glycine radical-dependent enzymes: this phenomenon has been observed for ribonucleotide reductases and glycerol dehydratases (19). Thus, this type of convergent evolution is beginning to emerge as a common theme for these radical enzymes and could perhaps reflect an adaptation by microbial organisms to bypass the different susceptibilities of radical enzymes to irreversible inactivation by molecular oxygen. Regardless of what pressures have shaped the evolution of choline TMA lyases, it is clear that they constitute a unique class of enzymes that have expanded the already impressive catalytic repertoire of glycine radical enzymes (17–19). With their study, Craciun and Balskus have answered a long-standing question regarding choline catabolism and opened up a unique branch of glycyl radical enzymology.