Saccharomyces cerevisiae SC‐8303 Reduces Fructans in White Pan Bread While Preserving Baking Quality

ABSTRACT

Wheat‐based bakery products are widely consumed; however, they cause gastrointestinal symptoms in some individuals, primarily attributed to fructans and fermentable carbohydrates that are abundant in wheat. This study investigated a yeast strain selection strategy to reduce fructan content during bread making and evaluated its effects on baking quality, compositional characteristics, and volatile profiles. Seventeen Saccharomyces cerevisiae strains isolated from nuruk, a traditional Korean fermentation starter, were screened for fructan‐reducing and phytate‐degrading activities using a dough fermentation model. Furthermore, the selected strain SC‐8303 was used for white pan bread production along with a commercial reference strain under identical processing conditions. SC‐8303 bread exhibited a significantly lower fructan content (0.13 ± 0.05 g/100 g DW) than the reference strain bread (0.37 ± 0.03 g/100 g DW), corresponding to an approximate 65% reduction, whereas key baking quality parameters, volatile profiles, and instrumental sensory measurements remained comparable between the two breads. SC‐8303 bread also exhibited selective increases in glutamine, glycine, glutamic acid, and lysine and in vitro protein digestibility (34.2 ± 3.4%) than in SC‐RS bread (27.6 ± 1.7%). These findings provide empirical support for yeast strain selection as a practical strategy for producing bread with reduced fructan content without adversely affecting product quality under standard bread‐making conditions.

Practical Applications

Fructan, a fermentable carbohydrate found in wheat, is a recognized trigger of gastrointestinal symptoms in sensitive individuals. Yeast strain selection based on fructan‐reducing activity can be directly integrated into conventional commercial bread‐making processes without requiring changes to existing formulations or equipment. This approach provides bakery manufacturers with a practical tool for developing low‐fructan bread products targeting consumers with fructan‐related dietary needs.

Abbreviations

CFU

colony‐forming unit

DW

dry weight

FAA

free amino acid

FMOC

9‐fluorenylmethyl chloroformate

FODMAP

fermentable oligo‐, di‐, monosaccharides, and polyols

HPLC

high‐performance liquid chromatography

HS‐SPME–GC/MS

headspace solid‐phase microextraction coupled with gas chromatography–mass spectrometry

IBS

irritable bowel syndrome

OPA

o‐phthalaldehyde

SC‐RS

Saccharomyces cerevisiae Red Star

SEM

standard error of the mean

TIC

total ion chromatogram

TTA

titratable total acidity

VOC

volatile organic compound

YPD

yeast extract–peptone–dextrose

1. Introduction

Bread is one of the most widely consumed cereal‐based foods worldwide and is a major source of dietary carbohydrates. However, gastrointestinal symptoms, including abdominal distension, gas production, and abdominal pain, following the consumption of wheat‐based products have been increasingly reported (Biesiekierski et al. 2011; Gibson and Shepherd 2010). These symptoms have been associated with fermentable carbohydrates present in wheat, particularly fructans, which are classified as fermentable oligo‐, di‐, monosaccharides, and polyols (FODMAPs) (Gibson and Shepherd 2010). Consequently, efforts have been made to reduce the fructan content in wheat‐based foods to improve gastrointestinal tolerance while maintaining product quality.

Recent evidence suggests that the gastrointestinal symptoms associated with wheat consumption may not be exclusively attributable to gluten, thereby shifting attention toward other compositional factors in wheat‐based foods (Biesiekierski et al. 2013; Skodje et al. 2018). Dietary interventions for irritable bowel syndrome (IBS) and other functional gastrointestinal disorders have increasingly focused on restricting fermentable carbohydrates, collectively termed FODMAPs (Halmos et al. 2014). Fructans are abundant in wheat and represent a major FODMAP component (Gibson and Shepherd 2010). Fructans are poorly absorbed in the small intestine and are instead fermented by the gut microbiota in the colon, leading to osmotic effects and gas production (Gibson and Shepherd 2010). This fermentation increases luminal fluid and intestinal distension, thereby exacerbating gastrointestinal symptoms. Consistent with this mechanism, fructan intake has been directly associated with symptom worsening in susceptible individuals (Skodje et al. 2018). These findings underscore the need for processing strategies to effectively reduce fructan levels in bread during production.

Several approaches have been proposed to reduce fructan content during bread making, including the modification of fermentation conditions, the extension of fermentation time, and the use of specific microorganisms (Fraberger et al. 2018). However, these approaches often involve changes in fermentation parameters that can directly affect dough rheology, gas production, and the structural properties of bread (Laurent et al. 2020; Pejcz et al. 2023). Consequently, fructan reduction strategies may compromise bread quality and flavor development, limiting their practical applicability in commercial products (Laurent et al. 2020; Pejcz et al. 2023). Moreover, fermentation‐induced compositional changes, including proteolysis, are closely associated with subsequent thermal reactions during baking and play a critical role in the formation of the final flavor profile (Thiele et al. 2002). Therefore, achieving fructan reduction without altering key quality attributes remains a major technological challenge in the development of bakery products. In this regard, the selection of yeast strains with intrinsically higher fructan‐degrading capacities is an alternative strategy that avoids the modification of fermentation conditions. Although strain‐dependent differences in fructan hydrolysis have been reported (Laurent et al. 2020; Fraberger et al. 2018), a comprehensive evaluation of this approach under standardized white pan bread conditions, encompassing the baking quality, aroma, and compositional characteristics of the final product, remains limited. Thus, it remains unclear whether fructan reduction can be achieved through strain selection alone without compromising key quality attributes.

In addition to fructan, wheat‐based foods contain phytic acid (phytate), an antinutritional factor that can form complexes with minerals, proteins, and starch, which can reduce mineral bioavailability and limit enzymatic nutrient digestion (Lopez et al. 2002). Importantly, phytate can be degraded during cereal fermentation and breadmaking through microbial and enzyme‐mediated processes (Haros et al. 2001). Therefore, phytate‐degrading potential should be considered in addition to fructan‐reducing activity in the selection of yeast strains for wheat‐based fermentation.

