Digestion and fermentation of bionic grain-resistant corn starch and its effects on glucose response in mice

Abstract

To address the rapid digestibility and nutritional imbalance of corn starch (CS), bionic grain-resistant corn starch embedded with gellan gum and soy protein isolate (CGS) was prepared by a co-encapsulation strategy to enhance enzymatic resistance and nutritional profile of CS. The proportion of resistant starch in cooked CGS increased from 9.3 to 22.6%. Structural characterization revealed that gellan gum formed a tight encapsulation structure with CS and soy protein isolate. Compared with CS, the CGS group in vitro fecal fermentation enhanced the yields of acetic, propionic, and butyric acids, and increased the abundance of Coprococcus comes, Bacteroides thetaiotaomicron, Prevotella copri, and Parabacteroides merdae by 17.6, 21.83, 3.17, and 6.68 times, respectively. Compared with CS, the CGS group exhibited a decrease in peak postprandial glucose from 11.98 to 9.58 mmol/L. This study offers a scientific foundation for creating starch-based foods to improve glucose metabolism disorders and promote intestinal health.

Introduction

Chronic diseases related to diet, including Type Ⅱ diabetes mellitus, hyperlipidemia, and obesity, are considered a significant and growing threat to human health. Starch constitutes the main energy-providing component of the human diet. However, the rapidly digestible starch (RDS) component in starch can lead to an elevation of blood glucose levels after meals, and studies have shown it to contribute to chronic diseases linked to diet¹. Based on digestive characteristics, starch can be categorized into RDS, slowly digestible starch (SDS), and resistant starch (RS). Previous reports have demonstrated that RS intake can smooth postprandial blood glucose and is metabolized by the microbiota serves as a prebiotic². Therefore, the development and utilization of RS in starch-based foods holds paramount importance in addressing diet-related chronic diseases.

The RS content of starch can be enhanced through various methods, including physical retrogradation, chemical modification, and enzymatic treatment. However, the current methods still face some issues, such as low preparation efficiency and long processing time. Recently, a new method has been developed to encapsulate starch using non-starch polysaccharides such as konjac glucomannan, xanthan gum (XG), agar³. A bionic grain structure is constructed by encapsulating starch with non-starch polysaccharides, thereby improving the digestion characteristics of the starch. Previous studies have shown that the addition of locust bean gum or XG significantly increased the solubility and RS content of rice starch⁴. Yam starch encapsulated with konjac glucomannan showed an increase in RS content from 11.9 to 43.5%, along with improved thermal stability and gel strength⁵. Most of the previous studies primarily investigated the effect of gum on the physicochemical properties and RS content of starch.

Corn starch (CS) exhibits issues such as rapid digestion, poor viscosity, and low nutritional value (e.g., inadequate content of lysine)⁶. Consequently, enhancing the nutritional value of CS and reducing its digestibility have been the primary focuses of food research. Unlike XG, Gellan gum (GG) exhibits excellent gelling properties and safety, the addition of GG can form a tighter barrier on the surface of the starch particles, potentially reducing the digestibility and glycemic index of starch during cooking⁷. While reducing the digestibility of CS, attention should also be paid to balance the nutrition of CS. Soy protein isolate (SPI) is a high-quality plant protein characterized by its nutritional abundance and cholesterol-free nature. SPI contains eight essential amino acids, such as lysine⁸. Therefore, bionic grain-resistant CS embedded with SPI and GG was proposed to be constructed in the study.

RS, as dietary fiber, remains undigested and unabsorbed by the human body after passing through the gastrointestinal tract, eventually being fermented by the intestinal microbiota in the colon. This fermentation process generates a variety of microbial metabolites, such as short-chain fatty acids (SCFAs), including acetic, propionic, and butyric acids. Alginate-inulin-chitosan microspheres facilitated the generation of SCFAs, with acetic acid as the predominant product⁹. Rice-konjac glucomannan complex can raise butyric acid level and promote the abundance of Bifidobacterium¹⁰. To date, no study has investigated the yield of SCFAs and their effect on the composition of gut microbiota in vitro fermentation, and postprandial blood glucose of bionic grain-resistant CS embedded with SPI and GG.

The aim of the study was to develop a bionic grain-resistant CS system, simultaneously enhancing enzymatic resistance and bolstering the nutritional profile, such as high essential amino acid content. A comprehensive investigation was conducted on the physical-chemical characteristics, in vitro digestion and fermentation properties, and the impact on postprandial blood glucose levels of bionic grain resistant CS obtained by encapsulating CS and SPI using GG. This research offers a rationale for developing novel starch-based foods with high resistance to digestion and holds significant importance in improving glucose homeostasis, modulating intestinal flora, and serving as a dietary intervention for various chronic diseases.

Results and discussion

Screen starch samples through digestive characteristics

Table 1 and Supplementary Table 1 present the levels of RDS, SDS, and RS in CS, CXL, and CGS groups. Compared with CS, CX exhibited a significant rise in RS content from 9.3 to 16.8%, which may be attributed to XG forming a physical shield around starch and hindering enzyme access to starch granules. It is widely recognized that the corn is deficient in the essential amino acid lysine, which limits its nutritional value. To address this, lysine was introduced into the CX system. The primary scientific basis for this group was nutritional fortification, and lysine might interact with CS and potentially influence digestibility. Compared to CS, CXL exhibited a higher content of RS, and the RS of 12% CXL reached 16.9%. Previous studies demonstrated the inhibitory effect of amino acids on the activity of porcine pancreatic α-amylase¹¹. However, the addition of lysine did not further increase the RS content of CX.

Table 1.

