A new study published in Food Chemistry: X offers a surprisingly elegant solution, and it comes from a part of the peanut that manufacturers routinely throw away. A research team led by Yuan Gao and Mingqing Wang of Shandong Academy of Agricultural Sciences turned its attention to the peanut embryonic axis, sometimes called the peanut heart. This tiny structure, which contains the hypocotyl, radicle, and plumule and represents only about 2.5 percent of the total kernel mass, is often removed during commercial processing because of its bitter taste. While peanut skins and hulls have attracted considerable attention as sources of valuable phenolics and proanthocyanidins, the embryonic axis has remained largely unexplored territory, despite earlier reports that it contains lower levels of the major allergen mRNAs Ara h 1 and Ara h 3 and higher concentrations of minerals such as magnesium, phosphorus, potassium, calcium, iron, and zinc compared with the cotyledon.

To understand what makes the embryonic axis chemically distinctive, the researchers conducted a comprehensive comparison between the axis and the cotyledon, the fleshy tissue that constitutes the bulk of the peanut kernel and the main raw material for peanut butter. Proximate analysis revealed meaningful compositional differences. The cotyledon contained 46.90 percent fat, whereas the embryonic axis held only 39.35 percent, a reduction of more than seven percentage points. The axis also carried more crude fiber, at 8.07 percent versus 6.31 percent in the cotyledon, and more carbohydrate, at 22.7 percent versus 16.4 percent. Protein content, by contrast, showed no significant difference between the two tissues. This profile of lower fat and higher fiber immediately suggested that the axis could serve as a functional ingredient capable of improving both the nutritional profile and the physical behavior of peanut butter.

The most striking findings emerged from untargeted metabolomics performed with ultra-performance liquid chromatography coupled to tandem mass spectrometry. Using a Thermo UHPLC-Q Exactive system operating in both positive and negative electrospray ionization modes, the team identified 419 metabolites that were significantly up-regulated and 201 that were down-regulated in the embryonic axis relative to the cotyledon. Among the plant secondary metabolites driving this separation, terpenoids dominated, accounting for more than half of the differential compounds, with triterpenoids forming the largest subgroup. Of the top 25 metabolites distinguishing the two tissues, 20 were triterpenoids, and all were markedly more abundant in the embryonic axis. These included compounds with well-documented biological activities, such as beta-elemonic acid, momordicinin, and maslinic acid, which have been associated with anti-cancer and anti-inflammatory effects, along with ginsenoside I, ginsenoside Rh7, and several soyasaponins known for antioxidant and antimicrobial properties.

Armed with this chemical portrait, the researchers formulated peanut butter in which roasted, peeled peanut kernels were partially replaced by embryonic axis at substitution levels of 2.5, 5, 7.5, and 10 percent by weight. The control spread was prepared from 100 percent roasted kernels with no added salt, sugar, oil, or stabilizers, ensuring that any observed effects could be attributed to the axis itself. Proximate analysis of the finished products showed that fat content declined progressively from 51.71 percent in the control to 49.23 percent at the 10 percent substitution level, while crude fiber rose significantly from 6.73 percent to 7.43 percent and carbohydrate climbed to 16.0 percent at the highest inclusion. Protein remained essentially unchanged across all formulations, hovering between 23.7 and 23.9 percent. Because lipid oxidation begins with the substrate itself, reducing the pool of oxidizable fat inherently slows the initiation and propagation of peroxidation reactions.

Texture and rheology measurements revealed that the embryonic axis did more than simply dilute the fat. Hardness increased from 0.79 newtons in the control to 0.94 newtons at 10 percent substitution, while adhesiveness rose from 0.96 to 1.51 newton-millimeters, and springiness, gumminess, and chewiness all climbed significantly as well. Rheological testing showed that all samples behaved as pseudoplastic, shear-thinning fluids, with apparent viscosity decreasing as shear rate increased from 0.1 to 100 reciprocal seconds. The flow curves fitted the power-law model with coefficients of determination between 0.989 and 0.996, and the consistency index rose from 37.29 pascal-seconds to the power of n in the control to roughly 50 at the two highest substitution levels. The researchers attribute this firmer, more cohesive structure to two mechanisms: the reduced free oil that would otherwise act as a lubricant, and the long molecular chains of dietary fiber, which intertwine into a three-dimensional network that binds oil and restricts its movement through the matrix.

