Remarkable features of the hotdog-fold thioesterases involved in phylloquinone (vitamin K1) biosynthesis

Phylloquinone (vitamin K1) is a bipartite molecule, consisting of a naphthoquinone ring attached to a phytyl side chain, that is synthesized by plants and certain cyanobacteria to serve as an electron carrier in photosystem I. The coupling of the ring and isoprenyl moieties relies on the cleavage of the CoA-thioester linkage with 1,4-dihydroxy-2-naphthoate (DHNA). It has long been a mystery if this hydrolysis is an enzymatic or chemical process. Using comparative genomics, protein biochemistry, genetics and metabolic profiling, we identified a cyanobacterial thioesterase responsible for the in vivo hydrolysis of DHNA-CoA. This enzyme bears a signature domain of the 4-hydroxybenzoyl-CoA thioesterase (4HBT) family of Hotdog-fold proteins. Surprisingly, plants, which obtained most of their phylloquinone biosynthetic genes with the acquisition of the plastid, do not contain orthologs of cyanobacterial DHNA-CoA thioesterase. We tested all of the predicted 4HBT Hotdog-fold proteins in Arabidopsis by functional complementation of the cyanobacterial mutant. We found two genes encoding functional DHNA-CoA thioesterases that display low percentages of identity and dissimilar catalytic motifs from their cyanobacterial counterparts. It appears that plant DHNA-CoA thioesterases originated from a horizontal gene transfer with a species of the Lactobacillales order. The cognate T-DNA knockout lines exhibit reduced DHNA-CoA thioesterase activity and phylloquinone content. Fluorescently tagging the Arabidopsis enzymes revealed that they are localized to the peroxisome. Subcellular fractionation assays confirmed this providing the first biochemical evidence for the involvement of peroxisomes in phylloquinone biosynthesis. Recent proteomics and GFP-reporter projects suggest that the two steps preceding DHNA-CoA thioesterase are also peroxisomal. Thus, the current model of phylloquinone biosynthesis reflects a split between plastids and peroxisomes, implying the movement of intermediates between the organelles. To assess the importance of the cognate transport steps, we have re-routed the peroxisomal branch of the pathway to plastids in Camelina sativa. Here we report the findings of our metabolic engineering strategy on the pool of phylloquinone