Genomic Determinants Encode the Reactivity and Regioselectivity of Flavin-Dependent Halogenases in Bacterial Genomes and Metagenomes

Abstract

Halogenases create diverse natural products by utilizing halide ions and are of great interest in the synthesis of potential pharmaceuticals and agrochemicals. An increasing number of halogenases discovered in microorganisms are annotated as flavin-dependent halogenases (FDHs), but their chemical reactivities are markedly different and the genomic contents associated with such functional distinction have not been revealed yet. Even though the reactivity and regioselectivity of FDHs are essential in the halogenation activity, these FDHs are annotated inaccurately in the protein sequence repositories without characterizing their functional activities. We carried out a comprehensive sequence analysis and biochemical characterization of FDHs. Using a probabilistic model that we built in this study, FDHs were discovered from 2,787 bacterial genomes and 17 sediment metagenomes. We analyzed the essential genomic determinants that are responsible for substrate binding and subsequent reactions: four flavin adenine dinucleotide-binding, one halide-binding, and four tryptophan-binding sites. Compared with previous studies, our study utilizes large-scale genomic information to propose a comprehensive set of sequence motifs that are related to the active sites and regioselectivity. We reveal that the genomic patterns and phylogenetic locations of the FDHs determine the enzymatic reactivities, which was experimentally validated in terms of the substrate scope and regioselectivity. A large portion of publicly available FDHs needs to be reevaluated to designate their correct functions. Our genomic models establish comprehensive links among genotypic information, reactivity, and regioselectivity of FDHs, thereby laying an important foundation for future discovery and classification of novel FDHs. IMPORTANCE Halogenases are playing an important role as tailoring enzymes in biosynthetic pathways. Flavin-dependent tryptophan halogenases (Trp-FDHs) are among the enzymes that have broad substrate scope and high selectivity. From bacterial genomes and metagenomes, we found highly diverse halogenase sequences by using a well-trained profile hidden Markov model built from the experimentally validated halogenases. The characterization of genotype, steady-state activity, substrate scope, and regioselectivity has established comprehensive links between the information encoded in the genomic sequence and reactivity of FDHs reported here. By constructing models for accurate and detailed sequence markers, our work should guide future discovery and classification of novel FDHs.

Keywords: active sites; genomic patterns; halogenase; mutagenesis; phylogeny; profile hidden Markov model; regioselectivity.

FIG 1

Known and putative halogenase sequences represented using t-SNE and a phylogenetic tree. (a) Embedding of the putative halogenases identified from the bacterial genomes and marine metagenomes, in conjunction with the known halogenases based on genomic patterns. (b) Phylogenetic tree of the halogenases.

FIG 2

Integrated structure and sequence analysis of tryptophan and nontryptophan halogenases. Tryptophan-binding sites determined in this study are denoted for each group in Fig. 1. (a) PrnA in A1 (2AQJ); (b) RebH in A1 (2OA1); (c) Thal in A2 (6H44); (d) SttH in B2 (5HY5); (e) PyrH in B1 (2WET); (f) BrvH in C (6FRL). PDB codes are in parentheses. Residues comprising Trp1 to -4 are colored yellow, cyan, orange, and purple, respectively. Tryptophan molecules and catalytic lysine residues are shown in green and magenta bars, respectively. (g) Multiple-sequence alignment and conservation sites of FAD1 to -4, Lys (K), Trp1 to -4, and halide sites of FDHs. Secondary structure for each motif site is represented as α-helices, β-sheets, and loops using PrnA and PyrH.

FIG 3

Halogenation of Trp-FDHs with tryptophan. (a) Chemical structures of the halogenated products. Product yields in chlorination (b) and bromination (c) reactions.

FIG 4

Altered regioselectivity of Trp-FDH by modification of Trp conservation sites. (a) Alteration between 6-Trp-FDH and 7-Trp-FDH. Total yields (top) and fractions (bottom) upon mutations of Hal2 to 7-Trp-FDH (left) and Hal7 to 6-Trp-FDH (right). (b) Alteration between 5-Trp-FDH and 6-Trp-FDH: Hal6 to 6-Trp-FDH (left) and Hal4 to 5-Trp-FDH (right). The fractions of the products in chlorination and bromination reactions are shown in left and right columns, respectively.

FIG 5

Halogenation of bacterial and metagenomic FDHs with indole. (a) Yields (%) of chlorinated and brominated indoles. (b) Reactivities of Hal1 variants at the Trp4 motif (YYEN) with tryptophan and indole. (c) Hydrogen-bonding interactions between Trp4 residues (YYEN) and a tryptophan shown in the structure of PrnA in A1 (2AQJ).

FIG 6

Catalytic activities of Trp-FDH identified from the bacterial genomes and metagenomes on diverse aromatic substrates.