Sweetpotato Leaves Revealed as Potent Fat-Blocking Compounds Vary by Variety

The sweetpotato is one of humanity’s most dependable staple crops, with global production exceeding 90 million tonnes a year. China dominates that harvest, growing 56.7 percent of the world’s supply and maintaining the largest germplasm database on record, with 423 registered varieties as of 2023. Yet for all the attention devoted to the starchy storage roots, the plant’s aerial parts—the leaves and stems—remain an enormous, largely wasted resource. Farmers typically discard this biomass after harvest, even though it is packed with polyphenols, anthocyanins, caffeic acid derivatives, and a family of unusual glycolipids called resin glycosides that have recently emerged as surprisingly powerful inhibitors of pancreatic lipase, the enzyme that breaks down dietary fat in the human gut.

A new study published in the Journal of Agriculture and Food Research has now mapped this hidden chemical diversity across 22 commonly cultivated Chinese sweetpotato varieties, and the results reveal a striking pattern: the anti-obesity potential of sweetpotato vines depends dramatically on which cultivar you grow. The research team, led by Qingtong Xie and Dejian Huang of the National University of Singapore together with collaborators in China, obtained fresh aerial parts from the National Sweetpotato In Vitro Genebank at the Xuzhou Institute of Agricultural Sciences. The 22 accessions spanned four functional categories—leafy vegetable varieties, fresh-consumption types, starch-processing cultivars, and anthocyanin-rich functional food lines—including widely planted staples such as Xushu 18, which alone covered 1.97 million acres in 2015.

The researchers freeze-dried and ground the entire above-ground biomass, then extracted the resin glycosides using dichloromethane followed by methanol dissolution and silica gel column chromatography. Extraction yields ranged from 0.42 to 0.87 percent of the dried biomass, with the ornamental and leafy variety Xucaiguanshu 1 producing the highest yield. To measure biological activity, the team turned to a standard pancreatic lipase assay based on p-nitrophenyl palmitate, a synthetic substrate whose cleavage releases a yellow product measurable at 410 nanometers. By tracking the initial rate of product formation in the presence and absence of each extract, the researchers calculated the concentration needed to inhibit half of the enzyme’s activity, the IC50 value, and expressed each sample’s potency as an orlistat equivalence—the amount of the pharmaceutical lipase blocker orlistat that would deliver the same inhibition per microgram of sample.

The spread in activity was remarkable. Yanshu 25, a fresh-consumption variety, was the clear champion, inhibiting lipase with an orlistat equivalence of 1.599 nanograms per microgram of extract and an IC50 of just 17.9 micrograms per milliliter. The starch cultivars Xushu 41 and Xushu 44 followed closely behind. At the other end of the spectrum, Xushu 22 and Sushu 8 were roughly four times weaker, with IC50 values above 53 micrograms per milliliter. To put these numbers in context, the best sweetpotato extracts outperformed many previously reported plant-derived lipase inhibitors, including carpesterol from Solanum stramonifolium and bamboo flavonoids, and even edged out resin glycosides from water spinach, a relative in the same plant family. Intriguingly, the researchers found no correlation between a cultivar’s intended agricultural use and its lipase-blocking power, a reminder that agronomic traits like starch content and phytochemical traits like resin glycoside production are governed by largely independent genetic networks.

To understand why some varieties are so much more potent, the team selected the three strongest and three weakest performers and profiled their resin glycosides using ultra-high-performance liquid chromatography coupled to quadrupole time-of-flight mass spectrometry in negative ionization mode. Against a reference library of 244 known resin glycosides, they tentatively annotated 106 glycoside-related signals, spanning one trisaccharide, 24 tetrasaccharide, and 81 pentasaccharide compounds. Sushu 8 showed the least chemical diversity with 87 compounds detected, while Xushu 44 topped the list with 105. Principal component analysis of the full profiles revealed four distinct chemical groupings among the cultivars but, crucially, no clean separation between high- and low-activity samples—strong evidence that overall composition matters less than the presence of specific molecular structures.

