INTRODUCTION
Lentils (Lens culinaris) are an ancient pulse originating from southwest Asia. In 2021, global production of lentils was 5.8 million tons, with Canada, India, and Australia as the primary producers (FAOSTAT, 2023). Lentils are inexpensive and rich in protein (24%–30% dry basis) (Wang & Daun, 2006); hence many refer to it as “poor man's meat” (Samaranayaka, 2017). In recent years, the push for sustainable protein sources and the diversification of plant protein products on the market has led to increased production of lentil flour- and lentil protein-based food products (Romano et al., 2021). Lentil proteins have been incorporated into a wide range of foods including salad dressing (Z. Ma et al., 2016), mayonnaise (Armaforte et al., 2021), bakery products (Eckert et al., 2018; Jarpa-Parra et al., 2017), and non-dairy yogurt (Boeck et al., 2021) for both functional and nutritional purposes.
The production of protein-rich extracts from pulses is typically accomplished through wet extraction processes of pulse flours conducted under alkaline conditions, otherwise known as alkaline extraction (AE) (Boye, Zare, & Pletch, 2010). Existing literature explores lentil protein extraction with an emphasis on yields as well as the physicochemical and functional properties of the extracted proteins. In these studies, extractions were performed in alkaline media (pH 8.0 to 11.0) and mild temperature conditions (22 to 40°C), with a solids-to-liquid ratio (SLR) in the range of 1:10 to 1:20, and total extraction time of 1–2 h (Jarpa-Parra et al., 2014; Joshi et al., 2011; Lee, Htoon, & Paterson, 2007). In particular, the work of Jarpa-Parra et al. (2014) leveraged response surface methodology to optimize all four of these parameters, resulting in optimized conditions (pH 9.0, 22°C, 1:10 SLR, 1 h) that yielded 14.5 g protein extract/100 g lentil flour following extraction and isolation (isoelectric precipitation at pH 4.2). Lentil protein extracts in previous works possessed desirable functional properties, particularly gelation and foaming capacity (Jarpa-Parra et al., 2014; Joshi et al., 2011). While these studies provided valuable insights into the effects of processing methods on protein properties, exploring the isolation, physicochemical, and functional properties of the individual protein classes (albumins, globulins, prolamins, and glutelins; referred to as Osborne fractions) in lentils can have far-reaching implications for the development of more effective extraction methods and the design of lentil protein-based products with tailored functionalities. Previous research has characterized lentil protein albumins, globulins, and glutelins (Chang et al., 2022; Ghumman et al., 2016; Osemwota et al., 2022), but these fractions have not been directly compared to protein extracts generated using commercially relevant extraction methods. Because protein functionality is heavily dependent on the extraction, fractionation, and quantification methods used (Kinsella, 1976), it is difficult to draw conclusions with respect to the functionality of the Osborne fractions compared to lentil protein extracts reported in different studies. Therefore, our approach was to sequentially fractionate lentil proteins based on solubility and assess the techno-functional properties of the Osborne fractions in parallel with lentil extracts produced using alkaline and enzymatic extraction.
Enzymatic extraction (EE) can improve protein yields by disrupting the cell matrix of flour particles, therefore increasing cellular porosity, and improving the extractability of intracellular components into the aqueous media. Carbohydrase-assisted extraction of lentil proteins has been performed using Viscozyme® L (cellulolytic enzyme mixture) (Miranda et al., 2022) and Dystizim® AG (glucoamylase) (Bildstein et al., 2008). However, protease-assisted extraction has not yet been investigated for lentil proteins with a specific focus on the effects of proteolysis on extraction yields and resulting physicochemical and functional properties. The addition of proteases provides a unique opportunity to not only improve protein extractability through the release of more soluble protein subunits and peptides, but also to concurrently modify protein structure and functionality to increase the value and potential food applications of the final ingredient. Protease-assisted extraction has been studied for a variety of leguminous crops including chickpeas, black beans, and soy, and has demonstrated to be an efficient and scalable strategy to improve extraction yields and protein solubility (de Moura et al., 2008; Machida, Dias, et al., 2022; Machida, Huang, et al., 2022; Yang et al., 2024). Notably, a Flavourzyme®/Celluclast® mixture (protease and carbohydrase) was employed for the extraction of antioxidative compounds from lentils, but this study did not explore the effects of this enzyme mixture on the proteins per se (Neta & de Castro, 2019).
Importantly, enzymatic hydrolysis of lentil proteins has been reported in the literature as a post-extraction strategy, and lentil hydrolysates have been shown to possess improved functional and/or biological properties (e.g., antioxidant, anti-hypertensive, anti-diabetic) (Avramenko et al., 2013; Boye, Roufik, et al., 2010; Garcia-Mora et al., 2014; Rezvankhah et al., 2021; Vogelsang-O'Dwyer et al., 2022; Xu et al., 2021). While there may be similarities between the properties of lentil hydrolysates generated during extraction (EE) and post-extraction, the substrate on which the enzyme is acting is inherently different (native proteins in lentil flour vs. lentil protein isolates that have undergone extraction and recovery). Specifically, the protein profiles of lentil flour proteins and lentil protein isolates are distinct due to the isoelectric precipitation process utilized in commercial pulse protein production that selectively precipitates globulins, resulting in the loss of water-soluble albumins in the final protein isolate (Karaca et al., 2011; Yang et al., 2022). Therefore, the overall aim of this study was to understand the contribution of Osborne fractions to the protein profile (evaluated by SDS-PAGE and proteomics), physicochemical (surface hydrophobicity, zeta potential), thermal, and functional (solubility, emulsifying, foaming) properties of lentil protein extracts obtained from alkaline and enzymatic extraction processes. The effects of proteolysis on the enzymatically extracted lentil proteins were also highlighted to explore the potential applications of AE and EE lentil proteins as food ingredients.
MATERIALS AND METHODS
Materials and starting material composition
Organic Green Lentil Flour (LF) was provided by The Annex by Ardent Mills (Denver, CO, USA). The proximate composition of the LF was 24.1 ± 0.1% protein, 1.85 ± 0.14% oil, 2.34 ± 0.03% ash, 8.34 ± 0.96% moisture, and 63.4 ± 0.9% carbohydrate (by difference). The lentil flour particle size distribution was as follows: D10 of 8.84 ± 0.54 μm, D50 of 30.0 ± 0.5 μm, D90 of 179 ± 2 μm, and D4.3 of 63.3 ± 1.1 μm. The determination of proximate composition and particle size is described in Section 2.5.
Ammonium bicarbonate (ABC) was purchased from Spectrum Chemical (Gardena, CA). Trichloroacetic acid (TCA), trifluoroacetic acid (TFA), and iodoacetamide (IAA) were purchased from MilliporeSigma (St. Louis, MO). Trypsin was purchased from Promega (Madison, WI). C18 solid-phase extraction (SPE) microplate was purchased from Glygen (Columbia, MD). Acetone, dithiothreitol (DTT), acetonitrile (LC–MS grade), and formic acid (FA) (LC–MS grade) were purchased from Fisher Scientific (Waltham, MA).
