The Capsaicin Biosynthesis Pathway
Capsaicin is not a single-enzyme product. It's assembled from two separate metabolic branches that converge at a final amide bond formation step. Understanding the full pathway is prerequisite to engineering heat into any plant.
Branch One: The Vanillyl Head Group
The aromatic vanillyl amine that forms the "head" of capsaicin comes from phenylalanine through the phenylpropanoid pathway — one of the most conserved metabolic routes in plant biology. The key steps are:
- PAL (phenylalanine ammonia-lyase): converts phenylalanine → trans-cinnamic acid
- C4H (cinnamate 4-hydroxylase): hydroxylates the ring → p-coumaric acid
- 4CL (4-coumarate-CoA ligase): activates the acid → p-coumaroyl-CoA
- HCT (hydroxycinnamoyl transferase) and downstream O-methyltransferases: lead to ferulic acid
- DOPA decarboxylase: converts an aromatic amino acid to form vanillylamine
Most dicot plants have a functional phenylpropanoid pathway, which is why tomato and tobacco are tractable hosts for the heat pathway — Branch One is already active.
Branch Two: The Acyl Tail
The branched-chain fatty acid that forms the "tail" of capsaicin begins from valine or leucine catabolism:
- BCAT (branched-chain amino acid aminotransferase): deaminates leucine → 4-methylpentanoic acid intermediate
- KAS (ketoacyl-ACP synthase) and KR (ketoacyl-ACP reductase): extend and reduce the acyl chain to the final 8- or 9-carbon length
- pAMT (putative aminotransferase): final transamination step producing the acyl amine substrate
The chain length produced by this branch determines which capsaicinoid is made: an 8-carbon chain gives capsaicin, a 9-carbon saturated chain gives dihydrocapsaicin, and shorter chains give nordihydrocapsaicin (less potent) or longer chains give homodihydrocapsaicin (more potent at TRPV1). Engineering the KAS/KR selectivity changes the capsaicinoid profile without changing total output.
The Convergence: AT3
The two branches meet at the final step: acyltransferase 3 (AT3, also called Capsaicin Synthase / CS in some literature) catalyzes the BAHD acyltransferase reaction that joins vanillylamine (Branch 1) with the acyl-CoA ester (Branch 2) via an amide bond to produce capsaicin.
AT3 is the gating enzyme. Its activity level sets the ceiling for total capsaicin output regardless of how much upstream substrate is available. This is why the Pun1 locus (which encodes AT3) determines pungency or non- pungency in Capsicum: a nonfunctional AT3 produces zero capsaicin regardless of substrate abundance.
Modeling the Pathway with FBA
Predicting how much capsaicin a given construct will produce in a new host requires knowing whether the host's metabolic network can supply sufficient flux through both branches simultaneously. Flux balance analysis (FBA) solves this by constraining a genome-scale metabolic model with stoichiometric and thermodynamic limits and maximizing biomass plus target compound production.
Every construct in the Scoville Splice catalog includes an FBA report modeling the host's metabolic context, substrate availability predictions, and expected capsaicinoid yield range. This report is generated by the third layer of the Intelligent Palate Architecture Platform: Metabolic Flux Analysis, run after ESM2 mutation scoring confirms the protein variants are functional.
Engineering Targets
Every step in both branches is a potential engineering lever:
- Overexpress PAL and C4H to increase phenylpropanoid flux
- Knock out competing branches (lignin biosynthesis) to redirect precursors
- Modify KAS selectivity to change the acyl chain length and capsaicinoid profile
- Use a high-activity AT3 variant to increase coupling efficiency
- Add a tissue-specific promoter to confine expression to target tissues
Each construct in our catalog is annotated with which levers were applied, what ESM2 score the protein variants achieved, and what FBA predicts for yield — giving researchers a fully transparent picture of the engineering decisions.