Sourdough fermentation and red lentil protein enrichment as sustainable valorization of pasta by-products to make new pasta

Abstract

This study explored the potential of using pasta by-products-based sourdough, with or without red lentil protein isolate (RLPI) enrichment, for pasta making. Fructilactobacillus sanfranciscensis A1, Lacticaseibacillus rhamnosus Fc3-6, and Lacticaseibacillus casei Fc1-13 were used as sourdough starters, while various fermentation times were screened during backslopping to achieve the target pH (approximately 4.6). Twenty percent of newly formulated sourdoughs were mixed with semolina flour for pasta making and evaluated for technological, textural, and biochemical properties. Confocal laser scanning microscopy showed that the new pasta had a similar microstructure compared to the traditional semolina counterpart. RLPI-enriched sourdough pasta showed superior amino acid profiles and optimal protein and starch digestibility. The enrichment of RLPI ensured a unique phenolic compound profile (catechin, salicylic acid, naringenin, and kaempferol), potentially enhancing the antioxidant value of new pasta. Although a slightly longer optimal cooking time, RLPI-enriched sourdough pasta showed favorable sensory attributes, including improved color perception, while preserving textural properties. Sourdough fermentation allowed RLPI to enrich pasta with recycled pasta by-products, promoting the sustainability of the food system.

Graphical abstract

Highlights

• •Pasta waste-based sourdough was used to develop innovative pasta.
• •Sourdough pasta was enriched with red lentil protein isolate.
• •The newly developed pasta improved amino acid profile and digestibility.
• •Phenolic compounds of newly developed pasta can enhance its antioxidant potential.
• •New pastas had favorable sensory attributes while preserving textural properties.

1. Introduction

The shift towards more sustainable food systems represents a critical global challenge for scientists, stakeholders and industries (Varzakas and Smaoui, 2024). Currently, around 30 % of global food production is wasted annually (Pontonio et al., 2024), causing significant economic and environmental impacts (Kohli et al., 2023). In Italy, this waste amounts to 20 million tons annually, with a 12 billion euro loss (Faggini et al., 2023). From the perspective of a sustainable food system, a shift towards the use of plant-based protein sources is also advisable (Lurie-Luke, 2024) (Semba et al., 2021) to reduce the global greenhouse gas emissions (Xu et al., 2021), and to cope with the increasing food demand (Deprà et al., 2022). Facing this background, an integrated approach that combines the repurposing of food by-products with this shift in protein sourcing offers a synergistic strategy for addressing these environmental challenges.

Pasta is a mainstay of many people's diets worldwide (López-Gómez et al., 2022), with 16.9 million tons produced worldwide in 2022 (International Pasta Organization). Italy leads in pasta production, contributing 3.9 million tons (Statistaa) and averaging 23 kg of pasta consumption per capita annually (Statistab), which in turn leads to significant food loss and waste (Faggini et al., 2023) (376,800 tons of waste annually) (López-Gómez et al., 2022). Pasta by-products (namely pasta lost during production phases or defected pasta) still retain high amounts of starch (ca. 60 %, w/w) and proteins (ca. 13 %, w/w), offering valuable reuse potential (Principato et al., 2019) (dos Santos et al., 2018). Indeed, it was reused through bioprocesses, including enzymatic treatment and fermentation for making new food ingredients (Milani et al., 2024) animal feed, compost, bio-based fuels, and production of key metabolites (e.g., lactic acid and volatile fatty acids) (Principato et al., 2019) (Marzo-Gago et al., 2022) (Possente et al., 2022). Nevertheless, the utilization of pasta by-products face significant technological limitations, consumer acceptability, marketability, and regulatory constraints (Rațu et al., 2023). For instance, the incorporation of pasta by-products as a mixture with durum wheat semolina for pasta manufacturing reduces the quality of the product, decreasing firmness and color, and increasing cooking loss, especially in long-cut pasta such as spaghetti (Donnellyl, 1980) (Fang’ and Khan’, 1996).

On the other hand, dietary proteins represent essential macronutrients for human nutrition and the urgent current need of a shift from animal to plant-based proteins source to support more sustainable food systems brought the attention on vegetable protein sources. Among alternatives, protein forms from legumes, such as concentrates and isolates, were considered (Tonini et al., 2024). Lentil is one of the pulse crops of Mediterranean area with the richest nutritional and functional profiles, with protein content ranging from 20 to 30 % (Tonini et al., 2024). Nevertheless, the incorporation of lentil-derived ingredients, such as flours, concentrates, and isolates, can introduce undesirable off-flavours, which hinder their use in various foods such as bakery products, salad dressing, pasta, and sausages (Chang et al., 2019). Furthermore, isolated ingredients often lack the broader nutritional profile, bioavailability and absorption rates found in less processed sources such as protein concentrates (Tonini et al., 2024).

In this perspective, sourdough fermentation emerges as a valuable biotechnological strategy to address the challenges associated with the use of pasta by-products and lentil protein isolates (Gobbetti et al., 2020). Indeed, although this approach still arises some challenges, such as a longer production process (due to possible in situ fermentation time) and the impact of fermented ingredients on the overall flavor of final food products (Montemurro et al., 2019), it is also known for the capacity to enhance the textural properties of food waste for potential use in pasta (Pontonio et al., 2024), and the technological and sensory characteristics of legume flours (Verni et al., 2022). Additionally, sourdough fermentation was effective in improving food digestibility, lowering starch hydrolysis index (Wang and Wang, 2024), enhancing consumer appeal for plant-based foods (Cera et al., 2024), and reducing antinutritional factors (Gabriele et al., 2019).

In this study, we aimed to prepare sourdoughs with pasta by-products (pasta by-products-based sourdough, PRSD) by using a consortium of lactic acid bacteria (LAB). This preparation was followed by back-slopping and fortification with red lentil protein isolate for novel pasta formulations. The comparison with traditional semolina pasta revealed the efficiency of the proposed biotechnological approach.

2. Materials and methods

2.1. Raw materials, microorganisms, and culture conditions

"Pasta regrind" (PR), a by-product generated during the cutting, drying, and milling stages of durum wheat pasta production, primarily composed of durum wheat semolina and water, was provided in accordance with current food legislation by Barilla G. and R. Fratelli S.p.A. (Parma, Italy). Briefly, pasta by-products were ground to a flour by using a mill. The final particle size was <300 μm, with a particle size distribution that was equivalent to that of the used semolina. Red lentil protein isolate (RLPI) was provided by Müller's Mühle (Gelsenkirchen, Germany). Before use, all ingredients listed with proximate composition (Table S1) were stored at room temperature. Fructilactobacillus sanfranciscensis A1, Lacticaseibacillus rhamnosus Fc3-6, and Lacticaseibacillus casei Fc1-13 from the Culture Collection of the Department of Soil, Plant and Food Science, University of Bari Aldo Moro (Bari, Italy) were used as starters. Strain cultures of LAB were stored in 20 % (v/v) glycerol at – 20 °C and routinely propagated in de Man, Rogosa and Sharpe (MRS) broth (Oxoid, Basingstoke, Hampshire, UK) for 24 h at 30 °C.