Alkaline extraction (
Lentil proteins were extracted using two extraction processes of commercial interest: alkaline extraction and enzymatic extraction. For the AE, LF was extracted in water at a 1:10 solids-to-liquid ratio (SLR) under alkaline conditions (pH 9.0) at 50°C for 1 h. For the EE, identical extraction conditions were utilized with the addition of a commercial protease (FoodPro® Alkaline Protease, Danisco, Rochester, NY, USA) to the extraction slurry (0.5% w/w; g enzyme/g LF). Following extraction, the slurry was centrifuged (3578g, 30 min, 4°C; Allegra X-14 R, Beckman Coulter, Brea, CA, USA) to separate the protein extract (supernatant) from the insoluble fraction. Due to the low lipid content, no emulsified cream phase was observed. Total protein extractability (TPE) was calculated using Equation 1 based on the amount of protein that was not extracted (present in the final insoluble fraction). Protein content was determined as described in Section 2.5.
Sequential protein fractionation
Osborne fractionation was used to separate protein fractions based on solubility in different extraction media (Osborne, 1924). LF was sequentially extracted with DI water, 1 M NaCl solution, 70% ethanol, and 0.05 M NaOH at 25°C for 1 h in each solvent (1:10 SLR) with constant stirring (120 rpm). Protein fractions at each stage in the sequence were decanted following centrifugation (3578g, 10 min, 4°C), and the precipitate was resuspended in the subsequent solvent. The sequential extraction process generated the Osborne fractions: albumin-rich (ALB), globulin-rich (GLO), prolamin-rich (PRO), and glutelin-rich (GLU). The inclusion of the suffix “-rich” acknowledges potential co-extraction of various protein classes in the fractions as no further purification was performed following extraction and decanting. The sequential extraction process was performed in triplicate to yield three independent processing replicates. Protein quantification of the extracts and insoluble fractions (unextracted material) was performed by the Dumas combustion method (Section 2.5) and used to calculate TPE (Equation 1). A protein mass balance was performed to calculate the distribution of protein in each Osborne fraction (Equation 2):
Protein fractions were directly freeze-dried (FreeZone, Labconco, Kansas City, MO, USA) for further analyses except for the PRO fraction, in which the ethanol was evaporated (SpeedVac Thermo Fisher Scientific, Waltham, MA, USA) prior to reconstitution with DI water and subsequent freeze-drying.
Protein molecular weight profile
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed under reducing conditions to visualize the molecular weight profile of the Osborne fractions and AE/EE protein extracts. A pre-cast 4%–20% Criterion™ TGX Precast Midi Protein Gel was used (Bio-Rad, Hercules, CA, USA). Gel images were captured using a Bio-Rad Gel Doc™ EZ Imager (Bio-Rad, Hercules, CA, USA).
Proximate composition and particle size
Protein content of each fraction was determined using the Dumas combustion method (Vario MAX Cube, Elementar Analysensysteme GmbH, Langenselbold, Germany) with a nitrogen conversion factor of 6.25 as previously reported (Jarpa-Parra et al., 2014; Miranda et al., 2022). Oil, ash, and moisture contents were determined according to AOAC methods 989.05, 923.03, and 925.09, respectively. The particle size of the LF was characterized using a Mastersizer 3000E (Malvern Instruments Ltd., Worcestershire, U.K.).
Proteomic analysis
Protein precipitation
For proteomics analysis, extraction replicates (n = 3) were pooled. The PRO fraction, consisting of molecules soluble in high organic conditions, was omitted from the following protein precipitation steps. Protein precipitation was carried out using ice-cold TCA/acetone (10/90) with 20 mM DTT. After incubation at −20°C for 1.5 h, samples were centrifuged (2470 g, 4°C, 30 min) to obtain a protein pellet, which was washed with ice-cold acetone, and dried.
Reduction, alkylation, and digestion
Protein pellets were dissolved in 50 mM ABC. Using a Qubit 3 Fluorometer (Thermo Fisher Scientific, Waltham, MA) to quantify protein, 50 μg of protein was used for enzymatic digestion. Samples were diluted to 1 μg/μL protein with additional 50 mM ABC. For the PRO fraction, ABC was added to an equivalent protein concentration. DTT at a concentration of 5 mM was added, and the samples were then subjected to heating at 60°C for 30 min. IAA at a concentration of 20 mM was added, and samples were incubated in darkness at room temperature for 30 min. Trypsin was added in a 1:50 ratio, and samples were subsequently incubated overnight at 37°C. Trypsin was then inactivated using 1% TFA to lower the pH to 2–3. The remaining insoluble material was removed by centrifugation (14,000g, 4°C, 15 min), leaving the supernatant containing the digested peptides.
Samples were purified by SPE using C18 in a microplate. Following activation and equilibration according to manufacturer recommendations, samples were loaded then washed with a solution of 1% acetonitrile and 0.1% TFA in water. Peptides were eluted with a solution of 80% acetonitrile and 0.1% TFA in water, dried by centrifugal evaporator (MiVac Quattro, Genevac Ltd., Ipswitch, Suffolk, UK), and reconstituted in a solution of 3% acetonitrile and 0.1% FA in water for LC–MS/MS analysis.
LC–MS/MS analysis was performed on an Agilent 6520 Accurate-Mass Q-TOF LC–MS with a Chip Cube interface (Agilent Technologies, Santa Clara, CA). Samples were injected onto a Zorbax 300SB-C18, 5 μm, 150 mm × 75 μm Chip with 0.1% FA in water at a flow rate of 4 μL min−1. Chromatographic separation was performed with a gradient consisting of 0.1% FA in water (A) and 90% acetonitrile/0.1% FA in water (B) at a flow rate of 0.3 μL min−1. The 70-min gradient was ramped from 0% to 30% B, 5 to 60 min; 30% to 100% B, 60 to 65 min; 100% B, 65 to 70 min. The capillary voltage was set to 1950 V. Analysis was done in positive ionization mode with scan ranges of m/z 275–1700 (MS) and m/z 50–1750 (MS/MS). All other ionization settings were as described by Huang et al. (2022).
Data analysis for protein identification
To sequence peptides and identify proteins, PEAKS Studio X+ (Bioinformatics Solutions Inc., Waterloo, ON, Canada) was used. Proteins were identified through spectral matching against NCBI databases. A cross-species search was performed to include other pulse species due to a lack of genome sequencing, which limited the existing NCBI database entries specific to lentils. As with any cross-species proteomics study, database searching of cross species introduces a higher degree of error (Wright et al., 2010). To minimize this, the identified lentil proteins with the highest sequence coverage were searched using NCBI BLASTP () to identify homologous protein sequences. The cross species presented in this work (pea, chickpea, and fava bean) showed protein homology to lentils and thus were selected for database searching.
The species Lens culinaris (lentil), Pisum sativum (pea), Cicer arietinum (chickpea), and Vicia faba (fava bean) were searched (, accessed Nov 2, 2022). The mass error tolerance for the precursor and fragment ions was limited to 20 ppm and 0.035 Da, respectively. Up to two missed cleavages per peptide and five variable modifications, including oxidation, phosphorylation, and deamidation, were allowed. Carbamidomethylation was set as a fixed post-translational modification. A false discovery rate of 1.0% was applied to filter the results, and proteins with at least one unique peptide were retained. Protein matches were selected and manually inspected to ensure accuracy.
