Development and Semi-Industrial Evaluation of Symbiotic Multi-Strain Starter Cultures for Sourdough Bread Production

Abstract

Sourdough fermentation is an important technological process in bread production, influencing acidification, product quality, and storage stability. The present study developed and comparatively evaluated two-strain and multi-strain bacterial starter cultures for application in different flour matrices and under semi-industrial bread-making conditions. The starter cultures comprised selected strains of Lactiplantibacillus plantarum, Lacticaseibacillus rhamnosus, Levilactobacillus brevis, Limosilactobacillus fermentum, Fructilactobacillus sanfranciscensis, and Propionibacterium freudenreichii subsp. shermanii. Baker’s yeast was added separately during final dough preparation and was not part of the bacterial starter combinations. Starter performance was assessed through viable cell counts, titratable acidity, in vitro antimicrobial activity, dough fermentation time, bread volume, descriptive sensory evaluation, and the onset of visually detectable microbial spoilage under the tested storage conditions. The selected combinations adapted to the investigated flour matrices, maintained high viable cell concentrations, supported acidification, and were successfully applied in semi-industrial bread production. Several formulations showed favorable technological and descriptive sensory characteristics and delayed the appearance of visible bacterial and fungal spoilage compared with the corresponding controls. Overall, the findings demonstrate the practical potential of the developed multi-strain starter cultures for application across different flour matrices and provide a strong basis for further technological refinement and mechanistic characterization of these sourdough systems.

1. Introduction

1.1. Sourdough Fermentation as a Biotechnological Process: The Role of Lactic Acid Bacteria in Sourdough Fermentation

Sourdough fermentation is a complex biotechnological process based on the interaction between microorganisms, endogenous flour enzymes and biochemical transformations occurring within the flour matrix. During fermentation, carbohydrates and proteins are metabolized by microorganisms, leading to the formation of organic acids, carbon dioxide, ethanol and numerous aroma-active compounds. These metabolites influence bread structure, flavor, nutritional value and overall product quality [1,2,3].

The sourdough microbial ecosystem is mainly composed of lactic acid bacteria (LAB) and yeasts, which coexist in stable or semi-stable microbial associations. LAB are primarily responsible for dough acidification through the production of lactic and acetic acids. The decrease in pH creates selective conditions that restrict the growth of many spoilage microorganisms and contributes to the microbiological stability of the product [2,3,4,5]. Yeasts, in turn, contribute mainly to gas formation and dough leavening, supporting the development of loaf volume and crumb structure [1,4].

Sourdough fermentation may be initiated either by spontaneous propagation of flour- and environment-associated microbiota or by the application of selected and standardized starter cultures. Spontaneous sourdoughs remain important in traditional and artisanal production, whereas defined starters are increasingly used when reproducibility, process standardization, and consistent product quality are required [2,5,6].

Sourdough starter cultures usually include selected LAB strains with high fermentative activity, good adaptation to the flour environment and the ability to produce organic acids, flavor compounds and antimicrobial metabolites. The application of carefully selected starters with optimized composition provides better control over fermentation dynamics, improves technological performance and contributes to the development of the characteristic sensory profile of sourdough bread [2,6,7].

The LAB species most commonly associated with sourdough ecosystems include Lactiplantibacillus plantarum, Lacticaseibacillus rhamnosus, Levilactobacillus brevis, Limosilactobacillus fermentum and Fructilactobacillus sanfranciscensis. These species are technologically important because of their acidifying capacity, metabolic activity and ability to contribute to flavor formation and microbial stability [3,4,5,8].

The metabolic activity of LAB also has a direct effect on dough rheology and bread quality. Through carbohydrate metabolism and proteolytic activity, LAB may influence gluten structure and release peptides and amino acids that serve as precursors of aroma compounds [1,3]. In addition, some LAB strains produce exopolysaccharides, which can improve crumb texture, moisture retention and freshness during storage [1,9].

Another important technological property of LAB is their ability to produce antimicrobial compounds, including organic acids, bacteriocins and other inhibitory metabolites that suppress the growth of spoilage microorganisms. This characteristic is especially relevant for clean-label bakery products, where natural preservation strategies are preferred to chemical preservatives [10,11,12].

Homofermentative and heterofermentative LAB strains contribute differently to sourdough fermentation. Homofermentative LAB mainly produce lactic acid, which lowers dough pH and contributes to bread safety and flavor development. Heterofermentative LAB produce lactic acid together with acetic acid, carbon dioxide, and other metabolites, thereby contributing to a more complex flavor profile and to textural development in sourdough bread [3,13].

The effectiveness of starter strains is largely determined by their ability to acidify the food matrix rapidly, produce desirable flavor and aroma compounds, and inhibit saprophytic and pathogenic microorganisms [7,8,10]. Balanced microbial associations allow for more precise regulation of pH, flavor profile and microbiological stability. These characteristics make LAB an important factor in the optimization of process parameters and final product quality. The interactions among LAB, yeasts and the flour matrix form a dynamic ecosystem whose efficiency depends strongly on the adaptability and compatibility of the individual microorganisms [4,5,14].

The balance among microorganisms participating in sourdough fermentation is important for acidification, gas formation, flavor development, and bread quality.

1.2. Influence of Flour Matrix on Fermentation

The composition of the flour matrix is one of the main factors determining sourdough fermentation dynamics. Flour provides fermentable carbohydrates, proteins, minerals and enzymes that support microbial growth and metabolic activity. Therefore, differences in flour type, milling process and extraction rate may influence microbial development, acidification kinetics and the formation of flavor compounds during fermentation [2,5,15]. Whole grain and rye flours generally contain higher levels of nutrients, minerals and endogenous enzymes than refined wheat flour. These characteristics may stimulate microbial growth and increase fermentation activity. At the same time, the buffering capacity and carbohydrate composition of the flour matrix can affect the rate of pH decrease and the balance between lactic and acetic acid production [3,5,15].

Understanding the interaction between microbial cultures and the flour environment is therefore essential for optimizing sourdough fermentation and for developing starter cultures adapted to specific raw materials and technological conditions.

1.3. Functional Ingredients and Enrichment with By-Products

In the food industry, cereal and nut-derived by-products can be used to enrich bakery products with dietary fiber, proteins, minerals, vitamins and bioactive compounds, thereby increasing their nutritional value and functional potential [15,16,17]. Malt bran and related brewing by-products are particularly interesting due to their fiber and protein content and their possible contribution to fermentation performance and bread quality [16,17]. Almond flour, on the other hand, provides plant proteins, lipids, minerals and antioxidant compounds and can be used as a functional ingredient in bakery formulations [18]. The fortification of bakery products with fiber-rich materials, plant ingredients and bioactive compounds is also associated with improved nutritional value, potential health-related effects and, in some cases, improved product stability [15,19].

1.4. Microbial Stability and Natural Preservation

Microbial spoilage remains a major challenge in bread production, particularly rope spoilage caused by members of the genus Bacillus and fungal spoilage during storage. Sourdough fermentation has long been recognized as a natural biopreservation strategy because LAB produce organic acids and other antimicrobial metabolites with inhibitory effects against spoilage microorganisms [10,11,12]. The antimicrobial activity of LAB can inhibit the growth and development of undesirable microorganisms and extend the shelf-life of bread products. Several studies have shown that selected LAB strains produce compounds with antifungal activity, supporting fungal inhibition and improving the microbiological stability of bakery products [11,12,20,21]. The selection of LAB strains with both technological and antimicrobial properties is therefore an important strategy for improving sourdough bread quality and safety. This highlights the potential role of symbiotic starters in sustainable technological approaches to bread production.

1.5. Study Rationale and Objectives

Previous studies indicate that selected sourdough microorganisms can influence acidification, technological performance, sensory properties, and resistance to spoilage. Nevertheless, the response of defined multi-strain combinations may vary considerably with flour composition and production conditions. A systematic comparative evaluation across different flour matrices and subsequent validation under semi-industrial conditions is therefore needed.

The aim of this study was to develop and comparatively evaluate two-strain and multi-strain bacterial starter cultures for sourdough fermentation in different flour matrices and to assess their technological performance, in vitro antimicrobial activity, descriptive sensory characteristics, and contribution to storage stability in selected bread formulations under semi-industrial production conditions. The study was designed as an application-oriented evaluation of microbiological and technological performance, providing a basis for the practical selection of starter–flour combinations.

The experimental workflow, linking starter composition, flour matrix, fermentation, and the evaluated bread quality indicators, is summarized schematically in Figure 1.