2.2. Pasta by-products-based sourdough (PRSD) preparation

Two fermentation phases were investigated to define the fermentation conditions for creating PR-based sourdough. Initially, a dough was prepared by mixing PR, water, and all starter cultures to obtain a dough yield (DY) of 160. Cells of LAB starters from the overnight cultures (ca. 15 h) were harvested by centrifugation at 10,000 rpm for 10 min at 4 °C, washed twice in sterile physiological solution (NaCl 0.9 %, w/v), and inoculated in the dough to a final cell density of ca. 7.0 Log CFU/g. The dough was fermented at 30 °C for 20 h. Throughout incubation, pH was monitored every 2 h using a pH-meter (SensIONTM + PH3, Hach, Italy) equipped with a food penetration probe. The target pH (ca. 4.6) was achieved after 12 h, establishing it as the required time for PR sourdough fermentation (M-PRSD) (Fig. 1A). In the second fermentation phase, backslopping was employed. Specifically, 30 % (w/w) of M-PRSD was used to inoculate water and PR, creating a mixture with the same DY. The percentage of inoculum was chosen based on preliminary assays (that adopted several different inoculum percentages) to quickly reach the target final pH of ca. 4.6. The resulting sourdough underwent fermentation at 30 °C for 12 h. Following the same previous approach, the incubation time necessary for backslopping was determined as 8 h, resulting in PRSD sourdough. This sourdough was subsequently used in further experimental investigations.

Fig. 1.

Schematic representation of biotechnological approach adopted for sourdough preparation (panel A) and pasta making (panel B).

2.3. Pasta by-products-based sourdough (PRSD) enriched with RLPI

The addition of RLPI into the PR-based sourdough was evaluated. Starting from PRSD, with 30 % (w/w) used as the inoculum, two sourdoughs enriched with RLPI were formulated, each maintaining a DY of 160. The first variable between the sourdoughs was RLPI added percentage: PRSD-L30 contained 30 % RLPI (w/w), while PRSD-L40 contained 40 % RLPI (w/w), with PR and water as remaining ingredients. The additions of 40 % and 30 % RLPI were chosen to produce a sourdough containing the highest percentage of RLPI while maintaining the DY of 160 and to investigate a formulation with a lower percentage of inclusion, respectively. The second variable, the incubation duration, was set to achieve a pH target of ca. 4.6, resulting in fermentations of PRSD-L30 for 24 h and PRSD-L40 for 48 h at 30 °C (Fig. 1A). PRSD without the addition of RLPI was used as the control.

2.4. Microbiological, physicochemical, and biochemical characterization of formulated sourdoughs

The cell density of LAB in PRSD, PRSD-L30, and PRSD-L40 after fermentation was evaluated through the plate count method. To this purpose, 10 g of sample was mixed with 90 mL of sterile physiological solution (NaCl 0.9 %, w/v) and homogenized with a Stomacher 400 lab blender (Seward Medical) followed by serial dilution. Mesophilic LAB were determined on MRS agar (Oxoid Ltd., Basingstoke, Hampshire, UK), supplemented with cycloheximide (0.1 %) (Sigma-Aldrich, USA), at 30 °C for 48 h. Total titratable acidity (TTA), expressed as the amount (mL) of 0.1 M NaOH to achieve the pH of 8.3 units, was measured on 10 g of samples homogenized with 90 mL of distilled water for 3 min in a bag mixer (400P; Interscience, St Nom, France). pH measurements were performed as previously described. Water-soluble extracts (WSEs) from all PR-based sourdough samples were prepared as reported by (Weiss’ et al., 1993). The concentration of maltose, glucose, and lactic and acetic acids was determined on WSEs using a high-performance liquid chromatography (HPLC) equipped with an Aminex HPX-87H column 300 × 7.8 mm (ion exclusion, Biorad, Richmond, CA), a PerkinElmer 200a refractive index detector (RI), and a UV detector operating at 210 nm. Elution was at 70 °C, with a flow rate of 0.6 mL/min, using 5 mM H₂SO₄ as mobile phase. Sugars and organic acids standards were purchased from Sigma-Aldrich (Steinheim, Germany). The fermentation quotient (FQ), defined as the molar ratio of lactic and acetic acid, was also determined.

2.5. Pasta making

Before pasta making, PRSD, PRSD-L30, and PRSD-L40 were freeze-dried using a pilot-scale lyophilizer (Epsilon 2-6D LSC plus freeze-drier, Germany) and ground to final particle size <100 μm (Fig. 1B). Five different pasta prototypes were manufactured to investigate the influence of formulated sourdoughs on pasta making. The experimental pasta (spaghetti) was prepared at semi-industrial plant of Barilla G. e R. Fratelli S.p.A. (Parma, Italy). Three different batches were produced for each pasta prototype. Three pasta doughs were formulated by substituting 20 % (w/w) of semolina flour with PRSD, PRSD-L30, and PRSD-L40 dried sourdoughs (P-PRSD, P-PRSD-L30, and P-PRSD-L40, respectively). Pasta made entirely with semolina flour (P-CTR) and another with 20 % (w/w) pasta regrind replacing semolina (P-PR) (Fig. 1B) were used as the controls. The 20 % substitution was chosen (based on preliminary trials) to reach in P-PR, P-PRSD, P-PRSD-L30, and P-PRSD-L40 the protein content of traditional pasta (P-CTR). Semolina flours and pasta regrind had particle size <300 μm. All pasta doughs had a DY of 143. Ingredients were mixed for 20 min at ca. 38 °C. Then, the final dough (moisture content, 32–33 %) was kneaded at 35 °C and extruded at 38–40 °C. The extruder (Ser.Com. S.r.l., Italy) had an extrusion pressure of 80 bar, an extrusion speed of 1.3 m/min, and six dies per insert for a total of four inserts. The die diameter was 42 mm. The extruded pasta was dried at 80–82 °C for ca. 8 h. Then, the process was completed with a moisture stabilization step at 83 °C for 70 min, followed by a 10-min cooling at 25 °C, according to the Barilla standard. The final pasta was characterized by moisture content of ca. 11 % and dry spaghetti diameter of 1.7 mm. pH and TTA measurements on pasta samples were performed as previously described.

2.6. Cooking time, hydration test, water absorption, texture, and color analysis

The optimal cooking time (OCT) of the pasta samples, corresponding to the disappearance of the white core, was defined by Barilla G. e R. Fratelli S.p.A. The hydration characteristics of pasta samples were determined according to the method of (Marti et al., 2011). Water absorption of pasta during cooking was determined by measuring the weight of the pasta before and after the cooking process. The results were expressed as $\left[\frac{{\mathrm{W}}_1–{\mathrm{W}}_0}{{\mathrm{W}}_0}\right]\mathrm{x}100$, where W₁ is the weight of cooked pasta and W₀ is the weight of the uncooked pasta.

Instrumental Texture Profile Analysis (TPA) was performed by TVT 6700 Texture Analyzer, equipped with a 35-mm diameter stainless steel cylindrical probe (probe P-CY35S). Pasta samples were cooked until the OCT, left to cool at room temperature, placed in a beaker (diameter, 60 mm; height, 85 mm), and filled to about half volume. The selected settings were the following: test speed 1 mm/s, 30 % deformation of the sample and two compression cycles (with an interval of 30 s). TPA was carried out using TexCalc 5 software, which measured hardness, chewiness, springiness, cohesiveness and resilience.

The color measurements of pasta samples were determined using a CR-400 Minolta Chroma Meter. The L∗, a∗, b∗ color space analysis method was used, where L∗ represents lightness (black-white), while a∗ and b∗ are the chromaticity coordinates (red-green and yellow-blue, respectively). The Euclidean distance (ΔE), a metric for quantifying the perceptually meaningful difference between two colors in the CIE L∗a∗b∗ space, was measured as follows: ΔE = [(ΔL∗)² + (Δa∗)² + (Δb∗)²]^1/2 where ΔL, Δa and Δb are the differences for L∗, a∗ and b∗ values between sample and reference (a white 175 ceramic plate having L∗ = 92, a∗ = - 0.35 and b∗ = 4.45).