Protein secondary structure
Circular dichroism (CD) spectroscopy was performed to estimate the secondary structure of the protein extracts (J-715 CD spectropolarimeter, JASCO Corp., Tokyo, Japan). Freeze-dried samples were dissolved in 0.005 M NaOH (0.5% w/v). Far-UV CD spectra were scanned in a 0.1 cm quartz CD cuvette (Hellma, Muellheim, Baden, Germany) at 25°C with a scanning speed, response time, and bandwidth of 10 nm/min, 8 s, and 1.0 nm, respectively (n = 3). The secondary structure content was quantified using the BeStSel software (Micsonai et al., 2015).
Thermal denaturation behavior
Differential scanning calorimetry (DSC-250, TA Instruments, New Castle, DE, USA) was performed to determine the thermal properties of the LF and protein extracts. Dry samples (~5 mg) were rehydrated (1 powder:3 water) overnight. A sealed empty aluminum pan was used as a reference. The transition was observed from 40 to 120°C with a heating rate of 10°C/min. Analyses were performed in triplicate.
Physicochemical properties
Surface hydrophobicity
Surface hydrophobicity (H0) was determined using 1-anilino-8-naphthalenesulfonate (ANS) according to Zhang et al. (2013). Briefly, liquid protein fractions and extracts were diluted to several concentrations (0.022–0.22 mg/mL protein) in sodium phosphate buffer (0.01 M, pH 7.0). Diluted protein samples (250 μL) were incubated with 1.25 μL ANS solution (8 mM in 0.01 M sodium phosphate buffer, pH 7.0) for 30 min, followed by measurement of fluorescence intensity (excitation 390 nm, emission 470 nm) (SpectraMax iD5 Multi-Mode Microplate Readers, Molecular Devices, San Jose, CA, USA). H0 was reported as the slope of the fluorescence intensity versus concentration line (n = 6).
Zeta potential
Freeze-dried samples were reconstituted in sodium phosphate buffer (0.01 M, pH 7.0) or sodium acetate buffer (0.01 M, pH 4.5) to 0.2% (w/v) and the zeta potential was measured using a Zetasizer Nano ZS (Malvern Instruments Ltd., UK). Measurements of each extraction replicate were performed in triplicate at 25°C (n = 9).
Functional properties
Protein solubility
Solubility was assessed with the method and calculations of Rickert et al. (2004) with minor modifications. Briefly, a 1% w/w dispersion of the freeze-dried samples was prepared in buffer (sodium phosphate, pH 7.0, 0.01 M, or sodium acetate, pH 4.5, 0.01 M) and stirred for 1 h (RT, 300 rpm). The dispersions were centrifuged (10,000g, 10 min, 20°C; Sorvall RC6+ Centrifuge, Thermo Scientific, Waltham, MA, USA) and the protein content of the supernatant was determined using the Dumas combustion method. Solubility was reported as the percent ratio of protein in the supernatant to the total protein in the dispersion.
Interfacial properties: Emulsification and foaming
For the determination of all interfacial properties, a 1% w/w sample suspension was prepared in sodium phosphate buffer (pH 7.0, 0.01 M) or sodium acetate buffer (pH 4.5, 0.01 M) with pH adjustments as necessary to achieve pH 7.0 and 4.5, respectively.
Emulsifying capacity (EC) was determined with the method of Bian et al. (2003) with a reduced volume. Protein dispersions (5 mL) were added to a 100 mL plastic beaker and were homogenized (Polytron PT 2500, Kinematica AG, Lucerne, Switzerland) while soybean oil (dyed with 4 ppm Sudan Red 7B) was added continuously at a 0.4 g/s flow rate until the point of emulsion breakage. EC was reported as the mass (g) of oil added before emulsion breakage per g of dry sample (Sathe & Salunkhe, 1981).
Foaming capacity (FC) and stability (FS) were determined according to Sathe and Salunkhe (1981) with reduced volumes. Five millliters of the protein dispersions were added to a 50 mL graduated centrifuge tube and homogenized for 1 min at 20,000 rpm (Polytron PT 2500, Kinematica AG, Lucerne, Switzerland). FC (percent increase in total volume) and FS (percent of foam remaining after 60 min) were calculated as described by Yang et al. (2024).
Statistical analyses
Statistical analyses (one and two-way analysis of variance with Tukey's post-hoc multiple comparisons test, Pearson's correlation coefficient) were performed using Statistica® Version 13.3 (TIBCO Software Inc., Palo Alto, CA, USA) with p < 0.05.
RESULTS AND DISCUSSION
Protein distribution and extractability
Protein extraction yields for the alkaline extraction (AE), enzymatic extraction (EE), and Osborne fractionation are presented in Figure 1a. Osborne fractionation achieved 97% total protein extractability (TPE) (i.e., 3% of the total protein remained unextracted in the insoluble fraction). Of the Osborne fractions, a protein mass balance revealed that most of the extracted proteins were in the albumin-rich (ALB, 43%) and globulin-rich (GLO, 37%) fractions, followed by the glutelin-rich (GLU, 14%) and prolamin-rich (PRO, 3%) fractions. With respect to protein content in the freeze-dried fractions (% w/w), the ALB fraction had the highest purity (42.3%), followed by the GLU (40.7%), GLO (23.9%), and PRO fraction (2.8%). It is generally accepted that globulins are the major protein class in lentils (42%–48% of total protein) (Bhatty et al., 1976; Neves & Lourenço, 1995); however, in the present work, direct extraction with water may have solubilized some salt-soluble proteins (globulins) due to the natural ions present in the lentil flour that can effectively form a weak salt solution in the extraction slurry (Bhatty et al., 1976). Therefore, the co-extraction of albumins and globulins in the ALB fraction likely explains the high protein extractability with just water. A similar protein distribution for Osborne fractions was reported by El-Nahry et al. (1980), in which 42%–50% of lentil proteins were extracted in the water-soluble fraction, and 28%–35% of proteins were extracted in 10% NaCl. The low distribution of protein in the prolamin fraction (3%) aligns with the findings of Bhatty et al. (1976) (3.1% prolamin); the authors remarked that these proteins are likely low-molecular weight nitrogenous compounds. Interestingly, Osemwota et al. (2022) reported that the glutelin fraction was the primary protein class in lentils; however, the fractionation sequence employed was slightly different compared to the present work, as the first extraction media was 2% NaCl and the supernatant was dialyzed against water to obtain the albumin fraction (supernatant) and globulin fraction (precipitate). The saline-insoluble residue was then extracted with 70% ethanol (prolamin fraction), and the ethanol-insoluble residue was extracted with 0.05 M NaOH (glutelin fraction) (Osemwota et al., 2022).
[IMAGE OMITTED. SEE PDF]
The AE and EE processes were highly effective in extracting proteins from lentil flour. A significant increase in TPE was observed when Alkaline Protease was added to the extraction slurry; specifically, TPE increased from 81% (AE, no enzyme) to 87% (EE, 0.5% w/w enzyme). While such increases have been demonstrated for the enzymatic extraction of other pulse proteins such as chickpea (from 63% to 84%) and common bean (from 75% to 81%) using the same commercial enzyme (FoodPro® Alkaline Protease) (Machida, Huang, et al., 2022; Yang et al., 2024), to the best of our knowledge, the enzyme-assisted extraction of lentil proteins has yet to be investigated. Increased protein extractability can be ascribed to molecular weight reduction and increased solubility of the partially hydrolyzed proteins compared to the intact proteins, as well as the overall increase in the porosity of the particles due to the solubilization of the surface proteins, which improves hydration and protein diffusion into the aqueous phase (extraction medium). The protein content of the freeze-dried AE and EE extracts were 53.0% and 56.3%, respectively.