2. Materials and Methods

2.1. Experimental Concept and Study Design

The study was designed as a comparative screening and semi-industrial evaluation of defined bacterial starter combinations under fixed fermentation and bread-making conditions. In the present study, the investigated starter combinations were bacterial; baker’s yeast was added separately during final dough preparation. Therefore, the term symbiotic starter is used here to describe compatible bacterial associations with complementary technological functions. The work did not include mathematical process optimization, response surface methodology, or kinetic modeling because the experimental design was intended for comparative evaluation and practical starter selection rather than formal mathematical optimization. The experimental factors included starter composition, flour type, sourdough inclusion level, and selected functional ingredients. The main response variables were viable cell concentration, titratable acidity, inhibition zone diameter, dough rise time, loaf volume, descriptive sensory scores, and time to visually detectable microbial spoilage.

Experimental groups and controls. The experimental sequence comprised (i) single-strain sourdoughs prepared with each flour; (ii) two-strain sourdoughs (Ph2 + X4); (iii) three multi-strain combinations and the respective sourdoughs; (iv) control breads prepared without bacterial starter sourdough but with the same flour, water, salt, baker’s yeast, and processing conditions; and (v) experimental breads containing the specified percentage of the respective 96 h starter sourdough with the respective multi-strain starter. Comparisons across flour types were made within the same starter category, whereas the effect of sourdough addition was assessed against the corresponding flour-matched control.

2.2. Microorganisms and Preparation of Starter Cultures

The starter cultures were prepared from selected bacterial strains with previously established technological potential for sourdough fermentation: Lp. plantarum Ph2, L. brevis X4, L. rhamnosus LBRC11, L. fermentum LBRH10, F. sanfranciscensis R, and Pr. freudenreichii subsp. shermanii NBIMCC 327 [1,2,8,22,23,24]. Baker’s yeast was not included in these bacterial starter combinations; commercial baker’s yeast was added separately during final dough preparation as described in Section 2.5. Saccharomyces cerevisiae ATCC 9763 was used only as an indicator microorganism in the antimicrobial assay.

The combinations were selected on the basis of preliminary experimental observations and previously reported technological properties of the component strains, with the aim of combining acidifying, heterofermentative, and propionic acid bacterial activities. The numerical ratios refer to culture suspension volumes. Before mixing, the viable cell concentration of each culture was determined by serial dilution followed by the spread-plate method and was approximately 1 × 10¹¹ CFU/cm³.

• Two-strain combination: Lp. plantarum Ph2:L. brevis X4 = 2:1.
• Multi-strain combination 1: Lp. plantarum Ph2, L. brevis X4, L. rhamnosus LBRC11, and L. fermentum LBRH10 in a ratio of 2:1:1:1, respectively.
• Multi-strain combination 2: Multi-strain combination 1 (Lp. plantarum Ph2, L. brevis X4, L. rhamnosus LBRC11, and L. fermentum LBRH10) and F. sanfranciscensis R in a ratio of 2:1, respectively.
• Multi-strain combination 3: Multi-strain combination 1 (Lp. plantarum Ph2, L. brevis X4, L. rhamnosus LBRC11, and L. fermentum LBRH10), F. sanfranciscensis R, and Pr. freudenreichii subsp. shermanii NBIMCC 327 in a ratio of 2:1:1, respectively.

The following microorganisms were used as indicators in the antimicrobial assays: Bacillus subtilis ATCC 6633, Bacillus mesentericus, Saccharomyces cerevisiae ATCC 9763, Aspergillus niger ATCC 16604, Penicillium chrysogenum ATCC 10106, and Rhizopus oryzae ATCC 11145.

The saprophytic microorganisms belong to the collection of the Department of Microbiology and Biotechnology at the University of Food Technologies, Plovdiv, and were used as indicator microorganisms in the antimicrobial activity assays. The strains were cultivated on LBG agar at 30 °C.

2.3. Nutrient Media

2.3.1. LAPTg10 Broth

Composition (g/dm³): peptone—15; yeast extract—10; tryptone—10; glucose—10. The pH was adjusted to 6.6–6.8, and Tween 80 was added at 1 cm³/dm³. The medium was sterilized at 121 °C for 20 min and used for cultivation of lactic acid bacteria.

2.3.2. LAPTg10 Agar

Composition (g/dm³): LAPTg10 broth supplemented with 20 g/dm³ agar. The medium was sterilized at 121 °C for 20 min and used to determine the viable cell concentration of lactic acid bacteria.

2.3.3. LBG Agar

Composition (g/dm³): tryptone—10; yeast extract—5; NaCl—10; glucose—10; agar—20. The pH was adjusted to 7.5. The medium was sterilized at 121 °C for 20 min and used to determine the viable cell concentration of pathogenic and saprophytic microorganisms.

2.3.4. Elective Medium for Propionibacterium sp.

Composition (g/dm³): tryptone—10; yeast extract—10; sodium lactate—10; KH₂PO₄—2.5; MnSO₄—0.005; agar—20. The sodium lactate solution was freshly prepared by neutralizing 7.0 g of lactic acid with 3.1 g of NaOH before addition of the remaining salts. The pH was adjusted to 6.8, and the medium was sterilized at 121 °C for 20 min.

2.4. Raw Materials, Flour Characterization, and Dough Formulation

Flours with different compositions and bread-making properties were used to evaluate the adaptation of the starter cultures to distinct flour matrices. Water addition was adjusted according to the measured water absorption of each flour, whereas the remaining processing variables were kept constant within each comparison. Differences among wheat, rye, spelt, and einkorn flours were therefore treated as a matrix effect and interpreted in relation to their physicochemical characteristics presented in

Table 1. Physicochemical parameters of wheat flour type 500, rye flour, einkorn flour, spelt flour. FN—falling number, WGY—wet gluten yield, and GR—gluten relaxation.

Flour Type	Protein	Moisture	Particle Size	Ash	Water Absorption	FN	WGY	GR
-	%	%	µm	%	%	S	%	mm
wheat flour type 500	12.3	12.9	150–180	0.60	56	307	27.0	5.0
rye flour	8.0	12.4	120–150	1.18	58	181	-	-
spelt flour	12.4	11.5	200–250	1.40	56	287	27.6	5.0
einkorn flour	9.6	12.5	200–250	1.12	54	279	16.0	6.0

.

The following flour types were used in the present study (Table 1): wheat flour type 500 and rye flour from the Dimitar Pilev mill, Konush, Bulgaria; einkorn and spelt flours from Ecosem Bulgaria Ltd., Sofia, Bulgaria; almond flour from KING NUTS & RAAPHORST, Groote Wetering, Spanjeweg, Bodegraven, Nederland; and malt bran supplied by Jagerhof Brewery, Plovdiv, Bulgaria. The physicochemical characteristics of wheat flour type 500, rye flour, einkorn flour, and spelt flour included protein, moisture, particle size, ash content, water absorption, falling number, wet gluten yield, and gluten relaxation. These parameters were determined using the ICC-standardized Inframatic 8600 system, PerkinEmler, Stockholm, Sweden, a universal near-infrared reflectance analyzer.

2.5. Fermentation Conditions and Two-Stage Bread-Making Procedure

Bread was produced using a two-stage procedure. In the first stage, inoculated sourdoughs were propagated by daily back-slopping for 96 h. In the second stage, the mature sourdough was incorporated into the final dough at the specified percentage relative to flour mass. Final doughs contained 1.5% salt, 2% commercial baker’s yeast, the specified percentage of sourdough, and drinking water adjusted according to flour water absorption. The components were mixed for 4 min at 40 rpm and for 8 min at 90 rpm, followed by a 10 min rest. Dough portions of 500 g were shaped and placed in rectangular baking tins (approximately 25 × 10 × 7 cm), then proofed at 35 °C and 80 ± 5% relative humidity for approximately 40–45 min.

2.6. Bread Production and Control Samples

After proofing, all samples were baked at 225 ± 5 °C for 22 min under semi-industrial conditions and cooled at room temperature for 120 min. For each basic bread formulation, the control was prepared with the same flour, water, salt, baker’s yeast, mixing, proofing, and baking conditions, but without the bacterial sourdough starter. For formulations containing almond flour or malt bran, no separate matrix-matched control containing the same level of the respective ingredient but without bacterial starter sourdough was included. Consequently, the effects of functional-ingredient incorporation and bacterial starter addition cannot be fully separated for these formulations.

