Borneo Fruit Reveals a Team Effort Against Fat-Digesting Enzyme

A little-known tropical fruit from the rainforests of Malaysian Borneo may hold a surprisingly sophisticated recipe for blocking fat digestion, and the secret, according to a new study, is not any single miracle compound but a coordinated team of molecules working together. Researchers led by Ahmad Iqbal Noordin and Mohd Fadzelly Abu Bakar of Universiti Tun Hussein Onn Malaysia, working with colleagues at Universiti Malaysia Sabah, Prince of Songkla University, and Universiti Putra Malaysia, have combined nuclear magnetic resonance spectroscopy, enzyme assays, and computer simulations to dissect how extracts of Garcinia parvifolia shut down pancreatic lipase, the enzyme that breaks down dietary fat in the human gut. The work, published open access in the journal Metabolomics, offers a textbook example of how modern metabolomics can move beyond the hunt for a single active ingredient and instead map the collective chemistry underlying a plant’s medicinal reputation.

Pancreatic lipase sits at the center of one of the most popular pharmacological strategies for weight management. The enzyme hydrolyzes triglycerides from food into absorbable fatty acids and monoacylglycerides, and blocking it reduces the fraction of dietary fat the body actually takes in. The only drug approved for this purpose, orlistat, works by forming a stable covalent bond with the catalytic residue serine 152 of the enzyme, permanently disabling it while its long hydrocarbon tail wedges into the enzyme’s hydrophobic groove, mimicking a triglyceride substrate. Although effective, orlistat is associated with gastrointestinal side effects, and pharmacovigilance analyses of adverse event databases have kept interest alive in gentler, plant-derived alternatives. That is where Garcinia species, famous for their polyphenols and xanthones, enter the picture, and where the lesser-known G. parvifolia, traditionally used in indigenous medicine across Southeast Asia, has remained almost entirely unstudied with respect to fat digestion.

The research team harvested fully ripe fruits in July 2024 from Penampang in Sabah, collecting samples from seven independent trees of comparable age to capture biological variability, with permits from the Sabah Biodiversity Centre and voucher specimens verified at the Borneensis Herbarium of Universiti Malaysia Sabah. The fruit peels were dried at 40 degrees Celsius, ground, sieved, and then extracted with four solvents of escalating polarity: 70 percent ethanol, acetone, chloroform, and hexane. The logic was straightforward. Because solvent polarity determines which classes of metabolites dissolve, comparing extracts across the polarity gradient would reveal whether the anti-lipase activity tracks with polar phenolics, lipophilic terpenes, or something in between. Ultrasonic-assisted extraction, which can reach efficiencies of 85 to 97 percent under mild conditions, was used to pull the chemistry out of the powdered peel.

The enzyme assays delivered a clear verdict. The 70 percent ethanol extract inhibited pancreatic lipase most strongly, with a half-maximal inhibitory concentration of 85.3 micrograms per milliliter, followed by acetone at 106.4 micrograms per milliliter, while the chloroform and hexane extracts performed far worse, never reaching fifty percent inhibition even at the highest tested concentration of 100 micrograms per milliliter. Orlistat, the positive control, inhibited the enzyme by 83 percent at just 25 micrograms per milliliter, with an IC50 of 17.5 micrograms per milliliter, validating the assay. The pattern pointed squarely at polar chemistry: whatever was blocking the enzyme dissolved best in aqueous ethanol, not in the greasy non-polar solvents. Inhibition rose steadily with concentration across all extracts, but the polar extracts dominated at every dose.

To see which molecules were responsible, the team turned to proton nuclear magnetic resonance spectroscopy, acquiring spectra on a 500 megahertz instrument and processing them with standardized binning across the chemical shift range of 0.5 to 10 parts per million. In total, thirty-four metabolites were annotated, drawing on the Chenomx database, the Human Metabolome Database, and two-dimensional J-resolved spectra to confirm assignments. The ethanol extract was by far the richest, showing intense signals from sugars such as glucose and xylose, amino acids, organic acids, and, critically, phenolic compounds. Distinctive peaks at 3.95, 4.60, and 6.99 parts per million were assigned to catechin, a flavonoid with a well-known track record against digestive enzymes, while signals in the downfield aromatic region hinted at phenolic acids such as protocatechuic and gallic acid, both common in Garcinia species. The chloroform and hexane spectra were comparatively barren, especially in the 2.5 to 5.5 parts per million region where hydrophilic metabolites cluster.