Content of rapidly digestible starch (RDS), slow digestible starch (SDS) and resistant starch (RS) of starch samples

Samples	RDS (%)	SDS (%)	RS (%)
CS	84.5 ± 1.2a	6.1 ± 2.5a	9.3 ± 2.0d
CG	83.8 ± 1.3a	4.4 ± 1.0a,b	12.1 ± 1.9c,d
3% CGS	84.4 ± 1.3a	1.1 ± 0.1c	14.4 ± 1.3c
6% CGS	75.3 ± 0.3c	6.7 ± 2.7a	18.0 ± 2.7b
9% CGS	73.6 ± 1.2c	3.9 ± 0.7a,b	22.6 ± 1.5a
12% CGS	79.4 ± 0.9b	2.3 ± 0.6c	18.3 ± 1.5b

CS corn starch, CG resistant corn starch using gellan gum encapsulation technology, CGS resistant corn starch by encapsulating CS and soy protein isolate (at 3, 6, 9, and 12% concentrations) using gellan gum. Values mean ± SD indicate the replicates of three experiments; within the same columns, values with different superscripts letters (a, b, c, and d) are significantly different (p < 0.05). One-way ANOVA was utilized to determine group differences, with Duncan’s test applied for multiple comparisons.

To address the broader amino acid imbalance of CS, we replaced the lysine with complete, food-grade SPI. Due to the high viscosity of XG, the CX solution was extremely difficult to mix uniformly. Therefore, SPI was added to the CG system, and the CGS group showed good anti-digestion. The RS content of CGS (14.4–22.6%) was significantly higher than that of CS (9.3%). As the concentration of SPI increases, the RS content of 9% CGS is the highest among all CGS samples, reaching up to 22.6%. The formation of SPI networks during the gelatinization process made the CGS structure more compact, which hindered the hydrolysis of CS. Meanwhile, proteins can reduce the activity of amylase or disrupt its binding with starch¹², thereby delaying the digestion of starch. A gel-like layer was formed on the surface of starch and protein by the film-forming property of GG¹³, which created a physical barrier to hinder the contact of digestive enzymes and further reduced the digestion rate of CS. Therefore, based on the high RS content and broad nutritional profile of CGS, the structure and fermentation properties of CGS were further investigated and analyzed.

Morphological observation

As shown in Fig. 1, CS granules exhibited a smooth surface and polygonal features. GG was typically too large to infiltrate the starch granules and was predominantly situated at the exterior of the starch granules¹⁴. The excellent gel properties of GG enabled it to uniformly encapsulate the surface of starch granules, forming a protective film on the outer layer of CS. A previous study has also demonstrated the formation of a gel layer on the surface of cooked rice grains with GG¹⁵. The addition of SPI and GG resulted in a larger size and rougher surface of CGS granules compared to CS granules. This can be attributed to the mild protein denaturation induced by vortices and oscillation, which enhanced protein-starch binding and facilitated protein encapsulation of starch granules. With the elevation in the concentration of SPI, the microstructure of CGS exhibited no discernible alteration.

Fig. 1. Morphological changes of starch samples observed under a scanning electron microscope.

CS corn starch, CG resistant corn starch using gellan gum encapsulation technology, CGS resistant corn starch by encapsulating CS and soy protein isolate (at 3, 6, 9, and 12% concentrations) using gellan gum.

Crystalline structure

As shown in Fig. 2, CS exhibited an A-type crystalline structure with a distinct diffraction peak at 15.1°, 17.1°, 17.8°, and 23.2° (2θ). The characteristic peaks observed in CG closely resembled those of CS, suggesting that they retain the same crystal structure. Similarly, no new diffraction peaks were observed in CGS, and the crystalline structure of the CGS sample remained type A. Compared with CS, the relative crystallinity of CG and CGS increased. This may be attributed to the addition of GG promoted the interaction of polysaccharide chains and starch molecular chains during heating and stirring, resulting in the generation of ordered crystal structures¹⁶. The tight binding formed between CS and SPI resulted in the weakening changes of CGS structure, thereby facilitating an enhancement in its crystallinity.

Fig. 2. X-ray diffractogram patterns and relative crystallinity of starch samples.

Note: The percentage values represent the starch crystallinity.

Rheological properties analysis

The variations in apparent viscosity of different starch samples with shear rate are shown in Fig. 3a. At the same shear rate, CG had higher viscosity than CS. After the addition of SPI, the apparent viscosity of CGS showed an initial rise and subsequent fall with the increase in SPI content, and the viscosity of 9% CGS was the highest. The reason may be that the molecular chains became entangled with each other, and a network structure was formed. When the SPI content exceeded the critical threshold, intermolecular forces such as hydrogen bonds were broken by shear forces, and the entangled polymer chains were dispersed, reducing the interactions between the flow layers and thereby inducing a decrease in apparent viscosity¹⁷.

Fig. 3. The static and dynamic rheological curve of starch samples.

Note: a Viscosity, b storage modulus (G'), c loss modulus (G''), d tan δ.

Dynamic rheological tests were commonly utilized to analyze the viscoelastic properties of starch samples. As shown in Fig. 3b, c, the value of storage modulus (G’) consistently exceeded the loss modulus (G”) for all samples, demonstrating characteristic weak gel-like structural behavior. Compared with CS, the CGS samples exhibited a consistent reduction in G’ values and an increase in G” values, indicating that the presence of SPI and GG resulted in a reduction in starch elasticity and an increase in viscosity. Tan δ is the ratio of G” to G’ and represents the ratio of elastic and viscous components (Fig. 3d). Systems exhibiting tan δ > 1 demonstrated liquid-like viscous dominance, whereas tan δ < 1 indicated solid-like elastic predominance. In CGS group, tan δ showed an increasing trend with the addition of SPI, indicating a reduction in solid-like behavior and a transition from solid-like viscoelastic behavior to liquid-like behavior¹⁸. Therefore, the CGS exhibited a typical weak gel structure.

Swelling power and solubility

With the increase in heat treatment conditions (60–90 °C), the swelling power of all samples gradually increased (Table 2). The swelling power of CS increased from 3.13 to 9.30 g/g across the thermal processing range of 60–90 °C‌, and the swelling power of CG was not significantly different from that of CS. However, the swelling power of CGS was significantly reduced at 70 °C and 80 °C compared to CS. The observed decrease in swelling power can be ascribed to a core-shell structure, in which the GG network encapsulated starch granules, thereby inhibiting their swelling¹⁹. At the same time, the hydrophobicity of SPI further inhibited the contact between water and starch, resulting in a lower swelling power of CGS.