Protein molecular weight profile
To better understand the inherent differences in the protein profiles of sequentially extracted and alkaline/enzymatically-extracted lentil proteins, SDS-PAGE in the presence of a reducing agent (β-mercaptoethanol) was performed (Figure 1c). The samples exhibited a variety of bands with molecular weights (MWs) between 10 and 100 kDa, resembling previously reported SDS-PAGE for lentil protein extracts (Alonso-Miravalles et al., 2019; Shrestha et al., 2023). Overall, the major bands observed in the ALB, GLO, and AE lanes could be attributed to the polypeptide constituents of the major storage proteins in lentils: legumin (~40 and 20 kDa subunits) and vicilin (~50 kDa). Legumin, an 11S globulin, is a hexameric protein formed by subunits with MWs of approximately 60 kDa; each subunit consists of an acidic (~ 40 kDa) and a basic (~ 20 kDa) subunit linked by a disulfide bond (Shewry et al., 1995). Therefore, in reducing conditions, the bands around 37–40 and 20–25 kDa may be the acidic and basic subunits of legumin, respectively (Alonso-Miravalles et al., 2019). Vicilin is a ~150 kDa 7S trimeric globulin with subunits that are not linked by disulfide bonds (Shewry, 1995); notably, vicilins in lentils have been demonstrated to be relatively heterogeneous in MW, which may explain the many bands observed around 50 kDa (Scippa et al., 2010). The ~97 and 70 kDa bands could correspond to lipoxygenase and convicilin, respectively (Chang et al., 2022; Shevkani et al., 2019).
The ALB and GLO fraction protein profiles were similar, with some changes in band intensity (i.e., thicker lipoxygenase band in ALB, thicker legumin bands in GLO), suggesting that the relative abundance of the proteins may vary. Interestingly, the GLO fraction also contained high-MW proteins/aggregates as evidenced by a dark band in the >250 kDa region. The shared bands between the ALB and GLO lanes further support the conclusion that globulins were co-extracted in the initial water-soluble phase as aforementioned. The PRO and GLU fractions exhibited different protein profiles with predominately low MW proteins/peptides (<20 kDa). In the GLU fraction, high-MW proteins/aggregates (>250 kDa) were also observed. The MW distribution of the AE proteins was similar to the ALB and GLO fractions, which was expected since the AE was performed under aqueous conditions. However, the lack of visible distinct subunits in the EE lane shows that the use of enzyme during the extraction extensively hydrolyzed the major storage proteins, resulting in the generation of smaller peptides with MW <15 kDa.
Proteomic analysis
A total of 129 proteins were identified among reviewed and unreviewed source databases (Table S1), of which 37 proteins were UniProtKB/Swiss-Prot reviewed, meaning they were manually annotated and are considered to be higher quality identifications (summarized in Table 1 with full data in Table S2). Most of the identified proteins were from the pea database (27), followed by fava bean (7), and lastly, lentil (3). As expected, the sequence coverage (%) of the protein species in Table 1 demonstrates that the proteins identified using the lentil database (i.e., lectin, Bowman-Birk protease inhibitor, lipid-transfer protein) had the highest coverage. The proteins shown in Table 1 were further categorized based on functional class (e.g., storage, disease/defense, metabolism) (Ialicicco et al., 2012; Scippa et al., 2010; Wang et al., 2016).
TABLE 1 UniProtKB/Swiss-Prot reviewed protein identifications in the Osborne fractions, AE, and EE lentil protein extracts by LC–MS/MS.
| Description | Accession | Database | Avg. mass | −10lgP | Coverage (%) | ALB | GLO | PRO | GLU | AE | EE |
| Storage-related proteins | |||||||||||
| Albumin-2 | sp|P08688.1|ALB2_PEA | Pea | 26,238 | 107.67 | 7 | ✓ | ✓ | ✓ | ✓ | ||
| Convicilin | sp|P13915.1|CVCA_PEA | Pea | 66,990 | 218.78 | 14 | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Ferritin-1 (chloroplastic) | sp|P19975.2|FRI1_PEA | Pea | 28,619 | 93.78 | 14 | ✓ | ✓ | ✓ | ✓ | ||
| Legumin A | sp|P02857.1|LEGA_PEA | Pea | 58,805 | 183.61 | 13 | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Legumin A2 | sp|P15838.1|LEGA2_PEA | Pea | 59,270 | 206.81 | 14 | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Legumin J | sp|P05692.1|LEGJ_PEA | Pea | 56,895 | 235.73 | 26 | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Provicilin | sp|P02855.1|VCLA_PEA | Pea | 31,540 | 200.88 | 21 | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| Vicilin | sp|P13918.2|VCLC_PEA | Pea | 52,231 | 312.25 | 39 | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
| Disease/defense-related proteins | |||||||||||
| 22.7 kDa class IV heat shock protein | sp|P19244.1|HSP41_PEA | Pea | 22,734 | 80.71 | 11 | ✓ | ✓ | ✓ | |||
| Bowman-Birk type proteinase inhibitor | sp|Q8W4Y8.2|IBB_LENCU | Lentil | 12,266 | 221.51 | 35 | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Heat shock 70 kDa protein (mitochondrial) | sp|P37900.1|HSP7M_PEA | Pea | 72,301 | 71.72 | 3 | ✓ | |||||
| Kunitz-type trypsin inhibitor-like 1 protein | sp|Q41015.2|PIP21_PEA | Pea | 23,792 | 74.57 | 17 | ✓ | |||||
| Lectin | sp|P02870.2|LEC_LENCU | Lentil | 30,352 | 273.47 | 29 | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Seed biotin-containing protein SBP65 | sp|Q41060.1|SBP65_PEA | Pea | 59,554 | 170.65 | 11 | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Superoxide dismutase [Cu-Zn] | sp|Q02610.2|SODC_PEA | Pea | 15,323 | 90.11 | 15 | ✓ | ✓ | ✓ | |||
| Superoxide dismutase [Cu-Zn] (chloroplastic) | sp|P11964.1|SODCP_PEA | Pea | 20,626 | 73.48 | 12 | ✓ | |||||
| Metabolism-related proteins | |||||||||||
| Alcohol dehydrogenase 1 | sp|P12886.1|ADH1_PEA | Pea | 41,155 | 150.37 | 14 | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Alpha-1 4 glucan phosphorylase L isozyme (chloroplastic/amyloplastic) | sp|P53536.2|PHSL_VICFA | Fava bean | 113,580 | 320.65 | 18 | ✓ | ✓ | ✓ | ✓ | ||
| Alpha-glucan phosphorylase (H isozyme) | sp|P53537.1|PHSH_VICFA | Fava bean | 95,924 | 90.37 | 3 | ✓ | ✓ | ||||
| Fructose-bisphosphate aldolase (cytoplasmic) | sp|P46257.1|ALF2_PEA | Pea | 38,491 | 128.89 | 12 | ✓ | ✓ | ✓ | |||
| Fructose-bisphosphate aldolase 1 (chloroplastic) | sp|Q01516.1|ALFC1_PEA | Pea | 38,657 | 91.53 | 8 | ✓ | ✓ | ||||
