Phylogenetic approaches to natural product structure prediction

Abstract

Phylogenetics is the study of the evolutionary relatedness among groups of organisms. Molecular phylogenetics uses sequence data to infer these relationships for both organisms and the genes they maintain. With the large amount of publicly available sequence data, phylogenetic inference has become increasingly important in all fields of biology. In the case of natural product research, phylogenetic relationships are proving to be highly informative in terms of delineating the architecture and function of the genes involved in secondary metabolite biosynthesis. Polyketide synthases and nonribosomal peptide synthetases provide model examples in which individual domain phylogenies display different predictive capacities, resolving features ranging from substrate specificity to structural motifs associated with the final metabolic product. This chapter provides examples in which phylogeny has proven effective in terms of predicting functional or structural aspects of secondary metabolism. The basics of how to build a reliable phylogenetic tree are explained along with information about programs and tools that can be used for this purpose. Furthermore, it introduces the Natural Product Domain Seeker, a recently developed Web tool that employs phylogenetic logic to classify ketosynthase and condensation domains based on established enzyme architecture and biochemical function.

Figure 8.1

Phylogenetic trees. Molecular phylogenetic analyses are usually displayed in the form of trees. Examples include (A) a rooted rectangular tree or (B) an unrooted radial tree. Both maximum likelihood trees were generated using MEGA.

Figure 8.2

Ketosynthase domain phylogeny. The three PKS types (I–III) are clearly resolved in this KS phylogenetic tree as is their close relationship to various FAS (fab) genes. Sequences classified as KS III form a distinct lineage that is involved in the initiation of aromatic polyketide biosynthesis. This maximum likelihood tree was generated with PhyML using a manually curated alignment generated with muscle.

Figure 8.3

Phylogenetic workflow and software.