The antioxidant evidence was equally compelling. Radical scavenging assays showed that spreads containing 7.5 and 10 percent embryonic axis displayed significantly higher activity against both DPPH and ABTS free radicals than the control, and the 10 percent formulation outperformed the 2.5 and 5 percent versions in the DPPH assay. Spearman correlation analysis confirmed that fiber content was strongly and positively associated with radical scavenging, with correlation coefficients of 0.903 for DPPH and 0.888 for ABTS. The enriched triterpenoid fraction likely contributed as well, since compounds such as maslinic acid, ginsenosides, and soyasaponins have demonstrated free radical neutralization and shelf-life extension in other food systems. Although roasting can partially degrade certain heat-labile saponins, the authors note that degradation is compound-specific and unlikely to have eliminated the overall antioxidant capacity of the enriched spreads.

Storage performance under accelerated conditions provided the decisive test. Samples packed in PET jars were held at 37 degrees Celsius for four weeks and monitored for oil separation and peroxide value. At the outset, all axis-fortified spreads showed markedly lower oil separation than the control, and a clear dose-dependent relationship emerged, with higher substitution levels corresponding to less oil loss over time. The researchers suggest that surface-active triterpenoid saponins may lower interfacial tension and form stable viscoelastic films that inhibit the coalescence and migration of oil droplets, while the denser fiber network physically entraps liquid oil. Peroxide values told a similar story: the control started at 5.25 millimoles per kilogram and climbed to 7.78 by week four, whereas the fortified spreads began between 3.28 and 4.77 and reached only 6.7 to 7.0 after the same period, demonstrating that the axis genuinely retarded lipid oxidation rather than merely masking its symptoms.

Crucially, none of these functional benefits came at the cost of appearance or palatability. Colorimetric analysis in the CIELab system showed no significant changes in brightness, redness, or yellowness across formulations, and the overall color difference relative to the control remained low, between 0.92 and 1.54, well within the range considered visually acceptable. A semi-trained sensory panel of twelve regular peanut butter consumers rated all spreads on a nine-point scale covering appearance, flavor, consistency, taste, spreadability, and overall acceptability. The 2.5 percent substitution earned the highest overall score of 8.0, compared with 7.4 for the control, and also received the best flavor rating at 7.8. Scores declined modestly at higher inclusion levels, likely because triterpenoid saponins impart a characteristic bitterness, but even the 10 percent formulation matched the control at 7.4, indicating that the axis never pushed the product beyond consumer tolerance.

The implications extend well beyond a single spread. The study demonstrates that a bitter, discarded fraction of the peanut, available essentially for free as a processing by-product, can simultaneously reduce fat, boost fiber, enhance antioxidant activity, suppress oil separation, and slow rancidity, all while preserving the clean-label, single-ingredient character that modern shoppers increasingly demand. The authors caution that their metabolomics results are semi-quantitative and based on limited biological replicates, so targeted absolute quantification will be needed to confirm the triterpenoid findings, and sensory testing was performed only at the initial time point rather than after storage. Future work may explore synergies between the embryonic axis and other natural antioxidants and extend the approach to other oil-rich food systems. For now, the humble peanut heart, once destined for the waste stream, has earned a place at the center of a smarter, more sustainable peanut butter.

Subject of Research: Use of peanut embryonic axis as a natural functional ingredient to improve the texture and oxidative stability of peanut butter

Article Title: Composition characterization of peanut embryonic axis and its application in peanut butter: Effects on texture and oxidative stability

Article References: Gao, Y., Fu, B., Wu, J., Yu, L., Bi, J., Song, Y., Wang, L., Jiang, C., & Wang, M. (2026). Composition characterization of peanut embryonic axis and its application in peanut butter: Effects on texture and oxidative stability. Food Chemistry: X, 39, Article 104532. https://doi.org/10.1016/j.fochx.2026.104532