| Glucose-1-phosphate adenylyltransferase small subunit 2 (chloroplastic) | sp|P52417.1|GLGS2_VICFA | Fava bean | 56,060 | 89.61 | 5 | ✓ | |||||
| Glyceraldehyde-3-phosphate dehydrogenase (cytosolic) | sp|P34922.1|G3PC_PEA | Pea | 36,609 | 139.47 | 22 | ✓ | ✓ | ✓ | ✓ | ||
| Lipoxygenase-3 | sp|P09918.1|LOX3_PEA | Pea | 97,629 | 257.33 | 28 | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Non-specific lipid-transfer protein 5 | sp|A0AT31.1|NLTP5_LENCU | Lentil | 11,686 | 191.9 | 34 | ✓ | |||||
| Nucleoside diphosphate kinase 1 | sp|P47922.1|NDK1_PEA | Pea | 16,463 | 125.35 | 23 | ✓ | ✓ | ✓ | ✓ | ✓ | |
| Phosphoglucomutase (cytoplasmic) | sp|Q9SM60.1|PGMC_PEA | Pea | 63,325 | 134.78 | 12 | ✓ | ✓ | ||||
| Probable sucrose-phosphate synthase | sp|Q43876.1|SPSA_VICFA | Fava bean | 118,204 | 107.45 | 2 | ✓ | |||||
| Ribulose bisphosphate carboxylase large chain | sp|P04717.3|RBL_PEA | Pea | 52,763 | 107.38 | 6 | ✓ | |||||
| RuBisCO large subunit-binding protein subunit alpha (chloroplastic) | sp|P08926.2|RUBA_PEA | Pea | 61,979 | 147.32 | 12 | ✓ | ✓ | ||||
| RuBisCO large subunit-binding protein subunit beta (chloroplastic) | sp|P08927.2|RUBB_PEA | Pea | 62,984 | 94.89 | 4 | ✓ | |||||
| Other | |||||||||||
| 14–3-3-like protein A | sp|P42653.1|1433A_VICFA | Fava bean | 29,420 | 96.07 | 10 | ✓ | |||||
| ATP synthase subunit alpha (mitochondrial) | sp|P05493.2|ATPAM_PEA | Pea | 55,045 | 78.16 | 4 | ✓ | ✓ | ||||
| Elongation factor 1-alpha | sp|O24534.1|EF1A_VICFA | Fava bean | 49,244 | 145.02 | 12 | ✓ | ✓ | ✓ | |||
| GTP-binding nuclear protein Ran/TC4 | sp|P38548.1|RAN_VICFA | Fava bean | 25,290 | 136.98 | 16 | ✓ | ✓ | ✓ | |||
| Histone H4 | sp|P62788.2|H4_PEA | Pea | 11,409 | 131.2 | 30 | ✓ | ✓ | ✓ | ✓ | ||
| Polyubiquitin | sp|P69322.2|UBIQP_PEA | Pea | 42,699 | 118.04 | 7 | ✓ | ✓ |
Of the Osborne fractions, the ALB fraction contained the most identified proteins (32), followed by the GLO (21), GLU (17), and PRO (6) fractions. This suggests that many proteins (particularly those in the metabolism-related functional class) were readily extracted in water and were therefore completely extracted in the first stage of sequential fractionation. The relatively smaller number of identified proteins in the PRO fraction may reflect the conditions of proteomic sample preparation. Proteins in the PRO fraction were extracted using high organic conditions, but proteomics sample preparation and protein digestion were performed in aqueous conditions. As such, it is possible that key proteins in the PRO fraction were precipitated over the course of proteomics sample preparation and were not in solution during protein digestion. To overcome this, subsequent studies may utilize alternative proteomics sample preparation techniques such as FASP (filter aided sample preparation) to minimize differences in protein physical properties.
In comparing the proteins identified, major storage proteins such as vicilin and provicilin were detected in all fractions, and legumin and convicilin were found in all fractions except PRO. The other protein species in the extracts were disease/defense-related proteins including lectins and trypsin inhibitors. Some differences were observed between the fractions for the metabolism-related proteins, but because a majority (~70%–80%) of the proteins in lentils are storage proteins (Ialicicco et al., 2012; Joshi et al., 2017), these minor differences may not be significant in the overall protein profile of the fractions.
Interestingly, the Kunitz-type protein inhibitor was not detected in any of the Osborne fractions but was detected in the AE extract. The apparent absence of Kunitz inhibitors in the fractions was somewhat surprising given the higher overall extractability of the sequential fractionation process (97%) compared to the AE (81%). This suggests that perhaps due to the nature of the sequential process, which employed milder conditions in the initial stages (25°C, approximately neutral pH), the Kunitz-type protein inhibitor was not as efficiently extracted, causing it to be distributed among all fractions in small amounts that were too low to detect. Notably, Kunitz-type proteins in lentils are inherently low in concentration as lentil protease inhibitors are primarily the Bowman-Birk type (Belitz & Weder, 1990; Weder & Kahleyß, 1998). However, there were a few confounding factors that made the definitive identification of this protein difficult. While there was a single peptide match for the Kunitz-type protein inhibitor in the ALB fraction, it was filtered out during data analysis due to a low confidence score (−10lgP) (Table S2). Additionally, the inherent uncertainty associated with using a homologous species database (pea, in absence of sufficient database entries specific to lentils) obscures the conclusions that can be made for this protein in lentils. Further protein characterizations specific to lentils are warranted to better understand the extractability of protein inhibitors in lentils.
Fewer proteins were identified in the EE protein extract compared to the AE (17 for EE vs. 25 for AE). This could be an artifact of sample preparation, since the EE sample contained many small proteins/peptides as evidenced by SDS-PAGE, which could have been removed during sample preparation (discarded supernatant of TCA/acetone precipitation) (Yang et al., 2024). This likely resulted in a less efficient proteomics analysis with fewer identifications of the parent proteins of the hydrolysates. Notably, the Kunitz-type protein inhibitor, a known antinutritional factor (ANF), was not detected in the EE. ANFs can reduce the metabolic utilization and/or digestion of plant foods, leading to impaired gastrointestinal functions and metabolic performance (Gemede & Ratta, 2014). These results suggest that enzymatic extraction of lentil proteins could contribute to the overall reduction of trypsin inhibitors in the resulting extracts, but as Bowman-Birk inhibitors were detected in both fractions, EE was not a sufficient strategy to completely reduce ANFs on its own. Proteomics results also revealed that two proteins were only found in EE and not AE (22.7 kDa class IV heat shock protein; elongation factor 1-alpha), which agrees with the higher extraction yields of the EE.
