EGCG accounts for roughly half of the total catechins in fresh tea leaves and is prized for antioxidant, anti-inflammatory and anticancer activities. Yet it is also notoriously unstable and poorly absorbed. During fermentation it is converted into a family of derivatives, including epigallocatechin, epicatechin, catechin, theasinensins, theaflavins, fuzhuanins and teadenols, many of which show better bioavailability and distinct biological activities. The research team, led by Zexin An and Mingzhi Zhu of Hunan Agricultural University, set out to map this transformation comprehensively, combining targeted quantification with untargeted metabolomics across four experimental systems: liquid fermentation with each fungus separately, solid-state fermentation of sterile raw dark tea, an in vivo rat study, and an in vitro anaerobic culture of rat gut microbiota.

In the liquid fermentation experiments, the fungi were grown for five days before EGCG was added to a final concentration of 1 milligram per milliliter, and samples were collected over 72 hours. In a sterile medium without fungi, EGCG and gallic acid remained stable, ruling out significant auto-oxidation. Once A. cristatus was present, EGCG became undetectable within 12 hours, whereas A. niger needed a full 24 hours to finish the job. First-order kinetic modeling quantified the difference: the apparent disappearance rate constant for A. cristatus was 0.326 per hour, roughly 3.05 times that of A. niger at 0.107 per hour, with corresponding half-lives of 2.13 and 6.49 hours. Gallic acid, the primary hydrolysis product of EGCG, accumulated steadily in both systems.

Untargeted metabolomics using ultra-high-performance liquid chromatography coupled to a Q-Exactive Orbitrap mass spectrometer detected 3,402 annotated metabolic features across the fungal cultures. After filtering for features that changed more than twofold relative to controls, the team identified 235 differential metabolites in the A. cristatus system and 114 in the A. niger system. By tracking how each metabolite’s abundance rose and fell relative to EGCG and gallic acid, the researchers organized them into early-, mid- and late-stage products. Early-stage metabolites included epigallocatechin itself and complex polyflavonoids such as samarangenin A and epigallocatechin-(4β→8)-epicatechin 3,3′-digallate, showing that fungal fermentation drives not only degradation but also polymerization from the very beginning. Mid-stage products shared by both fungi included theasinensin B, oolongtheanin, theaflavin 3,3′-digallate, prodelphinidin A2 3′-gallate and isoneotheaflavin 3-O-gallate, several of which have documented antihypertensive, cholesterol-lowering, antiviral or alpha-glucosidase-inhibiting activities.

The late stages diverged sharply between the two fungi. In A. cristatus cultures, 3-hydroxyflavone emerged as a dominant late-stage EGCG-associated metabolite, accumulating at more than 100 times the level seen with A. niger, a remarkable demonstration of the species-specific conversion efficiency. A. cristatus also generated a diverse set of gallic-acid-derived products, with database-assisted annotations pointing to compounds including ellagic acid, dihydroxyphaseic acid, camonagrel and lidoflazine. A. niger, by contrast, produced fewer but distinctive derivatives, with flavonoids such as quercetin 3-(6”-caffeylgalactoside), isotheaflavin, fuzhuanins A and prodelphinidin A1 peaking at 36 hours before declining, while pelargonidin 3-sambubioside and a methylenedioxy-substituted phenylcoumarin continued rising through 72 hours.

The most surprising result came when the team moved from the simplified liquid system to real tea. Sterile raw dark tea was fermented with each fungus for 15 days, mirroring the roughly two-week “flowering” stage of Fu brick tea production. Visually, the cultures diverged dramatically: A. cristatus eventually produced the characteristic golden “golden flowers” on the leaves, while A. niger grew vigorously and turned the leaves brownish. But the kinetics of EGCG disappearance reversed. In tea fermented with A. niger, EGCG dropped to 209.5 micrograms per gram by day 5 and was nearly undetectable by day 15, whereas A. cristatus fermentation showed a slower, more gradual decline that plateaued after day 10. The authors suggest several possible explanations, including the known ability of high EGCG concentrations to inhibit A. niger by suppressing chitinase and chitin synthase expression, and the possibility that the tea matrix redirects A. cristatus toward alternative metabolic pathways. They also caution that the tea was autoclaved before inoculation, a treatment that may itself alter catechins, proteins and their interactions, and that untreated raw tea was not directly compared.

