Fungal infections cause severe global mortality, yet the antifungal pipeline has changed little in decades. Amphotericin B and nystatin, polyene macrolides discovered in the mid-twentieth century, remain frontline drugs precisely because fungi rarely evolve resistance to them, but also because almost nothing better has reached the clinic. A Nature study breaks this long stalemate by showing that enzymes, not bespoke multi-step chemistry, can reshape these molecules into safer, more potent derivatives. The study recasts polyene optimization as a biosynthetic and glycosylation problem, reducing reliance on lengthy protecting-group chemistry, a shift that resonates with how modern Glycomics Assays are used to characterize complex glycans.

Introduction

Polyenes act as sterol sponges, extracting ergosterol from fungal membranes and rupturing them. The same membrane-disrupting behavior that underlies broad-spectrum activity also perturbs cholesterol-containing human cell membranes, which is why amphotericin B is notoriously toxic. Chemical derivatives that modify the molecule's sugars or its C16 carboxylate can reduce toxicity, but they demand long synthetic routes dominated by protecting groups and low yields. A renal-sparing amphotericin B analogue reported in 2023 required twelve chemical steps and was obtained in 0.7% overall yield from amphotericin B, a process that is difficult to call affordable for the regions that need it most.

The authors took a different path: let biology do the decorating. They mined genomes for the glycosyltransferases that install the second sugar on naturally occurring polyenes, then repurposed those enzymes to modify therapeutically relevant scaffolds.

Discovering the Polyene Glycosyltransferases

Selvamicin carries a rare second sugar, 4-O-methyl-L-digitoxose, and its candidate glycosyltransferase SelSV had not been experimentally characterized. Starting from that enzyme, the team identified homologs in several Streptomyces strains, overproduced them in Escherichia coli, and tested their ability to transfer sugars onto amphotericin B and nystatin A1. Among the homologues tested, KfuSV showed the highest activity with both scaffolds: it glycosylated amphotericin B and nystatin A1 and, across a panel of nineteen different TDP-sugar donors generated by tandem enzymatic assays, produced detectable products with eleven of them, as identified by LC–HRMS. This donor breadth makes KfuSV unusually versatile for polyene glycodiversification.

The key advance is controlled installation of a second sugar with defined donor selectivity. Because the enzyme accepts many TDP-activated donors, researchers can scout a landscape of glycosylated polyenes without synthesizing each donor chemically. LC–HRMS identifies glycosylated products, while NMR confirms the structures of lead derivatives, making enzymatic screening both scalable and quantifiable. Glycan Profiling by high-resolution mass spectrometry plays a central role in this analytical pipeline.

Fig. 1 Glycosylation of amphotericin B and nystatin A1 by KfuSV enables chemoenzymatic diversification of polyene antifungals (Mirza, et al., 2026).

Rewiring the Carboxylate with an Amidotransferase

The second sugar is only half the story. Polyene toxicity is partly driven by the C16 carboxylate, and synthetic amides at that position are known to be gentler. The study expanded the substrate scope of the amidotransferase PcsA, which amidated nystatin A1 and nystatin A3 with 98% and 96% conversion, respectively, and also introduced hydroxamate and, in small amounts, acyl-hydrazine functionality across several polyenes. The serinol amide derivative Nys34, however, was prepared by chemical modification rather than by PcsA. Combining a second sugar with C16 modifications produced nystatin A3 derivatives with higher antifungal activity and lower haemolytic toxicity than NysA3, the combination of properties clinicians value most. Notably, the parent NysA3 itself is more haemolytic than NysA1, so the safety benefit comes from the combined modifications.

Crucially, these molecules are reachable by fermentation or enzymatic steps rather than by the harsh reagents of total synthesis. A cleaner route lowers cost and environmental burden, both decisive for a drug intended for resource-limited settings where fungal disease burden is highest. The derivatives concurrently improve activity, reduce toxicity, and enable cleaner production. In a murine invasive aspergillosis model, Nys34 at 5 mg kg⁻¹ every 8 h for four doses reduced lung fungal burden by 87% (P = 0.0017) without overt signs of toxicity.