Table 2.

Swelling property and solubility parameters of starch samples

Samples	60 °C		70 °C		80 °C		90 °C
	S (%)	SP (g/g)	S (%)	SP (g/g)	S (%)	SP (g/g)	S (%)	SP (g/g)
CS	1.10 ± 0.16e	3.13 ± 0.14a,b	4.88 ± 1.80c	5.81 ± 0.23a	9.13 ± 1.44c	6.80 ± 0.16a	11.88 ± 2.29c	9.30 ± 0.69a
CG	4.13 ± 0.45d	3.08 ± 0.19a,b	8.75 ± 1.44b,c	5.33 ± 0.16a,b	8.63 ± 1.80c	6.34 ± 0.25b	20.75 ± 1.85a,b	8.98 ± 0.50a,b
3% CGS	10.00 ± 1.32c	3.07 ± 0.10a,b	12.38 ± 7.47a,b	5.11 ± 0.39b	13.88 ± 2.06b	6.22 ± 0.34b,c	18.63 ± 5.53a,b	9.10 ± 0.44a,b
6% CGS	13.83 ± 1.26b	3.01 ± 0.06b,c	12.75 ± 1.71a,b	4.92 ± 0.50b	13.88 ± 1.65b	5.92 ± 0.19 c,d	17.88 ± 1.97b	8.53 ± 0.50b,c
9% CGS	21.15 ± 1.47a	2.74 ± 0.07c	12.75 ± 4.05a,b	4.79 ± 0.45b	15.25 ± 2.40b	6.20 ± 0.28b,c	19.25 ± 2.63a,b	8.15 ± 0.34c
12%CGS	13.08 ± 1.08b	3.35 ± 0.26a	18.75 ± 4.66a	4.97 ± 0.29b	18.75 ± 1.19a	5.55 ± 0.28d	23.25 ± 1.19a	7.95 ± 0.12c

S solubility, SP swelling power. Values mean ± SD indicate the replicates of three experiments; within the same columns, values with different superscripts letters (a, b, c, d, and e) are significantly different (p < 0.05).

The solubility of starch primarily refers to the extent to which amylose and amylopectin dissolve in water after heating or gelatinization²⁰. The solubility of CS and CG exhibited a progressive increase with elevated temperatures exceeding 60 °C, demonstrating enhanced leaching of amylose molecules into the aqueous solution during thermal treatment‌. In comparison to CS, the solubility of CGS exhibited a significant increase at various temperatures. Moderate thermal treatment facilitated protein hydration by strengthening intermolecular interactions with water molecules. Prolonged heating facilitated the recombination of broken chemical bonds within the protein and induced a reorganization of the protein conformation, thereby enhancing its solubility²¹.

Texture analysis

The addition of GG reduced the hardness of the CS, and the hardness of the CGS gradually decreased with the increase in SPI concentration (Table 3). The hardness of starch samples was closely related to the swelling capacity of the starch granules. A lower swelling capacity prevented the starch granules from fully expanding and gelatinizing during the heating process, resulting in the inability to form a harder starch gel²². The elasticity of 9% CGS and 12% CGS was significantly reduced compared to CS. Stickiness and chewiness, respectively, quantify the energy expenditure required for disintegration and oral processing in semi-solid food matrices²³. The encapsulation of GG can modify the textural properties of CS, leading to reduced stickiness and chewiness, potentially through a reduction in hardness and an alteration network of CS. A compact encapsulation layer around CS was evidenced by SEM and XRD (Figs. 1 and 2). As a physical barrier, the encapsulation layer promoted the formation of a weak gel network and reduced the hardness of CS, and restricted the accessibility of digestive enzymes, thereby increasing starch anti-digestibility.

Table 3.

Texture analysis of cooked starch samples

Sample	Hardness1	Hardness2	Adhesiveness	Cohesiveness	Elasticity	Stickiness	Chewiness
CS	1.46 ± 0.02a	1.44 ± 0.01a	0.05 ± 0.00a	0.90 ± 0.00a	4.42 ± 0.12a	1.35 ± 0.01a	5.95 ± 0.17a
CG	1.34 ± 0.12b	1.27 ± 0.19b	0.05 ± 0.00b	0.90 ± 0.00a	4.10 ± 0.14b,c	1.21 ± 0.11b	4.95 ± 0.31b
3% CGS	1.27 ± 0.04b,c	1.21 ± 0.04b,c	0.06 ± 0.01b,c	0.88 ± 0.05a	4.66 ± 0.36a	1.10 ± 0.05b,c	5.13 ± 0.57b
6% CGS	1.20 ± 0.11c,d	1.17 ± 0.11b,c	0.06 ± 0.00b,c	0.90 ± 0.00a	4.36 ± 0.11a,b	1.07 ± 0.12c,d	4.68 ± 0.64b
9% CGS	1.09 ± 0.07d,e	1.07 ± 0.07c,d	0.06 ± 0.00c,d	0.90 ± 0.01a	3.92 ± 0.11c,d	0.96 ± 0.06d,e	3.75 ± 0.31c
12% CGS	0.99 ± 0.04e	0.96 ± 0.05d	0.06 ± 0.00d	0.90 ± 0.00a	3.76 ± 0.21d	0.86 ± 0.04e	3.20 ± 0.18c

Values presented as the mean of triplicate ± standard deviation. Values mean ± SD indicate the replicates of three experiments; within the same columns, values with different superscripts letters (a, b, c, d, and e) are significantly different (p < 0.05).

Morphological characterization of starch samples after in vitro human fecal fermentation

As shown in Fig. 4, the CS granules were degraded by the microbial flora after 24 h of fermentation, exhibiting a layered and porous morphology. CG displayed a lamellar structural morphology, and a large number of bacterial colonies were discernible on the surface of the sheet-like structure. CGS was decomposed by the enzymes of colonic bacteria, resulting in a morphology that was vastly different from its unfermented state, specifically manifesting as a thin and irregular sheet-like structure. Some incompletely decomposed starch granules were also observed in the CGS sample.

Fig. 4. Morphological changes of starch samples observed under a scanning electron microscope after 24 h in vitro fermentation.