2.7. Proteins, peptides, and amino acids quantification

The total nitrogen content of pasta samples was determined using the NDA 702 Dumas Nitrogen Analyzer (VELP Scientifica, 2014). Samples were weighed and carefully wrapped in tin foil before the analysis. Based on the nature of the sample, oxygen dosage was fixed at a rate of 400 mL/min to achieve the best combustion. Total nitrogen results were obtained using VELP DUMASoftTM 6.1.0 and converted to total protein content by multiplying with the standard conversion factor 6.25.

Peptides and free amino acid concentrations were quantified in phosphate buffer extracts (PBEs) of pasta samples (Rizzello et al., 2017). Briefly, 2 g of pasta were extracted with 8 mL of phosphate buffer (50 mM, pH 7.0, containing 0.1 M) for 1 h at 4 °C under shaking conditions, centrifuged at 12,000 rpm for 15 min, and syringe-filtered through 0.22-μm pore size LLG-Syringe filters SPHEROS (LLG-Labware, CA). Peptides concentrations were determined using the o-phthaldialdehyde (OPA) method. Total and individual free amino acids were analyzed by a Biochrom 30+ series Amino Acid Analyzer (Biochrom Ltd., Cambridge Science Park, England) with a Li-cation-exchange column (20 by 0.46 cm inner diameter). Proteins and peptides were precipitated by adding 5 % (v/v) cold solid sulfosalicylic acid, kept at 4 °C for 1 h and centrifuged at 12,000 rpm for 15 min. The supernatant was filtered again through a 0.22 μm pore size filter. Amino acids were post-column derivatized with ninhydrin reagent and detected by absorbance at 440 (proline) or 570 nm (all the other amino acids).

2.8. Free phenolic compounds analysis

Free phenolic compounds were assessed using an LC-MS platform composed of an Ultimate 3000 RSLCnano system coupled to a QExactive Plus Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). Analyses were conducted using a methanol/water-soluble extract (MWSE) obtained from pasta samples. The extraction process involved mixing 1 g of the sample with 5 mL of a methanol/water solution (80/20, v/v) acidified with hydrochloric acid (0.1 %, v/v). Sonication was performed for 1 min (two cycles, 30 s/cycle, 5 min interval between cycles) in an ice-bath, followed by an additional hour of extraction under stirring conditions at room temperature. The MWSE were then centrifuged at 12,000 rpm for 15 min and syringe-filtered through 0.22-μm pore size LLG-Syringe filters SPHEROS (LLG-Labware, CA). A Waters Acquity HSS T3 column (1.8 μm, 100 mm × 2.1 mm) (Milford, MA, USA) was used for chromatographic separations of phenolic compounds by applying the following binary gradient program: 0 min, 2 % B; from 0 to 3 min, linear gradient to 20 % B; from 3 to 4.3 min, isocratic 20 % B; from 4.3 to 9 min, linear gradient to 45 % B; from 9 to 11 min, linear gradient to 100 % B; from 11 to 13 min, wash at 100 % B; from 13.01 to 15 min, back to the initial conditions of 5 % B (solvent A = water + 0.1 % formic acid; solvent B = acetonitrile + 0.1 % formic acid). The flow rate was 300 μL/min, and the temperature of the column was set at 40 °C. The LC-MS system was equipped with a Higher Collisional energy Dissociation cell (HCD), and a HESI (Heated Electro Spray Ionization) interface was adopted for LC-HRMS coupling. MS detection following chromatographic separation was performed in Targeted-SIM/dd-MS² mode and in negative polarity using the following parameters: spray voltage, 2.80 kV; sheath gas flow rate, at 35 arbitrary units; auxiliary gas flow rate, at 10 arbitrary units; capillary temperature, at 300 °C; S lens RF level, at 50 arbitrary units; capillary gas heater temperature, 280 °C. The settings for the Q-ExactiveTM mass spectrometer were the following: mass scan range, 80–1200 m/z; resolution, 70.000 (FWHM); Automatic Gain Control (AGC) Target, 5∗10⁶ ions; maximum injection time (IT), 100 ms. Compounds were identified based on their reference standard, retention time, and [M− H]⁻ ions. Peak areas, obtained from eXtracted Ion Current (XIC) chromatographic traces, i.e., chromatograms created by extracting the ion current from HRMS spectra in a m/z interval, including the monoisotopic peak, were used as a measurement of MS response. They were thus employed to perform calibrations based on commercial standards acids and, afterwards, to quantify these phenolic compounds in pasta extracts. The Xcali bur™ v. 3.1 software (Thermo Fisher Scientific, Waltham, MA, USA) was used to control the Q-Exactive plus™ spectrometer and data elaboration.

2.9. In vitro antioxidant analysis

The in vitro antioxidant activity of MWSE obtained from pasta samples was estimated through the determination of the radical scavenging capacity on 1,1-diphenyl-2-picrylhydrazyl radical (DPPH) as described by (Yu, 2001).

2.10. Total and resistant starch, hydrolysis index (HI), and predicted glycemic index (pGI)

Total starch on raw and cooked (until the OCT) pasta (g/100g) and resistant starch on freeze-dried pasta samples (g/100g of dry weight (DW)) were determined using the Total Starch (K-TSTA) and the Resistant Starch (K-RSTAR) Assay kits (Megazyme, Ireland), respectively, according to the manufacturer's instructions. The percentage of starch loss during cooking was expressed as $\left[\frac{\mathrm{C}0–\mathrm{C}1}{{\mathrm{C}}_0}\right]\mathrm{x}100$, where C₀ is the concentration of starch in raw pasta and C₁ is the concentration of starch in the cooked sample. For comparative analysis among pasta samples, that may not reflect the absolute values, the predicted glycemic index (pGI) was determined by quantifying the hydrolysis index (HI) according to (Capriles and Arêas, 2013). The pGI was calculated using the equation pGI = 39.71 + 0.549 (HI), as previously proposed by (Gorii et al., 1997). A reference sample (white wheat bread) was used as a control to estimate the hydrolysis index (HI = 100).

2.11. Confocal laser scanning microscopy (CLSM) and image analysis

Pasta samples were cooked until the OCT, sectioned into 0.2 mm slices and placed over glass slides. Slices were stained with a solution 1 % w/v fluorescein isothiocyanate (FITC) (Sigma-Aldrich, USA) and 0.1 % w/v Rhodamine B (RhoB) (Sigma-Aldrich, USA) in NN-dimethylformamide (Honeywell, Germany), 1 h at room temperature, for labeling of starch and protein components (Parada and Aguilera, 2011). Slices were then rinsed with distilled water and observed using a Leica SP8LIA CLSM (Leica Microsystems, Germany) with 488 and 552 nm lasers. Fluorescence emission was between 503 and 540 nm (FITC) and 605–645 nm (RhoB). Images were captured with a 10X objective and analyzed with LAS X software (Leica, 5.1.0 version, 2022) (Da Ros et al., 2021).