Why Glycosylation Controls Polyene Behavior

Glycosylation is not decorative here; it is a principal determinant of polyene behavior. The native mycosamine sugar at C19 is essential for bioactivity, and a second sugar at C35 further modulates activity, although the molecular basis remains incompletely defined. For NysA3, the C35 sugar suppresses isomerization and favors the bioactive form; for Nys34, the modifications abolish detectable ergosterol binding altogether, indicating a mechanism distinct from the classical sterol-sponge model. As in glycoproteins, glycosylation here acts as a structure–activity determinant, now extended to small-molecule natural products.

These mechanistic insights support further engineering. Once the structure–function determinants of improved tolerability are defined, directed evolution or pathway engineering can push them further. Among the homologues tested, KfuSV showed the highest activity, and the T297A variant improved activity with non-native substrates, suggesting that enzyme engineering can broaden scope further. Unlike a fixed twelve-step synthesis, enzymatic platforms can be improved iteratively.

Fig. 2 Organization of polyene biosynthetic gene clusters and reverse glycosylation reactions confirming the functions of NysSV, MycS3 and KfuSV in polyene biosynthesis (Mirza, et al., 2026).

Analytical Demands of Enzymatic Libraries

Generating eleven glycosylated derivatives from one enzyme raises a measurement problem: how to tell them apart and rank them. The authors rely on liquid chromatography with high-resolution mass spectrometry, a core analytical platform in modern glycan analysis. Confirming the site and structure of each new glycosylated polyene derivative is exactly the domain of N-Glycan Profiling and related intact-glycopeptide methods, which resolve glycans at residue resolution rather than as bulk sugar content.

Broader Significance for Carbohydrate Chemistry

The paper is a case study in chemoenzymatic synthesis, using biocatalysts to install carbohydrates that chemistry struggles with. That philosophy underlies much of carbohydrate manufacturing, where enzymatic routes avoid the protecting-group overhead that makes sugars expensive. Isomer-Specific Glycan Profiling then assures that the products are the intended isomers, not anomeric mixtures.

For CD BioGlyco, the takeaway is that enzymatic glycosylation is moving from a service offered to customers into a discovery engine for new medicines. The same glycosyltransferases that decorate polyenes could, in principle, be screened against other hydrophobic natural products, opening a general platform for glycodiversification of under-exploited scaffolds. The boundary between contract glycobiology and antibiotic discovery is thinning.

Outlook

Antifungal resistance is rising and the pipeline is thin, so any route that makes better polyenes cheaply deserves attention. Enzymatic glycosylation and amidation will not replace all polyene chemistry, but they reset expectations about what is achievable. The next steps are clear: optimize the enzymes further, extend efficacy studies to additional systemic fungal infection models, and scale the fermentation routes.

If these advances translate, the decades-long amphotericin impasse could end not with a new synthetic triumph but with familiar enzymes repurposed for new chemistry, a reminder that the most powerful glycoscience often lets biology decorate its own molecules.

The enzymatic route also changes the economics of antifungal drug discovery in a way chemistry cannot. Because KfuSV and PcsA are proteins, they can be evolved, immobilized, and run in fermentation trains that manufacturers already operate, lowering the barrier between a promising derivative and a supply of it. That matters for neglected fungal diseases, where the market is too small to justify twelve-step syntheses but large enough that affordable, low-toxicity polyenes could save lives. The breadth of KfuSV, which accepts eleven sugar donors, means the glycosylation landscape of polyenes can be explored systematically, replacing analogue-by-analogue serendipity with systematic screening. Looking ahead, the same genome-mining logic could, in principle, be applied to other natural-product families whose bioactivity depends on an attached sugar, from macrolides to angucyclines, suggesting a general platform for glycodiversification. For a carbohydrate-focused organization, the message is that enzymatic glycosylation is becoming a discovery engine, not just a service, and that teams combining enzyme discovery with rigorous structural analytics will be best positioned to advance sugar-modified therapeutics.

Related Services & Products

• Glycan Release Service
• Glycan Separation and Purification Service
• Glycan Derivation Service
• Glycomic Quantitation
• Glycoprotein Enrichment
• Glycoprotein Quantification
• Glyco-biomarker Detection Service
• Custom Glycosylation Service
• Flow Cytometry Profiling

Reference

1. Mirza, S. N., et al. (2026). Enzymatic glycosylation and amidation reshapes polyene bioactivity. Nature. DOI: 10.1038/s41586-026-10834-8.