Note: 2, 5, 10 kx represent the magnifications of scanning electron microscope.

Owing to its optimal physicochemical properties, including the highest resistance to enzymatic digestion, enhanced viscosity, weak gel-like characteristics, and reduced hardness and chewiness that facilitate accurate oral administration, 9% CGS (subsequently, renamed as the CGS group) was selected as the representative bionic grain-RS for all subsequent in vitro fecal fermentation and in vivo postprandial blood glucose testing in mice.

Alterations in SCFAs concentration and gas production in vitro human fecal fermentation

An in vitro fecal fermentation model was set up to assess and compare the effects of various substrates on SCFAs production. The experimental design comprised five groups: a carbohydrate-free blank control (CON group), a recognized natural source of high RS was used as the positive control (HAMS group) and three test groups (GG, SPI, and 9% CGS group). Through the above five experimental groups, the SCFAs and gas production of the bionic grain-resistant CS were further compared and analyzed.

Figure 5 displays the gas production and SCFAs concentrations of the starch samples during in vitro fecal fermentation.‌ The gases produced by gut microbiota were an important component in determining their production and their impact on human health. The CGS group significantly enhanced total gas production relative to the CON group. This could be attributed to the full utilization of CGS by the microbial flora, which generated an increased amount of SCFAs during the fermentation process, resulting in increased gas production.

Fig. 5. Production of short-chain fatty acids of starch samples after 24 h in vitro fecal fermentation.

CON was the blank control group without the addition of carbohydrates, while the high amylose corn starch (HAMS) group served as the positive control group. GG gellan gum. SPI soybean protein isolate. Different letters indicate significant differences (p < 0.05), n = 6. One-way ANOVA was utilized to determine group differences.

Acetic acid is the major metabolite of dietary fiber and plays an important role in SCFAs. CGS showed excellent performance in the production of acetic acid by microbial fermentation. The propionic acid content in the CGS group reached 1699.32 mg/L, notably exceeding the 1246.53 mg/L in the CON group (p < 0.05). The CGS sample was utilized as a fermentation substrate by the intestinal microbiota in the colon, resulting in the production of SCFAs such as acetic, propionic, and butyric acids. HAMS and CGS were more effective at producing butyric acid compared to the CON group. This suggested that HAMS and CGS may have acted as prebiotics by balancing gut microbiota. During fermentation, CGS promoted the production of butyric acid. Butyric acid contributes to maintaining intestinal barrier function and alleviating inflammation²⁴, and has potential application value in improving intestinal health and preventing related diseases. Compared to the CON, HAMS significantly increased valeric acid content, while the SPI group showed higher levels of isobutyric and isopentanoic acids among the five groups. The results indicated that the CGS group exhibited greater production of SCFAs in comparison to the other four groups.

The focus of microbiota analysis is to explore the effect of bionic grain-resistant CS formed by GG encapsulated CS on the structure of gut microbiota. Therefore, no further in-depth microbial analysis of SPI was conducted. Subsequently, the gut microbiota composition will be analyzed for the CON, HAMS, GG, and CGS groups.

Effect of starch samples on gut microbiota diversity and richness during in vitro fermentation with human feces

Venn diagram analysis revealed shared and distinct OTU counts of the gut microbiota among the CON, HAMS, GG, and CGS groups (Fig. 6a). SPI was omitted from microbiota analyses due to its low percentage of dry weight of SPI in CGS and negligible impact on SCFAs production. The number of species shared by all groups was 150 at the Amplicon Sequence Variant level. The unique species counts of CON, HAMS, GG, and CGS groups were 1180, 721, 1976, and 1233, respectively. The number of unique species in the CGS group exceeded that of the CON and HAMS groups, suggesting that CGS may enhance the diversity of unique species in human gut microbiota.

Fig. 6. Effects of HAMS, GG, and CGS samples on the diversity and richness analysis of gut microbiota composition in different groups.

a Venn diagram. b α-diversity was evaluated from the Chao1 index, ACE index, Shannon index, and Simpson index. c β- diversity was evaluated from PCA and PCoA. Different letters indicate significant differences (p < 0.05), n = 6. One-way ANOVA was utilized to determine group differences.

Gut microbiota α-diversity indices are presented in Fig. 6b. The Chao 1 and ACE indices represent species abundance, while the Shannon and Simpson indices represent species diversity. The Chao1 and ACE indexes showed no significant differences among CON, GG, and CGS groups, but decreased significantly in the HAMS group. Compared with the CON, HAMS, and GG groups, the Shannon and Simpson indices in the CGS group exhibited a significant increase, indicating that CGS enhanced gut microbiota species diversity. β-diversity analysis clarified the inter-group variations among distinct gut microbiota. As shown in Fig. 6c, the microbiota structure of the CGS group was markedly separated from those of the CON, HAMS, and GG groups, highlighting its distinct composition. Notably, CON and GG groups exhibited similarity in gut microbiota structure.

Analysis of the relative abundance of gut microbiota at phylum and genus levels during in vitro fermentation with human feces

The relative abundance of gut microbiota at the phylum level in the CON, HAMS, GG, and CGS groups is shown in Fig. 7a. The gut microbiota across all groups mainly consisted of Bacteroidota, Firmicutes, Fusobacteriota, Proteobacteria, and Actinobacteriota. In contrast to CON group, HAMS group exhibited an increase in Bacteroidetes relative abundance, a contrasting decrease was observed in Firmicutes. The CGS group showed an upward trend in the relative abundance of both the Bacteroidota and Firmicutes compared with the CON group. The F/B (Firmicutes/Bacteroidota) ratio can serve as an indicator of gut microbiota health, with obese individuals typically exhibiting a rise in the F/B ratio. As shown in Fig. 7c, the F/B ratio in the CGS group was significantly reduced by 61.76% compared to the CON group.

Fig. 7. Effect of HAMS, GG, and CGS samples on gut microbiota composition after 24 h in vitro human fecal fermentation.