Image Credits: AI Generated

DOI: 10.1016/j.fochx.2026.104532

Keywords: peanut butter, peanut embryonic axis, lipid oxidation, triterpenoids, dietary fiber, natural antioxidants, oxidative stability, oil separation, peroxide value, food by-product valorization, rheology, clean-label foods

The research team, led by Nuly Z. Acosta-Acevedo of the Universidad Politécnica de Francisco I. Madero in Mexico, worked with viscera from commercial crossbred California × New Zealand rabbits raised at a farm in Hidalgo, Mexico. Six animals with an average live weight of about 2,047 grams were processed after routine commercial slaughter, and the researchers calculated a comprehensive viscera yield of 20.21 percent of live weight, a figure consistent with published ranges of 18 to 22 percent and notable because it included the full complement of organs rather than the selective subsets of edible offal that most studies report. Rabbit viscera are physiologically distinctive: the animal’s hindgut fermentation and cecotrophy shape a gastrointestinal tract that differs from those of fish, poultry and conventional livestock, potentially producing a unique protein matrix and amino acid profile when the tissue is broken down enzymatically.

Before any chemistry could begin, the raw material required careful preparation. The organs were cleaned, ground into a homogenate the authors call Fresh Rabbit Viscera, and then defatted using an adapted Bligh and Dyer method in which petroleum ether and methanol were stirred with the tissue in a 1:2 volume ratio before double centrifugation. Sterilization followed in an autoclave at 120 degrees Celsius for 15 minutes, and the material was then frozen at minus 45 degrees Celsius and freeze-dried for five days at minus 40 degrees Celsius and 0.210 millibar, yielding a stable powder designated Lyophilized Rabbit Viscera. Lyophilization drove moisture from 79.83 percent down to 2.40 percent and concentrated the protein content to 65.49 percent on a dry basis, placing the powder in the same league as porcine liver hydrolysates and chicken offal concentrates as a high-quality protein source.

Microbiological testing confirmed the pretreatment strategy worked. Fresh viscera carried aerobic mesophilic counts of 1.5 × 10⁴ colony-forming units per gram, comfortably below the 10⁶ threshold set by international quality standards, with no coliforms, molds or yeasts detected. After defatting, autoclaving and freeze-drying, the lyophilized powder showed no detectable microbial growth at all. The authors attribute this sterility to the combined disruption of microbial metabolism by organic solvents, spore inactivation by heat, and the collapse of cell membranes under low water activity, an outcome they emphasize is critical for ensuring that any subsequent proteolysis can be attributed solely to the enzymes the researchers added rather than to contaminating microbes.

The heart of the study was a head-to-head comparison of two industrial proteases with fundamentally different catalytic strategies. Alcalase, a serine endopeptidase from Bacillus licheniformis, cleaves internal peptide bonds and tends to generate medium-sized, hydrophobic peptides. Flavourzyme, a protease complex from Aspergillus oryzae, combines endopeptidase and exopeptidase activities, trimming peptides from their termini and pushing the reaction toward very short peptides and free amino acids. The team hydrolyzed the lyophilized viscera with each enzyme for 0, 2, 4, 6 and 8 hours at an enzyme-to-substrate ratio of 100:10, using pH 9 buffer and 55 degrees Celsius for Alcalase and pH 7 and 50 degrees Celsius for Flavourzyme, with three independent batches per condition to ensure biological replication.

Reverse-phase high-performance liquid chromatography revealed clearly divergent fragmentation patterns. Alcalase chromatograms showed a progressive shift toward shorter retention times as hydrolysis advanced, consistent with the endopeptidase first producing medium-sized peptides and then cutting them into smaller fragments; by six hours, more than 78 percent of the peptide signal corresponded to material below 1 kilodalton. Flavourzyme acted faster, generating strong signals for very small peptides and free amino acids almost immediately, with over 82 percent of the profile below 1 kilodalton by eight hours. The authors are careful to note that RP-HPLC estimates of molecular size are approximate, because retention depends on hydrophobicity and terminal composition as much as on mass, so the distributions should be read as relative hydrophobicity profiles, with mass spectrometry needed for definitive peptide identification.

