Dough pre-fermentation combined with refrigeration for selenium-enriched steamed bread: A balanced strategy to coordinate selenium biotransformation and quality.

Save

Se-enriched Chinese steamed bread (CSB) is a promising dietary Se source. This study prepared Se-enriched CSB using direct-fermented dough (DFD, 4–16 h) and pre-fermented refrigerated dough (PRD, 24–72 h) with Na₂SeO₃ at 500 or 1500 μg Se/kg. Selenomethionine predominated in DFD, whereas selenocystine dominated in PRD.

Both processes enabled selenoamino acid conversion, but conversion efficiency at 1500 μg Se/kg was approximately half that at 500 μg Se/kg. Compared with DFD-CSB, PRD-CSB generally showed lower hardness, remaining within 4.76–7.05 N. Refrigeration for 24–48 h improved dough extensibility and produced CSB with a relatively uniform pore structure, although slight separation appeared in PRD-48 CSB at 1500 μg Se/kg.

After 72 h, dough tenacity and extensibility declined, and all PRD-CSB exhibited crust–crumb separation, most pronounced at 1500 μg Se/kg. Thus, 24–48 h was suitable for PRD, whereas prolonged refrigeration and higher Se supplementation increased quality risks.

Introduction

Selenium (Se) plays a critical role in physiological processes such as antioxidant defense, immune regulation, and thyroid hormone metabolism (Chen et al., 2021). Chronic inadequate Se intake significantly increases the risk of chronic diseases and cognitive decline (Chen et al., 2024). Approximately one billion people worldwide face Se deficiency, and in China, about 50% of the population has a daily Se intake below the level recommended by the World Health Organization (Farooq, Zhang, Liu, et al., 2024).

Organic Se exhibits significantly higher bioavailability and markedly lower toxicity compared with inorganic forms, current mainstream nutritional guidelines generally recommend selenium-enriched agricultural products as important sources of human Se intake (Farooq, Zhang, Yuan, et al., 2024). It has been demonstrated that yeast can biotransform exogenous inorganic Se into organic forms during dough fermentation (Sánchez-Martínez et al., 2015). As the most widely consumed traditional fermented staple food in China (Peng et al., 2023), steamed bread (CSB) provides favorable conditions for this Se biotransformation through the fermentation step in its production and serves as an ideal carrier for daily dietary Se supplementation.

Probiotic microorganisms such as yeast and lactic acid bacteria hold great potential in the field of biological Se enrichment, and they are widely used as starter cultures in the food industry to produce Se-enriched foods (González-Salitre et al., 2021). Current research on the preparation of Se-fortified bread (such as CSB) using yeast fermentation mainly focuses on three approaches: Se-rich raw materials, Se-enriched yeast, and the exogenous Se addition. Wheat harvested after Se fertilizer application exhibits minimal selenium loss during milling and bread processing (Hart et al., 2011).

Some studies have used germinated Se-enriched legumes (Lazo-Vélez et al., 2021) or grains (Guardado-Félix et al., 2020) to replace part of the flour in bread making, showing that although this strategy can significantly improve nutritional value, the increase in Se content often relies on a high substitution ratio of raw materials, which may compromise product quality. Se-enriched yeast can produce more CO₂ during dough fermentation, thereby improving the internal pore structure of bread (He et al., 2023). However, continued increases in Se enrichment in yeast can suppress its fermentation activity, resulting in a reduced dough gas production rate and weakened viscoelasticity (Du et al., 2023; Du et al., 2024), which negatively affects the final product quality.

When Na₂SeO₃ is added to dough, fermentation time is the predominant factor affecting the production of selenomethionine (SeMet), whereas the addition level plays a minor role, and no significant quality differences were observed among the various fortified groups (Lazo-Vélez et al., 2013). Further studies have revealed that during dough fermentation, Na₂SeO₃ influences yeast metabolism, induces depolymerization of gluten proteins, alters dough viscoelasticity, and increases dough fermentation height, thereby altering CSB quality (Du et al., 2023). All the above studies employed a single continuous fermentation process at 30 °C, under which the efficiency of yeast-mediated biotransformation of Na₂SeO₃ into organic forms strongly correlated with fermentation duration (Kieliszek et al., 2025).

However, pursuing high levels of organic selenium conversion alone may lead to over-fermentation, compromising product quality. In pre-fermented refrigerated dough, yeast activity is suppressed under low-temperature conditions while still maintaining certain metabolic activity (Zeng et al., 2023). Therefore, the pre-fermentation followed by refrigeration process may represent a promising strategy to balance organic selenium conversion and product quality.

Research on Se-enriched CSB under different fermentation processes is limited, lacking systematic evaluation of selenium speciation, product quality, dough fermentation performance, and dough rheology. This study compared the effects of two Na₂SeO₃ levels (500 and 1500 μg Se/kg) on Se speciation, textural characteristics, and internal microstructure in CSB prepared from direct-fermented dough (DFD) and pre-fermented refrigerated dough (PRD). Dough fermentation performance and rheological properties were analyzed using rheofermentometer, 3D X-ray CT, and alveograph.

The objectives were (1) to elucidate selenoamino acid (SeAA) biotransformation patterns; (2) to reveal the intrinsic correlations among dough rheological properties, fermentation behavior, and CSB quality under the combined influence of selenium addition and refrigeration; and (3) to evaluate the feasibility of PRD as a potential processing method for Se-enriched CSB.

Materials and methods

Raw materials

The yeast (Saccharomyces cerevisiae) was obtained from Anqi Yeast Co., Ltd. (Yichang, Hubei, China). Organic wheat flour (71.81 ± 0.70% starch, 13.33 ± 0.16% protein, 1.02 ± 0.09% fat, 12.69 ± 0.07% moisture, 38.45 ± 0.35% wet gluten, 56.40 ± 2.24 μg/kg total Se) and Se-enriched wheat flour (73.24 ± 0.32% starch, 12.54 ± 0.05% protein, 1.43 ± 0.11% fat, 12.18 ± 0.06% moisture, 38.85 ± 0.64% wet gluten, 743.78 ± 35.86 μg/kg total Se) were provided by Xinjiang Fengyi Agricultural Development Co., Ltd.

(Changji, Xinjiang, China). Because the two flours differed in composition, the Se-enriched flour (SeR) was used only as a Se-speciation reference and not as a control sample for subsequent analyses of dough rheological properties, fermentation behavior, and CSB quality.

The Se standard solution was purchased from the China Nonferrous Metal Analysis and Testing Center (Beijing, China). The Standard solutions of selenocystine (SeCys₂), selenomethylcysteine (MeSeCys), and SeMet were obtained from the Chinese Institute of Metrology (Beijing, China). Proteinase XIV (5.5 U/mg) and proteinase K (30 U/mg) were purchased from Sigma-Aldrich (St.

Louis, MO, USA) and high-performance liquid chromatography (HPLC)-grade methanol were obtained from Merck (Poole, UK). Other reagents such as Na₂SeO₃, HCl, KBH4, HNO3, and others were of guaranteed grade and were obtained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China).

Preparation of dough and Chinese steamed bread

According to the previous experimental protocol Zhuang et al. (2025), aqueous solutions of Na₂SeO₃ at concentrations of 2.2 μg/mL and 6.6 μg/mL were prepared. Then, 250 g of wheat flour, 2.5 g of yeast, and 125 mL of the above Na₂SeO₃ solution were mixed in a dough mixer (DBCSJEF720-H, Foshan Xiuyuan Network Technology Co., Ltd., Foshan, Guangdong, China) and continuously stirred for 10 min to form a dough.