To untangle the chemistry of the tea matrix, the researchers applied weighted correlation network analysis to 5,144 detected compounds. In the A. cristatus fermentations, the data sorted into 15 color-coded modules, with the Meblue module of 651 metabolites showing a highly significant positive correlation with EGCG content. Within it, 585 core metabolites passed stringent thresholds, including 13 flavans such as catechin, epigallocatechin, taxifolin, neotheaflavin and methylated epigallocatechin gallates. A network built around EGCG linked it to 244 highly associated metabolites. Molecular networking on the GNPS platform then clustered structurally similar compounds sharing a common fragment ion at m/z 170.02, revealing a putative EGCG dimer at m/z 915.161, a potential epigallocatechin dimer at m/z 611.14, and other catechin relatives whose mass shifts hint at hydroxymethylation and related modifications. A. niger fermentations yielded their own module structure and unique core metabolites, including 4-gallocatechol, which accumulated significantly by day 15 and may represent a core late-stage product.

The study then followed EGCG into the body. Rats received a single oral dose of 500 milligrams per kilogram of body weight, and fecal samples were collected over 24 hours. EGCG first appeared in feces at 6 hours post-gavage, peaked at 440.27 micrograms per gram at 10 hours, and had fallen to just 6.14 micrograms per gram by 24 hours, pinpointing a 6-to-12-hour window as the key intestinal transit phase. Metabolomics of 106 available fecal samples yielded 14,931 annotated features, from which 683 EGCG-associated candidates were selected by chemical-class filtering and 59 core metabolites were prioritized using statistical models. Among 13 core flavans, the team found not only monomeric catechins and their derivatives but also a high-molecular-mass EGCG-associated product and putative theaflavin-related compounds, suggesting that oxidative transformations, potentially involving both non-enzymatic reactions and gut microbial interactions, occur within the intestine. A dimethylated catechin derivative also appeared, a modification that may alter polarity, stability and metabolic kinetics.

Perhaps the most instructive contrast was between the living gut and the petri dish. When rat gut microbiota were cultured anaerobically with EGCG at three concentrations, the bacteria reduced EGCG levels over time but unexpectedly failed to degrade it completely, even after 72 hours. Only nine flavans were detected among 5,942 annotated features, mostly monomeric catechins such as epigallocatechin, catechin and epicatechin, along with sulfated derivatives including epicatechin 4′-sulfate and 7′-sulfate. Seven of these nine metabolites also appeared in the fungal fermentations, and epicatechin 4′-sulfate showed similar concentration trends in vivo and in vitro. The far richer metabolic profile seen in living rats therefore cannot be explained by fecal microbiota alone; host enzymatic reactions, intestinal absorption, dynamic pH, bile, oxygen gradients and the mucus-associated microenvironment all likely contribute. 16S rDNA sequencing of the cultures showed Ligilactobacillus, Lactobacillus, Streptococcus and Escherichia-Shigella as dominant genera, with exploratory correlations between microbial abundance and EGCG levels that did not remain statistically significant after correction for multiple testing.

The authors are candid about the limits of the work: without isotope-labelled EGCG, the proposed pathways remain putative, since endogenous fungal metabolites, native tea constituents and host-derived compounds may change in parallel with EGCG. Future studies using carbon-13-labelled EGCG with high-resolution tandem mass spectrometry will be needed to confirm precursor-product relationships. Even so, the study delivers a clear message with practical implications. EGCG transformation is not a single uniform degradation pathway but a strain- and substrate-dependent process, meaning that fungal strain selection and fermentation conditions could be deliberately tuned to control how much EGCG is depleted and which catechin derivatives accumulate in dark tea. Given that many of the identified derivatives carry documented health-promoting activities, that tuning could shape not just the flavor of fermented tea but its functional value, while deepening our understanding of what happens to tea polyphenols between the cup and the colon.