Note: a phylum level, b genus level, c The ratio of Firmicutes to Bacteroidota, d Prevotella_9, e Bacteroides, f Sutterella, g Fusobacterium, h Lachnoclostridium, i Coprococcus relative abundance at the genus level. j Heatmap analysis of correlation between SCFAs and the abundance of gut microbiota at the genus level. Acetic: acetic acid, Prop: propionic acid, Butyric: butyric acid, Isobut: isobutyric acid, Isopent: isopentanoic acid, Valeric: valeric acid, Hexylic: hexylic acid. Different letters indicate significant differences (p < 0.05).

A genus-level assessment of gut microbiota (Fig. 7b) revealed that the CGS group exhibited increased relative abundances of Prevotella_9, Bacteroides, Holdemanella, Parabacteroides, and Dorea but decreased relative abundances of Fusobacterium, Escherichia Shigella, Lachnoclostridium, and Sutterella compared to the CON group. Prevotella_9 abundance was significantly increased in the CGS and HAMS groups in comparison with the CON group (Fig. 7d). Prevotella_9 has encompassed several known producers of succinate²⁵, and the succinate produced could be further converted into propionate by other intestinal bacteria. This process provided a plausible explanation for the higher concentration of propionate observed in the CGS group (Fig. 5). The abundance of Bacteroides in the CGS group increased compared to the other groups (Fig. 7e). Bacteroides constitutes a crucial segment of the human gut microbiota, being essential to gut ecology, metabolism, and the interplay between the host and microorganisms. Sutterella and Fusobacterium showed a downward trend in the CGS group compared to CON and GG groups (Fig. 7f, g). Previous studies have shown that Sutterella was associated with gut dysbiosis²⁶, while Fusobacterium may promote inflammation²⁷. The shifts in gut microbiota composition in the CGS group suggested that CGS may have the potential to contribute to reducing inflammatory processes or promoting overall intestinal homeostasis by restructuring the gut microbiota. CGS, as the substrate, reduced Lachnoclostridium abundance in vitro fermentation compared with the CON group (Fig. 7h).

Spearman correlation analysis was used to assess the association between gut microbiota abundance and SCFA concentrations (Fig. 7j). The concentrations of acetic, propionic, and butyric acids were correlated with Prevotella_9, Coprococcus, and Holdemanella at the genus level. Coprococcus has been demonstrated as a butyrate-producing beneficial bacterium²⁸. With the increase in Coprococcus abundance in the CGS group (Fig. 7i), the concentration of butyric acid also increased in comparison with the CON group. The abundance of Prevotella_9, Coprococcus, and Holdemanella showed a positive correlation with the levels of acetic, propionic and butyric acids, and a negative correlation with the level of isopentanoic acid. The abundance of E. Shigella, Sutterella, and Lachnoclostridium had a negative association with the concentrations of acetic, propionic and butyric acids.

The analysis of gut microbial relative abundance at species level and significant differences in gut microbial species

Figure 8a displays the species-level relative composition of the gut microbiota. CGS group exhibited that gut microbiota was primarily composed of Prevotella copri, Fusobacterium mortiferum, Holdemanella biformis, Dorea longicatena, Bacteroides thetaiotaomicron, Coprococcus comes, and Parabacteroides merdae. The relative abundance of P. copri in the CGS group was significantly increased compared to the CON and GG groups (Fig. 8b). P. copri can promote glucagon-like peptide-1 secretion, thereby stimulating insulin secretion²⁹. Furthermore, the high abundance of P. copri in the CGS group showed the potential of CGS to ameliorate hyperglycemia and insulin resistance. The relative abundances of H. biformis and B. thetaiotaomicron were increased, while relative abundances of F. mortiferum were reduced in CGS group compared with CON group (Fig. 8c–e). Elevated abundances of C. comes and P. merdae were observed in HAMS and CGS groups than in the CON group (Fig. 8f, g). C. comes, a prominent butyrate-producing bacterium³⁰, was found to be significantly elevated in the CGS group. This increase, accounting for the high concentration of butyrate, was observed in Fig. 5. B. thetaiotaomicron has a strong ability to degrade complex carbohydrates into small molecular sugars, which may serve as substrates for C. comes to promote butyrate production. In the CGS group, the availability of butyrate may be enhanced through high levels of B. thetaiotaomicron and the butyrate-producing C. comes, stimulating glucagon-like peptide-1 release and improving blood glucose homeostasis. And B. thetaiotaomicron was confirmed to suppress inflammation in IBD³¹. P. merdae displayed a negative association with the relative abundance and the activity of ulcerative colitis³². The increased relative abundance of P. merdae and B. thetaiotaomicron in the CGS group indicated that CGS may have beneficial effects in gut health by promoting bacteria associated with reduced inflammation. Therefore, CGS could serve as a promising prebiotic intervention in glucose homeostasis and related gastrointestinal disorders in the future.

Fig. 8. Effect of HAMS, GG, and CGS samples on gut microbiota composition at the species level after 24 h in vitro human fecal fermentation and LEfSe analysis.

Note: a Relative abundance at the species level, b Prevotella copri, c Fusobacterium mortiferum, d Holdemanella biformis, e Bacteroides thetaiotaomicron, f Coprococcus comes, g Parabacteroides merdae. h Heatmap analysis of correlation between SCFAs and the abundance of gut microbiota at the species level, i LEfSe Phylogenetic Tree, j Histogram of LDA scores. Acetic: acetic acid, Prop: propionic acid, Butyric: butyric acid, Isobut: isobutyric acid, Isopent: isopentanoic acid, Valeric: valeric acid, Hexylic: hexylic acid. Different letters indicate significant differences (p < 0.05).

We further analyzed the correlation between SCFAs and gut microbiota at the species level (Fig. 8h). The levels of acetic, propionic, and butyric acids were correlated with H. biformis, C. comes, P. copri, Escherichia coli, and Sutterella wadsorthensis. The abundance of H. biformis, C. comes, and P. copri showed a positive correlation with the concentrations of acetic, propionic and butyric acids, and a negative correlation with the levels of isobutyric and isopentanoic acid. The abundance of F. mortiferum, Lachnoclostridium edouardi, E. coli, and S. wadsworthensis showed a negative association with the levels of acetic and propionic acids. The abundance of E. coli and S. wadsorthensis showed a negative association with the levels of butyric acid.

