How You Extract a Protein Decides What It Can Do: Lessons from an Ancient Chinese Legume Seed

A humble seed that has been brewed into tea and simmered in soups across Asia for thousands of years is now stepping into the spotlight of modern food science. Astragali complanati semen, the mature dried seed of the legume Astragalus complanatus R. Br., has long been prized in traditional food culture, and more than ten ACS-based health foods are already registered in China. Yet the protein locked inside these kidney-shaped, light brown seeds has remained largely unexplored. A new study published in Food Chemistry: X set out to change that, asking a deceptively simple question: does the way you extract a protein determine what it can do? The answer, backed by an unusually comprehensive battery of structural, sensory, and functional measurements, is a resounding yes, and it carries important implications for anyone hunting for the next generation of plant protein ingredients.

The research team, led by Guoshi Cao and colleagues, compared three extraction methods applied to defatted ACS powder: alkaline extraction with acid precipitation, the workhorse of industrial plant protein production; salt extraction, a gentler wet technique that exploits the salting-in principle to keep proteins in their natural state; and ammonium sulfate precipitation, which forces proteins out of solution at very high ionic strength. Each method produced a strikingly different material. Alkaline extraction delivered the highest purity, with a protein content of 82.71 percent, but a modest protein yield of 18.85 percent. Salt extraction achieved the best recovery, with a protein yield of 29.71 percent and the highest extraction yield at 10.12 percent, though its protein content of 44.41 percent was far lower. Ammonium sulfate precipitation lagged on both counts, yielding a product with only 30.42 percent protein and a protein yield of just 10.78 percent.

The amino acid analysis revealed why the seed deserves attention as a food ingredient in its own right. The protein contains 17 amino acids, including seven essential ones, with glutamic acid, aspartic acid, and arginine dominating the profile. Glutamic and aspartic acids are well-known flavor enhancers, hinting that ACS protein could contribute umami character to foods. The alkaline-extracted protein showed the highest total amino acid and essential amino acid contents, with generous levels of leucine, phenylalanine, and histidine, which are key precursors of bioactive peptides. Both the alkaline and salt-extracted samples met FAO recommended intake values for essential amino acids, and the essential amino acid ratio of the alkaline sample aligned with WHO/FAO/UNU guidance for adult protein requirements, positioning ACS as a credible new dietary protein source rather than a mere curiosity.

Sensory quality, often the Achilles heel of novel plant proteins, was assessed with an electronic nose equipped with 18 gas sensors and a trained panel of ten assessors using quantitative descriptive analysis. Here the alkaline method shone brightest. Its product showed the lowest overall electronic nose response values, indicating the least volatile content, and panelists described a mild aroma with only a subtle beany note, an almost flavorless mouthfeel, and a sweet aftertaste. The likely explanation is chemical: alkaline conditions suppress lipoxygenase activity, the enzyme responsible for generating many of the aldehydes and alcohols behind grassy and green off-flavors, and isoelectric precipitation removes carbonyl-containing carbohydrates that contribute to undesirable tastes. In contrast, the salt-extracted and ammonium sulfate samples carried stronger green, grainy, spicy, and astringent attributes, with the latter’s astringency and saltiness traced to residual polyphenols and salts from the extraction process.

Beneath these surface differences lay profound structural divergence. The alkaline-extracted protein had the smallest average particle size, 239.50 nanometers, and the narrowest size distribution, because alkaline conditions partially unfold the protein, disrupting hydrogen bonds and electrostatic interactions. Salt extraction preserved the natural conformation but produced the largest particles, 695.97 nanometers, as salt bridges encouraged aggregation. Circular dichroism spectroscopy showed that alkaline extraction dismantled the ordered alpha-helical structure, leaving a protein dominated by beta-sheets at 39.36 percent and random coils at 28.51 percent. X-ray diffraction confirmed that the alkaline and salt-extracted samples were largely amorphous, the flexible state that favors hydration and solubility, while the ammonium sulfate sample was highly crystalline at 69.89 percent. Scanning electron microscopy completed the picture, revealing fine, smooth, wrinkled fragments for the alkaline protein and a porous, lamellar network for the salt-extracted one.

