Evaluation of protein composition and functionality of lupin protein isolates extracted from different blue lupin (Lupinus angustifolius) cultivars

Abstract

Lupin is a promising alternative protein source due to the climate resilience of this crop and its large protein content. Protein isolates (PIs) were obtained after alkaline extraction-isoelectric precipitation from 22 narrow-leaf lupin genotypes from commercial and experimental breeding lines cultivated under controlled environmental conditions. The isolates were screened for their protein content and ratio of the main proteins, namely α-conglutin (legumin-like) and β-conglutin (vicilin-like). The molecular mass of the proteins ranged from 320 to 400 kDa and 180–260 kDa for α- and β-conglutins, respectively. In all genotypes, the ratio of α-conglutin to the total (sum of α-conglutin + β-conglutin) exceeded 53 %. Six varieties were selected for further compositional and functional analysis. Protein isolation reduced alkaloids below 200 mg/kg, especially for isolates with higher protein content. Regardless the differences in the conglutins’ composition between the protein isolates, all protein isolates showed similar protein solubility, evaluated at a range of pH values. The produced protein isolates presented higher solubility than a commercially available lupin PI due to their milder extraction conditions. The most abundant essential amino acids were leucine and lysine, while the limiting one was methionine. The surface activity of the isolates at oil-water interfaces was evaluated using drop tensiometry, showing no differences with composition. The work clearly demonstrates the potential of PIs from lupin as a nutritionally-valuable ingredient to be safely used in foods.

Graphical abstract

Highlights

• •Lupin protein isolates obtained via alkaline extraction-isoelectric precipitation.
• •Lupin genotypes showed a higher ratio of α-conglutin (11 S) over β-conglutin (7 S).
• •Protein molar mass ranged from 320 to 400 kDa and 180–260 kDa for α- and β-conglutins.
• •Protein isolation resulted in reduction of alkaloids content below 200 mg/kg.
• •All protein isolates showed similar protein solubility vs pH and surface activity.

1. Introduction

A robust supply of nutritious and sustainable proteins is becoming an increasingly critical need, both for human and animal consumption. Various legume grains are being explored as a source of proteins for food in the temperate zones. Their nutritional value varies due to differences in their protein content, ratio amongst proteins, and their amino acid profile. Among legumes, lupin (Lupinus L.) stands out, as a non-starch leguminous seed with a high fiber (∼30–45 g/100 g) and high protein (∼25–45 g/100 g) content, and low oil content (∼5–12 g/100 g) (Mazumder et al., 2021; Roman et al., 2023; Sujak et al., 2006; Taberner-Pibernat et al., 2025). However, the use of lupin in foods is still very limited, mainly due to the lack of market demand for the crop, and the potential presence of toxic secondary metabolites, namely quinolizidine alkaloids, in the seed. Lupin is of great interest as it can be cultivated in diverse adverse climates and soil conditions, such as sandy and acidic soils, or low rain conditions, making it a highly attractive crop compared to other protein-rich sources (Lo et al., 2021; Valente et al., 2024).

There are mainly four domesticated species of lupins of agricultural interest, with selected agronomic varieties with low-content in toxic alkaloids; white lupin (Lupinus albus), blue or narrow leaf lupin (Lupinus angustifolius), yellow lupin (Lupinus luteus), and Andean lupin (Lupinus mutabilis). These lupin species are distributed in a wide range of geographical regions, with L. angustifolius being the most cultivated species, grown in Australia, Germany and Eastern European region (FAOSTAT, 2023).

The four main proteins found in lupin are called conglutins, and include three types of globulins, α-conglutin (legumin-like, 330–430 kDa), β-conglutin (vicilin-like, 143–260 kDa) and γ-conglutin (200 kDa) (Duranti et al., 2008; Foley et al., 2011). In addition to the storage proteins, an albumin fraction, δ-conglutin (∼13 kDa) is also present. This is a heterodimer composed of two subunits linked by a disulfide bridge, soluble at low pH (Foley et al., 2011). There are many reports on varying protein content and composition among lupin species and cultivars, with white lupins being richer in β-conglutin and blue lupins being richer in either α- or β-conglutins (Devkota et al., 2023; Duranti et al., 1981; Mazumder et al., 2021; Foley et al., 2011). The major storage globulins, α- and β-, and δ-conglutin can be precipitated at acidic pH, while the minor γ-conglutin presents a basic isoelectric point (Duranti et al., 2008; Shrestha et al., 2021). Therefore, taking advantage of the acidic isoelectric point of the major conglutins, protein isolates can be produced using the commonly employed alkaline or salt solubilization and subsequent isoelectric precipitation, with limited presence of albumin fractions (Foley et al., 2011; Muranyi et al., 2016). Furthermore, producing lupin protein isolates using wet extraction methods could also limit the presence of toxic alkaloids and other antinutritional components (Resta, Boschin, D'Agostina and Arnoldi, 2008), similarly to the soaking strategies used for debittering lupin seeds or flours (Jimenez-Martinez et al., 2010).

Studies have so far focused on the effect of different processing on the protein composition and functionality of the protein isolates (Devkota et al., 2023; Fontanari et al., 2012; Muranyi et al., 2016; Sirtori et al., 2010; Wong et al., 2013). Recent studies on other legumin protein isolates have also suggested the importance of protein composition on functionality (Husband et al., 2024; Zhang et al., 2025). In this sense, the hexamers to trimer ratios in soybeans and peas have been shown to be an important factor linked to the isolates’ functionality, for example, stiffer gels or better emulsifiers (Boye et al., 2010; Gravel and Doyen, 2023; Husband et al., 2024).