Species differences across multiple levels were conducted using LEfSe multi-level discriminant analysis, with the results presented in Fig. 8i. LEfSe analyzed microbiota from phylum to species levels, quantifying its impact on observed differences using LDA scores (Fig. 8j). Tannerellaceae, Bacillus, and Butyricicoccaceae were identified as dominant bacterial taxa in the CGS group through LEfSe analysis (LDA score > 4). Tannerellaceae were bacteria known to produce propionate and butyric acid³³, with propionate contributing to host health via the modulation of energy metabolism. Butyricicoccaceae were renowned for the ability to produce butyrate, which has recognized anti-inflammatory effects³⁴. The observation of Butyricicoccaceae and Tannerellaceae in the CGS group indicated that CGS may be a promising dietary intervention for alleviating obesity-related metabolic problems through modulation of gut microbial composition. Previous research has confirmed the significant efficacy of Bacillus in the treatment of ulcerative colitis³⁵. Utilizing CGS as a dietary intervention may ameliorate ulcerative colitis symptoms by increasing intestinal Bacillus.

KEGG metabolic pathway analysis

PICRUSt2 analysis based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database allows for the observation of functional shifts in gut microbiota composition. Figure 9 presents the differential analysis map of KEGG metabolic pathways under the second level. Compared with CON, KEGG pathways correlated with biosynthesis of metabolism, other secondary metabolites and Glycan biosynthesis, amino acid metabolism, and energy metabolism were upregulated in the CGS group. The boosted energy metabolism and amino acid metabolism directly support the increased SCFA production documented in the CGS group, as these pathways are fundamental to bacterial growth and fermentation. Consequently, the KEGG data linked the CGS-driven compositional shift to an enhanced metabolic output. These findings support the hypothesis that CGS functioned as a prebiotic by modulating the gut microbiota to amplify the synthesis of beneficial metabolites, such as butyrate, via the upregulation of these key metabolic pathways.

Fig. 9. KEGG metabolic pathway differential analysis map between the CON and CGS groups in vitro human fecal fermentation.

Note: Bars represent the proportion of different functions of gut microbiota in CON and CGS groups. The circle represents the proportion of differences in functional abundances within the 95% confidence interval, and the right value is the p value.

In vivo experiments

The CGS group exhibited a moderate increase in postprandial blood glucose levels, reaching its peak concentration of 9.58 mmol/L at 30 min (Fig. 10a). In contrast, the Glucose and CS groups reached their blood glucose peaks at 15 and 30 min, with peak concentrations of 16.3 mmol/L and 11.98 mmol/L, respectively. Notably, the administration of CGS to mice resulted in significantly lower blood glucose concentrations at 15, 45, 60, and 90 min compared to the CS group. This result demonstrated the efficacy of CGS in reducing postprandial blood glucose levels, and this effect was attributed to its lower digestibility. In order to account for inter-group variations in baseline glycemic levels, the maximum increment in blood glucose within each treatment group was determined by calculating the difference between the peak glycemic concentration and the baseline glycemic concentration at the outset (Fig. 10b). CGS group showed the minimal rise in glycemic concentration from baseline compared to Glucose and CS groups, with a mere rise of 3.83 mmol/L, whereas the increases in the Glucose and CS groups were 10.45 mmol/L and 6.35 mmol/L, respectively. These results suggested that CGS had a superior ability to regulate glucose homeostasis compared to CS. The disparity in the average area under the curve (AUC) could reflect variations in the rate and degree of starch digestion and glucose absorption³⁶. As illustrated in Fig. 10c, the AUC value of the CGS group was significantly reduced to 858.45 compared to the Glucose and CS groups. The results indicated that CGS obtained by the treatment methods of GG encapsulated SPI and CS reduced the digestibility of CS, thereby significantly improving glucose homeostasis in mice.

Fig. 10. Blood glucose levels in mice after oral administration of starch samples.

a Mean blood glucose concentration, b Mean increase in blood glucose, c Mean AUC of blood glucose in vivo. The values are expressed as mean ± SD, n = 6. Values with the different letter are significantly different at p < 0.05.

This study utilized excellent gelling properties of GG to simulate the grain cell wall structure, thereby further increasing the RS content of CS. The RS content of 9% CGS was increased from 9.3 to 22.6%. SEM images revealed that the surface of the CS granule was encapsulated by GG and formed a bionic grain structure. The addition of GG and SPI significantly ‌increased the viscosity of CS, forming a system with weak gel-like structural characteristics. And CGS exhibited a markedly improved textural quality by reducing the stickiness and chewiness of CS. The CGS group resulted in increased production of acetic, propionic, and butyric acids compared to the CON group after in vitro fecal fermentation. CGS increased the species diversity and changed the structure of gut microbiota by elevating the abundance of beneficial bacteria C. comes, B. thetaiotaomicron, P. copri, and P. merdae, while reducing the abundance of harmful bacteria F. mortiferum. The administration of the CGS group resulted in a remarkable reduction of up to 40% in the area under the blood glucose curve in mice, in addition to demonstrating lower postprandial peak concentration and maximum increase compared to the CS group. This study clarifies the potential of CGS in regulating postprandial blood glucose levels and adjusting gut microbiota structure, offering an important theoretical and practical foundation for the future development of starch-based functional foods aimed at blood glucose control and gut health enhancement.

Methods

Materials

CS (amylose content 24.0%, ash content 0.1%, moisture content 13.0%, fat content 0.1%, protein content 0.4%) was provided by Starpro Starch Company (Hangzhou, China). GG was purchased from Shanghai Yien Chemical Technology Co., Ltd. XG was purchased from Macklin Biochemical Technology Co., Ltd (Shanghai, China). Lysine was provided by Beijing Solarbio Science Technology Co., Ltd. (China). SPI was provided by Linyi Shansong Biological (Linyi, China). Porcine pancreatin (8×USP) and amyloglucosidase (≥260 U/mL) were purchased from Sigma-Aldrich (St. Louis, MO, USA). The remaining chemicals and reagents were in accordance with analytical grade criteria.

