Our starting material for damascones, the rose ketones, was zeaxanthin, a yellow pigment made by algae. Look at the zeaxanthin structure and you can see why. A long chain of conjugated double bonds connects two rings, and each ring already carries the methyl groups a damascone needs. The rings also carry hydroxyl groups that damascones don’t have. The work was figuring out where to break the chain and what to change afterward.
Plants do some of this with enzymes. Carotenoid cleavage dioxygenases cut pigments into smaller molecules called apocarotenoids. Zhipeng Qi and colleagues expressed one from sunflower in E. coli and found it cut several carotenoids at the 9,10 and 9′,10′ double bonds, producing β-ionone and hydroxylated ionones depending on the starting pigment. [1] β-Ionone, the violet note, has the same formula as β-damascone. [2] In ionone, the ketone sits two carbons out from the ring, and in damascone it sits against it. A cut at 9,10 leaves the oxygen in the violet position. Our process made its first cuts with ozone.
The bacteria in AGATE could hydrogenate the pigment, adding hydrogen across its double bonds. Ozone cuts at double bonds, so letting the bacteria go first would have used up the places we needed to cut. We broke the backbone with ozone, then fed the fragments into AGATE. It worked. The microbial stage mostly put the carbonyl next to the ring, the position that makes a damascone smell like a rose instead of a violet.
That fermentation ran at very low pH and produced hydrogen sulfide. It behaved more like a geobiology experiment than a flavor process. GC/MS confirmed damascones, but they came out alongside ionol, hydrogenated products, and sulfur-containing C13 compounds. My best explanation, which we never tested, is that the biology supplied the intermediates and the acid drove them to react with whatever was nearby, water in some molecules and hydrogen sulfide in others. We knew the product mixture better than we knew every reaction that made it.
We commercialized the process anyway. The damascones came out as a mixture, and turning them into a usable product took molecular distillation and a lot of downstream work. That recovery was the hardest part of the project. The bacteria put the oxygen where we wanted it, but we couldn’t control the proportions of what came out of the fermentation.
Sources
[1] Qi et al. (2022). Cloning and characterization of a novel carotenoid cleavage dioxygenase 1 from Helianthus annuus. Chemistry & Biodiversity.
https://doi.org/10.1002/cbdv.202100694
[2] Aloum et al. (2020). Ionone is more than a violet’s fragrance: a review. Molecules.
https://doi.org/10.3390/molecules25245822