2.7. Analysis of Microbiological, Fermentation, Technological, and Sensory Indicators

The sourdough and bread formulations were evaluated using viable cell counts, titratable acidity, agar well diffusion assays, dough rise time, loaf volume, descriptive sensory evaluation, and observation of spoilage onset during storage. Unless otherwise stated, analyses were performed in triplicate.

2.7.1. Biochemical Methods

A.

Determination of the titratable acidity of sourdough/yeast dough according to Vangelov and Karadjov (1993) [25].

B.

Determination of the titratable acidity of bread dough according to Vangelov and Karadjov (1993) [25]. 

2.7.2. Microbiological Methods

A.

Determination of viable microorganisms. Appropriate tenfold serial dilutions of each sample were prepared in sterile saline. Selected dilutions were then spread-plated or pour-plated on the corresponding agar medium. The inoculated Petri dishes and/or tubes were incubated for 3 days at the optimal growth temperature for the respective microorganism until the appearance of countable, isolated colonies.

B.

Determination of antimicrobial activity by the agar well diffusion method. Ninety-six-hour sourdoughs were diluted 1:1 (v/v) with sterile water. Indicator suspensions containing approximately 10⁶–10⁷ CFU/mL were incorporated into LBG agar. After solidification, 6 mm wells were prepared and filled with 0.06 mL of each sample, and the assays were performed in triplicate. Plates were incubated at 30 °C for 48 h. Antimicrobial activity was expressed as the total inhibition zone diameter in millimeters, including the 6 mm well. Sterile water was used as the negative control. The sourdough samples were tested at their native acidity; pH-neutralized samples were not evaluated.

2.7.3. Technological Methods: Preparation of Sourdough Starters and Testing Under Production Conditions

A.

Preparation of cell suspensions for inoculation. MRS broth (10 mL) was inoculated with 1% of the respective bacterial culture and incubated for 24 h at 30 °C or 37 °C, according to the strain. Biomass was harvested by centrifugation at 6000× g for 15 min at 4 °C, and the pellet was resuspended in sterile saline to the initial culture volume. Before mixing, the viable cell concentration of each individual culture suspension was determined by serial dilution followed by spread plating and was approximately 1 × 10¹¹ CFU/cm³. For multi-strain starters, the component suspensions were combined according to the predetermined ratios on a volumetric basis; the reported ratios therefore refer to culture-suspension volumes.

B.

Preparation of single-strain sourdough. Single-strain sourdough was prepared as follows. A flour suspension containing 44% flour and 56% tap water was prepared at 35 ± 1 °C and inoculated with 2% fresh 24 h culture of the respective strain. The mixture was homogenized with a spatula and incubated for 24 h at 30 ± 1 °C. The resulting single-strain sourdoughs were then subjected to daily back-slopping up to 96 h using 25% sourdough and 75% fresh flour/water mixture (44% flour and 56% water at 35 °C). Changes in viable cell concentration and titratable acidity were monitored daily throughout back-slopping and cultivation at 30 ± 1 °C.

C.

Preparation of multi-strain sourdough. Multi-strain sourdough was prepared as follows. A flour suspension containing 44% flour and 56% tap water was prepared at 35 ± 1 °C and inoculated with 2% fresh mixture of 24 h cultures combined in the predetermined ratio. The mixture was homogenized with a spatula and incubated for 24 h at 30 ± 1 °C. The resulting multi-strain sourdoughs were then subjected to daily back-slopping up to 96 h using 25% sourdough and 75% fresh flour/water suspension (44% flour and 56% water at 35 °C). Changes in viable cell concentration and titratable acidity were monitored daily throughout back-slopping and cultivation at 30 ± 1 °C.

D.

Determination of bread dough rise. The method was developed and is routinely employed at DAFA Ltd., where the laboratory baking tests were performed. It is based on the method for determining the lifting power of compressed baker’s yeast [25] and was used to assess the intensity of the dough fermentation through the associated increase in dough volume.

A 400 g portion was taken from the prepared bread dough, shaped into a baguette, and placed in a lightly greased rectangular mold (8 × 8 × 18 cm), where it was pressed to form a layer of uniform thickness. A transverse plate was positioned across the center of the mold, with its lower edge 40 mm above the dough surface. The mold was placed in a thermostat at 35 °C, and the time required for the dough to reach the transverse plate was recorded in minutes.

E.

Testing of sourdough starters for bread production under semi-industrial conditions at DAFA Ltd., Plovdiv, Bulgaria.

F.

Observation of bacterial spoilage in baked bread. Bread samples were stored at 37 °C and at room temperature and were evaluated by 10 production specialists using a four-level descriptive scale: I, barely perceptible; II, weak; III, moderate; and IV, strong spoilage. The assessment integrated visual, olfactory, and tactile indicators and was used to determine the time to first perceptible spoilage under the investigated storage conditions, providing a practical measure of storage stability relevant to production-oriented evaluation.

G.

Observation of fungal spoilage in baked bread. Bread samples were individually packed in polyethylene bags and stored at 30 °C and at room temperature under non-sterile conditions approximating ordinary storage conditions. Samples were visually inspected at 24 h intervals by 10 production specialists throughout the defined observation period. Fungal spoilage was considered detectable when visible fungal colonies or mycelial growth first appeared on the bread surface. Results were expressed as the time to first visually detectable fungal spoilage.

H.

Sensory evaluation. Dough pieces before and after final fermentation were evaluated visually and by touch for surface moisture and consistency. Baked breads were assessed by an internal panel of seven specialists experienced in bakery production and technological evaluation. Before the evaluation, the assessors were familiarized with the evaluated attributes and the use of the scoring scale to ensure a consistent interpretation of the descriptors. Samples were coded with random three-digit numbers and presented in randomized order to minimize presentation bias. The assessors were not informed of the starter formulation or sourdough inclusion level during evaluation. Each formulation was evaluated in separate sensory sessions under consistent laboratory conditions after the breads had cooled to room temperature. Water was provided for palate cleansing between samples. A 10-point scale ranging from 0 (lowest score) to 9 (highest score) was used. The evaluated attributes were loaf volume, overall aroma intensity, taste, crumb softness, elasticity, moisture, crumb color, crust color and taste, and aftertaste. Sensory scores were summarized as mean ± standard deviation across the seven assessors and were used primarily for comparative descriptive evaluation of the formulations [26].

2.8. Statistical Analysis

Statistical analysis was performed using Excel. Quantitative results were expressed as the mean ± standard deviation of the original independent replicates. For dough rise time, dough volume and bread acidity, the effects of functional ingredient level and starter formulation were evaluated separately for each flour matrix using a general linear model, with ingredient level as a continuous factor, starter formulation as a categorical factor, and their interaction. When replicate-level measurements from the same experimental series were available, a mixed-effects model with experimental batch as a random effect was preferred. Post hoc pairwise comparisons were adjusted using Tukey’s method. Statistical significance was accepted at p < 0.05.

2.9. Generative Artificial Intelligence (GenAI) Usage

Generative artificial intelligence (GenAI; OpenAI, San Francisco, CA, USA; GPT Image 2, model version gpt-image-2-2026-04-21) was used to assist in the preparation of Figures 1, 8 and 16–18. Figures 1 and 18 were created as schematic/conceptual illustrations, while Figures 8, 16 and 17 were prepared as summary visualizations of the authors’ experimental results. All scientific content, data, labels, and interpretations were provided and verified by the authors. GenAI was not used to generate or analyze experimental data. All remaining figures were prepared by the authors using Microsoft Excel 2016 (Microsoft Office Professional Plus 2016, Microsoft Corporation, Redmond, WA, USA).

3. Results

3.1. Development of Starter Cultures for Sourdough Bread Production Using Different Flour Types

3.1.1. Preparation of Single-Strain Sourdoughs to Evaluate Growth of Newly Isolated Lactic Acid Bacteria in Flour/Water Suspensions

Rapid accumulation of viable LAB cells is important for establishing a targeted and reproducible fermentation process. The growth and acidification capacity of L. brevis X4 and Lp. plantarum Ph2 were therefore evaluated. For each LAB strain, four single-strain sourdoughs were prepared using wheat, rye, spelt, or einkorn flour. Changes in viable cell concentration and titratable acidity during daily back-slopping for 96 h at 30 ± 1 °C are shown in Figure 2, Figure 3, Figure 4 and Figure 5. Both L. brevis X4 and Lp. plantarum Ph2 grew well in all flour/water suspensions, reaching viable cell concentrations of 10¹²–10¹³ CFU/mL by 96 h, while titratable acidity increased to 22–24 °N. By 24 h, each sourdough had developed a distinct aroma. The aroma profiles changed in type and intensity by 48 h and converged to a characteristic lactic acid aroma by 72 h, which persisted through 96 h.