The quantification of the percentage of area covered by the protein matrix of each pasta sample was conducted using the Image J software (Java, 1.52t version, 2020) (Anovitz and Cole, 2015). CLSM images were loaded into the software and subjected to RGB (red-green-blue) color splitting. The resulting image (corrected for shading) was then processed using automatic thresholding, converting the red image into binary (black and white) mode (Peighambardoust et al., 2006). Automatic thresholding was consistently applied across all images to ensure uniformity in comparisons. The fraction of area in each stripped image was then calculated and reported as the percentage of area covered by the protein matrix (PMA).

2.12. In vitro protein digestibility (IVPD) and protein digestibility corrected amino acid score (PDCAAS)

The IVPD was evaluated using the Protein Digestibility Assay kit (Megazyme, Ireland) according to the manufacturer's instructions. PDCAAS was calculated by multiplying the IVPD and the amino acid score of the limiting amino acid. To obtain a complete and totally free individual amino acids profile, all pasta samples underwent three separate hydrolysis sessions, as reported by (Tlais et al., 2023).

2.13. Antinutrients analysis

Phytic acid and raffinose concentrations were determined using the Phytic Acid (Phytate)/Total Phosphorus (K-PHYT) (based on the AOAC method 986.11) and the Raffinose/D-galactose (K-RAFGA) assay kits (Megazyme, Ireland), respectively, following the manufacturer's instructions. Condensed tannins were determined using the vanillin assay, as described by (Price et al., 1978) with the modifications proposed by (Krause et al., 2023). Concentrations of antinutrients in pasta samples were expressed in g/100g of pasta.

2.14. Sensory analysis

Sensory evaluation of pasta samples was carried out by a trained sensory panel (n = 10) from Barilla G. e R. Fratelli S.p.A. (Parma, Italy) using Profile Attribute Analysis (PAA). The samples were cooked (according to their OCT), presented in randomized order, and tasted within 5 min. The three following categories of attributes were used to score the samples on a 1 (lowest intensity) to 12 (highest intensity) scale: appearance (yellow color, greyish color, thickness, starch on surface, and liveliness in the plate), texture (hardness, chewiness, gumminess, cohesiveness, uniformity in mouth, surface in the mouth, consistency in mouth, astringent, starchy, tooth sticking, tooth packing, and easiness to swallow), and aroma & flavor (aroma balance, aroma fullness, aroma overall, Off-notes aroma, flavor balance, flavor fullness, semolina flavor, whole grain flavor, flour/starch flavor, milky note, toasted flavor, nuts, sweet, sour, salty, bitter, Off-notes flavor, and aftertaste). All participants provided written informed consent to participate in this study, which was approved by the Clinical Research Ethics Committee of the Cork Teaching Hospital (reference number: ECM 4 2023 & ECM 5 & 2023 ECM 3 2024).

2.15. Statistical analysis

All analyses were carried out considering three replicates. Data were subjected to one-way ANOVA and the pairwise comparison of treatment means was obtained by Tukey's post-hoc analysis at p-value (p) < 0.05. Analysis and graphs were done using the software RStudio version 4.2.2 (R Development Core Team). Data sets related to sensory analysis were analyzed through Principal Component Analysis (PCA) and Kiviat chart using the statistical software XLSTAT.

3. Results

3.1. Pasta by-products-based sourdough preparation enriched with RLPI

The initial pH of the first dough made from PR with a DY of 160 was 5.91 ± 0.01. After LAB inoculation, a consistent (p < 0.05) pH reduction was observed over time, reaching 4.69 ± 0.01 after 12 h at 30 °C. This sample, designated as M-PRSD, was used as a 30 % inoculum to prepare a new sourdough with the same DY, starting at a pH of 5.29 ± 0.02, which decreased to 4.55 ± 0.02 over 8 h of incubation (PRSD). Two additional sourdoughs enriched with different concentrations of RLPI (30 and 40 %) were then developed from PRSD. After 24 h of incubation, the pH of PRSD-L30 reached 4.69 ± 0.04, whereas PRSD-L40 required an additional 24 h to achieve a pH of 4.88 ± 0.06 (Table S2).

3.2. Microbiological, physicochemical, and biochemical characterization of formulated sourdoughs

As estimated on MRS agar, the cell density of LAB in PRSD, PRSD-L30, and PRSD-L40 sourdoughs was similar (p > 0.05) (from 8.96 ± 0.01 to 9.07 ± 0.04, Log CFU/g). Despite the comparable pH values among the three sourdoughs, TTA provided greater differentiation, with the highest TTA observed in PRSD-L30 (23.15 ± 0.75 mL 0.1M NaOH/10g), followed by PRSD-L40 (21.00 ± 0.40 mL 0.1M NaOH/10g), and the lowest in PRSD (7.05 ± 0.05 mL 0.1M NaOH/10g) (Table S2). HPLC analysis revealed maltose as the predominant sugar in PRSD (67.83 ± 1.75 mM), with significantly (p < 0.05) lower levels in PRSD-L30 and PRSD-L40 (Fig. 2). Glucose concentrations ranged from 11.37 ± 0.02 (PRSD-L30) to 5.90 ± 0.03 (PRSD-L40) mM. Backslopping significantly (p < 0.05) influenced the release of microbial metabolites, particularly lactic and acetic acids. The highest lactic acid concentration was observed in PRSD-L40, followed by PRSD-L30, while the lowest was in PRSD. However, acetic acid concentrations were similar between the RLPI-enriched sourdoughs but significantly higher than in PRSD. These variations were reflected in the fermentation quotient (FQ), which was the highest for PRSD-L40 (15.60 ± 0.49) and PRSD-L30 (13.70 ± 0.01) and the lowest in PRSD (6.69 ± 0.08) (Table S2).

Fig. 2.

Maltose, glucose, lactic and acetic acid concentration (mM) of sourdough made with “pasta regrind” (PR) inoculated with Fructilactobacillus sanfranciscensis A1, Lacticaseibacillus rhamnosus Fc3-6, and Lacticaseibacillus casei Fc1-13 (PRSD), sourdough composed of 30 % (w/w) PRSD as inoculum and enriched with 30 % (w/w) red lentils protein isolate (RLPI) (PRSD-L30), or 40 % (w/w) RLPI (PRSD-L40). PRSD, PRSD-L30, PRSD-L40 were incubated at 30 °C for 8, 24, and 48 h, respectively. Bars with different superscript letters differ significantly (p < 0.05). Details on the sourdough formulations are provided in material and methods.

3.3. Technological parameters, texture, and color determination

According to ISO 7304-1 (2016), the optimal cooking time (OCT) for the pasta samples was fixed as 9 min for P-CTR, P-PR, and P-PRSD, and 10.5 min for pasta enriched with 30 and 40 % RLPI. Initially, the water uptake kinetics were comparable across all samples during the 1st h (Fig. S1). Afterwards, significant (p < 0.05) deviations emerged, with P-PR and P-PRSD showing markedly higher water uptake compared to the other samples after 3 h (Fig. S1 and Table 1). After cooking, P-PRSD and P-PRSD-L30 resulted in the highest water absorption compared to other pasta samples.

Table 1.

Technological parameters, textural, and physical characterizations of pasta made entirely with semolina flour (P-CTR), pasta made with 20 % (w/w) pasta regrind replacing semolina (P-PR), and pastas made with 20 % substituting semolina flour with the different dried sourdoughs (P-PRSD, P-PRSD-L30, and P-PRSD-L40, respectively). Water uptake (%) at 25 °C was carried out on raw (uncooked) pasta samples.