Construction of bionic grain-resistant corn starch

Resistant CS using GG encapsulation technology (CG) was prepared as follows: GG was dispersed in distilled water and stirred with a magnetic stirrer (IKA C-MAG HS 7, Staufen, Germany) at 500 × g for 15 min to form a GG suspension (0.8%, w/v). CS was dispersed in distilled water and magnetically stirred at 500 × g for 15 min to form a CS suspension (18%, w/v). After adding 40 mL of CS suspension to 60 mL of GG suspension, the mixture was homogenized by high-pressure shear disperser (IKA T25 digital ULTRA-TURRAX, Staufen, Germany) at 5000 × g for 5 min. The obtained CG sample was oven-dried at 40 °C for 24 h using an electrothermal blowing dry box (BGZ-70, Shanghai, China).

The process of preparing resistant CS by encapsulating CS and SPI using GG (CGS) was as follows: GG was mixed with distilled water and magnetically stirred at 500 × g for 15 min at 50 °C to form a GG suspension (0.8%, w/v). The protein mixtures were obtained by mixing different concentrations of SPI (3, 6, 9, and 12%, based on the dry weight of CS) with 10 mL of distilled water. Finally, 30 mL of CS suspension (25%, w/v) and 10 mL of protein mixture were added to 60 mL of GG suspension, followed by homogenization at 5000 × g for 5 min. The 3% CGS, 6% CGS, 9% CGS, and 12% CGS samples were each obtained by oven-drying at 40 °C for 24 h.

In vitro digestibility

The in vitro digestibility of starch samples, including CS, CX, CG, CXL, and CGS samples, was assessed with prior methods³⁷. A 4 mL of sample suspension (5%, w/v) was heated at 95 °C for 15 min in a thermostatic water bath shaker (SHZ-82B, Changzhou, China), and then cooled to 25 °C. Sodium acetate buffer (4 mL, pH = 5.2) was added to the starch paste, which was then mixed with ten glass beads. The mixture was kept at 37 °C for 30 min using a thermostatic water bath shaker. Subsequently, a mixture of enzymes containing porcine pancreatin (8×USP) and amyloglucosidase (≥260 U/mL) was added to the mixture. At 0, 20, and 120 min, 0.1 mL aliquots were extracted from the mixture and added to a 90% ethanol solution to stop the digestion reaction. The mixture was centrifuged at 10,000 × g for 5 min using a Haier high-speed microcentrifuge (LX-165T2R, Qingdao, China), and the glucose level in the supernatant at each time point was measured using a glucose oxidase-peroxidase assay kit (Beijing Leadman Biochemistry Co., Ltd., Beijing, China). The levels of RDS, SDS, and RS of starch samples were determined as follows:

$$\mathrm{RDS}(\%)=({\mathrm{G}}_{20}-{\mathrm{G}}_0)\times 0.9\times 100/\mathrm{S}$$ 1

$$\mathrm{SDS}(\%)=({\mathrm{G}}_{120}-{\mathrm{G}}_{20})\times 0.9\times 100/\mathrm{S}$$ 2

$$\mathrm{RS}(\%)=1-\mathrm{RDS}-\mathrm{SDS}$$ 3

S represents the total starch content of the sample. A stoichiometric factor of 0.9 was used for starch in glucose calculations. The glucose levels measured at 0, 20, and 120 min were denoted as G₀, G₂₀, and G₁₂₀.

Scanning electron microscope (SEM)

The morphological characteristic images of starch samples were observed by scanning electron microscopy (Hitachi Regulus 8100, Japan). CS and encapsulated starch powders were adhered to the holder, then placed in the SEM chamber and coated with gold. The proper SEM image of the sample was obtained by adjusting the position, and observations were performed at 20 kV.

X-ray diffraction (XRD)

The crystalline of starch sample was characterized via XRD using a Bruker AXS D2 PHASER (40 kV, 40 mA, Cu–Kα radiation). Diffraction patterns were collected across a 2θ range of 4° to 40° at 0.5°/s. The degree of crystallinity was calculated as described in a previous study³⁸.

Rheological properties

The rheological properties of starch samples were determined using a rheometer (MCR302 rheometer, Anton-Paar, Graz, Austria). Starch suspension (6%, w/v) was maintained at 95 °C with constant stirring for 30 min on a magnetic stirrer, then cooled to room temperature, about 25 °C. An appropriate amount of sample paste was placed between two parallel plates (50-mm diameter) with a 0.2-mm gap. The dynamic rheological properties were measured at 0.1% strain within a linear viscoelastic range of 0.1–100 rad/s. Steady shear rate range was performed from 0.1 to 100 s⁻¹.

Textural profiling

The OmniTest texture analyzer, from PPT Group UK Ltd, West Sussex, UK, was used to determine the textural profile of starch samples. Cylindrical gel specimens (diameter: 3 cm; height: 2 cm) were subjected to compression testing using a P/36R cylindrical probe. The compression protocol entailed a deformation rate of 2 mm/s, a trigger threshold of 5 N, and a consistent test speed of 1.0 mm/s.

Swelling power and solubility

The 2% starch suspension (w/v) was then exposed to a water bath and heated at temperatures of 60, 70, 80, and 90 °C for 30 min, followed by cooling to 25 °C. The mixture was centrifuged at 4500 × g for 15 min using a high-speed refrigerated centrifuge (Thermo Scientific Multifuge X1R, Waltham, MA, USA), followed by collection of the supernatant and dried at 105 °C for 8 h. Both the precipitate after centrifugation and the supernatant after drying were measured to obtain their respective masses. The calculation formula for the swelling power and solubility was as follows:

$$\mathrm{Solubility}(\%)=(W_1/W_0)\times 100\%$$ 4

$$\mathrm{Swelling\; power}(\mathrm{g}/\mathrm{g})=W_2/W_0(1-\mathrm{Solubility})$$ 5

W₀ represents the initial mass of CS and encapsulated starch samples, while W₂ and W₁ correspond to the masses of precipitate and the dried supernatant, respectively.