Subject of Research: Microbial and gut microbiota biotransformation of the tea catechin EGCG during dark tea fermentation and intestinal metabolism

Article Title: Aspergillus Species and Gut Microbiota-Mediated Biotransformation of Epigallocatechin Gallate: From Dark Tea Fermentation to Intestinal Metabolism

Article References: An, Z., Ma, R., Huang, X., Zhang, Y., Huang, J., Liu, Z., Wang, K., & Zhu, M. (2026). Aspergillus Species and Gut Microbiota-Mediated Biotransformation of Epigallocatechin Gallate: From Dark Tea Fermentation to Intestinal Metabolism. Current Research in Food Science, Article 101587. https://doi.org/10.1016/j.crfs.2026.101587

Image Credits: AI Generated

DOI: Not provided

Keywords: EGCG, dark tea, Aspergillus cristatus, Aspergillus niger, fermentation, gut microbiota, metabolomics, catechins, gallic acid, theaflavins, bioavailability, biotransformation

Researchers at Huazhong Agricultural University began by asking whether the human gut microbiota as a whole could degrade licochalcone A. They recruited four healthy young Chinese men aged 24 to 27, excluding anyone with recent antibiotic use, probiotic consumption, gastrointestinal disorders, or metabolic disease, and collected fresh fecal samples that were processed entirely inside an anaerobic chamber to preserve the viability of oxygen-sensitive microbes. Each donor’s microbiota was then cultured in broth supplemented with licochalcone A and monitored for 48 hours. The results revealed striking interindividual variability: the most efficient community, from Donor A, degraded roughly 55 percent of the compound, while the least efficient, from Donor C, managed only about 27 percent. This spread matters, because it suggests that two people eating the same licorice-containing diet could experience substantially different exposure to the parent compound and its microbial derivatives.

The fermentation experiments produced a second, equally important observation. When licochalcone A was present, microbial biomass grew more vigorously than in controls, with optical density readings at 24 hours reaching 0.83 compared with 0.52 in untreated cultures, and the pH of the cultures dropped to around 5.5 versus roughly 6.4 in controls, a highly significant difference. Lower intestinal pH is generally considered a hallmark of a healthier microbial environment, favoring beneficial commensals while suppressing opportunistic pathogens. Gas chromatography–mass spectrometry confirmed the metabolic significance of this shift: concentrations of acetic, propionic, butyric, and isobutyric acids all rose significantly after 24 hours, with propionate and butyrate showing the most pronounced increases. These short-chain fatty acids are among the most consequential metabolites the gut microbiota produces, fueling colonocytes, reinforcing the intestinal barrier, enhancing insulin sensitivity, and exerting anti-inflammatory effects throughout the body.

Using liquid chromatography coupled with tandem mass spectrometry, the team then hunted for the chemical fingerprints of microbial transformation. Licochalcone A displays a protonated molecular ion at m/z 339.1; after 48 hours of anaerobic fermentation, a new ion appeared at m/z 355.2, a mass increase of 16 daltons that pointed to a single hydroxylation event. The new metabolite eluted slightly earlier than the parent compound, consistent with its increased polarity, and comparison with an authentic reference standard of 3-hydroxylicochalcone A confirmed matching retention time and fragmentation patterns. Remarkably, this metabolite, abbreviated 3-HLA, appeared consistently across all four donors, suggesting that the transformation is mediated by a conserved microbial capability rather than a quirk of one person’s microbiome. Because many dietary polyphenols must be converted by gut bacteria before they can reach systemic circulation, the identification of 3-HLA establishes microbial metabolism as an additional regulatory layer governing the bioavailability and physiological impact of licochalcone A.