Nuruk, a traditional Korean cereal‐based fermentation starter prepared by molding grains such as wheat, rice, or barley and allowing spontaneous fermentation, harbors a complex and dynamic microbial ecosystem composed mainly of filamentous fungi, yeasts, lactic acid bacteria, and other bacteria (Jung et al. 2012; J.‐E. Lee and Kim 2017). Representative taxa include amylolytic fungi such as Aspergillus oryzae and Lichtheimia corymbifera, fermentative yeasts such as Saccharomyces cerevisiae, Pichia spp., and Saccharomycopsis fibuligera, and lactic acid bacteria such as Lactobacillus, Pediococcus, and Weissella spp. During fermentation, these microbial groups contribute sequentially and cooperatively: fungi contribute to starch hydrolysis; lactic acid bacteria drive acidification and microbial succession; and yeasts utilize the generated fermentable sugars to produce alcohol, CO₂, and aroma compounds (Jung et al. 2012). In particular, S. cerevisiae plays a central role in sugar utilization, CO₂ production, and fermentation‐derived flavor development, which are key in bread fermentation (Hazelwood et al. 2008; Thiele et al. 2002). Therefore, the microbial diversity and fermentation‐adapted characteristics of nuruk make it a relevant source for selecting S. cerevisiae strains with bakery‐applicable carbohydrate metabolism and fermentation properties.

This study aimed to evaluate whether strain‐dependent fructan reduction in white pan bread can be achieved under standard processing conditions without adversely affecting product quality. To this end, we screened 17 S. cerevisiae strains isolated from nuruk, a traditional Korean fermentation starter, for fructan‐reducing and phytate‐degrading activities using a wheat flour dough fermentation model. A strain demonstrating superior fructan‐reducing activity, SC‐8303, was selected and applied to a standardized white pan bread process along with a commercial reference strain under identical processing conditions. The resulting bread was systematically evaluated for fructan content, baking quality, free amino acid composition, volatile compound profiles, and in vitro protein digestibility.

2. Materials and Methods

2.1. Yeast Strains and Culture Preparation

Seventeen yeast strains isolated from nuruk, a traditional Korean fermentation starter maintained at the LOTTE R&D Center (Seoul, Korea), were used in this study. For strain isolation, nuruk samples were suspended in sterile saline, serially diluted, inoculated onto yeast extract–peptone–dextrose agar (YPD agar; yeast extract 10 g/L, peptone 20 g/L, dextrose 20 g/L, and agar 15 g/L), and incubated at 30°C for 48 h. White circular colonies formed on YPD agar were selected and purified by three sequential single‐colony isolations on fresh YPD agar. Species‐level identification was conducted by internal transcribed spacer sequencing at Macrogen Inc. (Seoul, Korea), and the obtained sequences were analyzed using NCBI BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The 17 purified isolates identified as S. cerevisiae were preserved as glycerol stocks at −80°C and used for strain screening. A commercial baking yeast, Saccharomyces cerevisiae Red Star (SC‐RS; Lesaffre Group, USA), purchased from Jenico Co., Ltd. (Seoul, Korea), was used as a reference strain. All yeast strains were stored as glycerol stocks at −80°C. Before the experiments, the strains were activated by cultivation in YPD broth at 30°C for 24 h. The viable cell counts of the yeast cultures were determined using the plate count method on YPD agar (1.5% agar). After cultivation, the cell density was adjusted to approximately 1 × 10⁸ colony‐forming unit (CFU)/mL using sterile saline.

2.2. Screening of Yeast Strain

Yeast strains were screened to evaluate their ability to reduce fructan in a wheat flour dough fermentation model and to reduce phytate under in vitro conditions. Fructan reduction was evaluated using a modified flour‐based fermentation model, as described by Fraberger et al. (2018). Briefly, a flour suspension was prepared by mixing bread flour and distilled water at a 1:4 (w/w) ratio, and 100 mL aliquots were transferred into sterile containers. After inoculation with 1 mL of activated yeast culture, the samples were incubated at 30°C for 48 h. After fermentation, the samples were homogenized before analysis. For fructan extraction, 5 g of the fermented sample was mixed with 20 mL of distilled water, suspended, and heated at 100°C for 5 min; this extraction step was repeated twice (McCleary et al. 2000). The extraction conditions were optimized for the flour‐based fermentation matrix in which fructans were readily solubilized because of the high moisture content of the suspension. The samples were subsequently centrifuged at 8000 × g for 20 min, and the supernatant was collected. Fructan content was quantified using a commercial enzymatic assay kit (H‐FRUCHK; Megazyme, Wicklow, Ireland) based on the AOAC Method 999.03 (McCleary et al. 2000), and absorbance was measured at 340 nm. The fructan‐reducing activity was expressed as the relative percentage decrease in fructan content after fermentation compared with the initial fructan content before fermentation.

Phytate‐degrading activity was evaluated using a liquid culture model supplemented with phytate (Haros et al. 2001). Activated yeast strains were precultured in YPD medium and subsequently inoculated into a medium containing 2% (w/v) glucose and 0.05% (w/v) sodium phytate, followed by incubation at 30°C for 48 h. After incubation, the cells were removed by filtration, and the supernatant was collected. The residual phytate content in the supernatant was quantified using a commercial enzymatic assay kit (K‐PHYT; Megazyme) following the modified enzymatic protocol for fermented food matrices described by Rousta and Taherzadeh (2025), and the absorbance of the resulting colorimetric product was measured at 655 nm. Phytate degradation was calculated based on a standard curve generated using an inorganic phosphorus standard solution and expressed as the relative percentage decrease in residual phytate after incubation compared to the initial concentration.

Based on the results of fructan reduction and phytate degradation assays described above, candidate yeast strains were selected for subsequent baking experiments (Figure S1).