Samples	P-CTR	P-PR	P-PRSD	P-PRSD-L30	P-PRSD-L40
Technological parameters
Water uptake (%) on raw pasta	53.5 ± 0.9b	59.2 ± 0.9a	57.4 ± 0.4a	52.9 ± 1.2b	52.5 ± 0.2b
Water absorption (%)	107.4 ± 0.8b	105.2 ± 2.3b	113.6 ± 0.8a	111.7 ± 0.3a	106.5 ± 0.5b
Textural characteristics
Hardness (N)	5.0 ± 0.2a	3.1 ± 0.4b	2.9 ± 0.5b	4.3 ± 0.3a	4.4 ± 0.2a
Chewiness	132.0 ± 17.0c	175.5 ± 19.1c	134.0 ± 12.7c	470.5 ± 3.5b	633.0 ± 77.8a
Springiness	0.6 ± 0.0a	0.7 ± 0.1a	0.5 ± 0.1a	0.6 ± 0.1a	0.6 ± 0.0a
Cohesiveness	0.7 ± 0.1a	0.5 ± 0.1a	0.6 ± 0.0a	0.7 ± 0.1a	0.7 ± 0.0a
Resilience	0.2 ± 0.1a	0.1 ± 0.0b	0.1 ± 0.0ab	0.2 ± 0.1ab	0.2 ± 0.0ab
Color analysis
L∗	51.5 ± 1.1a	50.3 ± 0.7ab	50.0 ± 2.4ab	47.2 ± 0.5bc	46.1 ± 0.9c
a∗	−0.1 ± 0.2b	0.3 ± 0.2b	0.3 ± 0.2b	4.7 ± 0.4a	5.4 ± 0.7a
b∗	17.2 ± 0.3a	16.1 ± 0.8ab	14.9 ± 1.1b	17.4 ± 0.1a	17.8 ± 0.9a
ΔE	40.84 ± 0.88b	41.96 ± 0.72b	42.24 ± 0.81b	45.38 ± 0.39a	46.59 ± 0.28a

^a–c Means within the row with different letters are significantly different (p < 0.05).

As for the textural parameters, P-PR and P-PRSD had lower (p < 0.05) values of hardness compared to all the other pasta samples, with P-CTR, P-PRSD-L30, and P-PRSD-L40 displaying slightly but not significantly (p > 0.05) different hardness values. The presence of RLPI caused an increase of the chewiness in P-PRSD-L30 and P-PRSD-L40 compared to all the other pasta samples. Conversely, no significant differences (p > 0.05) in cohesiveness and springiness were observed among all cooked pasta samples. For resilience, significant differences (p < 0.05) were only observed between P-CTR and P-PR (Table 1). The color lightness (L∗) was substantially reduced due to the enrichment of the cooked pasta with sourdough containing RLPI compared to P-CTR. Whilst scaler quantity of red-green (a∗) significantly increased in P-PRSD-L30 and P-PRSD-L40 compared to P-CTR. The yellow-blue (b∗) values were similar between P-PRSD-L30, P-PRSD-L40, and P-CTR, but significantly higher than those of P-PRSD. The addition of RLPI to the pasta formulation (P-PRSD-L30 and P-PRSD-L40) had a substantial (p < 0.05) impact on the ΔE value, which was the greatest among all the samples evaluated.

3.4. Biochemical and physicochemical characterization of pasta

The pH of the control pasta (P-CTR) was 6.41 ± 0.01. Pasta enriched with PR, PRSD, PRSD-L30, or PRSD-L40 led to a significant (p < 0.05) decrease in pH, with the most pronounced reduction observed in P-PRSD-L40 (5.44 ± 0.01). In contrast, changes in TTA values were significantly higher only when RLPI-sourdough was used (Table S3). P-CTR had an initial concentration of protein, peptides, and starch of 16.2 ± 0.2 g/100 g, 3.9 ± 0.1 mg/g, and 78.5 ± 0.9 g/100 g, respectively. Likewise, the enrichment of pasta with RLPI-sourdoughs, particularly P-PRSD-L40, resulted in a significant increase in protein and peptide contents compared to P-CTR. Fortification with PR and PRSD showed similar protein levels but significantly (p < 0.05) lower peptide content. All fortified samples showed reduced starch content, especially P-PRSD and P-PR (Table S3).

3.5. Confocal laser scanning microscopy (CLSM) and image analysis of pasta

The distribution of protein components (red) and starch granules (green) had no discontinuity (Fig. 3). The comparison of percentages of area covered by protein matrix (PMA) by Image J software showed that the inclusion of RLPI (P-PRSD-L30 and P-PRSD-L40) resulted in the highest PMA values (22.3 ± 0.5 and 22.3 ± 0.5 %, respectively).

Fig. 3.

Confocal Laser Scanning Microscopy (CLSM) images of pasta stained with 1 % (w/v) fluorescein isothiocyanate (FITC) and 0.1 % (w/v) Rhodamine B (RhoB). Green fluorescence represents starch granules whereas red represents protein matrix. (a) Pasta made entirely with semolina flour (P-CTR); (b) pasta made with 20 % (w/w) pasta regrind replacing semolina (P-PR); (c) pasta made with 20 % (w/w) substituting semolina flour with PRSD sourdough (P-PRSD); (d) pasta made with 20 % (w/w) substituting semolina flour with PRSD-L30 sourdough (P-PRSD-L30); (e) pasta made with 20 % (w/w) substituting semolina flour with PRSD-L40 sourdough (P-PRSD-L40). Scale bars represent 80 μm.

3.6. Total and free amino acids (TFAA)

TFAA concentration in P-CTR was 1340.9 ± 24.0 mg/kg. Enriching pasta with PR or PRSD decreased TFAA concentrations while incorporating RLPI-enriched sourdough led to a significant (p < 0.05) increase, especially in P-PRSD-L40 (ca. 86 %) (Fig. 4). Amino acid profiling revealed similar trends for His and Met, with both showing increased levels in RLPI-fortified samples. Tyr and Gln followed a comparable pattern, with no significant differences between P-CTR and P-PR. Levels of Ser, Gly, Cys, Leu, Phe, Arg, and Pro in P-PR and P-PRSD were not significantly different from P-CTR but were markedly higher in P-PRSD-L40 and P-PRSD-L30. GABA was exclusively detected in P-PRSD-L40 and P-PRSD-L30. Ala, Lys, and Orn levels were similar between P-CTR and P-PR, increased significantly in P-PRSD, and were the highest in P-PRSD-L40 and P-PRSD-L30. Glu levels were similar in P-CTR and P-PRSD, significantly lower in P-PR, and significantly higher in P-PRSD-L40 and P-PRSD-L30. Only Val and Ile showed lower values in pasta made with RLPI-enriched sourdough than other pasta samples.

Fig. 4.

Quantification of the total and individual free amino acids (mg/kg) in pasta made entirely with semolina flour (P-CTR), pasta made with 20 % (w/w) pasta regrind replacing semolina (P-PR), and pastas made with 20 % substituting semolina flour with PRSD, PRSD-L30, and PRSD-L40 dried sourdoughs as described in the Materials and Methods (P-PRSD, P-PRSD-L30, and P-PRSD-L40, respectively).