With community-level degradation established, the researchers turned to the strain level. Seventeen bacterial species were randomly isolated from human fecal samples, deliberately without pre-selecting for degradation ability, and each was challenged with licochalcone A under anaerobic conditions. Degradation capacity varied enormously. Escherichia coli, Collinsella aerofaciens, Clostridium innocuum, and Clostridium butyricum all performed respectably, removing roughly 80 to 88 percent of the compound, while Enterococcus faecium, Enterococcus faecalis, and Bacteroides vulgatus managed less than 25 percent. But the standout was Bifidobacterium longum subsp. suillum, a subspecies never before implicated in natural product degradation and, according to the authors, never previously isolated from human feces in the context of chalcone metabolism. Strain AA-55, as the isolate is designated, degraded approximately 90 percent of licochalcone A within 48 hours, and its growth actually accelerated during the exponential phase when the compound was present.

To understand how this bacterium accomplishes the feat, the team sequenced its entire genome. The circular chromosome of strain AA-55 spans 2,250,570 base pairs with a G+C content of 59.76 percent and encodes 1,841 proteins, along with five CRISPR loci and a compact set of RNA genes. Functional annotation revealed substantial enrichment of genes associated with xenobiotic biodegradation and metabolism, secondary metabolite biosynthesis and catabolism, and oxidoreductase catalytic activity, all categories consistent with a capacity to chemically transform structurally diverse exogenous compounds. The genome also carries 71 carbohydrate-active enzymes, including 45 glycoside hydrolases and 14 glycosyltransferases. The authors are careful to note, however, that genomic annotation alone cannot prove which enzymes perform the hydroxylation; genetic knockouts, enzyme purification, and heterologous expression studies will be needed to pin down the exact molecular machinery.

Untargeted metabolomics added a dynamic dimension to the story. When cultures of strain AA-55 were exposed to licochalcone A and sampled at 0, 6, and 24 hours, principal component and discriminant analyses showed complete metabolic segregation between exposed and control groups, with the first two principal components explaining over 62 percent of the variance. At 6 hours, 227 metabolites were significantly upregulated and 582 downregulated; by 24 hours the figures were 280 and 675. Pathway enrichment centered on amino acid metabolism, nucleotide turnover, the TCA cycle, and ABC transporters, indicating that degrading licochalcone A is not an isolated enzymatic side reaction but a process woven into the bacterium’s broader metabolic reprogramming. Among the altered metabolites were glycoisoflavanone, licoagroside B, and deoxyphomalone, compound classes previously associated with biological activity, though their specific health effects remain uncharacterized.

Perhaps the most ecologically intriguing finding came from co-culture experiments. Akkermansia muciniphila and Bacteroides ovatus, two species of considerable interest in gut health research, could barely degrade licochalcone A on their own. But when the researchers grew them in culture supernatant harvested from strain AA-55 fermenting the compound, both species proliferated dramatically, with 2.3-fold and 3-fold increases in growth respectively compared with vehicle-conditioned medium. In other words, the primary degrader’s metabolic byproducts function as cross-feeding currency that sustains secondary consumers, echoing the well-documented prebiotic behavior of inulin and fructooligosaccharides, where fiber-fermenting pioneers feed the rest of the community. This positions licochalcone A not merely as a pharmacological flavonoid but as a potential prebiotic-like ecological modulator capable of reshaping gut community structure through cooperative metabolism.

In vivo validation in mice lent the laboratory findings physiological weight. Oral administration of licochalcone A at 50 milligrams per kilogram to conventional C57BL/6 mice, followed by 16S rRNA gene sequencing of fecal samples at baseline, 6, 24, and 48 hours, revealed transient shifts in microbial composition: Bacteroidota rose at 6 and 24 hours while Bacillota dipped early, and taxa including Dubosiella, Allobaculum, and Bifidobacterium increased in relative abundance at later time points. These results align with earlier reports that licochalcone A promotes beneficial taxa in models of colitis and type 2 diabetes, though the authors caution that the dose used here exceeds typical human dietary exposure and that the absence of a vehicle-gavage control leaves room for handling effects.