The statistical machinery of metabolomics then connected chemistry to biology. Principal component analysis separated the four extracts cleanly along the first principal component, which alone explained 89.1 percent of the variance and tracked the gradient from polar to non-polar constituents; together the first two components accounted for 97 percent of the total. Polar metabolites such as glucose, fructose, succinate, betaine, choline, and various amino acids loaded strongly on the positive side, where the ethanol and acetone extracts clustered, while the sterol beta-sitosterol and the sesquiterpene beta-caryophyllene associated with the non-polar side. A partial least squares regression then linked the spectral data directly to measured lipase inhibition, pairing each of the 28 spectral profiles with its own non-averaged inhibition value. The model performed impressively, with cumulative R-squared values of 0.97 for the X matrix and 0.967 for the Y variable, and a predictive Q-squared of 0.963. A 100-iteration permutation test confirmed the model was not overfitted, with all permuted values falling below the original and a negative Q-squared intercept of minus 0.266, a strong indicator of predictive validity.

The variable importance in projection analysis drove home the study’s central message: no single compound dominated. Catechin, catechol, mandelate, and indole-3-lactate scored high, but so did amino acids and derivatives such as aspartate, beta-alanine, and trans-4-hydroxy-L-proline, along with organic acids like succinate and isocitrate. In the PLS biplot, the ethanol and acetone extracts sat closest to the lipase inhibition vector, and the metabolites aligned with that vector formed a broad coalition of phenolics, aromatics, and primary metabolites rather than a lone star. The authors interpret this as evidence of additive or synergistic interaction: the primary metabolites are not direct inhibitors themselves, but their abundance signals a metabolically rich matrix that may amplify the effects of the true inhibitors, while betaine and choline, both linked to lipid metabolism regulation, may play supportive secondary roles.

Molecular docking with AutoDock Vina against the pancreatic lipase crystal structure provided a mechanistic coda. Catechin bound with an affinity of minus 6.09 kilocalories per mole, forming a network of hydrogen bonds with the catalytic triad residues serine 152 and histidine 263, effectively capping the active site and stabilizing a non-functional conformation of the enzyme. Beta-sitosterol, despite a slightly more favorable score of minus 6.34 kilocalories per mole, relies almost entirely on van der Waals forces, wedging into the greasy substrate channel and physically blocking triglyceride entry through steric hindrance. Beta-caryophyllene, at minus 5.19 kilocalories per mole, binds only hydrophobically and, being smaller, makes fewer contacts. The docking results align neatly with the wet-lab data: the lipophilic hexane extract, rich in sitosterol-type chemistry, performed poorly in the aqueous assay buffer, while the catechin-rich polar extracts excelled, suggesting that reversible hydrogen bonding to the catalytic machinery matters more in this system than non-specific hydrophobic plugging.

The authors are candid about the limits of the work. Bioactivity was assessed only in vitro and in silico, with no cellular or animal validation; docking was restricted to three representative metabolites rather than the full set of top-ranked variables; metabolite identification was qualitative rather than quantitative; and the samples were oven-dried rather than lyophilized. Docked complexes were not subjected to molecular dynamics simulations to test pose stability, and pure standards were not used to confirm the docking predictions. Even so, the study stands as a compelling demonstration that the anti-obesity potential of G. parvifolia is governed by a complex, multi-component synergy, and it positions this underutilized Bornean fruit as a candidate for functional foods and nutraceutical formulations. The researchers call for in vivo studies and clinical trials to establish safety, bioavailability, and real-world efficacy, but the broader lesson is already clear: when it comes to plant-based enzyme inhibition, the whole phytochemical orchestra, not just the first violin, may be what makes the music.

Subject of Research: NMR-based metabolomics of synergistic pancreatic lipase inhibition by Garcinia parvifolia metabolites

Article Title: Synergistic lipase inhibition by Garcinia parvifolia metabolites: a 1H NMR–based metabolomic deciphering of ligand interactions

Article References: Noordin, A. I., Abu Bakar, F. I., Saikim, F. H., Wichienchot, S., Isha, A., & Abu Bakar, M. F. (2026). Synergistic lipase inhibition by Garcinia parvifolia metabolites: a 1H NMR–based metabolomic deciphering of ligand interactions. Metabolomics, 22(5), Article 167. https://doi.org/10.1007/s11306-026-02536-z

Image Credits: AI Generated

DOI: 10.1007/s11306-026-02536-z

Keywords: Garcinia parvifolia, pancreatic lipase, metabolomics, NMR spectroscopy, molecular docking, catechin, anti-obesity, natural products, phenolic compounds, functional foods, Malaysian Borneo, synergy