2.3. White Pan Bread Making

White pan bread was manufactured using the sponge dough method based on previously reported laboratory‐scale white pan bread procedures with minor modifications (Kokawa et al. 2017; Kim et al. 2024). All yeast strains, including the commercial reference strain SC‐RS, were prepared using the same inoculum preparation procedure to ensure comparability between the treatments. Each strain was cultured in food‐grade YPD broth, beginning with a 10‐mL culture and subsequently scaled up to 200 mL and 5 L¸ and incubated at 30°C for 24–48 h. After cultivation, the yeast cells were harvested by centrifugation (8000 × g, 10 min) and washed twice with sterile distilled water to remove residual medium components, and the cell density was adjusted to a standardized viable cell count of approximately 1.5 × 10⁹ CFU/mL. The resulting yeast preparation was used as a paste inoculum for bread making, with the inoculum amount adjusted to achieve a final concentration of approximately 1 × 10⁷ CFU/g dough for all strains. The strong wheat flour used in this study had a protein content of 12.6 ± 0.3%, an ash content of 0.46%, a moisture content of 13.8%, and a farinograph water absorption of 68.0 ± 3.0%. White pan bread was prepared using a standard formulation based on AACC Method 10–11.01 (AACC International 2010), with minor modifications. The sponge dough was prepared by mixing 70 g of strong wheat flour, 40 g of distilled water, and 1.2 g of the prepared yeast inoculum, based on 100 g of total flour in the formulation. The sponge dough was then subjected to primary fermentation at 27°C and 75% relative humidity for 4 h. The final dough was prepared by incorporating the remaining 30 g of strong wheat flour, 25 g of distilled water, 6.0 g of sucrose, 2.3 g of sodium chloride, 3.0 g of skim milk powder, and 3.0 g of vegetable shortening into the fermented sponge dough, followed by a 15‐min rest period. This formulation was applied consistently across all treatment groups, enabling the comparison of yeast strain effects under identical breadmaking conditions. The dough was divided into equal portions, rounded, proofed, and then molded. The molded dough was placed in baking pans and proofed at 38°C and 85% relative humidity for 60 min, followed by baking in a deck oven at top and bottom temperatures of 180 and 210°C, respectively, for 25 min. After baking, the loaves were cooled at 25°C before subsequent analyses. Each treatment was performed in at least three independent bread‐making batches.

2.4. Determination of Fructan Content in White Pan Bread

The fructan content in baked white pan bread was determined using the crumb portion (crust removed). The crumbs were ground and homogenized before extraction. One gram of the homogenized sample was mixed with 25 mL of distilled water and heated in a 100°C water bath for 10 min with intermittent vortexing to facilitate the extraction of water‐soluble fructans. The extraction time was extended to 10 min relative to the flour‐based fermentation model used in the screening assay described above to account for the denser crumb matrix and reduced fructan extractability following baking‐induced structural changes. The extract was centrifuged at 8000 × g for 20 min, and the supernatant was collected. Fructan content was quantified using a commercial enzymatic assay kit (H‐FRUCHK; Megazyme) according to the AOAC Method 999.03 (McCleary et al. 2000), and absorbance was measured at 340 nm. Fructan content was calculated based on a calibration curve prepared using the kit standards and expressed as g/100 g dry weight (DW) based on the sample dry matter. All measurements were performed at least in triplicate using independently prepared samples.

2.5. In Vitro Protein Digestibility of White Pan Bread

In vitro gastrointestinal digestion of white pan bread was performed according to the standardized INFOGEST 2.0 static digestion method (Brodkorb et al. 2019), with minor modifications. All simulated digestive fluids were preheated to 37°C before use. Homogenized white pan bread crumb (10 g) was mixed 1:1 (vol/vol) with simulated gastric fluid containing pepsin from porcine gastric mucosa (Merck, Darmstadt, Germany) at 2000 U/mL in the final gastric mixture, adjusted to pH 3.0, and incubated at 37°C for 2 h. The resulting gastric chyme was subsequently mixed 1:1 (vol/vol) with simulated intestinal fluid supplemented with pancreatin from porcine pancreas (Merck) at a trypsin activity of 100 U/mL and bovine bile salts (Sigma‐Aldrich, St. Louis, MO, USA) at 10 mM in the final intestinal mixture, adjusted to pH 7.0, and incubated at 37°C for a further 2 h. The protein concentration of the bread sample was determined prior to digestion using the micro‐Kjeldahl method (Łysakowska et al. 2025). For free amino acid (FAA) extraction, 2 mL of the intestinal digesta was mixed with an equal volume of 16% trichloroacetic acid, shaken for 15 min, and centrifuged at 3000 rpm for 15 min (Ministry of Food and Drug Safety 2025). The resulting supernatant was used as the FAA extract, and FAA content was quantified using an amino acid analyzer. In vitro protein digestibility was calculated as the percentage of FAAs released after gastrointestinal digestion relative to the total protein content before digestion (Takagi et al. 2003; Moretton et al. 2023) as follows:

$$\begin{array}{ccc}& & \mathrm{In}\mathrm{vitro}\mathrm{protein}\mathrm{digestibility}\left(\%\right)\hfill \\ & & =\left(\frac{\mathrm{FAA}\mathrm{after}\mathrm{digestion},\mathrm{mg}/\mathrm{g}\mathrm{sample}}{\mathrm{Protein}\mathrm{before}\mathrm{digestion},\mathrm{mg}/\mathrm{g}\mathrm{sample}}\right)\times 100.\hfill \end{array}$$

All measurements were performed in at least triplicate using independently prepared samples.