Secondary structure and physicochemical properties
Circular dichroism (CD) was used to assess the effects of extraction media and proteolysis on the protein secondary structure composition. Estimations of the relative percentage of secondary structure elements revealed that generally, unordered structures were the predominant protein secondary structure (45%–49%) in the lentil protein extracts, followed by β-sheets (20%–35%), β-turns (14%–16%), and lastly α-helices (4%–15%) (Figure 2a). Similar secondary structure distributions were reported by Osemwota et al. (2022) for lentil protein albumin, globulin, and glutelin fractions. Among the Osborne fractions, the ALB fraction had the most similar protein secondary structure composition to the AE proteins, which is corroborated by their comparable protein profiles as identified by SDS-PAGE. Proteolysis (EE) significantly reduced the α-helix structures (15.3%–4.2%), increased β-sheet structures (20.3%–33.5%) and did not change the unordered structures (48.7%–48.7%) compared to the AE extract. For lentil proteins, a similar reduction in α-helix structure was reported for Flavourzyme®-treated red lentil proteins compared to the unhydrolyzed sample (22.5%–19.4%) (Lin et al., 2022). Additionally, the general trend that was observed in the present work (decrease in α-helix and increase in β-sheet structures) has been previously reported for hydrolyzed black bean and soy proteins (Chen et al., 2011; Yang et al., 2024; Zheng et al., 2019). Generally, β-sheets are considered to be more stable structures than α-helices (Damodaran, 2017; Shevkani et al., 2019). This suggests that following proteolysis, conformational changes may take place due to potential refolding or aggregation of the hydrolysates (Zheng et al., 2019).
[IMAGE OMITTED. SEE PDF]
Because some degree of surface hydrophobicity and charge are required to exhibit surface active properties (Avramenko et al., 2013), the surface characteristics of the fractions and extracts were also determined. Surface hydrophobicity (H0) provides a comparative measurement of the degree of hydrophobicity of the protein samples, which is dictated by the hydrophobic regions on the surface of the protein molecules (Sze-Tao & Sathe, 2000). Of the Osborne fractions, the GLO fraction had the highest H0, followed by ALB, GLU, and PRO (Figure 2b). This aligns with prior reports that globulins exhibit higher hydrophobicity than albumins (Osemwota et al., 2022; Papalamprou et al., 2009). The AE proteins had higher H0 compared to the Osborne fractions, which could be explained by the fact that alkaline conditions can dissociate vicilin and legumin subunits, resulting in newly exposed hydrophobic regions (Karaca et al., 2011). Interestingly, the GLU fraction that was extracted in the most alkaline conditions (~pH 12) exhibited lower H0; this has been similarly reported for other plant matrices and may be due to the complete denaturation of the proteins or dissociation of hydrophobic bonds due to the strong alkaline media (Akharume et al., 2020; Horax et al., 2010; Yang et al., 2024). Enzymatic extraction significantly reduced H0 from 4.1 × 106 (AE) to 1.6 × 106 (EE), which could be associated with the breakdown of hydrophobic amino acid clusters at the protein surface, or the aggregation of hydrolysates driven by hydrophobic interactions between newly exposed hydrophobic sites (Chen et al., 2011; Jung et al., 2005). This is corroborated by the selectivity of subtilisin-like proteases like the Alkaline Protease used in this study for cleaving peptide bonds with aromatic or hydrophobic amino acids in the P1 site (Gupta et al., 2002; Vogelsang-O'Dwyer et al., 2022). In addition, smaller peptides that are formed may not sufficiently bind to the fluorescent probe (ANS), contributing to the lower measured H0 (Wu et al., 1998). A decrease in H0 for lentil proteins treated with various commercial proteases (Alcalase®, Novozym®, Bromelain) compared to unhydrolyzed lentil protein isolates/concentrates has also been reported by several authors and has been attributed to the reburying of hydrophobic sites following hydrolysis (Avramenko et al., 2013; Vogelsang-O'Dwyer et al., 2022; Xu et al., 2021). This structural reorganization is supported by the observed changes in secondary structure as discussed previously.
Zeta potential provides a measure of the electrostatic potential at the electrical double layer on the surface of proteins in solution and can reflect the degree of stabilization of a system (Gerzhova et al., 2016). The results demonstrate that surface charge is influenced by both extraction media and proteolysis, as well as the pH of measurement (Figure 2c). The increase in measurement pH from 4.5 to 7.0 increased the absolute zeta potential (more negative) of all fractions and extracts. Of the fractions, only the GLO fraction exhibited appreciable surface charge at pH 4.5, which suggests that the isoelectric points of the ALB, PRO, and GLU fractions were around pH 4.5–5 in accordance with previous literature for the isoelectric points of lentil proteins (Chang et al., 2022; Jarpa-Parra et al., 2014). At pH 7.0, the PRO fraction had the highest absolute value for surface charge (−39 mV), followed by the ALB (−23 mV), GLU (−24 mV), and GLO (−18 mV) fractions. Yang et al. (2022) found that in general, the albumin fractions from mung bean, Bambara groundnut, and yellow pea, were less charged (lower absolute zeta potential) compared to globulin fractions. This contrasts with the findings in this study, but since the fractions generated in the present work were not purified in efforts to retain the native structure of the proteins, other charged species including soluble carbohydrates and phenolics may have impacted the measured zeta potential values (Jin et al., 2019).
The zeta potential of the AE proteins was similar to that of the ALB and GLU fractions at pH 7.0, which is expected since the proteins were extracted in aqueous/alkaline conditions. However, proteolysis (EE) increased the net surface charge at both pH values tested, which could be associated with the liberation of charged end groups (C- and N-termini) and/or free amino acids upon hydrolysis (de Souza et al., 2020). Vogelsang-O'Dwyer et al. (2022) also reported a decrease in isoelectric point (~pH 5 for unhydrolyzed vs. ~pH 4 for hydrolyzed) for lentil proteins treated with Alcalase® and Novozym®. Such a shift in isoelectric point could explain the increased surface charge of EE proteins at both pH 4.5 and 7.0. In addition, the increased absolute zeta potential values suggest that the hydrolyzed proteins are less likely to aggregate due to electrostatic repulsion of the proteins, therefore improving protein stability in solution (Gerzhova et al., 2016). Specifically, in neutral conditions, the EE extract exhibited a surface charge of −40 mV compared to −25 mV for the AE extract, suggesting very high stability/dispersibility as the surface charge is beyond the ±30 mV threshold for stability in solution (Dukhin & Goetz, 1998). When considering both physicochemical properties together at pH 7.0, the results show that higher H0 is correlated with lower absolute zeta potential (indicating less surface charge) for all samples (R2 = 0.82, p < 0.05), aligning with previous reports for various pulse proteins (Karaca et al., 2011).