Na₂SeO₃ was added to provide nominal Se supplementation levels of approximately 500 and 1500 μg Se/kg, calculated based on the amount of Na₂SeO₃ added relative to the flour mass. These values denote the added Se dose rather than the total Se concentration measured in the samples. Dough prepared with pure water instead of the Na₂SeO₃ solution served as the blank control group.

Direct-fermented dough (DFD): The mixed dough was placed in a plastic container, covered with plastic wrap, and evenly pierced with small holes using a toothpick. It was then fermented in a proofing box (GX-20, Guangdong Henglian Food Machinery Co., Ltd., Guangzhou, Guangdong, China) at 30 °C and 80% relative humidity for 4, 8, and 16 h. Each fermentation period was named as follows: DFD-4, DFD-8, DFD-16.

Pre-fermented refrigerated dough (PRD): The mixed dough was placed in a plastic container and fermented at 30 °C and 80% relative humidity for 2 h. It was then covered with plastic wrap and refrigerated in an incubator (BJ-250 L, Shanghai Yiheng Scientific Instrument Co., Ltd., Shanghai, China) at 4 °C for 24, 48, and 72 h. Each refrigeration period named as follows: PRD-24, PRD-48, PRD-72.

After refrigeration, the dough was transferred to 25 °C and allowed to relax for 30 min before degassing and shaping.

Preparation of CSB: The DFD and PRD were passed through a dough press (MP800, Shanghai Yechang Food Machinery Co., Ltd., Shanghai, China) 15 times to expel large pre-existing gas cells and standardize the dough before final shaping. The sheeted dough was divided into 60 g pieces, hand-shaped into smooth hemispherical shapes, and subjected to final proofing at 30 °C and 80% relative humidity for 50 min before steaming from cold water for 25 min. The CSB samples were cooled at 25 °C for 1 h before texture and 3D X-ray CT analyses.

The remaining samples were freeze-dried using a freeze dryer (LGJ-12D, Beijing Sihuan Hangxing Technology Co., Ltd., Beijing), ground into powder, and stored in self-sealing bags for subsequent tests.

Texture profile analysis (TPA) of CSB

Take the center portion of CSB and prepare a cubic sample with an edge length of 25 mm. According to the method described by Yue et al. (2020) with some modifications, a full TPA was conducted using a texture analyzer (Rapid TA, Shanghai Tengba Instrument Technology Co., Ltd., Shanghai, China).

Use a cylindrical flat-bottom probe with a diameter of P35. The compression strain was set to 50%, and the test speed was set to 1 mm/s.

X-ray CT scanning and image analysis

The dough after final proofing and the corresponding CSB were non-destructively examined using a desktop 3D X-ray CT scanner (NAOMI-CT 002 L, Rfamerica Co., Ltd., Nagano, Japan) in normal mode, following the method established by Zhuang et al. (2025). The proofed dough and CSB were scanned using identical acquisition settings.

The nominal voxel size was 160 × 160 × 160 μm3 (pixel size, 160 μm), and the reconstructed volume matrix was approximately 900 × 900 × 1100 voxels. The X-ray tube voltage and current were set at 50 kV and 10 mA, respectively. Each sample was positioned at the center of the rotating stage and scanned for 120 s over a complete 360° rotation.

The scanning field of view was Φ146 × 185 mm, and the CT data were recorded as 16-bit grayscale images.

The raw CT data were imported into Avizo software (version 9.0.1, Thermo Fisher Scientific, Waltham, MA, USA) for image processing and 3D reconstruction. After removal of the background and sample-stage regions, the gas-cell and solid phases were segmented based on the grayscale histogram using the same image-processing procedure for all samples. Porosity and pore-related parameters were then quantified from the reconstructed volumes.

The same reconstruction, segmentation, and quantitative analysis procedures were applied to both proofed dough and CSB.

Rheofermentometer test

The DFD and PRD fermentation characteristics were determined using a rheofermentometer (F4, Chopin Technologies, Villeneuve-la-Garenne, France). The dough (300 g) prepared in Section 2.2 was placed in the dedicated sample basket of the instrument. The test temperature was maintained at a constant 30 °C, and the restricted weight was 2000 g.

Changes in dough fermentation height and gas production capacity were recorded throughout the process, with continuous monitoring conducted for 120 min. The parameters obtained from this test include Hₘ (maximum dough expansion height), T₁ (time required to reach Hₘ), H'ₘ (maximum height of the gas release curve), T'₁ (time required for gas release to reach H'ₘ), Vₜ (total volume of evolved gas), and V₁/Vₜ (gas retention ratio).

Alveograph test

The Automatic Alveograph (Alveolab, Chopin Technologies, Villeneuve-la-Garenne, France) was used as a comparative bubble-inflation test to characterize the large-deformation behavior of the standardized dough immediately before final proofing. DFD and PRD samples were subjected to the same 15-pass sheeting treatment used for CSB preparation as described in Section 2.2. Five uniformly sized circular specimens were then prepared and rested at 25 °C for 5 min.

The “Demo: bubbles blowing only” program was then used to inflate each dough specimen until rupture. The resulting bubble-inflation curves were used to determine P (tenacity), L (extensibility), W (deformation energy), and SH (strain-hardening coefficient). Each sample was measured in triplicate, with five dough specimens analyzed per replicate.

Se speciation analysis by HPLC-ICP-MS

Se speciation was determined by HPLC-ICP-MS (1290 HPLC, 7800 ICP-MS, Agilent, Santa Clara, CA, USA) following the method of Zhuang et al. (2025). The freeze-dried sample powder (0.200 ± 0.001 g) was transferred into a 10 mL centrifuge tube, followed by the addition of 5 mL Tris-HCl buffer (75 mmol/L, pH 7.5), 10 mg proteinase XIV, and 10 mg proteinase K.

After vortex mixing, the samples were incubated at 37 °C with shaking at 200 rpm for 12 h in a constant temperature incubator (ZCZY-CGS8, Zhichu Biotechnology Co., Ltd., Shanghai, China). The extracts were centrifuged at 10000 ×g for 10 min, and the supernatants were filtered through a 0.22 μm aqueous membrane before HPLC-ICP-MS analysis.

Chromatographic separation was performed using a PRP-X100 column (250 mm × 4.1 mm, 10 μm). Mobile phase A consisted of 20 mmol/L citric acid adjusted to pH 4.0 with ammonia solution and containing 2% methanol, whereas mobile phase B consisted of 20 mmol/L citric acid adjusted to pH 6.0 with ammonia solution and containing 1% methanol. Both mobile phases were filtered through a 0.22 μm membrane and sonicated for 30 min before use.

The elution program was as follows: 0–2.5 min, 100% mobile phase A; 2.5–6.0 min, 100% mobile phase B; and 6.0–10.0 min, 100% mobile phase A. The flow rate was 1.0 mL/min and the injection volume was 10 μL. For ICP-MS detection, the sampling depth was 6.5 mm and the carrier-gas flow rate was 1.1 mL /min.

SeCys₂, MeSeCys, and SeMet were used as reference standards. The retention times were approximately 2.40, 2.97, and 4.63 min, respectively. Se species in the samples were assigned by matching their retention times with those of the corresponding standards and were quantified using the respective external calibration curves.

The chromatogram of the mixed standards is provided in Fig. S1.