The study’s limitations are candidly acknowledged: the human cohort comprised only four young men, the co-culture system simplifies the gut’s ecological complexity, and the biological function of 3-HLA in the body remains to be demonstrated. Even so, the work marks a genuine advance, revealing for the first time that bifidobacteria can drive chalcone biotransformation and identifying a specific strain whose metabolic appetite may determine how much active compound, and which derivatives, a licorice lover actually absorbs. As interest in microbiota-targeted functional foods accelerates, knowing which gut residents process licochalcone A, and how their activity ripples through the microbial food web, could prove essential for designing interventions that harness licorice’s ancient therapeutic promise with modern precision.

Subject of Research: Microbial degradation and biotransformation of the licorice flavonoid licochalcone A by the human gut microbiota

Article Title: Degradation and biotransformation of Licochalcone A by the human gut microbiota: Role of Bifidobacterium longum subsp . Suillum AA-55

Article References: Degradation and biotransformation of Licochalcone A by the human gut microbiota: Role of Bifidobacterium longum subsp . Suillum AA-55. (n.d.). https://doi.org/10.1016/j.jafr.2026.103299

Image Credits: AI Generated

DOI: 10.1016/j.jafr.2026.103299

Keywords: licochalcone A, gut microbiota, Bifidobacterium longum, licorice, 3-hydroxylicochalcone A, biotransformation, short-chain fatty acids, cross-feeding, polyphenols, metabolomics, prebiotic, functional foods

The research team, led by Imane E.L. Houssni, screened 1,055 records retrieved from Scopus, Web of Science, PubMed and ScienceDirect and retained 388 high-rigor studies after a PRISMA-guided triage that eliminated uncharacterized adsorbents, superficially descriptive biocontrol assays and reports lacking mass-balance verification. Inter-reviewer agreement was high, with a Cohen’s kappa of 0.88. The resulting synthesis organizes the field around a four-pillar framework: upstream inhibition of toxigenic fungi, passive physical sequestration, active enzymatic biotransformation, and downstream protection of the host’s intestinal barrier.

The first mechanistic pillar, biosorption, is deceptively simple. Toxin molecules bind to the microbial cell wall, a scaffold of peptidoglycan, teichoic acids, beta-glucans and surface-layer proteins, through electrostatic attraction, hydrophobic intercalation and steric entrapment. Crucially, this process does not require living cells. Heat-killed paraprobiotics work just as well, sidestepping risks of horizontal gene transfer in processed foods. Spectroscopic and microscopy studies show that peptidoglycan cross-linking and the charge density of teichoic acids dictate binding stoichiometry, while yeast beta-glucan triple helices form high-affinity pockets that capture zearalenone through hydrophobic insertion of its macrocyclic lactone ring.

Yet the review is blunt about biosorption’s Achilles heel: reversibility. Binding equilibria that look impressive in a static flask can collapse in the gut, where pH swings from gastric acidity to intestinal neutrality, bile salts micellize lipophilic toxins, digestive enzymes chew up cell-wall proteins, and peristaltic shear pries complexes apart. The authors argue that static isotherm parameters such as maximum adsorption capacity cannot be extrapolated to prove in vivo sequestration; definitive validation demands dynamic multi-compartmental simulators like TIM-1 and SHIME, plus fecal recovery assays.

The second pillar, enzymatic biotransformation, offers a permanent solution by chemically remodeling toxic pharmacophores. Lactonohydrolases of the ZHD102 family cleave zearalenone’s macrocyclic ester into hydrolyzed ZEN, a product devoid of estrogenic activity. Anaerobic gut commensals such as Eubacterium BBSH 797 express epoxide reductases that convert deoxynivalenol into the non-toxic DOM-1, disarming the molecule’s ability to inhibit ribosomal protein synthesis. For ochratoxin A, metalloproteases cleave the amide bond linking the phenylalanine moiety to the isocoumarin nucleus, yielding ochratoxin alpha, which is negligibly toxic and rapidly cleared by the kidneys. Fumonisin B1 detoxification proceeds in two steps: a carboxylesterase such as FumD strips the tricarballylic acid side chains, then an aminotransferase deaminates the C-2 amine, abolishing the toxin’s interference with ceramide synthase.