2.6. Physicochemical Analysis

The physicochemical properties of the white pan bread were analyzed to evaluate baking quality. Loaf weight was measured using an electronic balance (ML3002T; Mettler Toledo, Zürich, Switzerland), and loaf volume was determined using a Volscan Profiler 300 (Stable Micro Systems, Godalming, UK). The specific volume was calculated as the volume‐to‐weight ratio (mL/g). Texture analysis was performed using texture profile analysis with a texture analyzer (TA‐XT plus, Stable Micro Systems) (Moretton et al. 2023). For the analysis, bread crumb samples were sliced to a thickness of 10 mm, with two slices stacked before measurement, and analyzed under the following conditions: compression mode with a pre‐test speed of 1.0 mm/s, test speed of 1.0 mm/s, post‐test speed of 5.0 mm/s, compression distance of 50%, and a P45 probe. Hardness and springiness were calculated from the force–distance curve. The moisture content of the crumbs was determined using an infrared moisture analyzer (HX204; Mettler Toledo) according to the manufacturer's instructions. For pH and titratable total acidity (TTA) analysis, 10 g of the crumb sample was mixed with 90 mL of distilled water and homogenized. The homogenate was filtered, and the filtrate was used for subsequent measurements. The pH was measured using a pH meter, and the TTA was expressed as the volume of NaOH (0.1 N) required to titrate the filtrate to pH 8.3. All measurements were performed at least in triplicate using independently prepared samples.

2.7. Amino Acid Composition Analysis

The FAA composition of the white pan bread samples was analyzed by high‐performance liquid chromatography (HPLC) using the Agilent o‐phthalaldehyde (OPA)/9‐fluorenylmethyl chloroformate (FMOC) derivatization method. The crumb samples were lyophilized, ground, and homogenized prior to extraction. To extract FAAs, 1 g of sample was mixed with 10 mL of 70% (v/v) ethanol and stirred for 30 min at ambient temperature. The extract was centrifuged (10,000 × g, 10 min, 4°C), and the supernatant was filtered through a 0.22‐µm syringe filter prior to HPLC analysis. FAA was analyzed using an HPLC system equipped with an automated derivatization module (Agilent 1260 Infinity II LC system; Agilent Technologies, Santa Clara, CA, USA). Primary amino acids were pre‐column derivatized with OPA, whereas secondary amino acids were derivatized with FMOC (Soma et al. 2022). Derivatization was performed automatically in an autosampler according to the manufacturer's instructions. The injection volume was 5 µL. Amino acid separation was performed at 40°C using a reversed‐phase column (ZORBAX Eclipse AAA column, 4.6 × 150 mm, 5 µm; Agilent Technologies). The mobile phase consisted of solvent A (40 mM sodium phosphate buffer, pH 7.8) and solvent B (acetonitrile:methanol:water = 45:45:10, v/v/v) at a flow rate of 1.5 mL/min. The gradient program was as follows: 0% B initially, increased linearly to 57% B over 18.1 min, increased to 80% B from 18.1 to 18.6 min, maintained until 21.0 min, and then returned to initial conditions, with a total run time of 26 min. The derivatized amino acids were detected using a fluorescence detector, with OPA derivatives monitored at an excitation wavelength of 340 nm and emission at 450 nm, and FMOC derivatives monitored at an excitation wavelength of 266 nm and emission at 305 nm. The quantification of individual amino acids was performed using an external calibration curve generated from a commercial amino acid standard mixture (Sigma‐Aldrich), and FAA content was expressed as mg/kg DW based on the sample dry matter. All measurements were performed at least in triplicate using independently prepared samples.

2.8. Instrumental Sensory Analysis

The aroma and taste characteristics of baked white pan bread were evaluated using instrumental sensory analysis based on volatile organic compound (VOC) profiling and electronic tongue (e‐tongue) measurements. Volatile aroma compounds were analyzed using headspace solid‐phase microextraction coupled with gas chromatography–mass spectrometry (HS‐SPME–GC/MS) (E.‐J. Lee et al. 2025; Rodriguez‐Campos et al. 2011). The HS‐SPME–GC/MS analysis provided qualitative identification and semi‐quantitative comparison of VOCs based on GC‐MS peak areas, rather than absolute quantification. Three grams of homogenized crumb sample were placed in a 20‐mL headspace vial, sealed, and equilibrated at 40°C for 30 min. Volatile compounds were extracted using an 85‐µm CAR/PDMS fiber (Supelco, Bellefonte, PA, USA) at the same temperature for 30 min (E.‐J. Lee et al. 2025); the fiber was pre‐conditioned according to the manufacturer's instructions. The fiber was then thermally desorbed in the GC injection port at 250°C in splitless mode. The analysis was performed using an Agilent 7890 B gas chromatograph coupled with a 5977A mass‐selective detector (Agilent Technologies). Separation was carried out using a DB‐WAX capillary column (30 m × 0.25 mm i.d., 0.25 µm film thickness) with helium as the carrier gas at a constant flow rate of 1.0 mL/min. The oven temperature program was held at 40°C for 3 min, increased at 5°C/min to 230°C, and maintained for 5 min. Mass spectrometry was performed under electron ionization (70 eV) conditions with a scan range of m/z 35–450. Volatile compounds were identified using the Wiley12 (Wiley Science Solutions; https://sciencesolutions.wiley.com) and NIST20 (National Institute of Standards and Technology; https://chemdata.nist.gov) mass spectral libraries; compounds were retained when the match score exceeded the ≥80% threshold and were further confirmed by comparison of retention indices with published values where available. The identified volatile compounds were categorized into aroma groups (floral/rose, fruity, fermented, buttery/creamy, green/fatty/sour, and caramel/baked) based on their representative aroma descriptors as previously reported (Hazelwood et al. 2008; Thiele et al. 2002; Tian et al. 2020; Wang et al. 2023; S. Liu et al. 2022; Qi et al. 2025). The peak area of each identified VOC was used for semiquantitative comparison between treatments. Relative VOC distributions were calculated as the percentage contribution of each compound or aroma group to the total peak area of all identified VOCs.

Taste characteristics were analyzed using an e‐tongue (SA402B; Insent Inc., Atsugi, Japan) (Kobayashi et al. 2010; Woertz et al. 2011). Crumb samples were mixed with distilled water at a ratio of 1:3 (w/v), homogenized, and centrifuged at 3000 × g for 10 min at room temperature, and the resulting supernatant was used for analysis. The E‐tongue measurements were conducted using six sensors (GL1, CT0, C00, AAE, CA0, and AE1), each conditioned and washed according to the manufacturer's guidelines. A reference solution consisting of 30 mM potassium chloride and 0.3 mM tartaric acid was used to stabilize the sensor responses during the measurements. All analyses were performed at least in triplicate using independently prepared samples.