The analytical performance of the method was evaluated for SeCys₂, MeSeCys, and SeMet using matrix-spike recovery, limit of detection (LOD), and limit of quantification (LOQ). The LOD and LOQ were determined based on signal-to-noise ratios (S/N). The LODs, LOQs and recovery rate are provided in Table S1.

The SeAA conversion efficiency was calculated as follows:(1)SeAA conversion efficiency=CSeCys2+CMeSeCys+CSeMetCaddedSe×100where CSeCys2, CMeSeCys, and CSeMet represent the Se concentrations quantified as SeCys₂, MeSeCys, and SeMet, respectively, and CaddedSe represents the amount of Se supplied as Na₂SeO₃.

The number of yeast cells

Take 25 g of the prepared dough, place it into a sterile conical flask containing 225 mL of sterile physiological saline, and homogenize in a Clean Bench (SW-CJ-2FD, Suzhou Antai Air Technology Co., Ltd., Suzhou, Jiangsu, China) for 2 min. Then, 1 mL of the initial homogenate was pipetted and subjected to further serial tenfold dilution to obtain 10−4 and 10−5 dilution suspensions. Transfer 1 mL of each dilution homogenate into sterile Petri dishes, followed immediately by the addition of 15–20 mL of agar medium cooled to approximately 50 °C, and the plates were gently swirled to ensure thorough mixing.

Three replicate plates are prepared for each dilution level. After complete solidification, the plates were inverted and incubated at 30 °C for 48 h. then count the colony-forming units on the plates. For colony enumeration, the 10−5 dilution plates (corresponding to an overall dilution of 10−6 relative to the original dough after accounting for the initial 1:10 homogenization) containing 30–300 colonies were selected, and yeast counts were calculated and expressed as CFU/g of dough.

Statistical analysis

Data from three repeated experiments were analyzed to determine whether the variances were statistically homogeneous, and the results are expressed as the mean ± standard deviation. Statistical comparisons were performed using one-way analysis of variance, followed by the Duncan multiple range test using SPSS (version 19.0, IBM Corp., Armonk, NY, USA). Differences were considered statistically significant at P < 0.05.

Results and discussion

Se transformation in CSB under different fermentation methods and durations

As shown in Fig. 1, at both Na₂SeO₃ supplementation levels (500 μg Se/kg and 1500 μg Se/kg), the summed content of the three quantified selenoamino acids (SeAA) increased significantly with prolonged fermentation or refrigeration (P < 0.05). This trend is consistent with previous report that the accumulation of selenium-containing metabolites in selenized yeast is affected by Se exposure and cultivation time (LeBlanc & Mester, 2021).

At the same processing stage, the SeAA content in the 1500 μg Se/kg group was higher than that in the 500 μg Se/kg group. However, the SeAA conversion efficiency at 1500 μg Se/kg was approximately 50% of that observed at 500 μg Se/kg. Thus, the threefold increase in Se supplementation did not result in a proportional increase in SeAA content.

Yeast CFU counts showed only numerical, non-significant changes after Na₂SeO₃ supplementation under the measured conditions (Section 3.4.2). Therefore, reduced yeast proliferation is unlikely to be the primary explanation for the lower SeAA conversion efficiency at 1500 μg Se/kg. Previous studies have shown that selenite uptake and toxicity in Saccharomyces cerevisiae depend strongly on transport and culture conditions (Lazard et al., 2010), and that high selenite exposure can disturb cellular redox homeostasis and induce oxidative stress (Izquierdo et al., 2010; Kieliszek et al., 2015).

Such oxidative stress may partly contribute to the lower SeAA conversion efficiency at 1500 μg Se/kg. However, because ROS levels and selenium-related enzyme activities were not measured in the present study, this explanation remains a hypothesis.Fig. 1Selenium speciation in CSB under different fermentation methods and durations.

DFD-4, DFD-8, and DFD-16: CSB made from direct-fermented dough after 4, 8 and 16 h, respectively. PRD-24, PRD-48, and PRD-72: CSB made from pre-fermented refrigerated dough after 24, 48, and 72 h refrigeration, respectively. SeR: CSB made by natural Se-enriched flour under DFD-4 and used only as an external Se speciation reference.

MeSeCys: selenomethylcysteine. SeCys₂: selenocystine. SeMet: selenomethionine.

SeAA: sum of the three quantified selenoamino acids. Error bars represent standard deviations (n = 3) for each quantified Se species and SeAA. Different lowercase letters indicate significant differences among samples within the 500 μg Se/kg group, whereas different uppercase letters indicate significant differences among samples within the 1500 μg Se/kg group (P < 0.05).

SeR was not included in the statistical comparisons.Fig. 1

At 500 μg Se/kg, SeAA contents in PRD-48 and PRD-72 were significantly higher than those in DFD-8 and DFD-16, respectively (P < 0.05), indicating that the SeAA content continued to accumulate during PRD processing. Xie et al. (2023) reported increased GSH content in refrigerated dough during storage at 4 °C.

In addition, low-temperature exposure has been shown to induce redox-related responses in yeast (Ballester-Tomás et al., 2016; Zhang et al., 2003), while GSH can participate in selenite reduction during Se metabolism (Kieliszek et al., 2015). A related study also found that exogenous GSH increased the proportion of major selenoamino acids in Saccharomyces cerevisiae under selenite exposure, whereas high inorganic Se supplementation decreased both GSH levels and the proportion of selenoamino acids (Wang et al., 2024). These findings provide a plausible biochemical context for the continued increase in SeAA during PRD processing.

Nevertheless, the specific contribution of GSH in the present dough system remains to be verified. At 1500 μg Se/kg, no significant difference in SeAA content was observed between the two processing methods (P > 0.05). In all Na₂SeO₃-supplemented CSB samples, selenomethionine (SeMet) and selenocystine (SeCys₂) were the major quantified SeAA species, whereas no reliably quantifiable MeSeCys signal was observed at its expected retention time.

Previous work has shown that MeSeCys can be further metabolized to SeMet in yeast (Ogra et al., 2018), suggesting that the absence of quantifiable MeSeCys does not necessarily indicate inactive methylated-Se metabolism.

SeMet predominated in DFD, particularly at 1500 μg Se/kg, where it accounted for more than 60% of the SeAA content, consistent with previous reports of SeMet formation during yeast-mediated selenium enrichment of bread (Lazo-Vélez et al., 2013; Sánchez-Martínez et al., 2015). In contrast, SeCys₂ represented a higher proportion of the SeAA content in PRD, accounting for more than 70% at 500 μg Se/kg. However, the Se speciation results may be influenced by sample-preparation conditions.

Selenocysteine (SeCys) is highly reactive, and its analytical recovery can be affected by sample-preparation conditions, including proteolysis (Angaïts et al., 2024; Bierla et al., 2018). These studies, however, primarily address SeCys recovery rather than SeCys₂ recovery or formation under the present extraction conditions. Therefore, potential sample-preparation effects on the Se speciation results cannot be completely excluded.

Accordingly, the higher SeCys₂ content measured in PRD is interpreted as being associated with the dough processing method, rather than as direct evidence of enhanced SeCys₂ biosynthesis or SeCys₂ accumulation caused by a metabolic bottleneck.

For compositional context, SeMet was also the predominant quantified SeAA species in the SeR reference, whereas SeCys₂ and MeSeCys occurred at lower levels. However, because SeR was prepared from naturally Se-enriched wheat, it was used only as an external speciation reference and was not included in direct statistical or quality comparisons with DFD or PRD. The occurrence of targeted SeAA in both the Na₂SeO₃-supplemented fermented samples and the SeR reference indicates that SeAA-containing CSB can be obtained through distinct Se-enrichment systems.