These structural differences translated directly into function, the heart of the structure-function relationship the study set out to map. The alkaline-extracted protein was the most soluble across nearly the entire pH range from 3 to 11, peaking at 87.81 percent at pH 11, and it carried the highest zeta potential, reaching minus 31.72 millivolts at pH 11, which keeps particles repelling each other rather than clumping. Its water-holding capacity of 3.21 grams per gram exceeded that of alkaline-extracted pea protein, and it formed gels at the lowest concentration, 14 percent, comparable to soy protein isolate, thanks to beta-sheet-driven hydrogen bonding and its small particle size. It also digested best in a simulated gastrointestinal model, reaching 52.60 percent in vitro digestibility after 240 minutes, well above the salt-extracted and ammonium sulfate samples, whose rigid, compact structures resisted enzymatic attack.

The salt-extracted protein told a different story, one better suited to emulsified and fatty foods. It posted the highest oil-holding capacity at 4.91 grams per gram, surpassing chickpea protein isolates, and delivered the best emulsifying activity, 47.41 square meters per gram, and emulsion stability, 94.02 percent, at pH 11. Its compact alpha-helix and beta-turn-rich conformation and higher surface hydrophobicity allow it to anchor efficiently at the oil-water interface, forming dense films around droplets. The ammonium sulfate protein, meanwhile, excelled at foaming under acidic conditions, reaching 159.08 percent foaming capacity at pH 3, a trait the authors link to reduced electrostatic repulsion near the isoelectric point that speeds adsorption at the air-water interface. Each protein, in short, is a different ingredient with a different best use.

Antioxidant performance and storage stability added further weight to the case for alkaline extraction. The alkaline protein showed the strongest DPPH radical scavenging, 88.54 percent at 2.5 milligrams per milliliter, the best ABTS radical scavenging, and the highest reducing power, likely because unfolding exposed antioxidant amino acids such as lysine, leucine, and aspartic acid and created binding sites for co-extracted phenolic compounds. During four weeks of frozen storage, it accumulated the least lipid oxidation, measured as malondialdehyde, and the least protein oxidation, measured as carbonyl groups, at just 0.36 micromoles per gram compared with 2.19 for the ammonium sulfate sample. Its solubility dropped by less than 4 percent over the storage period, the smallest decline of the three, suggesting the alkaline protein would keep its functional properties on the shelf better than its rivals.

None of this makes alkaline extraction a flawless solution. The authors are candid that it produced a low yield and generated large volumes of acidic and alkaline wastewater, an environmental cost that future work must address, perhaps through eco-friendly techniques such as ultrasound- or enzyme-assisted extraction that could raise yields while limiting chemical inputs. Salt extraction, for its part, co-extracts non-protein components that dilute purity but preserves a protein superbly suited to emulsified products like meat analogues and mayonnaise. The broader lesson extends well beyond one seed: extraction method is not a neutral step in the supply chain but a design tool that shapes purity, flavor, conformation, digestibility, and shelf life. As the food industry searches for sustainable alternatives to animal protein, studies like this one show that unlocking an underutilized resource requires matching the extraction chemistry to the intended application, turning a discarded seed into a tailored ingredient.

Subject of Research: Structure-function relationships of Astragali complanati semen protein extracted by alkaline, salt, and ammonium sulfate precipitation methods

Article Title: Extraction method-dependent structure-function relationships of Astragali complanati semen protein: Implications for its application as a novel plant protein ingredient

Article References: Cao, G., Liu, J., Chen, Y., Shi, Y., Kong, L., Sun, Y., Xu, M., Zhang, H., & Yan, M. (2026). Extraction method-dependent structure-function relationships of Astragali complanati semen protein: Implications for its application as a novel plant protein ingredient. Food Chemistry: X, 39, Article 104607. https://doi.org/10.1016/j.fochx.2026.104607

Image Credits: AI Generated

DOI: 10.1016/j.fochx.2026.104607

Keywords: plant protein, Astragali complanati semen, protein extraction, alkaline extraction, salt extraction, ammonium sulfate precipitation, protein structure, food functionality, antioxidant activity, in vitro digestibility, sensory quality, storage stability