But the review insists on a critical caveat: biotransformation is not automatically detoxification. Some conversions backfire spectacularly. Zearalenone can be reduced to alpha-zearalenol, which binds estrogen receptors more avidly than the parent toxin. Genuine detoxification therefore requires full structural identification of every degradation product by high-resolution mass spectrometry or NMR, complete mass-balance closure, and bioassay-driven toxicological validation to rule out masked or paradoxically toxic intermediates.

The third and most surprising pillar reframes probiotics as biological shields rather than mere toxin sponges. Administered strains fortify the intestinal epithelium by upregulating tight junction proteins, including zonula occludens-1, occludin and the claudins, while suppressing the pathological hyperactivation of p38 MAPK and NF-kappa-B that deoxynivalenol and ochratoxin A trigger through ribotoxic stress. They also stimulate goblet cells to thicken the MUC2 mucus layer, promote secretory IgA secretion from gut-associated lymphoid tissue, and activate the host’s Nrf2 antioxidant pathway to restore the glutathione redox ratio. Perhaps most strikingly, postbiotic metabolites such as butyrate act as histone deacetylase inhibitors, a postbiotic-epigenetic bridge that remodels chromatin and transcriptionally reprograms barrier-restoring genes.

In the food industry, these mechanisms are converging on the concept of smart starters, bifunctional cultures that ferment the product while scrubbing it clean. Multi-strain lactobacilli consortia have achieved aflatoxin M1 reductions exceeding 60 percent in yogurt models, Lactiplantibacillus plantarum strains bind 65 to 85 percent of aflatoxin B1 in sourdough matrices, and the yeast Kluyveromyces marxianus removes 85 to 95 percent of zearalenone while enriching products with aromatic esters and B vitamins. The challenge is preserving taste and texture: over-acidification, off-flavors and proteolytic degradation can sink consumer acceptance regardless of molecular efficacy, and matrix components like milk fat can physically shield toxins from microbial contact.

The review also confronts the field’s methodological blind spots. Two strains of the same species can differ in binding capacity from under 10 percent to over 80 percent, real-world contamination involves multi-toxin cocktails that saturate binding sites, and masked mycotoxins such as deoxynivalenol-3-glucoside evade screening until gut hydrolases release the parent toxin in situ. Human trials remain scarce and often rely on indirect urinary biomarkers, while responder-versus-non-responder microbiome heterogeneity complicates regulatory validation by EFSA and the FDA.

Looking forward, the authors sketch a roadmap toward designer microbes: CRISPR-engineered chassis with densified adsorptive cell walls and heterologous detoxification enzymes, programmable biosensors that secrete enzymes only when luminal toxin spikes are detected, artificial intelligence-driven molecular docking to predict binding energies, gut-on-a-chip platforms and human intestinal organoids for physiologically faithful validation, and pH-responsive nano-encapsulation that delivers payloads exactly where absorption risk peaks. The smart starter concept currently sits at low technology readiness levels, but the trajectory is clear. Fermented foods, once viewed as a passive vector of dietary risk, are being reimagined as precision instruments of xenobiotic resilience, turning the trillions of microbes in our food and gut into an active, engineered line of defense against one of the oldest poisons in the human diet.

Subject of Research: Microbial mechanisms of mycotoxin detoxification in fermented foods and the gastrointestinal tract

Article Title: Microbial mycotoxin mitigation across the microbe–matrix–host continuum: mechanisms, fermented-food applications, and gastrointestinal translation

Article References: HOUSSNI, I. E., ZAHIDI, A., FAKIR, S. E., JAIMA, W., & HASSIKOU, R. (2026). Microbial mycotoxin mitigation across the microbe–matrix–host continuum: mechanisms, fermented-food applications, and gastrointestinal translation. Current Research in Biotechnology, Article 100421. https://doi.org/10.1016/j.crbiot.2026.100421

Image Credits: AI Generated

DOI: 10.1016/j.crbiot.2026.100421

Keywords: mycotoxins, probiotics, biosorption, biotransformation, lactic acid bacteria, fermented foods, gut barrier, aflatoxin, zearalenone, ochratoxin A, smart starters, food safety