Texture, specific volume, and porosity of CSB under different fermentation methods and durations

While investigating the biological transformation of Na₂SeO₃ by yeast during dough fermentation, the changes in CSB quality were simultaneously measured. As shown in Table 1, under the DFD process, the fermentation time significantly affected the textural properties, specific volume, and porosity of CSB (P < 0.05). With increasing fermentation duration, the hardness of CSB initially decreased and then increased, whereas resilience, specific volume, and porosity first increased and subsequently declined.

Among these, the DFD-4 sample exhibited the lowest hardness (5.35 N to 6.01 N), highest specific volume (2.75 mL/g to 2.77 mL/g), and highest porosity (26.92% to 29.15%), indicating that 4 h fermentation was optimal under the DFD process. These porosity values are comparable to the 23.5% reported for fresh conventional CSB using the same CT system (Yu et al., 2023). This suggests that at this stage, sufficient gas production occurred and gluten network development was well advanced, resulting in strong gas retention capacity, which contributed to ideal CSB quality (Yue et al., 2019).

When fermentation exceeded 8 h, resilience, specific volume, and porosity dropped sharply, while hardness increased significantly. This is primarily attributed to excessive hydrolysis of gluten proteins and weakened cross-linking caused by yeast metabolites during prolonged fermentation, thereby substantially reducing the gas retention capacity of dough (Heitmann et al., 2018). Previous studies have demonstrated that over-fermentation leads to structural collapse and quality degradation in bread (Castro-Reigía et al., 2024).

Therefore, selecting a 2 h pre-fermentation strategy in PRD is reasonable, as it ensures rapid yeast proliferation and metabolic activity while effectively maintaining the integrity and mechanical strength of the gluten network in the dough, thereby ensuring the structural tolerance of the dough to subsequent refrigeration stress.Table 1Texture properties, Specific Volume and Porosity of CSB.Dough typesTime (h)SampleTexture propertiesSpecific Volume (mL/g)Porosity (%)Hardness (N)ResilienceDFD4control6.01 ± 0.45b0.523 ± 0.021de2.75 ± 0.03g26.92 ± 0.63g5005.55 ± 0.53ab0.533 ± 0.003e2.77 ± 0.04g29.15 ± 1.56g15005.35 ± 0.15ab0.529 ± 0.007e2.77 ± 0.08g28.91 ± 1.47g8control8.58 ± 0.41d0.513 ± 0.013cde2.10 ± 0.06abc23.96 ± 1.89f5008.88 ± 0.42de0.497 ± 0.007abcd2.19 ± 0.05c22.96 ± 2.64def15009.00 ± 0.12de0.487 ± 0.007ab2.06 ± 0.06ab23.59 ± 1.21ef16control11.08 ± 1.03f0.479 ± 0.02a2.11 ± 0.07bc19.21 ± 0.70ab50010.64 ± 0.79f0.483 ± 0.012a2.02 ± 0.02a21.25 ± 0.86bcde15009.57 ± 0.55e0.516 ± 0.026de2.06 ± 0.06ab21.92 ± 0.61cdefPRD24control5.39 ± 0.38ab0.499 ± 0.004abcd2.66 ± 0.07f26.66 ± 0.74g5004.94 ± 0.13a0.499 ± 0.002abcd2.67 ± 0.05f27.70 ± 1.56g15004.76 ± 0.39a0.511 ± 0.001bcde2.65 ± 0.02f27.88 ± 0.21g48control5.22 ± 0.09ab0.503 ± 0.009abcd2.32 ± 0.02de20.34 ± 1.12abcd5005.39 ± 0.60ab0.518 ± 0.019de2.35 ± 0.08de20.27 ± 0.35abc15005.10 ± 0.15ab0.518 ± 0.004de2.29 ± 0.01d18.28 ± 0.79a72control5.15 ± 0.15ab0.488 ± 0.010abc2.40 ± 0.02e21.17 ± 2.60bcde5005.91 ± 0.40b0.502 ± 0.010abcd2.32 ± 0.02de19.02 ± 2.51ab15007.05 ± 0.57c0.519 ± 0.021de2.29 ± 0.02d19.01 ± 0.67abValues are presented as the mean ± standard deviation (n = 3). All treatment means within each column were compared by one-way ANOVA followed by Duncan's multiple range test. Means that do not share a common letter are significantly different (P < 0.05), whereas means sharing at least one letter are not significantly different (P > 0.05).

DFD: CSB made by direct-fermented dough; PRD: CSB made by pre-fermented refrigerated dough. 500 and 1500: based on the mass of flour, Na₂SeO₃ are added to the dough at concentrations of 500 μg Se/kg and 1500 μg Se/kg, respectively.

The quality characteristics of PRD-CSB were generally superior to those of DFD-CSB (Table 1; Fig. 2). This is consistent with the findings of Grgić et al.(Grgić et al., 2024) who reported that refrigerated treatment effectively improved bread texture and significantly reduced hardness.

Under different refrigeration durations, the hardness values of PRD-CSB remained stable within the range of 4.76–7.05 N, significantly lower than those of all other DFD-CSB samples except for DFD-4 (the optimal fermentation time group). For comparison, a TPA firmness of 8.1 N was reported for northern-type steamed bread prepared from refined wheat flour (Huang et al., 2021). However, CSB texture is affected by flour characteristics, formulation, processing conditions, and evaluation methods (Zhu, 2014).

Therefore, hardness values reported in different studies should be used for reference only. The refrigerated sponge doughs exhibit better gluten networks and softer bread textures (Zeng et al., 2023). Although the specific volume and porosity of PRD-CSB showed a slight decline with prolonged refrigeration, its hardness remained at a low and stable level, and resilience did not exhibit significant fluctuations.

This phenomenon can be attributed to the low-temperature stress during refrigeration modulating the interactions between yeast metabolites and gluten proteins, thereby influencing dough quality (Xie et al., 2023). Notably, (Grgić et al., 2024) also observed that after 24 h of retarded fermentation at 4 °C, both the specific volume and hardness of the final bread decreased as refrigeration time increased. Compared to DFD, the PRD was able to better balance the interaction between yeast metabolites and dough structure.

These results indicate that the benefits of PRD were most evident at moderate refrigeration durations, whereas longer refrigeration reduced specific volume and porosity.Fig. 2The 2D reconstructed images (wall thickness slices) with X-ray CT of CSB. The thickness and density exhibit a gradual increase from blue to green to red.

DFD-4, DFD-8, and DFD-16: CSB made by direct-fermented dough after fermentation for 4, 8, and 16 h, respectively. PRD-24, PRD-48, and PRD-72: CSB made by pre-fermented dough after refrigeration for 24, 48 and 72 h, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)Fig.

2

Based on the data in Table 1, analysis from the perspective of Na₂SeO₃ addition levels reveals the texture properties of CSB in the DFD process. The previous study found that the addition of Na₂SeO₃ reduced the gas production rate during the early stage of dough fermentation (Zhuang et al., 2025). When fermentation time was extended to 4 h, the hardness of DFD-500 and DFD-1500 became slightly lower than that of the control, indicating that the effect of Na₂SeO₃ on texture is time-dependent, short-term addition may slow down dough fermentation, whereas sufficient fermentation can promote a softening effect on product hardness.