Based on these results, a two-strain combination for sourdough bread production was developed by mixing Lp. plantarum Ph2 and L. brevis X4 at a 2:1 ratio (hereafter referred to as Ph2 + X4). The ability of this combination to grow in four flour suspensions—wheat flour type 500, rye flour, spelt flour, and einkorn flour—was then evaluated. The two LAB strains grew well together, reaching viable cell concentrations above 10¹²–10¹³ CFU/g, while the titratable acidity reached 21.0 °N in the rye flour sourdough (Figure 6 and Figure 7).

3.1.2. Development of Two-Strain Starter Cultures for Sourdoughs Prepared from Different Flour Types

All sourdoughs prepared with the two-strain combination developed a characteristic lactic acid aroma without detectable unpleasant notes according to the descriptive observations. The aroma was perceived as more intense in the spelt- and einkorn-based sourdoughs and weakest in the wheat flour type 500 sourdough. This matrix-dependent variation is consistent with differences in flour composition, fermentable carbohydrate availability, endogenous enzyme activity, mineral content, and buffering capacity, all of which can influence fermentation intensity and aroma development. The observed differences therefore highlight the important role of the flour matrix in shaping the sensory expression of the same starter combination [3,27,28]. The obtained results provide grounds for defining the two-strain combination for sourdough bread as a two-strain sourdough starter (2SSS). The antimicrobial activity of the 96-h sourdoughs with the two-strain starter against saprophytic microorganisms was determined (

Table 2. Antimicrobial activity of 96 h two-strain sourdoughs. Inhibition zone diameter, d (mm); well diameter = 6 mm.

Starter/Flour	B. subtilis ATCC 6633	B. mesentericus	S. cerevisiae ATCC 9763	A. niger ATCC 16604	Rh. oryzae ATCC 11145	P. chrysogenum ATCC 10106
Ph2+X4/Wheat	12.33 ± 0.47	9.17 ± 0.24	-	9.17 ± 0.24	9.17 ± 0.24	10.17 ± 0.24
Ph2+X4/Spelt	16.17 ± 0.24	12.33 ± 0.47	-	8.33 ± 0.47	9.33 ± 0.47	10.33 ± 0.47
Ph2+X4/Rye	14.17 ± 0.24	10.17 ± 0.24	-	9.33 ± 0.47	8.33 ± 0.47	9.17 ± 0.24
Ph2+X4/Einkorn	16.33 ± 0.47	11.17 ± 0.24	-	9.33 ± 0.47	8.33 ± 0.47	11.33 ± 0.47

). The two-strain starter did not inhibit the growth and development of the baker’s yeast Saccharomyces cerevisiae, but exhibited antimicrobial activity against the saprophytes included in the study, with a stronger inhibitory effect being found in sourdoughs with spelt flour and einkorn flour. The reported inhibition was probably due to the metabolic products synthesized by the LAB in the composition of the starters [10,11,29,30].

Testing of the two-strain starter under semi-industrial conditions. Bread variants were prepared using 96 h two-strain starter sourdough (2SSS) at inclusion levels of 7% and 15% relative to flour mass. All sourdoughs were successfully incorporated into the bread-making process, and no undesirable technological effects were observed. According to the internal technological assessment, 2SSS contributed to dough stability, elasticity, and strength during mixing and fermentation, while reducing surface adhesiveness. No fungal spores or wild yeasts were detected in the sourdoughs, which may reflect inhibition by metabolites produced by the LAB in the two-strain starter. Dough and bread characteristics were monitored to assess whether the tested sourdough inclusion levels adversely affected the established production process or product quality. The main qualitative technological trends across the investigated flour matrices are summarized in

Table 3. A summary of the common descriptive dough and bread characteristics and final proofing time observed across the investigated flour types for breads prepared with 7% and 15% of the 96 h two-strain sourdough (Lp. plantarum Ph2 + L. brevis X4).

-	Bread Control	Bread 7% Sourdough	Bread 15% Sourdough
Dough characteristics	The dough without 2SSS was wetter	The dough is more elastic, dry with good organoleptic properties	Stronger, more elastic dough with good organoleptic properties
Time for final fermentation, min	60	60	60
Pieces before baking	Relaxed slightly to the side	Good dimensional stability	More stable, with good dimensional stability
Bread loaf appearance	The bread loaf was smaller in volume	Bread with good volume and shape compared to the control	Bread with good volume and shape compared to the control
Softness and structure of the crumb	The control has larger, uneven pores	The crumb is softer, wetter and lighter, with more even pores	The crumb was softer, wetter and lighter, with more even pores. More pronounced indicators
Bread aroma	Yeast aroma	Pleasant, characteristic lactic acid aroma	Strong, pleasant, characteristic lactic acid aroma

The table summarizes the general qualitative trend observed across wheat, wheat–rye, spelt, and einkorn bread formulations. Flour-specific quantitative results are presented separately in Table 4. 

, whereas flour-specific quantitative parameters are presented in

Table 4. Sourdough acidity, dough rise, loaf volume, and bread acidity for bread variants prepared with 96 h two-strain starter sourdough. Sourdough inclusion levels were 7% and 15% relative to flour mass. 2SSS, two-strain starter sourdough.

	Acidity of the Dough, °N	Dough Rise, min	Loaf Volume, cm3	Bread Acidity, °N
Wheat bread—control	-	54	1720	1.2
Wheat bread—7% wheat sourdough with 2SSS Ph2+X4	11.6	48	1780	1.6
Wheat bread—15% wheat sourdough with 2SSS Ph2+X4	11.6	47	1740	2.1
Wheat-rye bread 70/30—control	-	61	1140	1.3
Wheat–rye bread 70/30—7% rye sourdough with 2SSS Ph2+X4	14.2	55	1360	1.8
Wheat–rye bread 70/30—15% rye sourdough with 2SSS Ph2+X4	14.2	56	1260	2.5
Spelt bread—control	-	52	1360	1.4
Spelt bread—7% spelt sourdough with 2SSS Ph2+X4	20.6	50	1480	2.8
Spelt bread—15% spelt sourdough with 2SSS Ph2+X4	20.6	50	1460	4.6
Einkorn bread—control	-	60	1210	1.6
Einkorn bread—7% einkorn sourdough with 2SSS Ph2+X4	21.0	53	1360	3.6
Einkorn bread—15% einkorn sourdough with 2SSS Ph2+X4	21.0	52	1320	4.2

.

Compared with the control, doughs containing the two-strain sourdough were described by the internal technological panel as less adhesive and more elastic. The corresponding breads showed greater loaf volume, improved dimensional stability, a more uniform crumb, and a characteristic lactic acid aroma.

Bread samples were stored at room temperature (20–22 °C) and at 37 °C for 96 h to monitor bacterial spoilage, and at 30 °C and room temperature for 120 h to monitor fungal spoilage. The endpoint was the first visually detectable sign of spoilage, assessed according to the descriptive procedures in Section 2.7.3.

The earliest visually detectable signs of bacterial spoilage occurred in the control breads after 48 h at 37 °C and after 72 h at room temperature. In contrast, no visible signs were observed in breads containing 7% or 15% sourdough throughout the 96 h observation period. These results demonstrate a marked delay in visually detectable bacterial spoilage under the investigated storage conditions. As the assessment was based on descriptive spoilage observations, the findings are interpreted specifically in terms of delayed visible spoilage rather than as a formal microbiological shelf-life validation.

The earliest signs of fungal spoilage were detected in the control bread at 72 h, both at 30 °C and at room temperature. With 7% two-strain starter sourdough, fungal spoilage was observed at 96 h at 30 °C, with the lowest degree of spoilage recorded in the spelt- and einkorn-based breads. When the sourdough inclusion level was increased to 15%, fungal spoilage in wheat bread was observed at 96 h at 30 °C and at 120 h at room temperature. No fungal spoilage was detected in the wheat–rye, spelt, or einkorn breads within the 120 h observation period (Table S2; Figure 8). Among the tested variants, an inclusion level of 7% sourdough was selected for delaying visible bacterial spoilage and inclusion level of 15% for delaying visible fungal spoilage without an apparent adverse effect on bread volume or flavor profile.

