Natural product biosynthesis

Three main pathways

Acetate (polyketide) pathway. Acetyl-CoA is carboxylated to malonyl-CoA. A starter acyl group then undergoes repeated decarboxylative Claisen condensations with malonyl units: loss of CO_2 generates the enolate nucleophile, which attacks a thioester, extending the chain by two carbons as a β-ketothioester. Fatty acid synthase reduces every β-keto group fully (ketoreduction, dehydration, enoyl reduction). Polyketide synthases may skip some or all of these steps, so the oxygen pattern of the product records which reductions happened. A fully unreduced chain cyclises by aldol or Claisen reactions to aromatic rings, which is why so many fungal and plant phenols have oxygen on alternate carbons, as in orsellinic acid.

Mevalonate pathway (terpenes and steroids). Three acetyl-CoA give HMG-CoA, which HMG-CoA reductase (the target of statins) reduces with NADPH to mevalonate. Phosphorylation and decarboxylative elimination give isopentenyl diphosphate (IPP), which isomerises to dimethylallyl diphosphate (DMAPP). Loss of diphosphate from DMAPP gives an allylic cation that is attacked by the alkene of IPP; deprotonation completes a head-to-tail C_5 addition. Repeating this gives geranyl (C_{10}), farnesyl (C_{15}) and geranylgeranyl (C_{20}) diphosphates. Terpene cyclases then ionise these and steer cation cascades (cyclisations, hydride and methyl shifts) to specific skeletons. Two farnesyl units join through their C1 carbons to give squalene (C_{30}), which is epoxidised and cyclised to lanosterol and on to cholesterol. Bacteria and plant plastids make IPP by the separate methylerythritol phosphate (MEP) pathway.

Shikimate pathway. Phosphoenolpyruvate and erythrose 4-phosphate are combined and cyclised to shikimic acid, then chorismic acid, the branch point to phenylalanine, tyrosine and tryptophan. Phenylalanine ammonia lyase gives cinnamic acid, the entry to lignin, flavonoids (with three malonyl units) and coumarins. Glyphosate kills plants by blocking the shikimate enzyme EPSP synthase.

Alkaloids

Alkaloids come from amino acids. A recurring step is a Mannich-type reaction between an amine and an iminium ion. In the Pictet–Spengler reaction catalysed by strictosidine synthase, tryptamine condenses with the aldehyde of secologanin (a terpenoid) and the indole attacks the iminium ion, giving strictosidine, the precursor of more than 2000 monoterpene indole alkaloids including quinine and vinblastine.

Labelling experiments

Pathways were worked out by feeding ^{13}C- or ^{14}C-labelled precursors. Feeding [1,2-^{13}C_2]acetate and finding intact ^{13}C–^{13}C couplings in the NMR spectrum of the product reveals which pairs of carbons came from the same acetate unit.

Modern biosynthesis research is driven by genomics. Genes for a natural product usually cluster together in bacteria and fungi, and tools such as antiSMASH predict biosynthetic gene clusters and even rough structures from sequence. Most sequenced Streptomyces genomes carry 20 to 40 clusters, far more than the compounds seen under lab conditions, so much effort goes into waking up these silent clusters by heterologous expression, promoter replacement or co-culture.

Modular type I PKSs and nonribosomal peptide synthetases (NRPSs) are the classic engineering targets because each module adds and tailors one unit. The erythromycin synthase (DEBS) is the model: three large proteins, six extension modules plus a loading module, a thioesterase that macrocyclises the product. Swapping acyltransferase domains to load different extender units, or deleting reductive domains, has produced new macrolides, but yields often collapse because domains interact through poorly understood interfaces. Cryo-EM structures of whole modules since the mid 2010s have changed how module boundaries are chosen for swaps.

Synthetic biology now reassembles entire plant pathways in microbes. Engineered yeast producing artemisinic acid, and later strains making opioids and the anticancer precursors of vinblastine, show that dozens of genes from several organisms can be combined, although titres are often too low for manufacture. Plant pathway discovery leans on transcriptomics, co-expression analysis and transient expression in Nicotiana benthamiana.

Enzymology has its own open questions: how terpene cyclases control carbocation cascades without making by-products, how radical SAM enzymes perform unusual C–C couplings, and how pericyclases catalyse Diels–Alder reactions in nature. Biocatalytic use of these enzymes in total synthesis, sometimes after directed evolution, is a fast-growing area.