In the PRD process, under 72 h refrigerated storage, the hardness of PRD-500 and PRD-1500 increased, with PRD-1500 showing a significant increase (P < 0.05). This may be due to excessive gluten protein depolymerization caused by higher Na₂SeO₃ concentrations during prolonged low-temperature storage, leading to increased hardness of CSB.

Reconstructed images of CSB under different fermentation methods and durations

The internal microstructure of CSB was non-destructively examined using a 3D X-ray CT scanner, and image reconstruction was performed with Avizo software to obtain 2D wall thickness slices along the XZ axis for each sample. In Fig. 2, a blue–green–red color gradient was used to linearly represent the variation in supporting wall thickness and local density of CSB from low to high, providing a more intuitive explanation of the differences in CSB quality characteristics described in Section 3.2.

Compared to DFD-CSB, PRD-CSB exhibits overall more uniform pore distribution, thinner and less dense supporting walls (predominantly represented by blue regions), and better morphology. PRD-24 and PRD-48 retained a relatively high proportion of these blue regions, consistent with their comparatively low hardness and more uniform internal structure. A previous study on refrigerated sponge dough reported greater starch hydration, a stronger gluten network, and a softer bread texture than those obtained with straight-dough fermentation (Zeng et al., 2023), supporting the view that appropriate refrigeration can favor dough structural organization.

Xie et al. (2023) also reported that refrigerating the fermented dough at 4 °C can increase the GSH content. GSH can weaken the interactions between gluten proteins and alter the gluten network structure, thereby affecting dough rheological properties (Du et al., 2024).

This may partly explain why PRD-CSB exhibited a softer texture and a lower-density pore structure after refrigerated storage. However, the structural benefit of refrigeration diminished when storage was prolonged. Under PRD-72 conditions, all three CSB groups showed crust–crumb separation, while slight separation was already observed in the 1500 μg Se/kg group at PRD-48.

The prolonged refrigerated storage progressively weakened gluten strength and loosened the dough network (Liu et al., 2025). This suggests that excessive refrigeration may reduce the ability of the dough matrix to constrain gas-cell expansion during steaming, thereby contributing to crust–crumb separation.

Observation of pore distribution revealed that pore size increased gradually from the center toward the surface layer. Although DFD-4 exhibited higher specific volume and porosity, its central regions still maintained a relatively dense green supporting wall structure (Fig. 2), which was consistent with its slightly higher hardness than PRD-24.

When DFD fermentation was extended to 8 h, pores accumulated in the outer layer and coalesced into larger cavities, while the green region in the center expanded. At 16 h, the connected pores became more pronounced and the central region shifted from bluish-green toward greenish-red, indicating thicker supporting walls and higher local density. The prolonged fermentation increased dough stickiness and markedly reduced biaxial extensional viscosity, indicating a progressive decline in the dough matrix resistance to deformation (Yue et al., 2019).

Yeast metabolites produced during extended fermentation can also modify gluten interactions and influence dough structural stability (Heitmann et al., 2018). DFD-8 and DFD-16 may have reached an over-fermented state, which weakened dough strength and reduced its ability to stabilize expanding pores. This may lead to pore coalescence and rupture, followed by local densification of the supporting walls.

Accordingly, increased hardness was observed in DFD-CSB after prolonged fermentation.

In general, fermentation level and refrigeration time are the main factors affecting the quality characteristics and structure of CSB, while different levels of Na₂SeO₃ addition can influence the quality properties of CSB under appropriate fermentation and refrigeration conditions. Notably, under DFD-4 and PRD-24 process, the 500 and 1500 μg Se/kg groups exhibited no statistically significant difference in specific volume relative to the control (P > 0.05), yet exhibits lower hardness and higher porosity numerically. As evidenced in Fig.

2, the control under DFD-4 and PRD-24 exhibited a larger high-density supporting wall area at the center of CSB compared to the Na₂SeO₃ addition groups, this phenomenon was particularly pronounced in DFD-4. Previous work has reported selenium-associated changes in yeast GSH metabolism and gluten interactions (Du et al., 2023), which may contribute to the observed structural differences.

Due to significant differences in processing methods between CSB and bread, their quality evaluation systems also differ, with CSB placing greater emphasis on pore structure uniformity and compactness (Peng et al., 2023). Two key factors influencing CSB texture properties are dough fermentability and rheological characteristics (Yue et al., 2020). Therefore, based on its relatively favorable measured quality characteristics, DFD-4 was selected for comparison with PRD to evaluate the effect of adding Na₂SeO₃ on dough gas retention capacity and rheological properties.

Fermentation performance of dough

Fermentation properties of dough

The gas production capacity, expansion characteristics, and gas retention capacity of the dough during final proofing were analyzed using a rheofermentometer. Comparing the fermentation height curves (Figs. 3A–C) and gas performance curves (Figs.

3E–G), the fermentation properties of DFD-4 were distinctly different from that of PRD. DFD-4 reached Hm earlier than PRD, with T1 values of 75.0–79.5 min (Table 2). A previous study has shown that low-temperature conditions can suppress microbial growth in dough (Liszkowska et al., 2024).

Accordingly, PRD required progressively longer to reach Hm as refrigeration time increased, and PRD-72 did not reach Hm within the 120 min test. It is worth noting that although DFD-4 reached Hm earlier, its Hm was significantly lower than that of PRD-24 and PRD-48 (Table 2). Observing the changes in the curves in Fig.

3A–C, DFD-4 entered a stable phase of dough fermentation height after approximately 50 min, whereas the dough height of PRD increased more gradually with continued fermentation. The gas-release curve of DFD-4 reached an early maximum within approximately 20 min and then declined, whereas PRD maintained a more gradual gas-release pattern after 20 min (Fig. 3E–G).

Table 2 shows that Vt of PRD decreased significantly with prolonged refrigeration time. The Vt of DFD-4 was significantly lower than that of PRD-24 (P < 0.05), but there was no significant difference between DFD-4 and PRD-48 (P > 0.05). Notably, although Vt of PRD-72 was significantly lower than that of DFD-4, Hm was not significantly different (P > 0.05), indicating that dough expansion and total gas production did not change in parallel.

Previous research on refrigerated dough also showed that different combinations of fermentation and 4 °C refrigeration altered yeast activity and fermentation properties (Xie et al., 2023). Consistent with these observations, the present results showed a progressive delay in T1 and decline in Vt as PRD refrigeration was extended from 24 to 72 h.Fig. 3Fermentation height curves of doughs in the control group (A), the 500 μg Se/kg group (B), and the 1500 μg Se/kg group (C).