The two-strain starter sourdoughs were incorporated at 7% and 15% under non-sterile production conditions. The baked breads were likewise stored under non-sterile conditions at the specified temperatures to approximate ordinary household storage, in contrast to the aseptic conditions used by Menteş et al. (2007) [31].

The results are consistent with previous reports [31,32,33,34]. Earlier studies showed that the inclusion of 15% or more sourdough in the final dough can inhibit bacterial and fungal spore development and contribute to longer bread shelf-life [31,33]. Acidification is also important for bread rise, enzyme activity, crumb elasticity, and storage stability, although excessive acidification may adversely affect some rheological properties [32,34].

3.2. Development of Multi-Strain Starter Cultures for Wheat, Wheat–Rye, Spelt, and Einkorn Sourdough Bread

3.2.1. The Influence of the Starter Microbial Composition on the Fermentation Kinetics

Building on the selected two-strain starter composed of Lp. plantarum Ph2, isolated from spontaneously fermented wheat sourdough, and L. brevis X4, isolated from spontaneously fermented Khorasan flour sourdough, three additional multi-strain combinations were developed by incorporating L. rhamnosus LBRC11, L. fermentum LBRH10, F. sanfranciscensis R, and Pr. freudenreichii subsp. shermanii NBIMCC 327 (582D):

✓

Multi-strain combination 1:  Lp. plantarum Ph2, L. brevis X4, L. rhamnosus LBRC11, and L. fermentum LBRH10 in a ratio of 2:1:1:1, respectively.

✓

Multi-strain combination 2:  Multi-strain combination 1 and F. sanfranciscensis R in a ratio of 2:1, respectively.

✓

Multi-strain combination 3:  Multi-strain combination 1, F. sanfranciscensis R, and Pr. freudenreichii subsp. shermanii NBIMCC 327 in a ratio of 2:1:1, respectively.

✓

The ability of the strains in the three multi-strain combinations to grow in flour suspensions was evaluated (Figure 9, Figure 10, Figure 11, Figure 12, Figure 13, Figure 14 and Figure 15).

The strains included in the three starter combinations grew well in the flour–water suspensions, reaching viable cell concentrations of 10¹³–10¹⁴ CFU/mL by 96 h (Figure 9, Figure 11, Figure 13 and Figure 14), while the titratable acidity of the resulting sourdoughs increased to values of up to approximately 25 °N (Figure 10, Figure 11 and Figure 15). An adaptation period of approximately 24–48 h was observed before the mixed cultures entered a phase of more intensive growth and acidification, indicating progressive stabilization of the microbial associations within the flour matrices. Differences in titratable acidity at comparable viable cell concentrations may reflect variation in strain-specific metabolic activity, substrate utilization, and the buffering capacity of the flour matrices. Because viable cell concentration is not a direct measure of acid production rate, similar population levels may therefore be associated with different titratable acidity values.

All sourdoughs developed a strong, characteristic lactic acid aroma without unpleasant off-notes. The strains in the three combinations grew well in the flour/water suspensions. Combination 2 produced a slightly milder, sweetish, yet rich and pleasant lactic-acid aroma, resembling the profile typically associated with whole grain breads or breads prepared from higher-ash flours. Overall, the growth, acidification, and sensory observations supported the suitability of the three combinations as sourdough starter cultures.

3.2.2. Relationships Between Acidification and In Vitro Antimicrobial Activity

The antimicrobial activity of the twelve 96 h multi-strain starter sourdoughs against the selected indicator microorganisms is presented in

Table 5. Antimicrobial activity of 96 h sourdoughs prepared with three multi-strain starters. Inhibition zone diameter, d (mm); well diameter = 6 mm; incubation time = 48 h.

Starter/Flour	B. subtilis ATCC 6633	B. mesentericus	S. cerevisiae ATCC 9763	A. niger ATCC 16604	Rh. oryzae ATCC 11145	P. chrysogenum ATCC 10106
Starter 1/wheat	10.17 ± 0.24	10.33 ± 0.47	-	10.33 ± 0.47	11.17 ± 0.24	10.33 ± 0.47
Starter 1/einkorn	11.33 ± 0.47	11.17 ± 0.24	-	10.33 ± 0.47	13.33 ± 0.47	10.33 ± 0.47
Starter 1/spelt	12.17 ± 0.24	12.33 ± 0.47	-	10.17 ± 0.24	14.17 ± 0.24	12.17 ± 0.24
Starter 1/rye	11.33 ± 0.47	14.67 ± 0.47	-	12.17 ± 0.24	9.17 ± 0.24	9.17 ± 0.24
Starter 2/wheat	9.17 ± 0.24	8.17 ± 0.24	-	9.17 ± 0.24	13.17 ± 0.24	9.33 ± 0.47
Starter 2/einkorn	10.17 ± 0.24	12.17 ± 0.24	-	10.33 ± 0.47	16.33 ± 0.47	11.67 ± 0.47
Starter 2/spelt	12.67 ± 0.47	13.33 ± 0.47	-	10.33 ± 0.47	16.33 ± 0.47	10.17 ± 0.24
Starter 2/rye	9.17 ± 0.24	10.33 ± 0.47	-	9.17 ± 0.24	11.33 ± 1.47	8.17 ± 0.24
Starter 3/wheat	9.17 ± 0.24	9.17 ± 0.24	-	9.17 ± 0.24	15.17 ± 0.24	10.33 ± 0.47
Starter 3/einkorn	10.17 ± 0.24	11.33 ± 0.47	-	9.33 ± 0.47	15.33 ± 0.47	10.33 ± 0.47
Starter 3/spelt	10.17 ± 0.24	13.67 ± 0.47	-	10.33 ± 0.47	14.17 ± 0.24	10.17 ± 0.24
Starter 3/rye	10.33 ± 0.47	10.17 ± 0.24	-	10.17 ± 0.24	14.33 ± 0.47	10.17 ± 0.24

. None of the sourdoughs inhibited the baker’s yeast indicator, whereas all showed inhibitory activity against the bacterial and fungal spoilage indicators included in the study. Stronger inhibition was observed for several rye- and spelt-based sourdoughs and for formulations prepared with Starter 3. The observed inhibition is likely associated with metabolites produced by the lactic acid and propionic acid bacteria during co-cultivation.

The multi-strain sourdoughs differed in both titratable acidity and inhibition zone diameter. In several formulations, greater acidification was accompanied by stronger inhibition of the tested indicator microorganisms, suggesting a parallel relationship between these responses. As no formal correlation analysis was undertaken, this pattern is interpreted as an association rather than evidence of a direct causal relationship.

The observed inhibitory activity is likely to reflect the combined contribution of acidification and other antimicrobial metabolites formed during fermentation. The present results therefore support the antimicrobial potential of the developed starter systems, while more detailed characterization of individual organic acids and other inhibitory compounds, together with pH-neutralized assays, could further clarify the relative contribution of the underlying mechanisms.

3.2.3. Technological and Descriptive Dough Characteristics Under Semi-Industrial Conditions

Bread was prepared from each 96 h sourdough produced with the three multi-strain starters and the investigated flour types, using sourdough inclusion levels of 5%, 7%, or 10%. All three starters were successfully applied in bread preparation and baking, and no undesirable technological effects were observed. According to the technological assessment, the sourdoughs contributed to dough stability, strength, and elasticity and facilitated handling during final dough preparation and fermentation.

Dough and bread characteristics were monitored to determine whether the tested sourdough inclusion levels affected the physical and sensory properties of the products or the established production process (

Table 6. Selected sourdough addition levels for the investigated bread types produced with 96 h fermented sourdough.

