Microbial flavour & fragrance production

Flavour and fragrance molecules are low-volume, high-value, and structurally intricate — exactly the profile that suits biosynthesis over chemical synthesis or extraction. Most targets fall into two families: terpenoids, which give citrus, mint, woody and floral notes, and phenylpropanoid derivatives, of which vanillin is the commercial giant.

Building a terpene#

All terpenoids are assembled from two five-carbon units, IPP and DMAPP. Cells make these by one of two routes: the mevalonate pathway, native to yeast and animals, or the MEP pathway, native to most bacteria and to plant plastids. Prenyltransferases condense the units into C10, C15 and C20 backbones, and a terpene synthase then cyclises the chain into a specific skeleton — this is the enzyme that determines which of hundreds of possible products is made.

Most valuable terpenoids then require oxidation: a hydroxyl or ketone in a defined position, installed by a cytochrome P450. This is usually the bottleneck. P450s are membrane-associated, require a redox partner to deliver electrons, and are frequently poorly expressed and poorly coupled in a microbial host, wasting reducing power on unproductive cycles. Much of the engineering effort in this field is spent on P450 expression, partner pairing and membrane engineering rather than on the pathway upstream.

The product is a poison and a vapour#

Terpenoids partition into membranes and disrupt them; many are antimicrobial precisely because of this. A cell engineered to make one is therefore engineered to poison itself, and titre is capped by toxicity well below what the pathway could support. The same molecules are volatile, so product is lost to the exhaust gas of an aerated fermenter.

Both problems have one answer: remove the product as it forms. A second, immiscible organic phase — a biocompatible solvent such as dodecane — is layered into the fermenter, and the hydrophobic product partitions into it continuously. This keeps the aqueous concentration below the toxic threshold, prevents volatilisation, and delivers the product already concentrated in a separate phase. In-situ product removal is not a downstream refinement here; it is what makes the process work at all.

Vanillin, and the meaning of “natural”#

Vanillin illustrates a labelling point that is otherwise easy to get wrong. The molecule is identical whatever its origin — extracted from cured vanilla pods, synthesised from petrochemical guaiacol, or made by microbial bioconversion of ferulic acid from rice bran or by de novo fermentation from sugar.

In major jurisdictions, whether a flavour may be called “natural” depends on the process, not on the molecule: production from a material of biological origin by a biological or physical process qualifies, chemical synthesis from a petrochemical does not. Fermentation-derived vanillin is therefore natural in the regulatory sense while being chemically indistinguishable from the synthetic. That is a definitional fact worth stating plainly, since it is often mistaken for a claim about quality or composition.

Extracted vanilla is a different product from vanillin — the pod contains hundreds of minor compounds — so the comparison is between vanillin sources, not between fermentation and vanilla.

Redox and carbon-yield limits apply as they do to any fermentation: highly reduced products cost more electrons per carbon, and that sets the ceiling on yield from sugar.