Natural Product Drug Discovery

Natural product drug discovery is the identification, structural elucidation, and semisynthetic or synthetic optimization of secondary metabolites—specialized small molecules biosynthesized by terrestrial plants, soil and marine bacteria (particularly filamentous Streptomyces actinomycetes and myxobacteria), fungi, and marine invertebrates—for therapeutic use. Unlike primary metabolites (amino acids, nucleotides, glucose, and lipids) required for core cellular growth and replication across all organisms, secondary metabolites are lineage-specific chemical weapons and signaling molecules shaped by billions of years of natural selection to cross biological membranes and modulate macromolecular targets (such as the bacterial ribosome, tubulin cytoskeleton, cell-wall synthases, and ion channels) in competing organisms.

Comprehensive analyses of global regulatory approvals by David J. Newman and Gordon M. Cragg (spanning 1981–2019) demonstrate that natural products remain a foundational source of human medicines. Across the 1,881 small-molecule and biological drugs approved worldwide over that 39-year period, approximately 50% to 65% of all small-molecule new chemical entities (NCEs)—including >70% of antibacterial agents and >60% of small-molecule anticancer drugs—fall into natural-product-derived categories: unaltered natural products (N, ~4.6%), defined botanical mixtures (NB, ~0.8%), semisynthetic natural product derivatives (ND, ~21%), or totally synthetic drugs constructed around a natural product pharmacophore or acting as natural substrate mimics (S\* and S\*/NM, ~24%).

Three-Dimensional Complexity, Fsp³, and Stereogenic Centers

Because secondary metabolites are assembled by chiral enzymatic machinery rather than planar achiral cross-coupling reactions, natural products occupy regions of chemical space largely absent from conventional combinatorial high-throughput screening (HTS) libraries:

1. High Fraction of sp³-Hybridized Carbons (Fsp³) and Chiral Centers: Natural products exhibit an average Fsp³ ≈ 0.45–0.68 and typically contain 3 to 15+ tetrahedral stereogenic centers, compared with an average Fsp³ ≈ 0.22–0.28 and 0 to 1 chiral centers in synthetic HTS decks. This three-dimensional architecture increases target selectivity and lowers off-target promiscuity.
2. Bridged, Spirocyclic, and Macrocyclic Ring Systems: Whereas synthetic libraries rely heavily on 1,4-disubstituted benzenes and simple monocyclic heterocycles, natural products frequently feature fused aliphatic polycycles (steroids, taxanes), spiroketals, and conformationally pre-organized 12- to 36-membered macrocycles (macrolides, ansamycins) that span broad protein–protein interaction surfaces.
3. Oxygen-Rich Polarity and Metabolite-Likeness: Natural products possess a higher ratio of oxygen atoms (ethers, alcohols, lactones, epoxides) to nitrogen, sulfur, and halogen atoms than synthetic drugs. Although many complex natural products exceed the molecular weight and hydrogen-bond thresholds of Lipinski's rule of five ("beyond Rule of Five" or bRo5 space), they achieve cellular entry either through conformational "chameleonic" intramolecular hydrogen bonding or by hijacking endogenous membrane drug transporters (SLC uptake carriers) evolved to recognize structurally related primary metabolites.

Polyketides (PKS) and Non-Ribosomal Peptides (NRPS)

Microbial and fungal secondary metabolites are predominantly assembled by giant multimodular megasynthases that operate like molecular assembly lines:

• Polyketide Synthases (PKS Types I, II, and III): Catalyze iterative Claisen condensations of malonyl-CoA and methylmalonyl-CoA extender units, followed by selective ketoreductase, dehydratase, and enoylreductase processing. Landmark polyketide medicines include 14- and 15-membered macrolide antibiotics (erythromycin from Saccharopolyspora erythraea and its semisynthetic azalide derivative azithromycin, which bind the 50S ribosomal exit tunnel); the macrocyclic immunosuppressants rapamycin (sirolimus) (from Streptomyces hygroscopicus, a molecular glue bridging FKBP12 to mTOR) and tacrolimus (FK506) (bridging FKBP12 to calcineurin); the antiparasitic macrocyclic lactones avermectins and ivermectin (isolated by Satoshi Ōmura and William Campbell from Streptomyces avermitilis, awarded the 2015 Nobel Prize in Physiology or Medicine for eradicating river blindness and lymphatic filariasis); the anthracycline topoisomerase II poison doxorubicin; and the fungal polyketide lovastatin (mevinolin) (discovered by Akira Endo from Aspergillus terreus), whose β-hydroxy-δ-lactone warhead mimics the HMG-CoA transition state and inspired the entire class of cholesterol-lowering statins (simvastatin, atorvastatin, rosuvastatin).
• Non-Ribosomal Peptide Synthetases (NRPS): Assemble linear, branched, or cyclic peptides independent of mRNA templates, incorporating non-proteinogenic D-amino acids, N-methylated residues, β-hydroxy acids, and heterocyclic thiazoline/oxazoline rings. NRPS pathways yield the β-lactam antibiotics (penicillins from Penicillium chrysogenum and cephalosporins from Acremonium), the glycopeptide cell-wall inhibitor vancomycin (from Amycolatopsis orientalis), the lipopeptide daptomycin, the anticancer glycopeptide bleomycin (hybrid NRPS–PKS), and the orally bioavailable cyclic undecapeptide immunosuppressant cyclosporine A (from Tolypocladium inflatum).

Terpenoids and Plant Alkaloids

Higher plants synthesize vast repertoires of terpenoids (isoprenoids)—constructed from five-carbon isopentenyl pyrophosphate (IPP) and dimethylallyl pyrophosphate (DMAPP) units via the mevalonate or methylerythritol phosphate (MEP) pathways—and nitrogenous alkaloids derived from amino acids:

• Artemisinin: Isolated in 1972 by Tu Youyou (co-recipient of the 2015 Nobel Prize) from sweet wormwood (Artemisia annua) via low-temperature diethyl ether extraction inspired by Ge Hong's 4th-century CE medical text, artemisinin is a sesquiterpene lactone (C₁₅H₂₂O₅) containing a bridged 1,2,4-trioxane endoperoxide ring. When cleaved by ferrous heme iron (Fe²⁺) inside Plasmodium falciparum-infected erythrocytes, the endoperoxide generates carbon-centered free radicals that alkylate parasite proteins, forming the basis of artemisinin-based combination therapies (ACTs) using water-soluble semisynthetic artesunate and lipid-soluble artemether.
• Paclitaxel (Taxol): Isolated by Monroe Wall and Mansukh Wani (1971) from the bark of the Pacific yew (Taxus brevifolia) and solved for sustainable production via semisynthesis from 10-deacetylbaccatin III in English yew needles (Taxus baccata) by Robert Holton and Pierre Potier, paclitaxel is a polyoxygenated diterpenoid containing a strained oxetane ring that binds the lumen of polymerized β-tubulin to hyperstabilize microtubules and arrest mitosis.
• Alkaloids: Include the isoquinoline analgesic morphine (Papaver somniferum), the antimalarial quinoline quinine (Cinchona bark), the Catharanthus roseus bisindole antimitotics vinblastine and vincristine, and the quinoline alkaloid camptothecin (Camptotheca acuminata), which traps DNA topoisomerase I cleavage complexes and serves as the parent scaffold for irinotecan, topotecan, and the potent antibody–drug conjugate payload deruxtecan (DXd).