Antioxidant testing produced the study’s most striking numbers. The DPPH radical-scavenging assay, a standard hydrogen-atom-transfer screen, was compromised by a visible precipitate that formed when the hydrolysates met the alcoholic DPPH reagent, a known artifact for hydrophobic, protein-rich samples, so those results were treated as tentative and likely underestimates. The team instead leaned on the oxygen radical absorbance capacity assay, run in both hydrophilic buffer and a lipophilic system containing 7 percent randomly methylated beta-cyclodextrin, which solubilizes hydrophobic peptides so they can neutralize peroxyl radicals. Alcalase hydrolysates climbed steadily from 31.86 micromoles of Trolox equivalents per gram of protein at time zero to a peak of 249.50 at six hours, a statistically significant increase, before declining at eight hours as over-hydrolysis shredded the active fragments. Flavourzyme peaked earlier and plateaued near 40 micromoles per gram, far below the Alcalase maximum.

The explanation, the authors argue, lies in peptide architecture. Short peptides rich in hydrophobic and aromatic residues donate hydrogen atoms efficiently to peroxyl radicals, and Alcalase preserves precisely this class of medium-sized hydrophobic fragments, whereas Flavourzyme’s exopeptidase activity grinds them down into free amino acids with little antiradical power. The pattern mirrors published results for porcine liver, pork by-products and freshwater carp hydrolysates, and the peak value for rabbit viscera compares favorably with reported figures for fish and poultry by-product hydrolysates. In both enzyme systems, prolonged hydrolysis diminished activity, underscoring that hydrolysis time is a tunable dial: stop too early and the bioactive fraction has not been released, stop too late and it has been destroyed.

The cell biology experiments added a provocative dimension. Using the MTT assay, which measures mitochondrial metabolic activity, the researchers exposed SW-480 human colon adenocarcinoma cells to hydrolysate concentrations of 25, 35, 40 and 45 percent for 24 hours. Both enzyme systems reduced viability in a concentration-dependent manner, but Alcalase was consistently stronger, with all five hydrolysis durations driving viability below 50 percent. Surprisingly, the greatest reduction came not from a fully digested sample but from the non-hydrolyzed, zero-hour Alcalase preparation, which cut viability to 24.20 percent at the highest concentration. The authors suggest that pasteurization and defatting alone may have exposed bioactive components in the native protein matrix, and they caution that the MTT assay cannot distinguish apoptosis from necrosis or nonspecific metabolic interference, nor can it speak to selectivity against healthy cells, which were not tested in this exploratory screen.

The researchers are candid about the limits of their work. No peptide sequences were identified, so no causal link between specific fractions and the observed effects can yet be drawn; sensory qualities, bitterness, solubility and stability in real food matrices remain unexamined; and the DPPH data are provisional. Still, the study makes a credible case that an underexplored slaughter by-product, one that rabbit production generates at a rate of roughly 860,000 tonnes of meat annually worldwide, can be upcycled into functional ingredients within a circular bioeconomy framework. The next steps the authors propose are peptide sequencing by liquid chromatography tandem mass spectrometry, testing against non-tumor cell lines to establish selectivity, in vivo validation, and ultimately industrial scale-up. If those efforts succeed, the humblest parts of the rabbit could find a second life not as waste, but as nutraceuticals on the shelf.

Subject of Research: Enzymatic hydrolysis of rabbit viscera to produce antioxidant bioactive peptides

Article Title: Valorization of rabbit viscera by enzymatic hydrolysis: bioactive peptides with antioxidant and cell viability properties

Article References: Acosta-Acevedo, N. Z., Lopez-Perea, P., Herrera-Hernández, M. G., Vergara-Castañeda, H. A., Pérez-Escalante, E., Martínez-Carrera, D., & Olloqui, E. J. (2026). Valorization of rabbit viscera by enzymatic hydrolysis: bioactive peptides with antioxidant and cell viability properties. Food Science of Animal Resources, 46(1), Article 72. https://doi.org/10.1007/s44463-026-00096-3

Image Credits: AI Generated

DOI: 10.1007/s44463-026-00096-3

Keywords: rabbit viscera, enzymatic hydrolysis, bioactive peptides, antioxidant capacity, ORAC assay, DPPH assay, Alcalase, Flavourzyme, cell viability, SW-480 colon cancer cells, food by-product valorization, circular economy