Bread Type	Starter Culture	Sourdough Acidity, °N	Control Volume, cm3	Selected Sourdough Inclusion Level, %	Selected Bread Volume, cm3	Volume Increase vs. Control, %	Bread Acidity, °N	General Technological and Sensory Effect
Wheat bread	Starter 1	13.6	1700	7	1800	+5.9	2.0	Best volume and shape; very pleasant lactic acid aroma
Wheat bread	Starter 2	13.6	1720	7	1880	+9.3	2.2	Highest volume; well-developed bread with characteristic aroma
Wheat bread	Starter 3	14.0	1680	7–10	1780	+6.0	2.0–2.6	Good volume; stronger aroma at 10%, slightly sharper
Wheat–rye bread	Starter 1	12.6	1160	10	1300	+12.1	3.2	Best volume and shape; strong pleasant lactic acid aroma
Wheat–rye bread	Starter 2	13.2	1020	7	1220	+19.6	2.8	Best balance between volume and sensory quality
Wheat–rye bread	Starter 3	13.0	1180	10	1300	+10.2	3.6	Best volume and shape; strong characteristic aroma
Spelt bread	Starter 1	24.0	1380	7–10	1500	+8.7	3.4–3.8	Highest volume; very pleasant to strong characteristic aroma
Spelt bread	Starter 2	22.2	1280	7	1370	+7.0	3.6	Better developed bread; pleasant characteristic aroma
Spelt bread	Starter 3	23.0	1350	7	1450	+7.4	3.8	Best development and good sensory profile
Einkorn bread	Starter 1	22.4	1320	7	1440	+9.1	4.0	Best volume; very pleasant characteristic aroma
Einkorn bread	Starter 2	23.8	1300	7–10	1420	+9.2	3.4–4.2	Good volume; stronger aroma at 10%
Einkorn bread	Starter 3	22.8	1230	10	1420	+15.4	4.4	Highest volume; strong pleasant lactic acid aroma

Notes: Starter 1: L. rhamnosus LBRC11/Lp. plantarum Ph2/L. brevis X4/L. fermentum LBRH10 = 1:2:1:1. Starter 2: Starter 1/F. sanfranciscensis R = 2:1. Starter 3: Starter 1/F. sanfranciscensis R/Pr. shermanii 582D = 2:1:1. All sourdoughs were fermented for 96 h. Final proofing time was 60 min for all variants. The selected sourdough inclusion level was based primarily on maximum bread volume; when values were similar, sensory quality and acidity were also considered.

). The sourdough-containing doughs generally fermented more rapidly and were described as more elastic and stronger, while several bread variants showed greater loaf volume. The baked breads had a softer, lighter-colored crumb and a pleasant characteristic lactic acid aroma. No fungal spores or wild yeasts were detected in the sourdoughs.

All twelve 96 h sourdoughs were evaluated at different inclusion levels to identify the most suitable starter and formulation range for wheat, wheat–rye, spelt, and einkorn breads. Selection considered the delay in visible bacterial and fungal spoilage together with technological and sensory performance. The production specialists described several sourdough-containing doughs as more stable, with improved handling and dimensional stability compared with the corresponding controls.

These observations are consistent with literature describing the contribution of controlled proteolysis and exopolysaccharide production to gluten-network modification and water retention. In the present study, the technological observations indicate that fermentation influenced dough structure and handling in addition to supporting gas production. This is relevant to the development of clean-label bakery formulations that rely on fermentation rather than additional technological additives.

The 96 h sourdoughs were successfully incorporated into all investigated bread types under the tested semi-industrial conditions. Compared with the corresponding controls, several formulations showed higher loaf volumes, improved shape retention, and more favorable descriptive aroma characteristics. Based on the combined technological and sensory observations, specific sourdough inclusion levels were selected as the most suitable among the formulations evaluated in the present study. These levels therefore represent the preferred formulation ranges identified under the tested conditions and provide a practical basis for further application and refinement.

3.2.4. Sensory Observations and Time to Visually Detectable Spoilage

The baked breads were stored at room temperature (20–25 °C) and at 37 °C for 96 h to monitor bacterial spoilage, and at room temperature and 30 °C for 168 h to monitor fungal spoilage. The earliest signs of bacterial spoilage were detected in the control wheat and spelt breads at 96 h at 37 °C. No bacterial spoilage was observed within 96 h in the breads containing 5%, 7%, or 10% multi-strain starter sourdough, either at 37 °C or at room temperature (Figure 16 and Figure 17). The earliest fungal spoilage in wheat bread was detected at 168 h at 30 °C in the control and in breads containing 5% or 7% sourdough prepared with Starter 1 or Starter 2. With 10% sourdough from Starter 1 or Starter 2, fungal spoilage at 168 h was observed only in some variants, whereas no visible fungal spoilage was detected in wheat breads containing Starter 3 within the 168 h observation period (Figure 16 and Figure 17).

In einkorn bread, fungal spoilage appeared earliest in the control at 96 h at 30 °C. At 168 h, spoilage was also observed in several variants with lower sourdough inclusion levels, whereas breads containing 7% or 10% sourdough remained free of visible fungal spoilage. In spelt bread, fungal spoilage was detected at 168 h in the control and in several breads containing 5%, 7%, or 10% sourdough from Starters 1 or 2; breads containing 7% or 10% Starter 3 showed no visible spoilage at either storage temperature. In wheat–rye bread, the control showed fungal spoilage at 96 h at 30 °C, whereas breads containing 10% sourdough from any of the three starters showed no visible fungal spoilage within 168 h at either temperature (Figure 16 and Figure 17).

Among the evaluated variants, 7% sourdough was selected as the most suitable level for delaying visible bacterial spoilage, whereas 10% was selected for delaying visible fungal spoilage, without an apparent adverse effect on loaf volume or the descriptive flavor profile. These levels represent the preferred formulation choices identified under the tested conditions and provide practical guidance for application of the selected sourdough systems.

The sensory characteristics of the bread products varied with the fermentation conditions and the sourdough formulation. Breads produced with the symbiotic starter cultures generally showed a more balanced flavor profile and a more pronounced aroma, consistent with the contribution of organic acids, volatile compounds, and other fermentation metabolites.

The sensory differences were observed under the specified formulation and fermentation conditions and reflect the combined influence of starter activity, flour composition, and sourdough inclusion level. The resulting aroma differences therefore provide a useful product-level indication of formulation-dependent sensory performance. Further volatile-compound profiling could complement these observations and help clarify the metabolic basis of the aroma patterns.

The sourdough starters were incorporated at different levels under non-sterile production conditions, and the baked breads were stored under non-sterile conditions at the specified temperatures to approximate ordinary household storage. This differs from the aseptic experimental conditions described by Menteş et al. (2007) [31].

The results are consistent with previous studies. Inclusion levels of 10–15% or more sourdough in the final dough has been reported to inhibit bacterial and fungal spore development and to contribute to longer bread shelf-life [31,33]. Acidification is also important for bread rise, enzyme activity, crumb elasticity, and storage stability, although excessive pH reduction may adversely affect some rheological properties [32,34,35].

Breads produced with selected starters showed a later onset of visually detectable spoilage than the corresponding controls under the investigated storage conditions, demonstrating the practical potential of the selected cultures to enhance bread storage stability under semi-industrial conditions. This effect is likely associated with the combined influence of acidification and antimicrobial metabolites generated during sourdough fermentation, while the relative contribution of these mechanisms remains an interesting subject for further investigation.

3.3. Application of the Three Multi-Strain Sourdough Starters in Wheat and Spelt Breads Enriched with Almond Flour and Malt Bran Flour

The three starter combinations were propagated in flour/water suspensions prepared with wheat flour type 500 and spelt flour (Figure 9, Figure 10, Figure 11, Figure 12, Figure 13, Figure 14 and Figure 15), and the antimicrobial activity of the resulting 96 h sourdoughs was evaluated. None of the multi-strain starters inhibited the baker’s yeast indicator, whereas all showed antimicrobial activity against the spoilage indicators included in the study (Table 5). The three starters generally showed higher antimicrobial activity when propagated in spelt flour sourdough. This inhibition is likely associated with metabolites produced during co-cultivation by the lactic acid bacteria and, where present, propionic acid bacteria in the multi-strain starters.

Testing of Sourdough Starters Under Semi-Industrial Bread-Making Conditions

Bread was prepared with sourdoughs produced using each of the three multi-strain starters. The selected inclusion level was 15% for Starter 1 and 10% for Starters 2 and 3. All three multi-strain starters were successfully applied in bread-making when propagated in either wheat or spelt flour sourdough.

For each multi-strain starter, breads were also prepared with 5%, 10%, 15%, 20%, or 25% malt bran flour (MBF) or almond flour (AF), relative to the main flour content of the formulation. No undesirable technological effects that would prevent production were observed. The multi-strain starter sourdoughs contributed to dough stability, strength, and elasticity during processing and fermentation, thereby facilitating bread production.

Physical and sensory characteristics of the dough and bread were monitored to evaluate the technological effects of almond flour and malt bran flour incorporation (Tables S1–S8). Increasing the proportion of almond flour slightly prolonged fermentation. Control formulations without almond flour generally showed greater dough stability and elasticity, larger loaf volume, and good dimensional stability. The finished breads had a softer and more uniform crumb. Additional test breads were prepared from spelt flour using each multi-strain starter and 5%, 10%, 15%, 20%, or 25% MBF or AF, as detailed in Table S1–S8.

