Introduction
Berberine bridge enzyme (BBE, Enzyme Commission number 1.21.3.3) catalyzes a key reaction in the biosynthetic pathways toward a major and diverse class of plant alkaloids—the benzophenanthridines. These molecules have been widely studied because of their many potential applications that range from hypotension treatment to anticancer activity1. Consistently, there is a growing demand to understand and engineer cellular and biocatalytic systems that can efficiently synthesize these compounds1. BBE shows a highly refined selectivity and reactivity, as it specifically recognizes the S enantiomer of reticuline, which is converted into (S)-scoulerine by a regioselective cyclization reaction (Fig. 1a). This atypical oxidative cyclization generates a chemical link ("berberine bridge") between the substrate isoquinoline and benzyl rings, which results in the formation of the tetracyclic skeleton of berberine-related alkaloids. A crucial feature of the reaction is that such a selective modification cannot be achieved by synthetic organic chemistry. Therefore, BBE represents an invaluable biocatalytic tool for synthetic biology approaches aiming at the production of bioactive alkaloids2. The structural and biochemical analysis of the enzyme reported by Winkler et al.3 in this issue of Nature Chemical Biology represents an important advance, providing conclusive evidence in support of an unprecedented mechanism of catalysis.
Figure 1: Biochemistry of BBE and homologous enzymes.
(a) BBE catalyzes the conversion of (S)-reticuline to (S)-scoulerine, a cyclization reaction that involves the formation of the so-called berberine bridge that links the benzyl and isoquinoline rings of the substrate. Reticuline is synthesized from L-tyrosine, whereas the product of the BBE reaction represents the precursor of a wide and diverse group of alkaloids that have various biological activities. (b) Simplified scheme of the catalytic mechanism of BBE proposed by Winkler et al.3. (c) Plant genomes exhibit a number of BBE homologs whose biological activities remain mostly to be uncovered. Among the BBE homologs, there are enzymes that take part in the biosynthesis of cannabinoids. Remarkably, these synthases act on molecules that substantially differ from reticuline, highlighting the biocatalytic potential and diversity of the family of BBE-related enzymes.
Full size image (80 KB)Though BBE activity was already known for several decades4, the mechanism by which the berberine bridge is formed remained elusive5. The combination of mutagenesis of several active site residues, kinetic analyses and a crystal structure of the enzyme from Eschscholzia californica in complex with (S)-reticuline supports a mechanism in which a hydride anion is transferred from the N-methyl group of the substrate isoquinoline ring to the flavin group of the FAD cofactor. Furthermore, the authors demonstrate that the two-electron oxidation of the substrate is coupled to a regioselective proton abstraction from the substrate benzyl group by an essential glutamate residue (Glu417) (Fig. 1b). This concerted "proton and hydride" abstraction gives rise to a Friedel-Crafts–like alkylation, eventually yielding the ring-closed (S)-scoulerine. Key to this elegant solution for a seemingly complex reaction is the proper orientation of the substrate with respect to the two catalytic groups; the flavin cofactor, the N-methyl group of the substrate isoquinoline ring, the hydroxyl group of the substrate benzyl ring and the carboxylate of Glu417 are precisely aligned within the active site to promote catalysis3. This binding mode is at the heart of the catalytic mechanism uncovered by this study.
The structural and mechanistic insights into BBE function are relevant for the analysis of related enzymes. In fact, plant genomes appear to be rich in BBE homologs. For instance, 12 homologs have been observed in the Arabidopsis thaliana genome6. This indicates that enzymes of the BBE family are likely to be involved in diverse and yet-unknown processes of plant metabolism, thus representing a reservoir of catalytic potential that remains to be explored. This notion is exemplified by the BBE homologs that have already been shown to take part in the biosynthesis of another class of plant alkaloids—the cannabinoids1. 
-Tetrahydrocannabinolic acid synthase and cannabidiolic acid synthase from Cannabis sativa catalyze a BBE-like oxidative cyclization reaction while acting on compounds with a different scaffold with respect to that of reticuline (Fig. 1c)7, 8. As both enzymes show high sequence homology to BBE (40% sequence identity), including conservation of the FAD-linking residues and the active site glutamate, these synthases are likely to use a similar concerted reaction that couples oxidation with proton abstraction.
A striking structural feature of BBE and the above-mentioned synthases is the mode in which the FAD cofactor is bound: it is covalently tethered to the polypeptide chain via two residues. Such an 8
-histidyl-6-S-cysteinyl-FAD cofactor has only been observed in a small number of recently reported oxidases9. Except for this characteristic mode of cofactor binding, members of this subclass of bicovalent flavoproteins also share another feature: they act on relatively complex molecules, often secondary metabolites. This may also relate to the rationale behind the double covalent FAD linkage. To create an active site that accommodates a bulky substrate molecule, while assuring appropriate positioning of residues and the flavin cofactor, a tight anchoring of the cofactor may be beneficial. Future studies on the recently discovered bicovalent flavoproteins will clarify the exact functional role of the double covalent anchoring of the cofactor.
Structural and mechanistic knowledge enables dedicated enzyme redesign by which new synthetic routes can be developed. In fact, one of the BBE mutants reported by Winkler et al.3 already shows formation of a new ring closure product, (S)-coreximine. Adapting BBE and related proteins toward other substrate types (by directed evolution for example) will aid synthetic biology approaches aiming at generation of new pharmaceuticals (see for example ref. 10). In fact, nature already illustrates such flexibility in selectivity, as BBE homologs of the cannabinoid metabolism act on compounds that are chemically and biologically unrelated to reticuline and similar alkaloids (Fig. 1). The results presented by Winkler et al.3 and the recent successful attempts of metabolic pathway engineering1 indicate that the biocatalytic potential of BBE and similar enzymes is still largely unexplored and will represent a major topic for future research.