2.9. Statistical Analysis

All experiments were performed in at least triplicate, and the data are expressed as mean ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism (version 10.3; GraphPad Software, San Diego, CA, USA). Differences between the two groups were analyzed using an unpaired two‐tailed Student's t‐test. Statistical significance was set at p <0.05.

3. Results and Discussion

3.1. Yeast Strain Selection Based on Fructan Reduction and Phytate Degradation

Seventeen S. cerevisiae strains isolated from nuruk were screened for fructan‐reducing and phytate‐degrading activities under dough fermentation conditions. Fructan‐reducing activity varied substantially among strains, with relative fructan levels ranging from approximately 13% to 75% of the non‐inoculated control (Figure S1). Among all tested strains, SC‐8303 exhibited the lowest residual fructan level, corresponding to approximately 13% of the non‐inoculated control, and was therefore selected as the primary candidate for subsequent baking experiments. Phytate‐degrading activity also differed among strains, with SC‐8303 showing the greatest reduction in residual phytate at approximately 54% of the non‐inoculated control compared to the positive control at approximately 66%.

Collectively, these screening results demonstrate that fructan‐reducing activity during dough fermentation varies substantially depending on the yeast strain used. Fructans are recognized as major FODMAP carbohydrates in wheat‐based foods and are associated with gastrointestinal symptoms owing to their rapid fermentation in the colon (Gibson and Shepherd 2010; Skodje et al. 2018). In this context, the lower fructan content observed in SC‐8303 fermented dough may provide a compositional basis for improving the gastrointestinal tolerance of wheat‐based bakery products. Similar strain‐dependent variability in fructan degradation has been previously reported. Laurent et al. (2020) showed that selected high‐invertase S. cerevisiae strains degraded 82.3%–92.3% of wheat grain fructans after 3 h of dough fermentation, whereas low‐invertase strains degraded 63.6%–64.5% under the same conditions. In the present study, residual fructan levels after dough fermentation were approximately 45% and 13% of the non‐inoculated dough control for SC‐RS and SC‐8303, respectively, corresponding to reductions of approximately 55% and 87%. These results are consistent with previous reports showing that fructan degradation during dough fermentation is strongly influenced by yeast strain characteristics and demonstrate the superior fructan‐degrading capacity of SC‐8303. Phytate‐degrading activity provided an additional basis for distinguishing strain‐dependent compositional properties at the screening stage. Among the tested strains, SC‐8303 showed the lowest residual phytate level, suggesting a higher phytate‐reducing potential under the assay conditions. Phytate is an antinutritional factor that can form complexes with minerals, proteins, and starch, thereby reducing mineral bioavailability and limiting enzymatic nutrient digestion, which may contribute to reduced digestive suitability of cereal‐based foods (Lopez et al. 2002). Microbial fermentation, including yeast‐ and mixed culture‐mediated processes, has been reported to contribute to phytate degradation, although the extent of degradation depends on the microbial composition and fermentation conditions (Fraberger et al. 2018). Accordingly, the phytate‐degrading activity observed in SC‐8303 was interpreted as supplementary strain characterization information. However, as phytate content was not quantified in the final bread matrix, no conclusions can be drawn regarding its effects on mineral bioavailability, nutrient digestibility, or digestive tolerance in the baked product. Taken together, SC‐8303 was selected primarily on the basis of its superior fructan‐reducing activity, whereas phytate‐degrading activity was used only as supplementary strain characterization information during screening. These findings support targeted yeast strain selection as a practical approach for reducing fructan content in cereal‐based products without relying solely on the modification of processing conditions.

3.2. Effects of SC‐8303 on Fructan Content and in Vitro Digestibility

The fructan content and in vitro protein digestibility of white pan bread prepared using SC‐8303 and SC‐RS are shown in Figure 1. SC‐8303 bread exhibited a significantly lower fructan content (0.13 ± 0.05 g/100 g DW) compared with SC‐RS bread (0.37 ± 0.03 g/100 g DW), representing an approximately 65% reduction (p < 0.05). In vitro protein digestibility was significantly higher in SC‐8303 bread (34.2 ± 3.4%) than in SC‐RS bread (27.6 ± 1.7%) (p < 0.05).

FIGURE 1.

Fructan content and in vitro protein digestibility of white pan bread prepared using different Saccharomyces cerevisiae strains. (A) Fructan content of white pan bread. (B) In vitro protein digestibility of white pan bread. SC‐RS, commercial (control) yeast strain Saccharomyces cerevisiae Red Star; SC‐8303, Saccharomyces cerevisiae SC‐8303. Protein digestibility was calculated as the ratio of free amino acid concentration after in vitro digestion to the protein concentration before digestion. Values are presented as mean ± SEM. * and ** indicate significant differences between groups at p < 0.05 and p < 0.01, respectively.

The lower fructan content in the SC‐8303 bread was consistent with the screening results obtained under dough fermentation conditions, confirming a strain‐dependent fructan reduction in the final baked product. The in vitro protein digestibility was evaluated under simulated gastrointestinal conditions. This parameter reflects the relative release of free amino acids during the in vitro digestion. Previous studies have shown that the fermentation of cereal‐based systems can influence protein digestibility through proteolysis and structural modification of the matrix, leading to enhanced susceptibility to enzymatic hydrolysis (Moretton et al. 2023; Torcello‐Gómez et al. 2020). However, our analysis was based on free amino acid release after digestion, rather than the direct measurement of protein digestibility or matrix structure. Therefore, the higher in vitro protein digestibility observed in SC‐8303 bread reflects a greater measurable release of free amino acids under in vitro conditions rather than direct evidence of enhanced protein digestibility mechanisms. Fermentation is also known to increase the release of low‐molecular‐weight compounds, including free amino acids and peptides, through proteolytic activity during dough fermentation (Thiele et al. 2002). The higher in vitro protein digestibility observed in SC‐8303 bread was consistent with these general trends, although the underlying mechanisms were not directly evaluated.