Among the evaluated formulations, 10% and 15% additions of malt bran flour or almond flour provided the most favorable balance between technological performance and sensory characteristics. Higher inclusion levels, up to 20%, remained technologically feasible in some formulations, although their suitability depended on the desired sensory profile. Accordingly, 10–15% was identified as the preferred inclusion range under the conditions evaluated in the present study.

The baked breads were stored at room temperature and at 37 °C to monitor bacterial spoilage (Tables S9 and S10), and at room temperature and at 30 °C to monitor fungal spoilage (Tables S11 and S12).

In the control wheat bread, bacterial spoilage was detected at 72 h at 37 °C and at 96 h at room temperature. In wheat bread formulations containing multi-strain starter sourdough and different levels of almond flour or malt bran flour, no bacterial spoilage was detected within the 96 h observation period. Thus, the starter-containing formulations showed a delayed onset of visible bacterial spoilage, while incorporation of almond flour or malt bran flour within the tested ranges did not adversely affect storage stability (Tables S9 and S10).

Fungal spoilage in the control wheat bread was detected at 72 h at 30 °C and at 96 h at room temperature. In wheat bread formulations containing the three multi-strain starter sourdoughs and different levels of almond flour or malt bran flour, fungal spoilage was first observed at 120 h. In formulations prepared with Starter 3, no visible fungal spoilage was detected within 120 h at either 30 °C or room temperature. These results demonstrate delayed visible fungal spoilage in the starter-containing formulations, while almond flour and malt bran flour incorporation within the tested ranges did not adversely affect storage stability (Tables S11 and S12).

The incorporation of bacterial starter sourdough into the final dough enabled consistent fermentation performance across the investigated formulations and contributed to a later onset of visually detectable spoilage in several bread variants. The observed effects were reproducible under the applied semi-industrial processing and storage conditions, demonstrating the practical potential of the selected starter combinations for improving dough performance and bread storage stability.

Overall, the developed starter combinations were successfully applied across different flour matrices and semi-industrial bread formulations. Their performance was characterized through microbiological, technological, sensory, and storage-related parameters, allowing for the identification of starter–flour combinations with particularly favorable technological behavior. The results therefore provide a practical basis for selecting suitable starter cultures and sourdough inclusion levels for different bread formulations.

The present study establishes a solid experimental and technological foundation for further refinement of the developed systems. Future investigations may extend this work through detailed fermentation kinetics, profiling of organic acids and volatile compounds, instrumental rheological characterization, and quantitative monitoring of microbial populations during storage, thereby providing additional insight into the mechanisms underlying the observed technological and preservation effects.

3.4. Statistical Analysis

Across all flour–ingredient matrices, increasing the level of almond flour or malt bran flour was associated with a marked reduction in dough volume. In the exploratory models fitted to the reported treatment means, the concentration effect was significant in all four matrices (p < 0.001), whereas the concentration × starter interaction was not significant (p > 0.60), indicating comparable volume reduction slopes among starter formulations. At 25% inclusion, dough volume was 30–36% lower than the corresponding starter control (

Table 7. The effects of additive percentage, starter formulation and their interaction. Values are based on models fitted to the reported table means.

Flour	Ingredient	Outcome	Effect	F	p	R2	Interpretation
Spelt	Almond flour	Dough rise time	Addition level	23.74	0.0004	0.837	Significant
Spelt	Almond flour	Dough rise time	Starter	16.66	0.0003	0.837	Significant
Spelt	Almond flour	Dough rise time	Level × starter	2.39	0.1336	0.837	Not significant
Spelt	Almond flour	Dough volume	Addition level	424.39	0.0000	0.973	Significant
Spelt	Almond flour	Dough volume	Starter	3.40	0.0678	0.973	Not significant
Spelt	Almond flour	Dough volume	Level × starter	0.28	0.7584	0.973	Not significant
Spelt	Almond flour	Bread acidity	Addition level	28.70	0.0002	0.818	Significant
Spelt	Almond flour	Bread acidity	Starter	3.27	0.0735	0.818	Not significant
Spelt	Almond flour	Bread acidity	Level × starter	9.26	0.0037	0.818	Significant
Spelt	Malt bran flour	Dough rise time	Addition level	39.08	0.0000	0.865	Significant
Spelt	Malt bran flour	Dough rise time	Starter	12.93	0.0010	0.865	Significant
Spelt	Malt bran flour	Dough rise time	Level × starter	5.86	0.0168	0.865	Significant
Spelt	Malt bran flour	Dough volume	Addition level	515.91	0.0000	0.977	Significant
Spelt	Malt bran flour	Dough volume	Starter	1.81	0.2061	0.977	Not significant
Spelt	Malt bran flour	Dough volume	Level × starter	0.51	0.6114	0.977	Not significant
Spelt	Malt bran flour	Bread acidity	Addition level	119.94	0.0000	0.915	Significant
Spelt	Malt bran flour	Bread acidity	Starter	0.73	0.5001	0.915	Not significant
Spelt	Malt bran flour	Bread acidity	Level × starter	3.74	0.0548	0.915	Not significant
Wheat	Almond flour	Dough rise time	Addition level	120.79	0.0000	0.936	Significant
Wheat	Almond flour	Dough rise time	Starter	23.26	0.0001	0.936	Significant
Wheat	Almond flour	Dough rise time	Level × starter	4.65	0.0319	0.936	Significant
Wheat	Almond flour	Dough volume	Addition level	243.13	0.0000	0.954	Significant
Wheat	Almond flour	Dough volume	Starter	3.75	0.0542	0.954	Not significant
Wheat	Almond flour	Dough volume	Level × starter	0.19	0.8266	0.954	Not significant
Wheat	Almond flour	Bread acidity	Addition level	0.26	0.6199	0.473	Not significant
Wheat	Almond flour	Bread acidity	Starter	5.00	0.0263	0.473	Significant
Wheat	Almond flour	Bread acidity	Level × starter	0.26	0.7758	0.473	Not significant
Wheat	Malt bran flour	Dough rise time	Addition level	33.88	0.0001	0.824	Significant
Wheat	Malt bran flour	Dough rise time	Starter	6.87	0.0103	0.824	Significant
Wheat	Malt bran flour	Dough rise time	Level × starter	4.36	0.0377	0.824	Significant
Wheat	Malt bran flour	Dough volume	Addition level	204.70	0.0000	0.945	Significant
Wheat	Malt bran flour	Dough volume	Starter	1.48	0.2658	0.945	Not significant
Wheat	Malt bran flour	Dough volume	Level × starter	0.06	0.9440	0.945	Not significant
Wheat	Malt bran flour	Bread acidity	Addition level	28.54	0.0002	0.725	Significant
Wheat	Malt bran flour	Bread acidity	Starter	0.22	0.8032	0.725	Not significant
Wheat	Malt bran flour	Bread acidity	Level × starter	1.36	0.2926	0.725	Not significant

).

The time required for dough rise generally increased with the inclusion level. The concentration effect was significant in wheat bread with almond flour, wheat bread with malt bran flour, spelt bread with almond flour, and spelt bread with malt bran flour (all p < 0.001). Starter formulation also influenced rise time, and significant concentration × starter interactions were detected in three of the four matrices, suggesting that the magnitude of the delay was formulation-dependent.

Bread acidity responded differently according to flour and ingredient type. No concentration-dependent effect was detected for almond flour in wheat bread (p = 0.620), whereas malt bran flour increased bread acidity in wheat bread (p < 0.001). In spelt bread, both almond flour and malt bran flour were associated with increasing bread acidity (p < 0.001). A significant concentration × starter interaction was observed for spelt bread enriched with almond flour (p = 0.0037).

Dough volume. The additive percentage had a highly significant negative association with dough volume in all four matrices (all exploratory ANCOVA p-values < 0.001). The starter × concentration interaction was not significant in any matrix (p = 0.612–0.944), indicating broadly parallel volume reduction trends among the three starter formulations.

Dough rise time. Increasing almond flour or malt bran flour generally prolonged dough rise time. The concentration effect was significant in all four matrices (p < 0.001). Starter formulation also affected rise time (p = 0.0003–0.0103). Significant starter × concentration interactions were observed for wheat–almond flour, wheat–malt bran flour and spelt–malt bran flour, whereas the interaction was not significant for spelt–almond flour.