3.7. Nutritional characterization

The total free amino acid composition was analyzed to evaluate the in vitro protein digestibility (IVPD). Acid, performic oxidation, and alkaline hydrolysis showed high variations among the pasta samples (Table S4). The variations in Lys, His, Pro, and Arg mostly reflected the differences in protein digestibility among the samples. The inclusion of PRSD-L30 and especially PRSD-L40 sourdough in pasta samples significantly (p < 0.05) enhanced the IVPD compared to other pasta samples (Table 2). Essential amino acid scores and the first limiting amino acid were determined based on FAO guidelines. Except for the P-PRSD-L30 and P-PRSD-L40 pasta samples, where Trp had the lowest ratio, Lys was identified as the first limiting amino acids in all samples. As expected, including RLPI in pasta samples increased the PDCAAS values compared to P-CTR. P-PRSD showed the lowest significant resistant starch values, followed by P-PRSD-L30 and P-PRSD-L40, compared to control samples (P-CTR and P-PR). While almost all pasta samples had significantly lower values compared to P-CTR, P-PRSD-L40 showed the lowest hydrolysis index (HI) and predicted glycemic index (pGI) (Table 2).

Table 2.

Nutritional characterization of pasta made entirely with semolina flour (P-CTR), pasta made with 20 % (w/w) pasta regrind replacing semolina (P-PR), and pastas made with 20 % substituting semolina flour with the different dried sourdoughs (P-PRSD, P-PRSD-L30, and P-PRSD-L40, respectively).

Samples	P-CTR	P-PR	P-PRSD	P-PRSD-L30	P-PRSD-L40
In vitro protein digestibility (IVPD) (%)	87.8 ± 0.6b	86.9 ± 1.5b	87.6 ± 2.3b	90.6 ± 0.2ab	92.4 ± 0.8a
First Limiting Amino Acid	Lys	Lys	Lys	Trp	Trp
Amino Acid Score	0.27	0.28	0.27	0.28	0.28
PDCAAS	0.23	0.25	0.24	0.26	0.26
Resistant starch (g/100 g DW)	1.79 ± 0.01a	1.87 ± 0.01a	1.40 ± 0.02c	1.57 ± 0.08b	1.58 ± 0.04b
Hydrolysis index (HI)	74.04 ± 0.27a	68.54 ± 0.46d	72.27 ± 0.28b	69.84 ± 0.50c	63.66 ± 0.37e
Predicted glycemic index (pGI)a	80.36 ± 0.29a	77.34 ± 0.71c	79.39 ± 0.09ab	78.05 ± 0.19bc	74.66 ± 0.54d

^a–e Means within the row with different letters are significantly different (p < 0.05).

^aPredicted values, based on predictive mathematical model (Gorii et al., 1997), were used for comparative analysis of pasta samples but may not accurately reflect absolute values.

3.8. Quantitative LC-HRMS analysis of free phenolic compounds

Forty analytical standards were used to identify and subsequently quantify main free phenolic compounds in pasta samples using LC-HRMS. Out of these, ten standards were detected (Table 3). In the control pasta (P-CTR), ferulic acid was the most abundant phenolic compound, followed by p-coumaric acid and 4-hydroxybenzaldehyde. The addition of PR to the pasta did not significantly (p > 0.05) alter the phenolic compounds profile compared to P-CTR, except for significant (p < 0.05) increase in syringaldehyde, which showed the highest concentration among all samples. When pasta by-products-based sourdough (PRSD) was added to the formulation, there was significant hydrolysis of p-coumaric and protocatechuic acids compared to P-CTR and P-PR, and ferulic acid was completely metabolized. On the contrary, the enrichment of pasta with RLPI substantially affected the profile of phenolic compounds compared to other samples. Specifically, catechin, salicylic acid, naringenin, and kaempferol were exclusively found in P-PRSD-L30 and P-PRSD-L40. The same samples were characterized by a significant reduction in protocatechuic acid, p-coumaric acid, and ferulic acid compared to the P-CTR (Table 3). Although genistein was present in all samples, only P-PRSD-L40 showed significantly higher concentrations than P-CTR.

Table 3.

Quantification of phenolic compounds (μg/kg) through HPLC-HRMS in pasta made entirely with semolina flour (P-CTR), pasta made with 20 % (w/w) pasta regrind replacing semolina (P-PR), and pastas made with 20 % substituting semolina flour with the different dried sourdoughs (P-PRSD, P-PRSD-L30, and P-PRSD-L40, respectively).

Samples	P-CTR	P-PR	P-PRSD	P-PRSD-L30	P-PRSD-L40
Protocatechuic acid	31.01 ± 7.85a	26.88 ± 5.99a	4.63 ± 1.48b	7.69 ± 0.23b	n.d.
Catechin	n.d.	n.d.	n.d.	112.55 ± 4.79a	120.60 ± 1.71a
4-hydroxybenzaldehyde	96.33 ± 11.87abc	46.17 ± 20.25c	54.00 ± 14.14bc	141.06 ± 9.84a	101.98 ± 20.93ab
p-coumaric acid	160.80 ± 13.80ab	177.90 ± 17.63a	88.89 ± 1.66cd	124.43 ± 12.83c	74.39 ± 4.62d
Syringaldehyde	10.71 ± 1.85b	25.40 ± 6.99a	10.44 ± 0.98b	12.47 ± 1.82b	11.24 ± 2.35b
Ferulic acid	670.02 ± 104.56a	430.89 ± 137.35a	n.d.	152.97 ± 38.54b	161.38 ± 14.76b
Salicylic acid	n.d.	n.d.	n.d.	211.60 ± 17.46b	302.44 ± 15.79a
Naringenin	n.d.	n.d.	n.d.	24.17 ± 1.35b	37.53 ± 1.22a
Genistein	4.47 ± 0.70bc	4.33 ± 0.49bc	3.03 ± 0.36c	5.57 ± 0.32ab	6.47 ± 0.47a
Kaempferol	n.d.	n.d.	n.d.	7.33 ± 1.70b	13.86 ± 2.67a

^a–d Means within the row with different letters are significantly different (p < 0.05).

n.d. means not detected.

3.9. In vitro antioxidant analysis

The antioxidant activity of the P-CTR, as determined by DPPH assay, was 4.27 ± 0.38 mmol BHT/kg. No significant (p > 0.05) differences compared to the P-PR were observed. Conversely, other fortifications significantly enhanced the radical scavenging activity (p < 0.05), with values ranging from 5.78 ± 0.81 mmol BHT/kg in P-PRSD to 7.30 ± 0.23 mmol BHT/kg in P-PRSD-L30.

3.10. Antinutrients analysis

The impact of enriching pasta with sourdough ingredients on the antinutritional profile was assessed (Fig. 5). The control pasta (P-CTR) contained phytic acid (0.10 ± 0.01 g/100 g), raffinose (0.12 ± 0.01 g/100 g), and condensed tannins (0.01 ± 0.00 g/100 g). The variations of phytic acid and raffinose content among the samples after the fortification followed almost identical patterns. PR fortification significantly (p < 0.05) elevated phytic acid and raffinose levels, which were subsequently reduced to control levels by PRSD fermentation. For both factors, P-PRSD-L40 and followed by P-PRSD-L30 showed the highest levels. In contrast condensed tannins increased significantly only in sourdough-fortified samples, ranging from 0.05 ± 0.01 g/100 g in P-PRSD to 0.08 ± 0.01 g/100 g in P-PRSD-L40.

Fig. 5.

Antinutritional factors values (g/100 g) in pasta made entirely with semolina flour (P-CTR), pasta made with 20 % (w/w) pasta regrind replacing semolina (P-PR), and pastas made with 20 % substituting semolina flour with PRSD, PRSD-L30, and PRSD-L40 dried sourdoughs as described in the Materials and Methods (P-PRSD, P-PRSD-L30, and P-PRSD-L40, respectively). Bars with different superscript letters differ significantly (p < 0.05).