Gas performance curves of doughs in the control group (E), the 500 μg Se/kg group (F), and the 1500 μg Se/kg group (G). Numerical values of dough height and gas production at 50 min are provided in Supplementary Table S2.Fig. 3Table 2Fermentation parameters and yeast cell count of dougha.Dough typesTime (h)SampleFermentation propertiesYeast cell counts (×106 CFU/g)Hm (mm)T1 (min)H'm (mm)T'1 (min)Vt (mL)V1 (mL)V1/Vt (%)4°C30 °CDFD4control35.80 ± 0.10a79.50 ± 4.50a71.13 ± 2.22e18.77 ± 0.75a759 ± 15b737 ± 8b97.10/8.83 ± 0.90c50036.73 ± 0.75ab77.17 ± 2.25a79.90 ± 8.40f19.50 ± 1.50a780 ± 11b758 ± 7b97.18/8.27 ± 1.17bc150037.23 ± 1.25ab75.00 ± 9.00a78.70 ± 7.30ef18.27 ± 1.25a777 ± 7b752 ± 9b96.78/8.17 ± 0.29bcPRD24control42.00 ± 3.58c92.33 ± 2.25b62.50 ± 3.15d61.17 ± 6.25bc1008 ± 78c916 ± 67c90.125.97 ± 0.50cd8.10 ± 1.11bc50040.67 ± 2.55c92.33 ± 0.76b59.80 ± 6.10d67.00 ± 3.61c972 ± 102c896 ± 95c92.185.76 ± 0.59abcd7.13 ± 1.20ab150039.07 ± 1.72bc96.83 ± 0.76b55.57 ± 4.86cd59.50 ± 3.12b989 ± 12c930 ± 15c94.086.13 ± 0.81d7.33 ± 0.67abc48control41.57 ± 0.85c111.67 ± 3.75cd47.40 ± 1.60b83.50 ± 8.00e799 ± 26b780 ± 21b97.625.77 ± 0.91abcd6.53 ± 0.90a50040.80 ± 0.40c114.67 ± 2.25de48.03 ± 0.45bc76.50 ± 4.50d812 ± 3b794 ± 4b97.725.93 ± 0.90bcd6.57 ± 0.99a150040.77 ± 0.15c107.33 ± 5.25c47.15 ± 1.35b78.00 ± 3.00de788 ± 34b768 ± 30b97.465.53 ± 0.59abcd6.40 ± 0.56a72control37.27 ± 2.37ab120.00 ± 0.00e34.83 ± 4.65a20.33 ± 2.25a607 ± 77a604 ± 75a99.514.63 ± 0.40ab6.07 ± 0.55a50036.93 ± 1.11ab120.00 ± 0.00e36.10 ± 3.73a19.50 ± 1.50a613 ± 66a610 ± 64a99.514.73 ± 0.80abc6.13 ± 0.21a150037.13 ± 1.40ab120.00 ± 0.00e35.83 ± 3.25a19.50 ± 1.50a592 ± 60a589 ± 58a99.554.53 ± 0.32a5.92 ± 0.47aValues are presented as the mean ± standard deviation (n = 3).

All treatment means within each column were compared by one-way ANOVA followed by Duncan's multiple range test. Means that do not share a common letter are significantly different (P < 0.05), whereas means sharing at least one letter are not significantly different (P > 0.05). DFD: direct-fermented dough; PRD: pre-fermented refrigerated dough.

500 and 1500: based on the mass of flour, Na₂SeO₃ are added to the dough at concentrations of 500 μg Se/kg and 1500 μg Se/kg, respectively; Hm: maximum expansion height of dough; T1: time for dough to reach Hm; H'm: maximum height of gas release curve; T'1: time for gas to reach H'm; Vt: total volume of gas release; V1/Vt: gas retention rate. 4°C: total colony count of PRD after refrigerated storage; 30 °C: total colony counts of DFD and PRD after final proofing at 30 °C for 50 min and before steaming.

Under the same fermentation process, the dough fermentation height curves and gas performance curves of the two Na₂SeO₃ addition groups were consistent with those of the control, and there were no significant differences in the corresponding fermentation performance indicators in Table 2 (P > 0.05). Thus, within the tested range, Na₂SeO₃ supplementation did not significantly alter Hm, T1, or Vt under the same fermentation process. Lazo-Vélez et al.

(2013) similarly reported that Na₂SeO₃ addition did not significantly change CO₂ generation during sponge-dough bread fermentation. In contrast, Zhuang et al. (2025) observed altered gas-production behavior in Na₂SeO₃-supplemented CSB without a significant change in total gas production, whereas He et al.

(2023) reported higher Hm and total gas volume when Se-enriched yeast was used as the starter. These differences suggest that the effect of Se fortification on dough fermentation depends on the Se source, supplementation level, and fermentation system. For comparison at the end of the 50 min final-proofing period, the orange dotted line in Fig.

3 marks 50 min, and the corresponding dough-height and gas-production values are provided in Supplementary Table S2. At this time point, the differences among DFD and PRD treatments were larger than those caused by Na₂SeO₃ addition. PRD-72 showed lower gas production than PRD-48, whereas no significant differences in the specific volume or porosity of the final CSB were observed between the two treatments.

The relatively large pores formed in the outer layer of PRD-72 CSB (Fig. 2) may partly explain this apparent discrepancy. Overall, refrigeration duration had a greater influence on fermentation behavior than Na₂SeO₃ supplementation under the present conditions.

The yeast cell counts of dough

The number of yeast cells in different dough samples was determined using the plate counting method to investigate the effects of dough fermentation processes (DFD and PRD) and the addition of Na₂SeO₃ on the fermentation ability of dough. As shown in Table 2, after refrigeration at 4 °C, mean yeast counts were 5.76–6.13 × 106 CFU/g for PRD-24, 5.53–5.93 × 106 CFU/g for PRD-48, and 4.53–4.73 × 106 CFU/g for PRD-72. Although these differences were not statistically significant (P > 0.05), the numerical decline with storage time provides the baseline context for the lower counts at PRD-72.

This is similar to the research results of Tao et al. (2018) who reported that yeast cells in refrigerated dough continued to maintain growth, and a significant decrease in yeast cell counts only occurred after 5 d of refrigeration. After 50 min of final proofing at 30 °C, mean yeast counts were 8.17–8.83 × 106 CFU/g for DFD-4, 7.13–8.10 × 106 CFU/g for PRD-24, 6.40–6.57 × 106 CFU/g for PRD-48, and 5.92–6.13 × 106 CFU/g for PRD-72 (Table 2).

DFD-4 was slightly higher than PRD-24 (P > 0.05) and significantly higher than PRD-48 and PRD-72 (P < 0.05). This is highly consistent with the changes in gas production of each dough sample after 50 min of fermentation shown in Table S2. Previous study has shown that CO₂ production during dough fermentation is highly positively correlated with the maximum yeast population (Albertin et al., 2011).

Although the gas production of DFD-4 was relatively high, it decreased rapidly after 20 min of fermentation. However, the yeast cell counts of PRD maintained a stable gas production characteristic, especially the gas production of PRD-24 still rose slowly. But the prolonged refrigeration time impaired the recovery of yeast cell counts to the level observed in DFD-4.

In both DFD-4 and PRD-24, Na₂SeO₃ addition resulted in a numerical decrease in yeast cell counts; however, this decrease was not statistically significant relative to their respective controls (P > 0.05). This numerical tendency is consistent with previous reports that selenium can inhibit yeast growth (An et al., 2025); however, the effect was not statistically significant under the present conditions. In this study, the inhibitory effect of low-concentration Na₂SeO₃ was limited.

Further analysis revealed no significant difference between PRD-48 and PRD-72 (P > 0.05). Overall, refrigeration duration was associated with larger changes in yeast counts than Na₂SeO₃ supplementation under the tested conditions.

Pore characteristics of proofed dough prior to steaming

The CO₂ produced by yeast together with gluten forms pore nuclei. During steaming, the pores expand upon heating, resulting in a loose and porous structure of CSB. Therefore, the morphological characteristics of the pore during the dough fermentation directly affect the quality of CSB (Peng et al., 2023).

To further investigate the effects of fermentation process and Na₂SeO₃ on CSB quality, 3D X-ray CT was used to scan the proofed dough at the end of final proofing, followed by quantitative analysis of the pore structural parameters. As shown in Table 3, under identical fermentation conditions, no significant differences in specific volume and porosity were observed between the Na₂SeO₃ addition groups and the control (P > 0.05). At the same Na₂SeO₃ addition level, the specific volume of DFD-4 was significantly higher than that of PRD-48 and PRD-72 (P < 0.05), and slightly higher than that of PRD-24 (P > 0.05).