That hypothesis was confirmed by a supervised statistical approach. Orthogonal partial least squares-discriminant analysis cleanly segregated the high- and low-activity groups with excellent fit and predictive power, and flagged 16 differential resin glycoside features. Fifteen of these were enriched in the potent cultivars, including leptophyllin A, woodrosin I, batatoside J, and operculin XV, while a single compound, intrapilosin III, predominated in the weak performers. All of the differential molecules share a characteristic architecture: a sugar core of four or five units built from glucose, rhamnose, and fucose, linked to jalapinolic acid, an 11-hydroxyhexadecanoic acid that cyclizes with the sugar to form a signature macrolactone ring. The remaining hydroxyl groups carry variable fatty acyl decorations—2-methylbutanoyl, decanoyl, dodecanoyl, cinnamoyl, tigloyl, and others—that give the family its extraordinary structural variety.

Three acyl substituents stood out among the active compounds. The 2-methylbutanoyl group appeared in 11 of the 15 high-activity markers, consistent with earlier work showing that swapping this group for tiglate in related molecules shifts their predicted binding away from lipase’s known sites. Dodecanoyl, a medium-chain fatty acid, appeared in six markers and had previously been implicated in competitive lipase inhibition by purified pescaprein III. Most intriguingly, the cinnamoyl group showed up in five markers—the first time this aromatic substituent has been linked to lipase inhibition in sweetpotato aerial parts. The connection is chemically plausible, since cinnamic acid itself inhibits pancreatic lipase with an IC50 of about 39 micrograms per milliliter and reduced serum lipase activity in high-fat-diet-fed rats, while caffeoylquinic acids block the enzyme by hydrogen-bonding to its catalytic triad of serine, aspartate, and histidine residues.

Computer simulations added a structural dimension to these correlations. Docking leptophyllin A, the top statistical marker, into human pancreatic lipase yielded a favorable binding energy of −6.555 kilocalories per mole and revealed hydrogen bonds between the sugar moiety and six amino acid residues, including Thr292, Arg339, and Asp387. The fatty acyl chains made extensive hydrophobic contacts with surrounding residues, and notably, the molecule bound outside the enzyme’s catalytic center—consistent with the noncompetitive inhibition pattern previously observed for sweetpotato resin glycoside mixtures. Interestingly, related resin glycosides from Calystegia hederacea that carry only short, branched acyl chains are predicted to bind within the catalytic site instead, suggesting that different acylation patterns may steer these molecules toward distinct inhibitory mechanisms, several of which can converge on potent activity.

As a practical screening shortcut, the team also hydrolyzed the resin glycosides with hot alkali to release their acyl components and analyzed the products by gas chromatography–mass spectrometry. The two most active cultivars, Yanshu 25 and Xushu 44, contained the highest levels of decanoic acid, dodecanoic acid, and cinnamic acid, while the weak performers had far less—though Xushu 41 broke the pattern, apparently compensating with an abundance of muricatin and obrizabin-type glycosides instead. The authors caution that their assay is a rapid in vitro screen, and that cell-based and animal studies, plus work with purified individual compounds, will be needed to confirm physiological relevance. Still, the study delivers an actionable message: millions of tonnes of discarded sweetpotato vines are not uniform waste but a variable reservoir of natural fat-blocking molecules, and choosing the right variety—Yanshu 25 at the top of the list—could turn an agricultural by-product into a functional food ingredient for fighting obesity.

Subject of Research: Cultivar-dependent resin glycoside profiles and pancreatic lipase inhibitory activity in sweetpotato aerial parts

Article Title: Unraveling Cultivar-Dependent Resin Glycoside Profiles and Pancreatic Lipase Inhibitory Activity in Aerial Parts of Commonly Cultivated Sweetpotatoes ( Ipomoea batatas ) in China

Article References: Xie, Q., Lin, Y., Song, Z., Lu, Y., Shi, X., Cao, Q., Liu, T., & Huang, D. (2026). Unraveling Cultivar-Dependent Resin Glycoside Profiles and Pancreatic Lipase Inhibitory Activity in Aerial Parts of Commonly Cultivated Sweetpotatoes (Ipomoea batatas) in China. Journal of Agriculture and Food Research, Article 103354. https://doi.org/10.1016/j.jafr.2026.103354

Image Credits: AI Generated

DOI: 10.1016/j.jafr.2026.103354

Keywords: sweetpotato, resin glycosides, pancreatic lipase, anti-obesity, Ipomoea batatas, cultivar variation, UHPLC-QTOF-MS, molecular docking, functional foods, food waste valorization, natural inhibitors, China