Thermal properties
Differential scanning calorimetry (DSC) provides useful information on the thermal transition behavior of proteins during heat-induced denaturation. Knowledge of parameters including the transition onset temperature (To), denaturation temperature (Td), and transition enthalpy (ΔH) can provide evidence of potential partial denaturation due to processing. This information is critical to understand since pasteurization or sterilization may be necessary for commercial production, and protein denaturation may influence heat exchanger fouling and lead to undesirable thermal aggregation. Table 2 shows the thermal properties of the LF (protein transition only), Osborne fractions, and AE/EE protein extracts. Of the Osborne fractions, the Td of the GLO fraction was the highest, followed by ALB. Higher Td for globulins compared to albumins has also been reported for soapnut seeds (Yin et al., 2011), but the opposite trend has been reported for peas and lentils (Kornet et al., 2021; Osemwota et al., 2022). This discrepancy may be explained by the residual salt content of the GLO fraction in this work, which has been demonstrated to stabilize pea proteins against denaturation, therefore increasing Td (Sun & Arntfield, 2010). The PRO and GLU fractions had no discernable thermal transition, likely due to the low protein content of the PRO fraction and possible denaturation of the GLU fraction due to the highly alkaline extraction pH (~pH 12) (Ma & Harwalkar, 1984). Notably, Osemwota et al. (2022) did observe a thermal transition for lentil glutelins, but the transition enthalpy was relatively low compared to the other fractions, which indicates a high degree of denaturation (Arntfield & Murray, 1981). The AE and EE protein extracts exhibited thermal transitions at lower temperatures compared to the Osborne fractions, which could be attributed to partial unfolding of the proteins under alkaline extraction conditions (pH 9.0) (Lee, Htoon, Uthayakumaran, & Paterson, 2007). Likewise, Barbana and Boye (2013) reported lower Td (79.62°C) for green lentil protein concentrates compared to green lentil flour (84.44°C). Proteolysis did not affect the Td of the protein extracts, as similarly reported for black bean AE and EE extracts (Yang et al., 2024). The similarity in thermal transition is interesting in light of the different secondary structure observed between the AE and EE extracts (hydrolyzed sample containing more β-sheet structures), which suggests that the EE extracts may contain more stable protein structures (Damodaran, 2017; Shevkani et al., 2019). However, Td can be affected by a number of factors including secondary, tertiary, and quaternary structure, and therefore cannot be adequately explained using only secondary structure composition (Arntfield & Murray, 1981). Notably, proteolysis (EE) did significantly decrease the transition enthalpy, suggesting partial loss of structural integrity due to hydrolysis (Ahmed et al., 2019).
TABLE 2 Protein thermal transition behavior of lentil flour (LF), Osborne fractions, AE (alkaline extraction), and EE (enzymatic extraction) proteins.
| To (°C) | Td (°C) | ΔH (J/g) | |
| LF | 87.6 ± 0.3b | 93.2 ± 0.4a | 0.35 ± 0.09a |
| ALB | 84.9 ± 0.7b | 89.2 ± 0.6b | 0.17 ± 0.00b |
| GLO | 90.8 ± 1.0a | 94.5 ± 0.9a | 0.03 ± 0.00c |
| PRO | - | - | - |
| GLU | - | - | - |
| AE | 79.4 ± 0.6c | 85.9 ± 0.2c | 0.19 ± 0.01a |
| EE | 79.8 ± 0.9c | 84.9 ± 0.4c | 0.12 ± 0.01b |
Functional properties
Solubility
The effects of extraction media and hydrolysis on protein functionality were also assessed. Protein solubility is a function of hydrophilicity and electrostatic repulsion and is usually the functional property that is most affected by enzymatic hydrolysis (Panyam & Kilara, 1996). Solubility strongly influences other techno-functional properties such as thickening, water retention, foaming, emulsifying, and gelling properties (Wouters et al., 2016). In the present work, lentil protein solubility was significantly affected by extraction media, proteolysis, and measurement pH (Figure 3a). Overall, solubility was higher for all samples at pH 7.0, which is further from the isoelectric point of lentil proteins. The solubility of the Osborne fractions followed the order of ALB > GLO > GLU > PRO at both pH values. The high and low solubilities of the ALB and PRO fractions, respectively, are expected based on the nature of their respective extraction solvents (water and ethanol). The AE proteins exhibited similar solubility to the ALB fraction at both measurement pHs. These results are interesting considering the significantly lower denaturation temperatures (Section 3.4) that suggested partial protein unfolding in the AE proteins due to the alkaline extraction conditions. This could indicate that the extraction medium was sufficiently mild such that any pH-related structural modifications would be reversible or have negligible impacts with respect to protein solubility. Proteolysis significantly improved solubility at pH 4.5 (33%–58%), corresponding well to the significant increase in surface charge that likely enhanced electrostatic repulsion, and therefore, dispersibility in solution. These results agree with the literature in which protease-assisted extraction resulted in the production of almond proteins with higher solubility in acidic conditions (de Souza et al., 2020; Dias & de Moura Bell, 2022). No significant differences were observed at pH 7 (93%–99%) between AE and EE, which could be attributed to the naturally high solubility of lentil proteins in neutral conditions.
[IMAGE OMITTED. SEE PDF]
Interfacial properties
Proteins can act as emulsifiers because of the presence of hydrophobic and hydrophilic regions on the surface of a protein molecule (Damodaran, 2005). The emulsifying capacity (EC) of the Osborne fractions ranged from 259 to 518 g oil/g dry sample at pH 4.5 and 356 to 595 g oil/g dry sample at pH 7.0, with the ALB and PRO fractions exhibiting the highest and lowest EC at both pHs, respectively. These results generally corresponded well with the observed trends for solubility, which is a known prerequisite for emulsification (Damodaran, 2005). The exception was the PRO fraction, as it was able to form emulsions despite insolubility at pH 4.5 and 7.0. This could be explained by the presence of non-protein surface-active components in the PRO fraction (e.g., phospholipids, carbohydrates), which may exhibit emulsion-stabilizing abilities (Sivapratha & Sarkar, 2018), as well as potential effects of the homogenization process that could alter the surface properties of the proteins to favor emulsification (Du et al., 2020; Xu et al., 2018).
The EC of the AE fraction was similar to the GLO fraction, but lower than the ALB fraction at both pH values. Since the AE proteins contain a mixture of albumins and globulins, these results suggest that the higher MW proteins present in the GLO fraction (observed through SDS-PAGE) as well as the higher structural rigidity of globulins compared to flexible albumins (Malomo & Aluko, 2015), may impair emulsion formation. Our findings contrast with previous reports that lentil albumins have poor emulsifying properties due to their low-MW and hydrophilic nature that prevent them from forming cohesive films around oil droplets (Chang et al., 2022).
Proteolysis (EE) increased the EC at both measurement pHs compared to the AE extract, which is likely a function of the higher solubility, higher surface charge, and lower surface hydrophobicity of the EE proteins. High charge and low hydrophobicity contribute to stability in solution (i.e., resistances to aggregation) that allow for rapid migration of proteins to the oil/water interface; however, some degree of hydrophobicity allows for protein-oil interactions that could enhance emulsion stability (Burger & Zhang, 2019; Karaca et al., 2011). Improved emulsifying properties upon protein hydrolysis were also reported by Jung et al. (2005) for soy protein and by Barac et al. (2012) for pea proteins. However, Avramenko et al. (2013) reported that trypsin hydrolysis of lentil proteins reduced the emulsifying activity and stability indices, even with degrees of hydrolysis as low as 4%. These results emphasize that unique functionalities that may arise from the use of different enzymes and hydrolysis conditions.