3.3. Effects of SC‐8303 on Baking Quality and Textural Properties

Table 1 summarizes the baking quality parameters of white pan bread prepared using SC‐8303 and the commercial strain SC‐RS. Loaf volume and weight did not differ between the two treatments. The specific volume was slightly higher in the SC‐8303 bread, but the difference was not significant (p = 0.0732). Crumb hardness was significantly lower in the SC‐8303 bread than in the SC‐RS bread (p = 0.0139), whereas springiness did not differ significantly between the two. The moisture content of the bread crumb was 6.47 ± 0.94% in SC‐RS and 6.41 ± 0.42% in SC‐8303 bread. The pH and TTA were comparable between the two breads.

TABLE 1.

Physicochemical and baking quality characteristics of white pan bread prepared using different Saccharomyces cerevisiae strains.

Contents	SC‐RS	SC‐8303	p value
Loaf volume (mL)	2,151.27 ± 209.48	2,219.42 ± 164.63	0.8111
Weight (g)	452.31 ± 1.73	453.68 ± 2.31	0.7465
Specific volume (mL/g)	4.76 ± 0.45	5.10 ± 0.38	0.0732
Hardness (force, g)	6,115.08 ± 449.76	5002.74 ± 613.15 a	0.0139
Springiness	0.93 ± 0.02	0.96 ± 0.03	0.4232
TTA (mL)	0.21 ± 0.02	0.19 ± 0.01	0.5145
pH	4.70 ± 0.19	4.75 ± 0.22	0.5145
Moisture (%)	6.47 ± 0.94	6.41 ± 0.42	0.8147

Note: Values are presented as mean ± SEM.

Abbreviation: SC‐RS, commercial (control) yeast strain Saccharomyces cerevisiae Red Star; SC‐8303, Saccharomyces cerevisiae 8303; specific volume, loaf volume/weight; TTA, titratable total acidity.

^a Significant difference between SC‐RS and SC‐8303 (p < 0.05).

Specific volume reflects the ability of the dough to retain fermentation gases within the gluten network (Kokawa et al. 2017). There was no difference in loaf volume or specific volume between breads produced using SC‐8303 and the commercial strain, indicating comparable gas production and retention capacities. Although specific volume was slightly higher in SC‐8303 bread, this difference was not statistically significant, indicating that loaf expansion was not significantly affected by the yeast strain. In contrast, crumb hardness is influenced by the internal structure of the bread matrix, including the crumb cell structure and protein–starch interactions during baking (Moretton et al. 2023). Because moisture content, pH, and TTA were similar between treatments, the lower hardness observed in SC‐8303 bread is unlikely due to differences in water content or acidity and likely reflects the differences in the crumb structure formed during fermentation and baking. The significant decrease in crumb hardness therefore indicates a strain‐associated textural difference, while the overall baking performance remained largely unchanged. Previous studies have reported that fermentation‐based FODMAP reduction can alter bread structure and textural properties depending on the extent of carbohydrate degradation and fermentation conditions (Fraberger et al. 2018; Pejcz et al. 2023). In contrast, our findings indicate that the SC‐8303‐mediated reduction in fructan content did not adversely affect the major baking quality parameters of white pan bread. Collectively, these results suggest that despite the SC‐8303‐mediated reduction in fructan content, key baking quality parameters were maintained in SC‐8303 bread, indicating that strain selection enabled the modification of carbohydrate composition without measurable changes in overall baking performance.

3.4. Effects on FAA Composition

Table 2 summarizes the FAA composition of white pan bread prepared using SC‐8303 and SC‐RS. Histidine and leucine were not detected in either, whereas alanine, proline, valine, and tryptophan levels were comparable in both breads. In contrast, glutamine, glycine, and glutamic acid levels were significantly higher in the SC‐8303 bread than in the SC‐RS bread (p < 0.0001), with lysine levels also elevated (p = 0.0014). Arginine and threonine levels varied between the breads without reaching statistical significance.

TABLE 2.

Free amino acid composition of white pan bread prepared using different Saccharomyces cerevisiae strains.

Amino acids (mg/kg dry weight)	SC‐RS	SC‐8303	p value
Histidine	0.0 ± 0.0	0.0 ± 0.0	>0.9999
Glutamine	47.6 ± 0.4	86.1 ± 1.1 a	<0.0001
Arginine	36.3 ± 1.0	39.6 ± 0.8	0.2450
Glycine	35.8 ± 1.2	49.2 ± 0.8 a	<0.0001
Glutamic acid	84.8 ± 1.2	99.6 ± 2.0 a	<0.0001
Threonine	16.7 ± 0.7	13.6 ± 0.5	0.3261
Alanine	40.3 ± 1.1	39.7 ± 1.4	>0.9999
Proline	37.5 ± 1.7	40.4 ± 1.6	0.4221
Lysine	10.5 ± 0.5	16.4 ± 0.9 a	0.0014
Valine	13.3 ± 0.9	13.2 ± 1.0	>0.9999
Leucine	0.0 ± 0.0	0.0 ± 0.0	>0.9999
Tryptophan	15.6 ± 0.7	15.9 ± 0.7	>0.9999

Note: Values are presented as mean ± SEM.

Abbreviations: SC‐RS, commercial (control) yeast strain Saccharomyces cerevisiae Red Star; SC‐8303, Saccharomyces cerevisiae 8303.

^a Significant difference between SC‐RS and SC‐8303 (p < 0.05).