Moreover, the porosity of DFD-4 was significantly greater than that of PRD-24, PRD-48, and PRD-72 (P < 0.05). Both the specific volume and porosity of PRD decreased significantly with prolonged refrigeration time (P < 0.05). The concurrent decline in Vt (Table 2) suggests that reduced gas production may have contributed to the lower degree of dough expansion.

However, gas production alone cannot fully explain the differences in pore morphology.Table 3Pore parameters of proofed CSB dough, and Alveograph parameters of dough.Dough typesTime (h)SampleSpecific Volume (mL/g)Porosity (%)Pore parametersAlveograph parametersArea (mm2)Volume (mm3)Diameter (mm)Cell numberP (mm H2O)L (mm)W (10−4 J)SHDFD4control1.46 ± 0.02c18.63 ± 0.77d0.883 ± 0.013b0.101 ± 0.001b0.413 ± 0.010b32,488 ± 721d187 ± 6fg35 ± 1a227 ± 10ab1.13 ± 0.08a5001.45 ± 0.03c18.35 ± 1.13d0.878 ± 0.009b0.096 ± 0.002b0.401 ± 0.008ab32,047 ± 507d190 ± 4g36 ± 1ab238 ± 10abc1.19 ± 0.06a15001.46 ± 0.02c18.55 ± 0.58d0.865 ± 0.029b0.093 ± 0.004b0.402 ± 0.004ab32,899 ± 846d179 ± 5ef35 ± 1a215 ± 10a1.22 ± 0.05aPRD24control1.43 ± 0.02c16.25 ± 0.50c0.754 ± 0.046a0.078 ± 0.006a0.385 ± 0.009a22,522 ± 2155c177 ± 5de34 ± 4a261 ± 26bcd2.08 ± 0.09e5001.44 ± 0.02c15.95 ± 0.44c0.812 ± 0.024ab0.092 ± 0.003b0.403 ± 0.001bc22,095 ± 626c175 ± 8de38 ± 2abc277 ± 13d1.98 ± 0.04cde15001.44 ± 0.01c16.06 ± 0.48c0.816 ± 0.021ab0.090 ± 0.011ab0.391 ± 0.026a21,925 ± 1644c172 ± 5de43 ± 4c322 ± 41e1.87 ± 0.08cd48control1.37 ± 0.03b11.63 ± 0.69b0.776 ± 0.028a0.102 ± 0.004b0.391 ± 0.006a8531 ± 232b172 ± 3de39 ± 1abcd274 ± 6d1.96 ± 0.04cde5001.33 ± 0.02b11.37 ± 0.48b0.766 ± 0.035a0.103 ± 0.011b0.386 ± 0.006a8258 ± 350b168 ± 6cd41 ± 1bc272 ± 14d1.86 ± 0.02cd15001.38 ± 0.02b11.83 ± 0.83b0.788 ± 0.04a0.117 ± 0.005c0.394 ± 0.010ab8335 ± 340b154 ± 1ab51 ± 6d267 ± 9cd1.63 ± 0.21b72control1.20 ± 0.04a3.94 ± 0.36a1.189 ± 0.021c0.197 ± 0.007d0.461 ± 0.004c2468 ± 138a162 ± 8bc34 ± 3a235 ± 23abc2.00 ± 0.03de5001.18 ± 0.06a4.00 ± 0.25a1.324 ± 0.054d0.236 ± 0.005e0.475 ± 0.007cd2406 ± 106a154 ± 3ab36 ± 2ab232 ± 10ab1.94 ± 0.01cde15001.19 ± 0.03a3.92 ± 0.67a1.44 ± 0.096e0.254 ± 0.015f0.491 ± 0.014d2008 ± 157a149 ± 7a34 ± 2a206 ± 18a1.84 ± 0.02cValues are presented as the mean ± standard deviation (n = 3). All treatment means within each column were compared by one-way ANOVA followed by Duncan's multiple range test. Means that do not share a common letter are significantly different (P < 0.05), whereas means sharing at least one letter are not significantly different (P > 0.05).

DFD: direct-fermented dough; PRD: pre-fermented refrigerated dough. 500 and 1500: based on the mass of flour, Na₂SeO₃ are added to the dough at concentrations of 500 μg Se/kg and 1500 μg Se/kg, respectively; Area: mean cell area; Volume: mean cell volume; Diameter: mean cell diameter; P: dough tenacity; L: dough extensibility; W: dough deformation energy; SH: strain-hardening coefficient.

Analysis of the pore parameters in proofed dough using Avizo software revealed that DFD-4 exhibited the highest number of pores (Table 3), which is associated with its strong gas-producing capacity. In comparison, the number of pores in PRD decreased, and this reduction became significant over prolonged refrigeration time (P < 0.05). Examination of pore size parameters showed that DFD-4 had a higher mean cell area than PRD-24 and PRD-48, while no significant differences were observed in mean cell volume or diameter (P > 0.05).

Interestingly, although PRD-72 had the fewest pores, its mean cell area, volume, and diameter were significantly larger than those of the other samples (P < 0.05). In PRD, the addition of Na₂SeO₃ led to an overall increase in pore size compared to the control, with this trend being particularly evident in PRD-72.

As shown in Fig. 4E–Q, the pore morphology in proofed dough exhibits a distinct flattening characteristic, differing from the typical irregular elliptical pores reported in previous studies on CSB (Zhuang et al., 2025). Given this morphological specificity, pore volume was selected as the indicator for analyzing the pore size distribution of each sample.

Fig. 4A-D presents the volume distribution and volumetric proportion of pores in proofed dough. Compared to the control, the addition of Na₂SeO₃ significantly reduced the number of pores larger than 0.05 mm3 in DFD-4 and PRD-48 (P < 0.05), whereas the opposite trend was observed in PRD-24, where the number of pores within this size range increased after Na₂SeO₃ addition (P < 0.05).

In all samples, pores larger than 0.1 mm3 accounted for a low number proportion, yet their cumulative volume represented over 70% of the total pore volume. In PRD samples, as refrigeration time increased, the number of pores continued to decline, but the cumulative volume proportion of pores exceeding 0.5 mm3 rose, while the cumulative volume proportion of pores between 0.1 and 0.5 mm3 correspondingly decreased.Fig. 4Histogram of the pore volume distribution and proportion in proofed dough under DFD-4 (A), PRD-24 (B), PRD-48 (C), and PRD-72 (D).

The 3D pore (volume > 0.5 mm3) distribution of proofed CSB dough: control group (E-H), 500 μg Se/kg group (I-L), and 1500 μg Se/kg group (M-Q).Fig. 4

To more clearly visualize the spatial distribution of pores, three-dimensional reconstruction was performed using Avizo software for pores larger than 0.5 mm3, resulting in a 3D pore structure model of the proofed dough (Fig. 4E–Q). Under DFD-4 conditions, pores in the Na₂SeO₃ addition groups were significantly more concentrated in the central region of the proofed dough compared to the control.

Under PRD-24 conditions, pores in all three groups were predominantly distributed in the central region, but the control contained slightly fewer large pores. Detailed comparisons in Fig. 4F, J, and N revealed that the Na₂SeO₃ addition groups still contained a considerable number of discrete pores in the outer layer, whereas the control exhibited notably lower pore density in the periphery.