Overcoming Rediscovery via BGC Mining and MS/MS Networking

During the 1990s and 2000s, many pharmaceutical firms scaled back phenotypic natural product screening due to three technical bottlenecks: (1) slow bioassay-guided fractionation of complex crude extracts, (2) frequent rediscovery of known "nuisance" compounds (such as streptothricin or actinomycin), and (3) scarce supply and synthetic intractability of complex leads. Three technological advances have resolved these barriers:

1. Genome Mining of Silent Biosynthetic Gene Clusters (BGCs): Whole-genome sequencing revealed that a single Streptomyces or Aspergillus strain harbors 25 to 50 distinct BGCs, yet >80%–90% of these clusters remain transcriptionally silent (cryptic) under standard laboratory monoculture. Bioinformatics tools such as antiSMASH identify novel PKS, NRPS, and RiPP gene clusters directly from genomic DNA, allowing researchers to awaken silent pathways via CRISPR promoter engineering, epigenetic histone deacetylase modifiers, microbial co-cultivation, or heterologous expression in chassis hosts (Streptomyces coelicolor, Saccharomyces cerevisiae).
2. MS/MS Molecular Networking Dereplication: Platforms such as GNPS (Global Natural Products Social Molecular Networking) cluster high-resolution tandem mass spectrometry (LC–HRMS/MS) fragmentation spectra from thousands of crude extracts in minutes, flagging known scaffolds before isolation and pinpointing novel structural analogs.
3. Function-Oriented Synthesis (FOS): Pioneered by Paul Wender, FOS distills a complex natural product down to its minimal active pharmacophore for practical synthesis. A celebrated triumph is the marine sponge polyether macrolide halichondrin B (MW = 1,110 Da, 32 stereocenters): Yoshito Kishi's total synthesis revealed that all antimitotic tubulin-destabilizing activity resided in the macrocyclic right half, enabling Eisai chemists to synthesize eribulin (Halaven)—a simplified synthetic ketone analog with 19 chiral centers manufactured on commercial multi-kilogram scale in ~62 steps.

What percentage of approved small-molecule drugs originate from natural products?

According to the comprehensive FDA and global approval analyses by David Newman and Gordon Cragg (1981–2019), approximately 50% to 65% of all approved small-molecule drugs are unaltered natural products, semisynthetic natural product derivatives, or synthetic compounds designed around a natural-product pharmacophore. In infectious disease and oncology, more than 70% of small-molecule antibiotics and over 60% of anticancer agents trace their origins to natural products.

Why do natural products have higher Fsp³ and stereochemical complexity than synthetic screening libraries?

Synthetic combinatorial libraries are predominantly built using fast, achiral, palladium-catalyzed sp²–sp² cross-coupling and amide bond-forming reactions, yielding flat aromatic rings (Fsp³ ≈ 0.25). Natural products are synthesized by three-dimensional chiral enzyme active sites (such as polyketide synthases, terpene cyclases, and P450 oxygenases) that fold linear carbon chains into stereochemically dense sp³ cages (Fsp³ ≈ 0.55) optimized by evolution to fit chiral protein cavities.

What is a silent (cryptic) biosynthetic gene cluster in microbial drug discovery?

Soil bacteria (such as Streptomyces) and filamentous fungi possess 25 to 50 biosynthetic gene clusters (BGCs) in their genomes that encode pathways for secondary metabolites, yet they express only 2 to 5 of them in standard laboratory Petri dishes because the remaining 80%–90% require specific ecological stress or competitor signals. Genome mining uses tools like antiSMASH to locate these silent BGCs and activate them via promoter editing or heterologous expression.

1. Newman, D. J., & Cragg, G. M. (2020). "Natural products as sources of new drugs over the nearly four decades from 01/1981 to 09/2019." Journal of Natural Products, 83(3), 770–803. https://doi.org/10.1021/acs.jnatprod.9b01285
2. Atanasov, A. G., Zotchev, S. B., Dirsch, V. M., & Supuran, C. T. (2021). "Natural products in drug discovery: advances and opportunities." Nature Reviews Drug Discovery, 20(3), 200–216. https://doi.org/10.1038/s41573-020-00114-z
3. Tu, Y. (2011). "The discovery of artemisinin (qinghaosu) and gifts from Chinese medicine." Nature Medicine, 17(10), 1217–1220. https://doi.org/10.1038/nm.2471
4. Yu, M. J., Zheng, W., & Seletsky, B. M. (2013). "From micrograms to grams: scale-up synthesis of eribulin mesylate." Natural Product Reports, 30(9), 1158–1164. https://doi.org/10.1039/C3NP70051H