These results indicate that fermentation with SC‐8303 resulted in selective changes in specific amino acids rather than uniform increases in total FAA levels. In cereal‐based fermentation systems, proteins are hydrolyzed into peptides and free amino acids by endogenous enzymes and microbes (Moretton et al. 2023; Thiele et al. 2002). However, the final FAA composition is influenced by various factors during fermentation, and the present results reflect the differences in the resulting amino acid profiles. Selective changes in FAA composition have also been reported in wheat‐based fermented products, depending on the fermentation system and processing stage (Valerio et al. 2017; Y. Liu et al. 2023). Valerio et al. (2017) reported that the total FAA content of dough increased from 572 ± 48.7 mg/kg in the reference dough to 776 ± 13.2 mg/kg in dough containing a Lactobacillus plantarum fermentation product, although this difference was reduced after baking. Y. Liu et al. (2023) further showed that L. plantarum–yeast co‐fermentation selectively modified digestion‐released amino acid profiles in steamed bread, with increased lysine, alanine, and methionine levels, but decreased glutamic acid and total amino acid levels compared with yeast fermentation alone. These previous findings provide quantitative context for interpreting the present results, suggesting that amino acid‐related profiles in wheat‐based fermented products may vary depending on the yeast strain or starter system, fermentation conditions, processing stage, and analytical target. Free amino acids also function as key precursors for thermal reactions during baking, participate in the Maillard and Strecker reactions, and contribute to the formation of aroma‐active compounds (Ali et al. 2021). However, because only a limited number of amino acids differed significantly, these compositional changes are more likely to reflect subtle variations in amino acid profiles than major alterations in the overall product characteristics. In addition, lysine is a limiting amino acid in cereal proteins, and its increase in the free form may have nutritional relevance (Shewry and Hey 2015). However, because the present analysis was limited to FFA, further evaluation, including total amino acid composition and post‐digestion release, is required to assess the changes in nutritional availability.

3.5. Effects on Volatile Profile and Instrumental Taste Attributes

A total of 22 VOCs were identified in white pan bread prepared using SC‐RS and SC‐8303, and their relative peak areas are summarized in Table 3. The identified compounds were categorized into six aroma groups: floral/rose, fruity, fermented, buttery/creamy, green/fatty/sour, and caramel/baked (Figure 2A,B). As shown in Figure 2A, the overall VOC distribution patterns were highly similar between the two groups, with floral/rose compounds, predominantly phenethyl alcohol, accounting for the largest proportion of the total VOC profile in both treatments. Aroma category‐level comparisons revealed no significant differences in the floral/rose, fruity, fermented, buttery/creamy, or green/fatty/sour fractions between SC‐RS and SC‐8303 breads (Figure 2B). In contrast, the caramel/baked fraction was significantly higher in SC‐8303 bread than in SC‐RS bread (p < 0.01), which was primarily attributed to the exclusive detection of furfural in SC‐8303 bread. Among the individual compounds, phenethyl alcohol was detected at lower levels in SC‐8303 bread than in SC‐RS bread (p < 0.0001). Ethyl octanoate levels were lower in SC‐8303 bread; however, this difference was not significant (p = 0.0742). E‐tongue analysis indicated comparable taste profiles between the treatments, with no significant differences in sweetness, sourness, bitterness, saltiness, umami, or aftertaste (Figure 2C).

TABLE 3.

Volatile organic compound profiles of white pan bread prepared using different Saccharomyces cerevisiae strains.

Volatile compounds	GC‐MS peak area (×103 a.u.)		p value
	SC‐RS	SC‐8303
Phenethyl acetate	764.39 ± 36.47	802.54 ± 65.13	>0.9999
Phenethyl alcohol	69,446.96 ± 4352.37	63,180.84 ± 2021.84*	<0.0001
Limonene	1184.74 ± 77.04	1726.93 ± 89.80	>0.9999
Ethylhexanol	585.71 ± 43.78	497.05 ± 23.79	>0.9999
Ethyl decanoate	2184.19 ± 129.37	3379.87 ± 300.53	0.9993
Ethyl octanoate	13,465.60 ± 994.55	10,154.86 ± 655.22	0.0742
Ethyl hexanoate	1454.10 ± 100.59	1167.02 ± 88.10	>0.9999
Benzaldehyde	436.29 ± 23.25	2355.10 ± 204.84	0.8598
Isobutyl alcohol	2896.89 ± 220.25	2596.52 ± 265.78	>0.9999
Isoamyl alcohol	15,597.55 ± 1128.97	13,766.98 ± 567.30	0.9029
Diacetyl	1635.57 ± 90.16	1468.05 ± 122.39	>0.9999
Acetoin	5417.44 ± 349.49	4163.77 ± 432.92	0.9986
Nonanal	1673.07 ± 116.01	1087.75 ± 85.33	>0.9999
Heptenal	443.40 ± 23.95	623.93 ± 69.14	>0.9999
Heptyl alcohol	294.85 ± 14.17	266.27 ± 13.10	>0.9999
Hexanol	1388.58 ± 88.16	1557.90 ± 85.46	>0.9999
Furfural	0.00 ± 0.00	1251.68 ± 94.18****	<0.0001
Pentylfuran	1053.22 ± 62.25	1018.92 ± 87.82	>0.9999
Acetic acid	1028.08 ± 70.94	1374.14 ± 98.03	>0.9999
Dodecanol	777.15 ± 46.02	745.05 ± 49.82	>0.9999
Dodecenal	296.91 ± 14.13	354.75 ± 27.75	>0.9999
Decadienal	166.37 ± 8.35	154.05 ± 3.19	>0.9999

Note: Values are presented as mean ± SEM. The GC‐MS peak areas were obtained from the total ion chromatogram and are expressed as arbitrary units (a.u.).

Abbreviations: SC‐RS, commercial (control) yeast strain Saccharomyces cerevisiae Red Star; SC‐8303, Saccharomyces cerevisiae SC‐8303.

*p < 0.05, ****p < 0.0001 indicate significant differences between SC‐RS and SC‐8303.

FIGURE 2.

Volatile organic compound profiles and instrumental taste attributes of white pan bread prepared using different Saccharomyces cerevisiae strains. (A) Relative distribution of individual volatile compounds. (B) Relative peak area of volatile compounds grouped by aroma category. (C) Instrumental taste profiles measured using an electronic tongue system. SC‐RS, commercial reference strain Saccharomyces cerevisiae Red Star; SC‐8303, nuruk‐derived Saccharomyces cerevisiae SC‐8303. Volatile compounds are expressed as relative peak areas (% vs. total GC‐MS peak area). Values are presented as mean ± SEM (n = 3). Significant differences between treatments are indicated by **p < 0.01; ns, not significant.