3.11. Sensory characterization

According to PCA analysis, sensory descriptors for appearance, texture, and aroma and flavor characteristics categorized the samples into three distinct clusters: one cluster comprised solely of P-CTR, the second containing P-PR and P-PRSD, and the third including P-PRSD-L30 and P-PRSD-L40 (Fig. 6 and Fig. S2). In terms of appearance, pasta fortified with RLPI enriched sourdoughs generally showed higher ratings across descriptors, except for thickness when cooked, which was similar to P-CTR, and grayish color when cooked, where P-CTR scored the highest (Table S5). P-PR and P-PRSD received the lowest scores in this category. P-PRSD-L30 and P-PRSD-L40 consistently (p < 0.05) received the highest scores across most descriptors, with occasional high scores shared by P-CTR. Panelists consistently rated P-PR and P-PRSD with the lowest scores for texture attributes. For aroma and flavor, pasta with RLPI enriched sourdoughs showed significantly higher ratings for toasted, nutty, and bitter attributes compared to other samples. On the contrary, semolina and whole grain attributes were notably higher in P-CTR, P-PR, and P-PRSD compared to P-PRSD-L30 and P-PRSD-L40. Attributes like floury/starchy notes, milky undertones, and off-flavors did not significantly differ from P-CTR in fortified pastas. Overall aroma and other remaining attributes received the highest scores in P-CTR, P-PRSD-L30, and P-PRSD-L40 compared to P-PR and P-PRSD, which consistently scored lower (Table S5).

Fig. 6.

Spider web chart displaying sensory average scores based on Profile Attribute Analysis (PAA) of pasta made entirely with semolina flour (P-CTR), pasta made with 20 % (w/w) pasta regrind replacing semolina (P-PR), and pastas made with 20 % substituting semolina flour with PRSD, PRSD-L30, and PRSD-L40 dried sourdoughs as described in the Materials and Methods (P-PRSD, P-PRSD-L30, and P-PRSD-L40, respectively).

4. Discussion

In a global context where food biotechnology focuses on sustainable and innovative practices, a growing demand for dignified food waste and using cost-effective and resource-efficient plant proteins for human consumption is evident. In our innovative approach, based on the use of sourdough fermentation, we exploited the synergistic interplay between pasta regrind and red lentil protein isolate to develop new pasta formulations, thereby enhancing functionality while responding to the modern vision of the circular economy.

To build a robust microbial ecosystem, effective sourdough starters were required. Three LAB strains belonging to F. sanfranciscensis, L. rhamnosus, and L. casei, renowned for their acidifying and diverse enzymatic properties (Rogalski, 2021) (Stefanovic et al., 2017) (Waśko et al., 2012), were combined to form a microbial consortium to develop the PR mother sourdough. This consortium showed rapid acidification, achieving a pH reduction in the mother sourdough to ca. 4.6, aligning with the optimal performance range (3.4–4.9) reported by (Arora et al., 2021). Its effectiveness continued through first backslopping, accelerating pH reduction and the maturation of the final PR-based sourdough. In contrast, the addition of RLPI in the sourdough formulations delayed pH reduction, likely due to its buffering effect, which neutralizes some of acids produced during fermentation and slows the rate of pH decline (Poznanski et al., 2013). Despite similar pH values to PRSD, prolonged fermentation in PRSD-L30 and PRSD-L40 (24 and 48 h, respectively) led to significant reduction of fermentable sugars (Wang and Wang, 2024), particularly maltose, driving increased lactic and acetic acid production and higher TTA levels (Boeck et al., 2021).

Freeze-dried PRSD, PRSD-L30, and PRSD-L40 replaced 20 % of semolina in pasta formulations, with control samples made from 100 % semolina (P-CTR) and 20 % pasta regrind (P-PR). The higher water uptake of P-PR and P-PRSD is probably attributed to the higher PR concentration compared to the other samples, as PR can have a higher water retention capacity than semolina flour depending on its particle size, microstructure and temperatures during the pasta drying process (Tagliasco et al., 2021). However, P-PR showed a lower hardness and resilience compared to P-CTR, which resulted in a lower firm texture. This is due to the high drying temperature of PR that reflects in higher starch damage and proteins denaturation in the final dough (Tagliasco et al., 2021). Simultaneously, the absence of a significant difference in pasta hardness following the addition of RLPI compared to P-CTR suggests that the isolate does not affect the protein-starch matrix. The lower lightness of fortified pasta compared to the control sample can be attributed both to the higher content of ash, and to the lentils pigments which act as coloring components decreasing lightness (Teterycz et al., 2020). The increased redness in P-PRSD-L30 and P-PRSD-L40 is caused by carotenoids of red lentils, which impart an orange-red hue (Teterycz et al., 2020), affecting the color of the RLPI enriched sourdough pasta. The increase observed in the ΔE parameter of pasta following the addition of red lentil flour agrees with (Teterycz et al., 2020).

Fortifying semolina pasta with PR and RLPI-based sourdoughs induced significant biochemical changes. Consistent with previous study, the sourdough pasta showed a lower pre-cooking pH compared to conventional semolina pasta (Fois et al., 2018). Additionally, the RLPI-enriched pasta had higher protein content, in agreement with (Bresciani et al., 2022). Despite differences in protein and starch content, microscopy confirmed structural uniformity across all samples, with RLPI-enriched pasta showing the greatest protein matrix coverage, correlating with its elevated protein levels. This high protein content, combined with the proteolytic activity of the initial microbial consortium which dominated the microbial ecosystem of the sourdough, justified the high release of peptides in P-PRSD-L30 and P-PRSD-L40 (Tonini et al., 2024) (Ramos-Pereira et al., 2021) (Solieri et al., 2022).

Given the role of amino acids in the nutritional and sensory qualities of pasta (Messia et al., 2021), the impact of PR and RLPI-based sourdoughs in modifying the amino acid composition was considerable. As legumes represent a primary source of amino acids for human nutrition but often lack of some essential amino acid (Carbonaro and Nucara, 2022), their use in combination with semolina in the form of RLPI-enriched sourdoughs improved the amino acid profile, leading to a significant rise in most free amino acids. Among these, Gln, the primary amino acid in red lentil proteins (Lee et al., 2021), can boost the nutritional profile and texture of pasta by interacting with the protein network. Similarly, other increased amino acids can improve flavor, protein structure, and elasticity, while offering functional benefits like antioxidant activity (Cys), counteracting the risk of developing type 2 diabetes (Glu), immune support (Arg and Lys), and muscle repair (Pro) (Najafi et al., 2023) (Jankowski et al., 2020) (Wu et al., 2011). The elevated inhibitory neurotransmitter GABA levels observed exclusively in P-PRSD-L30 and P-PRSD-L40 may be attributed either to the addition of RLPI or to the sourdough fermentation enhancing GABA release from RLPI (Torino et al., 2013), which in turn might regulate neural signaling, blood pressure, and provide calming effects (Almutairi et al., 2024). In comparing P-PR and P-PRSD, it became more evident the impact of microbial starters on some amino acids. The observed reductions in Thr, Ser, Gln, Val, Met, and Arg observed in P-PRSD resulted from the different metabolic activity of the starters (Fernández and Zúñiga, 2006). For instance, F. sanfranciscensis strains used Met as a carbon source (De Angelis et al., 2007), Thr, Met, and Arg were used by L. rhamnosus (Sun et al., 2019), and Val was catabolized by several L. casei strains, contributing to the production of branched chain amino acid derivatives (Bancalari et al., 2017), responsible for the acidic, sour, cheesy, and buttery notes.

