Paralogous genes in Arabidopsis thaliana contribute to diversified phenylpropanoid metabolism

Significant evidence supports the idea that gene duplication drives the evolution of new gene function. Besides being silenced, duplicated genes can either neofunctionalize or subfunctionalize under selective pressure or neutral drift. Understanding the trajectory of how each gene is fixed and presumably provides added fitness remains difficult. Plant specialized metabolism provides an attractive platform to study the fixation of genes post duplication and how this process leads to the chemical diversity seen today. Specialized metabolites by definition are thought to be dispensable under normal growth conditions. Thus deleterious mutations occurring in paralogous genes that otherwise would be selected against in primary metabolism may be more tolerated in specialized metabolism. Using the Arabidopsis phenylpropanoid pathway as a model, in this thesis I describe two cases where neofunctionalization and subfunctionalization of duplicated genes contributed to metabolite diversification. The phenylpropanoid pathway intersects with primary metabolism at phenylalanine. Recently we identified a new set of phenylalanine derived compounds in Arabidopsis which we named arabidopyrones (APs) which include arabidopyl alcohol, iso-arabidopyl alcohol, arabidopic acid and iso-arabidopic acid. CYP84A4 is a paralog of CYP84A1, a well-characterized enzyme in the phenylpropanoid pathway, and CYP84A4 has neofunctionalized relative to its ancestral function. CYP84A4 3-hydroxylates p-coumaraldehyde, a phenylpropanoid intermediate, to generate caffealdehyde. Caffealdehyde can be used by a conserved ring cleavage dioxygenase, AtLigB, in a step required to make the heterocyclic APs. Understanding AP biosynthesis may provide a unique opportunity to learn the broader biological function of LigB homologs. To do so, we tested the hypothesis that enzymes and intermediates in the phenylpropanoid pathway leading to p-coumaraldehyde are involved in AP biosynthesis. The general phenylpropanoid pathway gives rise to p-coumaryl CoA via phenylalanine ammonia lyase (PAL), cinnamate 4-hydroxylase (C4H) and 4-coumarate: CoA ligase (4CL). Cinnamoyl CoA reductase (CCR) then converts p-coumaryl CoA to p-coumaraldehyde, the substrate of CYP84A4. Through the analyses of mutants that are defective in these genes, stable-isotope labeling studies, and chemical complementation experiments, we conclude that the activities of these enzymes are required for AP biosynthesis. In addition, we found that cinnamyl alcohol dehydrogenase C and D (CAD C and D), known enzymes in later steps of the phenylpropanoid pathway, are involved in AP biosynthesis in that they may convert caffealdehyde to caffeoyl alcohol which then can be used by AtLigB to generate arabidopyl alcohol and arabidopic acid. Four isoforms of 4CLs have been identified in Arabidopsis. 4CL generates p-coumaryl CoA and caffeoyl CoA from their respective acids which are required for the major products of this pathway. Phylogenetic analysis reveals that 4CL1, 4CL2 and 4CL4 are more closely related to one another than to 4CL3. Promoter-GUS analysis shows that 4CL1 and 4CL2 are expressed in lignifying cells. In contrast, 4CL3 is expressed in a broad range of cell types, indicating that 4CLs have subfunctionalized with regard to expression patterns. We found that 4cl3 mutants have an over-all reduction in flavonoid biosynthesis, suggesting that 4CL3 has acquired a distinct role in phenylpropanoid metabolism. Sinapoylmalate, the major hydroxycinnamoyl ester found in Arabidopsis is greatly reduced in a 4cl1 4cl3 mutant, showing that 4CL1 and 4CL3 function redundantly in its biosynthesis. The 4cl1 4cl2 double mutant and the 4cl1 4cl2 4cl3 triple mutant are both dwarf and contain less lignin than wild type, indicating that 4CL1 and 4CL2 are important for plant growth and that 4CL3 has a role in lignin biosynthesis in addition to its function in soluble metabolism. We could not find an important role for 4CL4 in any of the organs examined, consistent with its limited expression profile. Together, these data show that the four paralogs of 4CLs in Arabidopsis diverged in their expression patterns, resulting in their overlapping yet distinct roles in phenylpropanoid metabolism