Overall, VOC profiles were comparable between SC‐8303 and SC‐RS breads; no significant differences were observed in the floral/rose, fruity, fermented, buttery/creamy, or green/fatty/sour aroma fractions, which collectively accounted for the majority of the total VOC profiles in both treatments. VOCs in baked products can arise from various sources, including raw breadmaking ingredients, baking‐induced reactions such as the Maillard reaction and lipid oxidation, and microbial metabolism during fermentation (Ali et al. 2021; S. Liu et al. 2022). Because identical formulations and baking conditions were applied to both treatments, the observed differences between SC‐RS and SC‐8303 breads could be compared directly. However, the potential contribution of raw material‐derived precursors to the final VOC profile cannot be completely excluded because the raw ingredients were not analyzed separately. The major alcohols, esters, and carbonyl compounds were consistently detected in both treatments, with phenethyl alcohol being the dominant compound. Phenethyl alcohol and isoamyl alcohol, the primary contributors to the floral/rose and fermented fractions, respectively, are well‐established products of amino acid catabolism via the Ehrlich pathway in S. cerevisiae (Hazelwood et al. 2008), and their comparable relative proportions between treatments suggest that the overall fermentative metabolism of SC‐8303 did not differ markedly from that of SC‐RS. At the individual compound level, phenethyl alcohol was detected at lower levels in SC‐8303 bread, and no significant differences were observed in other esters, alcohols, aldehydes, or ketones between treatments. Although benzaldehyde showed a higher peak area in SC‐8303 bread, this difference was not significant (p = 0.8598), likely reflecting high intersample variability rather than a treatment effect. The buttery/creamy fraction, represented by diacetyl and acetoin, showed no significant differences between treatments, indicating that pyruvate metabolism during fermentation was similar in both strains (Tian et al. 2020). The caramel/baked fraction was significantly higher in SC‐8303 bread, primarily driven by the exclusive detection of furfural in SC‐8303 treatment. Furfural is not a direct yeast fermentation metabolite but is primarily generated during baking through the Maillard reaction and thermal degradation of reducing sugars (S. Liu et al. 2022; Qi et al. 2025). Therefore, its exclusive detection in SC‐8303 bread should be interpreted as a difference in the final baked bread matrix rather than direct evidence of strain‐specific VOC production. This difference may be associated with strain‐dependent changes in precursor availability during dough fermentation, but further analysis of raw materials and baking‐stage precursors is required to clarify its origin. These compound‐ and category‐level findings collectively indicate that fermentation with SC‐8303 induced only minor changes in the selected volatile components while preserving the overall VOC profile and instrumental taste characteristics of white pan bread. The overall similarity of VOC profiles between the two treatments further suggests that SC‐8303 maintained the aroma characteristics of standard white pan bread, with detectable differences limited to a small number of compounds. Similar strain‐dependent differences in bread aroma profiles have been reported in previous studies, in which fermentation with different yeast or starter cultures altered specific alcohols, esters, and Maillard‐related volatiles without causing significant changes in the overall sensory profile (Kokawa et al. 2017). Our findings are consistent with this pattern and support the suitability of SC‐8303 for standard breadmaking conditions without markedly altering the overall VOC and instrumental taste profiles.

In summary, our findings indicate that fermentation with a specific S. cerevisiae strain reduced fructan content, while maintaining baking quality and aroma‐related properties. However, the interpretation of these findings is subject to methodological limitations. Fructan reduction was used as a screening criterion and subsequently confirmed in the final bread, whereas phytate‐degrading activity was assessed only at the screening stage and was not quantified in the final product; therefore, the contribution of phytate degradation to the nutritional properties of baked bread remains to be established in future studies. The mechanisms underlying strain‐dependent fructan reduction, including potential differences in invertase activity or fermentation kinetics, were not directly investigated in the present study, and further mechanistic analyses are warranted. In addition, digestion‐related parameters were derived from an in vitro model and represented relative changes under controlled conditions rather than physiological digestion. Accordingly, the observed increase in in vitro protein digestibility should be interpreted with caution because it cannot be directly extrapolated to enhanced digestion in vivo. Further validation, using animal models or human studies, is required to determine the physiological relevance of our findings.

4. Conclusion

In this study, we investigated the effects of S. cerevisiae LRCC 8303 fermentation on fructan content in white pan bread and its associated effects on baking quality and digestion‐related parameters. Fermentation with SC‐8303 resulted in lower fructan levels than in the control strain, while key baking quality attributes, including specific volume, moisture content, and crumb texture, were preserved. SC‐8303 bread showed higher in vitro protein digestibility and selective changes in free amino acid composition, although localized differences were observed in selected individual volatile compounds, overall aroma category distributions, and instrumental taste characteristics remained comparable between the treatments. Although phytate‐degrading activity was identified during strain screening, this effect was not quantified in the final bread matrix and warrants further investigation. Because our findings are based on in vitro analyses, their physiological relevance remains to be established. Overall, our findings provide empirical support for yeast strain selection as a practical strategy for producing bread with reduced fructan content while maintaining product quality under standard bread‐making conditions.

Author Contributions

Jaeyool Jang: writing – original draft, investigation, methodology, validation. Eunju Lee: conceptualization, visualization, formal analysis, writing – original draft. Minju Seo: methodology, formal analysis, software, data curation. Insup Yun: investigation, visualization, data curation. Hyun Cho: project administration, validation, conceptualization. Wonseok Jung: supervision, resources, funding acquisition. Seokmin Yoon: writing – review and editing, project administration, data curation.

Conflicts of Interest

The authors declare no conflicts of interest.

Funding

No external funding was received for this study.

Supporting information

Supplementary Material: jfds71284‐sup‐0001‐FigureS1.docx

Data Availability Statement

The data generated and analyzed in this study are available from the corresponding author upon reasonable request.