Bread acidity. Almond flour did not produce a concentration-dependent change in bread acidity in the wheat matrix (p = 0.620). In contrast, malt bran flour increased bread acidity in wheat bread (p < 0.001), while both almond flour and malt bran flour increased bread acidity in spelt bread (p < 0.001). For spelt bread with almond flour, the response differed among starter formulations, as indicated by a significant interaction (p = 0.0037).

The magnitude of the technological effect. At 25% addition, dough volume decreased by approximately 30–36% relative to the corresponding starter control. The combined quantitative and qualitative evidence supports selecting lower inclusion levels, especially 5–10%, for further validation.

4. Discussion

This study provides an extensive comparative dataset on the behavior of defined bacterial starter combinations in different flour matrices and their application under semi-industrial bread-making conditions. Its main contribution is the selection and practical evaluation of starter formulations through an integrated assessment of viable cell concentration, titratable acidity, in vitro inhibitory activity, bread quality indicators, descriptive sensory characteristics, and storage stability. Figure 18 synthesizes these findings into an integrated conceptual framework, linking the experimentally observed responses with mechanistic relationships supported by the literature [36].

4.1. Starter Performance in Different Flour Matrices

The selected starter combinations maintained high viable cell concentrations throughout propagation and showed a progressive increase in titratable acidity, confirming their robust performance in the investigated flour matrices. Differences in acidity were observed among wheat-, spelt-, rye-, and einkorn-based sourdoughs even when viable cell concentrations were comparable, reflecting the influence of both starter metabolism and flour matrix characteristics on acidification. Such variation is consistent with differences in buffering capacity, nutrient availability, and enzymatic activity among cereal substrates.

Previous studies have similarly demonstrated that flour type, feeding regime, fermentation time, and temperature influence sourdough microbial development and acidification [27,28]. The present study extends these observations by directly comparing several defined starter combinations under a common propagation procedure and across multiple flour matrices. This comparative approach provides a practical basis for evaluating starter adaptability and selecting combinations with stable performance under standardized fermentation conditions.

From a technological perspective, the ability of the selected starters to perform consistently in more than one flour matrix represents an important advantage for their application in bread production. The comparative results further demonstrate that starter composition, flour type, fermentation time, and enrichment level can jointly influence the technological characteristics of the resulting dough and bread, providing useful criteria for formulation selection and process adaptation.

Among the enrichment levels evaluated, increasing replacement with malt bran flour or almond flour produced the clearest effect on dough volume, particularly at the highest inclusion levels. The absence of a significant starter × concentration interaction for dough volume indicates that this response was primarily associated with the increasing proportion of the enrichment ingredient rather than with a specific starter formulation. The accompanying qualitative observations of reduced elasticity, lower dimensional stability, and a more compact crumb at 20–25% inclusion were consistent with the quantitative trend. Taken together, the results identified 5–10% inclusion as the most technologically suitable range among the tested formulations, providing the best overall balance between enrichment and dough processing performance.

4.2. In Vitro Antimicrobial Activity and Storage Observations

Several sourdough preparations exhibited clear in vitro inhibitory activity against the selected bacterial and fungal indicator microorganisms, while the baker’s yeast indicator was not affected. The observed inhibition demonstrates the antimicrobial potential of the developed starter systems and is consistent with the combined effects of sourdough acidification and fermentation-derived metabolites. Although the relative contribution of individual antimicrobial compounds was not resolved in the present study, the results clearly differentiate the inhibitory performance of the evaluated starter combinations.

This in vitro activity was complemented by the storage observations, in which several sourdough breads showed a later onset of visible spoilage than the corresponding controls. These findings are consistent with previous reports describing the biopreservative potential of sourdough fermentation [20,29,37] and demonstrate that the selected starter cultures can contribute to improved storage stability under the investigated conditions. The present results therefore provide both microbiological and practical evidence supporting the use of the developed cultures as functional starter systems for bread production. Further characterization of organic acids and other antimicrobial metabolites, together with quantitative monitoring of microbial populations during storage, could provide additional insight into the mechanisms underlying these effects.

4.3. Technological and Sensory Performance

The semi-industrial trials demonstrated that the selected sourdoughs could be successfully incorporated into breads produced from different flour matrices. Several formulations showed favorable changes in loaf volume together with improved technological and descriptive sensory characteristics, including dough handling, crumb structure, aroma, and taste. The observed differences in dough strength, elasticity, gas retention, and overall handling behavior provide useful technological evidence of formulation-dependent performance under semi-industrial processing conditions.

The aroma differences identified by the internal panel are consistent with the established contribution of sourdough fermentation to the formation of organic acids and volatile metabolites [1,7]. In the present study, these effects were assessed through the resulting sensory characteristics of the breads, allowing for clear differentiation among formulations at the product level. Further targeted profiling of volatile compounds and individual organic acids could complement these findings and provide additional mechanistic detail on the metabolites contributing to the observed aroma profiles.

4.4. Practical Implications and Future Perspectives

The tested starter combinations provide a practical basis for selecting bacterial cultures and sourdough inclusion levels adapted to different flour matrices and bread formulations. Their comparative evaluation under semi-industrial conditions enabled the identification of starter–matrix combinations and inclusion ranges showing favorable microbiological, technological, sensory, and storage-related performance. The ability of the selected starter systems to maintain high viable cell concentrations, promote acidification, perform across different cereal matrices, and contribute to desirable bread characteristics supports their potential application in standardized bread production.

The selected sourdough and functional ingredient inclusion levels represent recommended formulation ranges among the variants evaluated under the processing conditions applied in the present study. In this context, the results provide practical guidance for formulation development while allowing for further adaptation according to flour type, desired bread characteristics, and technological requirements. Importantly, the semi-industrial validation strengthens the applicability of the findings beyond laboratory-scale starter screening and provides a basis for further technological refinement and potential scale-up.

4.5. Perspectives for Further Research

Future research can build on the present comparative and application-oriented findings by extending the characterization of the most promising starter–flour combinations. Factorial experimental designs and fermentation kinetic modeling could be applied to further refine formulation and processing parameters and to examine interactions among starter composition, flour matrix, fermentation conditions, and inclusion level.

Additional characterization of organic acids, residual sugars, volatile compounds, CO₂ evolution, microbial population dynamics, and dough rheological behavior would complement the microbiological, technological, sensory, and storage data obtained in the present study. Integrating these approaches would provide deeper insight into the mechanisms underlying fermentation performance, bread quality development, and the observed biopreservative effects, while supporting further refinement and broader application of the developed starter systems.

4.6. Interpretation of Figure 18

Figure 18 provides an integrated conceptual representation of the relationships between starter culture activity, sourdough fermentation, technological performance, antimicrobial effects, and bread quality outcomes. The framework combines variables directly evaluated in the present study with biochemical and physiological mechanisms that are well established in the sourdough literature. Experimentally evaluated outcomes include viable cell concentration, titratable acidity, in vitro inhibitory activity, technological bread characteristics, descriptive sensory properties, and the onset of visually detectable spoilage. Mechanistic elements such as the formation of specific organic acids, exopolysaccharides, peptides and other fermentation metabolites, as well as phytic-acid degradation and potential health-related effects, are included as literature-supported pathways that provide biological context for the observed responses [38,39,40]. These elements are distinguished graphically from the experimentally measured variables to clarify their conceptual role within the proposed framework.

5. Conclusions

Defined two-strain and multi-strain bacterial starter cultures were successfully developed and evaluated in wheat, rye, spelt, and einkorn flour matrices, as well as in selected enriched bread formulations. Under the applied propagation conditions, the cultures maintained high viable cell concentrations and promoted progressive acidification. Several sourdough preparations exhibited clear in vitro inhibitory activity against selected bacterial and fungal spoilage indicators without inhibiting the baker’s yeast indicator. The selected starters were successfully implemented under semi-industrial bread-making conditions and demonstrated favorable technological and descriptive sensory performance in several formulations. Their application was also associated with a later onset of visually detectable bacterial and fungal spoilage compared with the corresponding controls, supporting their potential contribution to improved storage stability. The results demonstrate the practical applicability and adaptability of the developed starter combinations across different flour matrices and provide a basis for selecting suitable starter–flour combinations and sourdough inclusion levels for bread production. This study therefore establishes a strong technological platform for further refinement and broader application of defined multi-strain sourdough systems. Future research may build on these findings through more detailed kinetic, metabolite, rheological, and microbial characterization to further elucidate the mechanisms underlying the observed technological and biopreservative effects.