Greater plant defence is predicted to evolve at lower latitudes in response to increased herbivore pressure. However, recent studies question the generality of this pattern. In this study, we tested for genetically based latitudinal clines in resistance to herbivores and underlying defence traits of Oenothera biennis. We grew plants from 137 populations from across the entire native range of O. biennis. Populations from lower latitudes showed greater resistance to multiple specialist and generalist herbivores. These patterns were associated with an increase in total phenolics at lower latitudes. A significant proportion of the phenolics were driven by the concentrations of two major ellagitannins, which exhibited opposing latitudinal clines. Our analyses suggest that these findings are unlikely to be explained by local adaptation of herbivore populations or genetic variation in phenology. Rather greater herbivory at high latitudes can be explained by latitudinal clines in the evolution of plant defences.

Metagenomics has made accessible an enormous reserve of global biochemical diversity. To tap into this vast resource of novel enzymes, we have screened over one million clones from metagenome DNA libraries derived from sixteen different environments for carboxylesterase activity and identified 714 positive hits. We have validated the esterase activity of 80 selected genes, which belong to 17 different protein families including unknown and cyclase-like proteins. Three metagenomic enzymes exhibited lipase activity, and seven proteins showed polyester depolymerization activity against polylactic acid and polycaprolactone. Detailed biochemical characterization of four new enzymes revealed their substrate preference, whereas their catalytic residues were identified using site-directed mutagenesis. The crystal structure of the metal-ion dependent esterase MGS0169 from the amidohydrolase superfamily revealed a novel active site with a bound unknown ligand. Thus, activity-centered metagenomics has revealed diverse enzymes and novel families of microbial carboxylesterases, whose activity could not have been predicted using bioinformatics tools.