The substrate in question, ent-pimara-8(14),15-dien-19-oic acid, is a pimarane-type diterpene carboxylic acid, a molecule built around a rigid tricyclic core decorated with two carbon-carbon double bonds, an internal one at position 8(14) and an external vinyl group at position 15. When the researchers added this compound to growing liquid cultures of the fungus, the organism responded by transforming it into four oxidized products. Two of these, 15,16-epoxy-ent-pimar-8(14)-en-19-oic acid and 15,16-dihydroxy-ent-pimar-8(14)-en-19-oic acid, were known compounds, but the other two, 17-hydroxy-15,16-epoxy-ent-pimar-8(14)-en-19-oic acid and 15-oxo-16-hydroxy-ent-pimar-8(14)-en-19-oic acid, had never been described before. In every case, the fungus attacked the vinyl group and left the internal double bond intact.

That selectivity is what makes the result remarkable. Monosubstituted double bonds like the vinyl group are typically more reactive toward chemical oxidants, yet when chemists treated a related pimaradiene with the classic epoxidation reagent meta-chloroperbenzoic acid, they obtained only epoxides at the internal 8(14) position along with rearranged byproducts, the opposite outcome. Previous microbial studies pointed the same way: a strain of Aspergillus niger oxidized the tricyclic ring system of the same substrate at carbons 1, 6, 7, 11 and the internal double bond while ignoring the vinyl group, and the endophytic fungus Preussia minima also hydroxylated positions away from the vinyl moiety. The new work shows that A. sclerotiorum does essentially the reverse, a biotransformation pattern the authors describe as at least uncommon.

To probe whether this apparent regioselectivity was real, the team challenged the fungus with a second diterpene, abietic acid, a resin acid bearing two endocyclic double bonds at positions 7(8) and 13(14). If the fungus simply preferred external double bonds, abietic acid, which has none, should be metabolized differently. Instead, the organism oxidized a methyl group attached at carbon 13, outside the ring system, producing 16-hydroxyabieta-7(8),13(14)-dien-19-oic acid. This mirrors what it did with the pimarane substrate, where the vinyl group and the methyl group both hang off carbon 13, suggesting the enzymes recognize the spatial region around that carbon rather than a particular bond type.

The structural detective work relied on one- and two-dimensional nuclear magnetic resonance spectroscopy and high-resolution mass spectrometry. In the epoxide product, the characteristic ABX spin system of vinylic hydrogens near 5.4 and 4.9 parts per million in the proton spectrum vanished, replaced by new signals between 2.5 and 2.8 parts per million, while carbon-13 signals for the sp2 carbons at 147.2 and 112.9 parts per million gave way to aliphatic signals at 58.9 and 42.5. For the diol product, the carbon-15 resonance appeared at 80.4 parts per million, a deshielded position that literature correlations assign to the 15S configuration. Because an epoxide hydrolase opening the epoxide by an SN2-type attack at carbon 16 would yield only one epimer, whereas an SN1 mechanism through a carbocation would yield both, the researchers could infer that the epoxide products share that same 15S configuration. The hydroxyketone product, carrying a carbonyl signal at 214.8 parts per million, appears to arise from further oxidation of the diol’s C-15 hydroxyl group.

Hunting for the enzyme behind this chemistry, the team scanned publicly available whole-genome shotgun contigs of A. sclerotiorum using tblastn, searching for cytochrome P450 sequences similar to CYP105A1 from Streptomyces griseolus, a bacterial enzyme previously shown to oxidize the C-15 isopropyl group of abietic acid and to epoxidize the vinyl group of an isopimarane diterpene. The search returned two candidate open reading frames, one of 1083 nucleotides encoding 360 amino acids and another of 1641 nucleotides encoding 546 amino acids, with sequence identities of 25 and 24 percent and query coverages of 37 and 35 percent respectively. Crucially, both proteins carry the canonical P450 signature motif FxxGxxxCxG, including the cysteine residue that binds the heme iron, confirming they belong to the cytochrome P450 superfamily.