Foaming properties reflect the ability of proteins to migrate to the gas/water interface and reorient themselves to form a film around gas bubbles that can resist rupture for a certain amount of time (Dickinson, 1999). Foaming plays important textural and structural roles in a variety of food products such as chocolate mousse, milk foams, cakes, bread, and beer (Wouters et al., 2016). The impact of the extraction media and proteolysis on foaming properties at pH 4.5 and 7.0 is shown in Figure 3c,d. The GLU fraction yielded the highest overall foaming capacity (FC), but the foam produced was not stable (2%–3% foaming stability, FS). At pH 4.5, the ALB, GLO, AE, and EE proteins all exhibited similar FC values. However, at pH 7.0, the ALB fraction had significantly higher FC (84% FC) than the GLO fraction (57% FC). This could correspond to the significantly higher surface charge that was observed for the ALB fraction compared to the GLO fraction under neutral conditions; these findings are also supported by similar results of Chang et al. (2022) that lentil albumins had higher FC than globulins. As observed for EC, the PRO fraction was also able to form a foam following homogenization despite minimal solubility; however, the FC was significantly lower than the other protein fractions and exhibited no FS after 60 min. In acidic conditions, there was no significant difference in FC or FS between the AE or EE extracts despite all the physicochemical and structural differences as previously discussed, as well as significant differences in protein solubility. Conversely, in neutral conditions, proteolysis significantly improved FC (65% for AE vs. 90% for EE) but decreased FS (78% for AE vs. 60% for EE), which aligns with the results of Ahmed et al. (2019). These results highlight the interactive effects between pH and degree of hydrolysis in protein functionality. Vogelsang-O'Dwyer et al. (2022) reported no significant differences in FC or FS under neutral conditions for lentil protein concentrates and hydrolysates and discussed the delicate balance between attaining a sufficient degree of hydrolysis to increase molecular flexibility and solubility, while not over-hydrolyzing the proteins and impairing the formation of a stable interfacial layer. Therefore, a major conclusion from analysis of the interfacial properties of AE and EE proteins is that generally, the enzyme type and concentration tested (Alkaline Protease, 0.5% w/w) provided a sufficient level of hydrolysis to improve the emulsifying and foaming properties of the extracts.
CONCLUSION
Protease-assisted extraction (EE) was utilized for the extraction of proteins from lentil flour for the first time. The composition, physicochemical and functional properties of the resulting extract were compared with major lentil protein fractions isolated by Osborne fractionation, revealing albumin-rich (ALB) and globulin-rich (GLO) fractions as the major protein fractions in lentil flour. The addition of protease significantly enhanced protein extractability in the AE process from 81% to 87%. Proteomic analysis revealed similarities between ALB and GLO fractions and the AE extract. However, Alkaline Protease extensively hydrolyzed proteins in EE, resulting in fewer protein identifications. The ALB fraction exhibited superior solubility and interfacial properties at neutral pH, suggesting its potential in commercial applications. These findings suggest that alternative strategies, other than isoelectric precipitation, should be considered to retain the highly functional albumin proteins. Importantly, EE yielded protein fractions with solubility similar to that of the ALB fraction in both acidic and neutral pH, underscoring the efficacy of this extraction approach in generating highly soluble protein fractions. While AE proteins shared similarities with ALB and GLO fractions, they exhibited a lower thermal denaturation temperature. Overall, proteolysis resulted in protein fractions with improved emulsifying and foaming capacities compared with the AE, reaching levels comparable to those observed in the ALB-rich fraction. AE and EE proved effective for single-stage extraction of high-functionality lentil protein extracts. This study unveiled how efficient and scalable AE and EE strategies impact the composition and physicochemical properties of lentil proteins and how these changes subsequently impact their techno-functional properties, a determining aspect in establishing their potential use in industrial food applications.
AUTHOR CONTRIBUTIONS
Fernanda F. G. Dias: Formal analysis, Investigation, Methodology, Writing—review & editing. Jasmin S. Yang: Formal analysis, Investigation, Methodology, Writing—original draft. T. Truc K. Pham: Formal analysis, Investigation, Methodology, Writing—original draft. Daniela Barile: Supervision, Resources, Writing—review & editing. Juliana M. L. N. de Moura Bell: Supervision, Conceptualization, Project administration, Resources, Funding acquisition, Writing—review & editing. All authors contributed to and approved the final draft of the manuscript.
ACKNOWLEDGMENTS
This study was funded by the USDA Agricultural Research Service, Pulse Crop Health Initiative (PCHI) (grant number 58-3060-0-044). The graphical abstract was created with BioRender.com (641D4ECE-0003).
CONFLICT OF INTEREST STATEMENT
The authors declare that they have no conflict of interest.
ETHICS STATEMENT
This article does not contain any studies with human participants performed by any of the authors.
Ahmed J, Mulla M, Al‐Ruwaih N, Arfat YA. Effect of high‐pressure treatment prior to enzymatic hydrolysis on rheological, thermal, and antioxidant properties of lentil protein isolate. Legume Sci. 2019;1: [eLocator: e10]. [DOI: https://dx.doi.org/10.1002/leg3.10]
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
© 2024. This work is published under http://creativecommons.org/licenses/by/4.0/ (the "License"). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Abstract
Lentil proteins are gaining popularity as food ingredients, serving both functional and nutritional purposes. To better understand the properties of lentil proteins extracted using commercially relevant methods (alkaline and enzymatic), sequential fractionation by solubility (Osborne fractionation) was performed and the physicochemical, thermal, and functional properties of the extracts were characterized. Fractionation revealed that 43% of lentil proteins were water‐soluble (ALB, albumin‐rich), 37% salt‐soluble (GLO, globulin‐rich), 14% alkaline‐soluble (GLU, glutelin‐rich), and 3% ethanol‐soluble (PRO, prolamin‐rich). Protein extraction yields of 81% and 87% were achieved by alkaline (pH 9.0, 50 °C, 1:10 solids‐to‐liquid ratio, 60 min) and enzymatic extraction (same conditions with 0.5% (w/w) Alkaline Protease), respectively. Proteomic analysis allowed for the identification of 129 proteins among all extracts, and the ALB and GLO fractions exhibited similar protein profiles as the alkaline‐extracted proteins. The secondary structure of the protein fractions was dominated by β‐sheets (20%–35%) and unordered structures (45%–48%). Surface hydrophobicity and absolute zeta potential were negatively correlated (
You have requested "on-the-fly" machine translation of selected content from our databases. This functionality is provided solely for your convenience and is in no way intended to replace human translation. Show full disclaimer
Neither ProQuest nor its licensors make any representations or warranties with respect to the translations. The translations are automatically generated "AS IS" and "AS AVAILABLE" and are not retained in our systems. PROQUEST AND ITS LICENSORS SPECIFICALLY DISCLAIM ANY AND ALL EXPRESS OR IMPLIED WARRANTIES, INCLUDING WITHOUT LIMITATION, ANY WARRANTIES FOR AVAILABILITY, ACCURACY, TIMELINESS, COMPLETENESS, NON-INFRINGMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Your use of the translations is subject to all use restrictions contained in your Electronic Products License Agreement and by using the translation functionality you agree to forgo any and all claims against ProQuest or its licensors for your use of the translation functionality and any output derived there from. Hide full disclaimer
Details
; Yang, Jasmin S. 1
; Pham, T. Truc K. 1 ; Barile, Daniela 2 ; L. N. de Moura Bell, Juliana M. 3
1 Department of Food Science and Technology, University of California, Davis, California, USA
2 Department of Food Science and Technology, University of California, Davis, California, USA, Foods for Health Institute, University of California, Davis, California, USA
3 Department of Food Science and Technology, University of California, Davis, California, USA, Department of Biological and Agricultural Engineering, University of California, Davis, California, USA