During steaming, pores expand upon heating, and a relatively uniform spatial arrangement may contribute to a more homogeneous porous structure in CSB (Xie et al., 2024). Consequently, under DFD-4 and PRD-24 conditions, CSB produced with Na₂SeO₃ addition showed a more uniform pore distribution and slightly lower hardness. Although total gas production (Vt) decreased under PRD-48 and PRD-72 conditions, pores exceeding 0.5 mm3 were still observed, indicating that reduced gas production did not prevent the formation of localized large pores.

These pores were evenly distributed throughout the dough, but as the concentration of Na₂SeO₃ increased and refrigeration time extended, sheet-like large pores developed in the outer layers. This phenomenon was particularly evident in PRD-72 (Fig. 4L, Q), where the Na₂SeO₃ addition groups displayed significantly more sheet-like large pores in the outer layer than the control.

Samples with more pronounced sheet-like pores in the outer layer also exhibited more evident crust–crumb separation after steaming (Fig. 2). Notably, in the 1500 μg Se/kg group of PRD-48, distinct sheet-like large pores were also observed, accompanied by slight crust–crumb separation.

Related study has shown that during the early stages of fermentation, pores grow freely, while in the second phase, pores begin to interconnect, forming complex deformed chambers (Turbin-Orger et al., 2015). Previous studies have shown that pore characteristics in dough and the porous structure of CSB are associated not only with dough gas-production capacity but also with its rheological properties (Qian et al., 2025; Yue et al., 2020). In particular, although PRD-72 had the lowest Vt, relatively large pores still formed after final proofing.

Therefore, gas production alone cannot fully explain the differences in pore morphology, which should be further interpreted in combination with changes in dough rheological properties.

Rheological properties of dough determined by Alveograph

The rheological properties of dough under different fermentation conditions showed significant differences (Fig. 5). Compared with DFD-4, all PRD samples showed a significantly higher strain hardening (SH, Table 3).

A previous study has reported that increased SH positively correlated with a uniform texture in the final product (Ktenioudaki et al., 2011), which is consistent with the finer and more uniform pore structure observed in PRD-CSB (Fig. 2).Fig. 5The alveograph curves of doughs under DFD-4 (A), PRD-24 (B), PRD-48 (C), and PRD-72 (D).Fig.

5

Under DFD-4 conditions, the alveograph curves of the Na₂SeO₃ addition groups and the control nearly overlapped (Fig. 5A), and Table 3 shows a slight decrease in P value for the 1500 μg Se/kg group. This may be because the fermentation time was relatively short, and the effect of Na₂SeO₃ on the rheological properties of the dough had not yet fully manifested.

Under PRD conditions, P values decreased as refrigeration time extended and Na₂SeO₃ concentration increased, while L values initially increased then declined. In PRD-24, P values were significantly lower than in DFD-4. Although no significant differences were found among the P values of three sample groups (P > 0.05), L values increased significantly with increasing Na₂SeO₃ concentration (P < 0.05).

The increase in L indicates greater dough extensibility and was accompanied by differences in pore distribution in the proofed dough (Fig. 4J, N). Studies have reported that refrigeration causes dough softening (Xie et al., 2023) and alters its rheological properties (Simsek et al., 2010).

In PRD-48, the 1500 μg Se/kg group showed a further decrease in P and an increase in L, while sheet-like large pores were concurrently observed after final proofing (Fig. 4P). These large pores were accompanied by slight crust–crumb separation after steaming (Fig.

2). The alveograph curve of the 500 μg Se/kg group was close to that of the control, indicating that after prolonged refrigeration, the effect of Na₂SeO₃ at this concentration on dough rheological properties became less apparent.

In PRD-72, both P and L decreased in all three groups, showing reduced dough tenacity and extensibility after prolonged refrigeration. The gluten network is a major determinant of the large-deformation properties of dough (Li et al., 2025). Relevant research has demonstrated that prolonged refrigeration significantly weakens gluten strength, negatively affecting the final product (Liu et al., 2025).

Consistently, all PRD-72 CSB samples exhibited crust–crumb separation and varying degrees of structural collapse (Fig. 2), indicating deterioration of dough rheological properties after excessive refrigeration. Considering that PRD-72 had the lowest Vt (Table 2), the pore defects observed in the dough and CSB may have been mainly influenced by the deterioration of dough rheological properties rather than by gas production.

Therefore, an appropriate refrigeration duration after pre-fermentation is necessary to maintain the quality of Se-enriched CSB.

Conclusions

DFD and PRD yielded distinct measured selenium profiles, with SeMet predominating in DFD and SeCys₂ in PRD. Increasing Se supplementation from 500 to 1500 μg Se/kg increased SeAA content but approximately halved its conversion efficiency, indicating diminishing returns from higher supplementation. Extending DFD from 4 to 16 h increased CSB hardness (from 5.35 to 6.01 to 9.57–11.08 N) and reduced specific volume (from 2.75 to 2.77 to 2.02–2.11 mL/g), whereas PRD at 24–48 h maintained relatively low hardness (4.76–5.39 N) and generally uniform pore structures.

Compared with DFD-4, PRD-24 and PRD-48 exhibited delayed gas-release peaks and higher strain-hardening coefficients. These findings highlight the need to consider both fermentation kinetics and dough rheology when selecting refrigeration conditions. However, crust–crumb separation in all PRD-72 groups and in the 1500 μg Se/kg group after 48 h limits the suitability of these conditions for direct industrial application.

Neither process was universally superior, given their different selenium profiles and processing constraints. Within the tested range, supplementation at 500 μg Se/kg combined with 24–48 h refrigeration offered a favorable balance between SeAA accumulation and CSB quality, highlighting the importance of controlling refrigeration duration and supplementation level. Future studies should establish a comprehensive selenium mass balance, elucidate the mechanisms underlying yeast–gluten interactions, and evaluate selenium bioavailability and product sensory acceptance to validate the proposed processing strategy.

CRediT authorship contribution statement

Kun Zhuang: Writing – original draft, Software, Investigation, Data curation. Wenping Ding: Writing – review & editing, Methodology, Conceptualization. Lan Yao: Supervision, Project administration.

Xiong Chen: Writing – review & editing, Resources, Project administration, Conceptualization. Jun Dai: Writing – review & editing, Project administration, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Related articles

PubMed

Chitosan functionalized selenium nanoconjugates of Mucuna pruriens as a nanotherapeutic approach for Parkinson's disease.

Nature’s mother holds many untapped treasures of phytoconstituents that can cure and exhibit unexpected therapeutic activity against several diseases. Seeds of velvet beans, popularly known as Mucuna pruriens (MP), are the richest source of Levodopa (L-DOPA). L-DOPA is the precursor for dopamine, a key neurotransmitter deficient in the brain of patients suffering from Parkinson’s disease (PD). The peripheral degradation of L-DOPA before crossing the blood-brain barrier (BBB) is the greatest chal...

PubMed

Biosynthesis of selenium nanoparticles with orange peel waste in microwave: characterization, and spectral properties.

The objective of this study was to reuse agroindustrial waste from orange peels to obtain an orange peel aqueous extract (OPEa), used in the biosynthesis and characterization of selenium nanoparticles (SeNPs). The OPEa acted as a reducing and stabilizing agent in the synthesis of SeNPs, due to its polyphenolic compounds, such as flavonoids, alkaloids, terpenoids, and coumarins, identified by UHPLC-HRMS/MS. The synthesis of SeNPs was conducted in a microwave reactor, at 120 °C and 150 °C, for 30...