To visualize how such an enzyme might steer the substrate, the researchers turned to molecular docking using the crystal structure of a bacterial cytochrome P450 as a proxy, since no structure exists for the Aspergillus enzyme. In the docked complex, the diterpene sits in the active site with its vinyl group and the C-13 methyl group positioned closest to the heme catalytic center, exactly where an oxygenating iron-oxo species would deliver its atom. Basic amino acids, including arginine and asparagine, appear to anchor the molecule through hydrogen bonding and electrostatic interactions with the carboxylic acid at carbon 4, holding the substrate in the orientation that favors oxygenation at the side chain rather than the ring system. The docking picture closely resembles calculations published for CYP105A1 acting on abietic acid, reinforcing the idea that a similar P450 in the fungus drives the observed regioselectivity.

The practical payoff came from antiparasitic testing. The parent diterpene and its four derivatives were evaluated against promastigote forms of Leishmania amazonensis and epimastigote forms of Trypanosoma cruzi, the parasites responsible for cutaneous leishmaniasis and Chagas disease, both neglected tropical diseases that affect millions of people in Latin America and beyond. Oxidation consistently improved activity. Against L. amazonensis, the epoxide and the hydroxyketone derivatives showed IC50 values less than half that of the parent compound. Most strikingly, the hydroxyketone inhibited the trypomastigote stage of T. cruzi with an IC50 of 21.4 micromolar, slightly better than the reference drug benznidazole, which came in at 34.5 micromolar under the same assay conditions. Equally important, cytotoxicity assays against healthy Vero cells showed very low toxicity, particularly for the hydroxyketone, yielding favorable selectivity indices.

The structure-activity picture that emerges suggests that chemical modifications at the vinyl side chain matter more for antiparasitic potency than variations on rings A and B of the ent-pimarane skeleton. Earlier studies had found that hydroxylation near ring A did not substantially enhance activity against another T. cruzi strain, while ent-pimaranes modified at ring B showed IC50 values in the 15 to 20 micromolar range. The new derivatives, functionalized precisely at the side chain, now extend that trend and hint that the side chain is a promising handle for drug design.

Beyond the immediate antiparasitic results, the study validates a broader strategy: mining endophytic fungi from diterpene-producing plants for enzymes that recognize the very molecules their hosts biosynthesize. The intimate evolutionary relationship between endophytes and plant biochemistry may predispose their enzymatic machinery to accept plant terpenes as substrates, and expanding the pool of such microorganisms helps overcome the classic unpredictability of whole-cell biotransformation. With two new diterpenes in hand, a candidate P450 enzyme identified, and a docking model that explains the selectivity, the researchers have laid out a complete chain of evidence from genome to molecule to bioactivity, offering a template for turning humble tree-dwelling fungi into factories for medicinally relevant chemistry.

Subject of Research: Regioselective epoxidation of a plant diterpene by the endophytic fungus Aspergillus sclerotiorum and the antiparasitic activity of the resulting oxidized products.

Article Title: Epoxidation of ent-pimara-8(14),15-dien-19-oic acid by whole cells of the endophytic fungus Aspergillus sclerotiorum

Article References: Din, Z. U., de Medeiros, L. S., Abreu, L. M., Lazarin-Bidóia, D., Scariot, D. B., Garcia, F. P., de Paula, J. C., Nakamura, C. V., Fill, T. P., & Rodrigues-Filho, E. (2026). Epoxidation of ent-pimara-8(14),15-dien-19-oic acid by whole cells of the endophytic fungus Aspergillus sclerotiorum. International Microbiology. https://doi.org/10.1007/s10123-026-00895-0

Image Credits: AI Generated

DOI: 10.1007/s10123-026-00895-0

Keywords: Aspergillus sclerotiorum, biotransformation, diterpene, ent-pimaradienoic acid, epoxidation, cytochrome P450, endophytic fungus, Cupressus lusitanica, Leishmania amazonensis, Trypanosoma cruzi, antiparasitic activity, regioselectivity