Protein and starch digestibility parameters were analyzed to further assess the nutritional value of the formulated pasta. While IVPD reflects protein breakdown efficiency, essential amino acid balance and PDCAAS provide a more comprehensive protein quality evaluation (Lee et al., 2016). The addition of RLPI-enriched sourdoughs in pasta increased the IVPD, as fermentation enhances proteins hydrolysis and bioavailability (Sá et al., 2020). In accordance with a previous study, Lys was identified as the first limiting amino acid in semolina-based pasta (Berrazaga et al., 2019). Based on our findings, RLPI-enriched sourdoughs effectively mitigated this deficiency. The observed enhancement in PDCAAS can be attributed to the protein-rich RLPI and its diverse amino acid profile, as well as sourdough metabolic processes, wherein starters facilitate the synthesis of essential and non-essential amino acids through glycolytic intermediates and protease activity (Ghumman et al., 2019) (Shimizu and Matsuoka, 2022) (Mockus et al., 2024). Resistant starch is a type of dietary fiber that resists digestion in the small intestine and reaches the colon. Higher levels of resistant starch are typically advantageous for gut health, glycemic control, and satiety (Guo et al., 2022) (Chen et al., 2024). The addition of dried sourdough into pasta formulations, mainly in P-PRSD, negatively impacted on the resistant starch content. The higher resistant starch content observed in RLPI-enriched sourdough pasta compared to P-PRSD under identical fermentation conditions might be related to the higher protein content present in the RLPI-enriched sourdough pasta, since proteins can act as a physical barrier to starch-digestive enzymes, and thus prevent the starch hydrolysis (Zhang et al., 2022). Resistant starch is generally linked to a lower HI (Rizzello et al., 2017). Based on our findings, an opposite trend was observed with one exception. Lower HI is advantageous for glycemic control and metabolic health (Chung et al., 2008). In sourdough pastas, the reduced HI might be ascribed to the microbiological acidification, which induced protein-starch interactions and hindered enzymatic access to starch (Fois et al., 2021), though this does not fully explain the low HI in P-PR. As the predicted glycemic index (pGI) is directly influenced by the hydrolysis index (Koseoglu and Celikel, 2022), similar patterns were noted across pasta samples. The combined effect of high RLPI enrichment, ingredients texture, starch type and degree of gelatinization, and physical entrapment of starch molecules within pastas might be responsible for the lowest pGI in P-PRSD-L40 (Petitot et al., 2010).

Enriching pasta with PR and RLPI-based sourdoughs not only improved its nutritional profile but also provided a strategic approach to formulate functional pasta products with potent antioxidant and bioactive properties. In pasta where fermentation was employed, microbial starters played a key role in metabolizing phenolic acids, p-coumaric, protocatechuic, and ferulic acids using phenolic acid decarboxylase and reductase (Ricci et al., 2019) (Yang et al., 2023). This reduction might support the release of phenolic acid derivatives and volatile aromatic compounds (Kumar and Goel, 2019), which generally possess higher antioxidant activity and distinct sensory attributes (Pinto et al., 2021). Catechin, salicylic acid, naringenin and kaempferol, phenolics commonly associated with lentils (Zou et al., 2011) (Xu et al., 2007), were uniquely identified in RLPI-enriched sourdough pasta. These compounds are known to support cardiovascular health, reduce oxidative stress and inflammation and regulate metabolism (Čižmárová et al., 2023).

Lentils are unfortunately well known for their antinutritional factors, such as phytic acid which binds to essential minerals in the digestive tract reducing their absorption and thus affecting their bioavailability, that pose significant challenges to their dietary exploitation (Hall et al., 2017). This reflected in higher residues of phytic acid, raffinose and condensed tannins in pasta fortified with RLPI. The absence of a control pasta enriched with RLPI but without starters did not allow to define the effect of microbial fermentation on these antinutritional factors in RLPI-enriched pasta. However, the literature supports that sourdough fermentation likely reduces these compounds compared to initial flour levels (Gobbetti et al., 2020). Based on our findings, this was evident in P-PRSD, where starters activity reduced phytic acid and raffinose compared to P-PR through the action of phytase and α-galactosidase (Gabriele et al., 2019) (Curiel et al., 2015), respectively, thereby enhancing nutrient bioavailability (Perri et al., 2021).

Sensory and cooking time are the most important attributes for the pasta consumers. Modifications in sensory features can enhance the consumer eating experience, making the pasta more distinctive and appealing. The addition of either PR (with and without fermentation) or RLPI (mixed with PR and fermented), each with distinct profiles of proteins, starch, sugars, organic acids, amino acids, and phenolic compounds, markedly altered the organoleptic and texture properties of pasta samples compared to the control (Bustos et al., 2015) (Raina et al., 2005). RLPI-enriched sourdough pastas showed the best sensory profile, with some attributes comparable to the control and others displaying more desirable and distinctive attributed. For instance, the yellow color of the cooked pasta, a critical indicator of pasta quality and consumer preference (Marinelli et al., 2015) (Debbouz et al., 1995), was more enhanced in RLPI enriched sourdough pastas. Our findings aligned with other studies where the positive impact of lentils enrichment had on the nutritional value and sensory quality of spaghetti was demonstrated (Stefano et al., 2020).

5. Conclusions

Our investigation highlighted the synergetic interactions between sourdough fermentation, PR, and RLPI in developing new, functional, and sustainable pasta formulations. First, the selected microbial consortium showed robust efficiency in producing mature sourdough initially from PR and subsequently fortified with RLPI. Incorporating RLPI-enriched sourdoughs led to the highest improvements in the nutritional profiles of pasta, including increased protein and peptide content, reduced starch levels, a rich amino acid spectrum, and optimal protein and starch digestibility parameters. Despite the presence of anti-nutritional factors, the same sourdoughs enriched pasta with bioactive compounds such as catechin, naringenin, salicylic acid, and kaempferol, offering high antioxidant capacity. These changes, together with favorable sensory attributes and acceptable textural aspects, marked a substantial progression in sustainable food production, waste management, and nutritional quality enhancement.

CRediT authorship contribution statement

Alessandro Stringari: Conceptualization, Methodology, Investigation, Data curation, Formal analysis, Writing – original draft. Ali Zein Alabiden Tlais: Conceptualization, Methodology, Investigation, Data curation, Formal analysis, Writing – review & editing. Andrea Polo: Methodology, Investigation, Writing – review & editing. Hana Ameur: Methodology, Investigation. Tiziana De Micheli: Conceptualization, Validation, Resources, Writing – review & editing. Nicoletta Aquaro: Conceptualization, Validation, Resources, Writing – review & editing. Emanuele Zannini: Conceptualization, Funding acquisition, Writing – review & editing. Marco Gobbetti: Supervision, Project administration. Raffaella Di Cagno: Supervision, Validation, Project administration, Writing – review & editing.

Ethical statement

This work did not involve animal or human studies for experimentation.

Data availability

Majority of the data are displayed in the main manuscript but also in the supplementary material. Further data will be made available on request.

Funding

The work for this publication has been undertaken as part of the SMART PROTEIN project https://smartproteinproject.eu/objectives/. This project has received funding from the European Union's Horizon 2020 research and innovation program under grant agreement No. 862957. The APC was funded by the Open Access Publishing Fund of the Free University of Bozen-Bolzano.

Declaration of competing interest

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

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Data availability

Data will be made available on request.