In vitro human fecal fermentation

In vitro fecal fermentation proceeded according to previous method³⁹ with modifications. Prior to sample collection, written informed consent was obtained from participants, and their privacy rights were strictly observed. The study protocol and consent form were supervised and approved by the Ethics Committee Medical College of Qingdao University (protocol ID: QDU-HEC-2025343, approved on March 17, 2025). Human fecal samples were obtained from three healthy participants (BMI between 18.5 and 24 kg/m², aged 20–25) who were free from antibiotic or probiotic use for 3 months prior to sample collection. To minimize variations in fecal microbiota composition among the three participants, equal amounts of feces were collected and combined. The bacterial suspension was prepared in a mini vinyl glovebox (Coy Lab Products Inc, Michigan, USA) by dispersing 0.2 g of feces in 20 mL of sterilized PBS solution (pH = 7.4), and then filtering through four layers of sterile gauze.

High-amylose maize starch (HAMS), GG, SPI, and 9% CGS samples were heated at 95 °C for 15 min and processed as described in Section 2.3 after 120 min in vitro digestion to obtain a starch residue, respectively. The starch digestion residue was lyophilized at −60 °C for 20 h in a CTFD-10S freeze dryer (Qingdao Yonghe Chuangxin Technology Co., Ltd., Qingdao, China) before further analysis. Then 0.1 g samples of HAMS, 9% CGS, GG, and SPI were transferred into individual fermentation flasks and subsequently UV-irradiated for 30 min in a clean bench (HR1360-IIA2, Qingdao, China). Subsequently, 10 mL of sterile brain heart infusion broth medium (pH 7.0–7.2) and 0.1 mL of bacterial suspension were added to the fermentation flask in the anaerobic chamber. The mixture was incubated at 37 °C for 24 h under constant temperature fermentation. Six biological replicates of in vitro fecal fermentation were performed for each sample. Gas production of different starch samples was measured after 24 h. The fermentation broth of the starch sample underwent centrifugation (12,000 × g, 5 min). The resulting supernatant and the precipitate were utilized for SCFAs quantification and 16S rRNA bacterial sequencing, respectively.

Measurement of SCFA levels

The fermentation supernatant (1 mL) was added with 200 μL of 50% sulfuric acid. Subsequently, 100 μL of an internal standard solution containing cyclohexanone (1000 mg/L) was added, followed by 2 mL of ethyl ether. The mixture was vortexed for 1 min and subjected to centrifugation (12,000 × g, 10 min) at 4 °C. The clear supernatant was carefully collected for subsequent analysis. The separation and detection of SCFAs were performed using a gas chromatograph equipped with a DB-WAX capillary column (30 m × 0.25 mm × 0.25 μm) and a mass spectrometry system.

16S rRNA sequencing amplicon

Amplification of the bacterial 16S rRNA gene was achieved through polymerase chain reaction using specific primers designed for the variable region v3 + v4. Primer information includes the forward primer 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and the reverse primer 806R (5′-GGACTACHVGGGTWTCTAAT-3′). The libraries for sequencing were subsequently prepared, with sequencing carried out on the Illumina NovaSeq6000 platform (provided by Biomarker Technology Co. Ltd., Beijing, China). Trimmomatic (version 0.33) was used to quality filter the raw data, and Cutadapt (version 1.9.1) was used to identify and remove primer sequences. Subsequently, paired-end reads were merged and chimera-filtered using USEARCH v10 in combination with UCHIME v 8.1, yielding refined sequences for downstream analyses.

In vivo blood glucose experiments

Twenty-four healthy male C57BL/6J mice (6–8 weeks) were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. Housing for the mice was provided at the Experimental Animal Center of Qingdao University Medical Science Center under controlled environment conditions. Mice were fed AIN93G for one week before being randomly divided into three groups: Glucose, CS, and CGS groups. Since the study aimed to improve the anti-digestibility of CS (used as the raw material) via encapsulation, we compared the effects of the encapsulated product (CGS) with CS on blood glucose levels. Ethical approval for all animal experiments was granted by the Qingdao University Animal Experimentation Ethics Committee (protocol ID: QDU-AEC-20241211C571820241213113), and procedures complied with the National Research Council Guide for the Care and Use of Laboratory Animals.

CS and CGS suspensions (0.2 g/mL) were heated at 95 °C for 15 min with a magnetic stirrer, and then cooled to 25 °C. After a 14 to 16 h fast, glucose, cooked CS, and cooked CGS solutions were administered to mice via gavage at a dose of 2 mg/g body weight. Blood samples were then collected from mice tail veins at 0, 15, 30, 45, 60, 90, and 120 min after gavage. The blood glucose levels in the tail veins of mice were determined using the Accu-Chek dynamic blood glucose monitoring system, and a functional relationship with time was plotted accordingly. The peak-to-baseline difference in glucose concentrations was utilized to determine the maximum increase in blood glucose levels in respective groups. The AUC increment for each animal was calculated and analyzed with medical statistical software, GraphPad Prism 9.0. At the conclusion of the experiment, all mice were humanely euthanized via carbon dioxide inhalation. Mice were processed in batches within a euthanasia chamber, where CO₂ was initially delivered at a rate of 20% of the chamber volume per minute. Once the mice had lost consciousness, the gas was progressively increased while maintaining pressure below 0.5 KPa. The animals were monitored until all movement and respiration ceased, and the gas supply was discontinued upon confirmation of pupillary dilation. A subsequent 2 min observation period was implemented to verify death.

Statistical analyses

Each experiment was repeated a minimum of three times to ensure accuracy. IBM SPSS Statistics 26 was used for statistical analyses. One-way ANOVA was utilized to determine group differences, with Duncan’s test applied for multiple comparisons. The level of significance was set at p < 0.05.

Author contributions

Jie Liu: Writing—Original draft preparation, Conceptualization, Data curation. Yanchen Meng: Investigation. Aiguo Ma: Supervision. Ranran Chang: Conceptualization, Methodology, Investigation, Validation, Supervision.

Data availability

Data will be made available on request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41538-025-00647-w.