To the best of our knowledge, no studies have focused on evaluating how varietal differences in narrow-leaf lupins can affect the composition and functionality of the protein isolates, specifically in relation to their protein composition. Only a few studies have focused on evaluating the genotypical effect on the compositional traits (including antinutritional and proteomic analysis) and functionality of lupin seeds flours (Mazumder et al., 2021; Roman et al., 2023; Spina et al., 2022), without further evaluating the isoelectric protein fractions. The predominance of different conglutin fractions should result in differences in the functionality, and, hence, in future applications of lupin protein isolates.

This research aims at evaluating whether varietal differences of blue lupin crops grown under the same environmental and soil conditions can affect the properties of their extracted protein isolates, including protein content, amino acids profile and protein composition, allowing the identification of lupin varieties richer in 11S/7 S (α- to β-conglutin) fraction. To do so, 22 lupin seeds from commercial varieties and pre-breeding lines were grown under the same controlled environmental and soil conditions and the resulting grains were used for protein extraction using an alkaline extraction-isoelectric precipitation procedure; the most widely adopted extraction procedure in the food industry to obtain protein isolates. The obtained lupin protein isolates were evaluated for their protein content, composition, thermal transitions, reduction in toxic quinolizidine alkaloids, interfacial properties, and compared to the initial lupin flours. We hypothesized that, even if grown under similar conditions, the distinct lupin genotypes will result in protein isolates with different protein concentration and composition, with consequences to their nutritional and techno-functional properties.

2. Materials and methods

2.1. Chemicals

Sodium dihydrogen phosphate (NaH₂PO₄), disodium hydrogen phosphate (Na₂HPO₄), sodium chloride (NaCl), sodium azide (NaN₃), dithioerythritol (DTE), 2-mercaptoethanol, and bovine serum albumin (BSA, A2153, lyophilized powder, ≥96 % purity) were purchased from Sigma Aldrich (Steinheim, Germany). The ultrapure water (18.2 MΩ) was obtained from a Milli-Q system (Millipore, Bedford, MA). LC-MS grade acetonitrile, methanol (≥99.8 %), formic acid and ammonium formate were bought from Sigma Aldrich (Steinheim, Germany). All other chemicals were at least of analytical grade.

2.2. Lupin varieties

Lupin varieties were obtained from Nordic Seed (Odder, Denmark). The 22 varieties of blue lupins were grown in the greenhouse as previously reported (Roman et al., 2023). In brief, plants were grown under same conditions of light, temperature, soil and irrigation. Light was natural light and LED (250 μmol m⁻² s⁻¹, FL300 GROW LED lighting, Senmatic, Søndersø, Denmark) for 14 h total. Each protein cultivar was grown (15 plants) in plastic pots (Teku MXC11, Poppelmann, Lohne, Germany) in peat mixture (Pindstrup no. 2) and watered using an ebb/flood irrigation system (AMI Penta irrigation and fertilizer mixer, Senmatic, Søndersø, Denmark). Nutrient composition of lupin seeds was as previously published (Roman et al., 2023).

2.3. Preparation of lupin flours and protein isolates from blue lupin seeds

Dry seeds were harvested from the pods, milled into flours using an ultracentrifugal mill. The composition of the flour is reported elsewhere (Roman et al., 2023) (see also Table S1). From the whole lupin flours, protein isolates were obtained following an alkaline extraction-isoelectric precipitation (Roman et al., 2023). Protein content in the flours and isolates was based on crude nitrogen measured following AACC method 46–30.01 (AACC, 2015) on an automated Dumatherm Nitrogen analyzer (Gerhardt GmbH & Co. KG, Königswinter, Germany), with non-distinction between protein and non-protein nitrogen, using a 6.25 conversion factor from nitrogen to protein. For comparison purposes, a commercial lupin protein isolate (number 10600 obtained from ProLupin, Grimmen, Germany) was used. Commercial protein isolate (C-LPI, 92.3 ± 0.6 % protein content; N factor 6.25) was the protein fraction obtained after aqueous extraction and spray-drying of blue lupins seeds (L. angustifolius) according to the manufacturer’ specifications (following patent US6335044B1).

2.4. Protein profiling using size exclusion chromatography

The composition, average-weight molecular mass (M_w) and peak ratios for the 22 lupin protein isolates were determined using a high-performance size exclusion chromatography (HPSEC) system equipped with a Diode Array Detector (DAD, Agilent 1260 series, Agilent Technologies, Waldbronn, Germany) operating at 280 and 214 nm, a refractive index detector at 35 °C (RI, Shodex RI-501EX, Munich, Germany) and a multiangle light scattering detector (MALS, Dawn, Wyatt Technology, Santa Barbara, CA, USA). Briefly, protein isolates (6.0 mg protein/mL) were solubilized in 50 mM sodium phosphate buffer (38 mM Na₂HPO₄ and 12 mM NaH₂PO₄) containing 150 mM NaCl and 0.02 % sodium azide (pH 7.2) by stirring in a magnetic stirring plate set at 500 rpm and 25 °C for 2 h for solubilization. Subsequently, samples were vortexed, and filtered through 0.45 μm Whatman syringe filters (PVDF w/GMF). The protein recovery after the filtration step was ∼80 % based on N₂ content determined in Dumas before and after filtration in a selected samples subset. The filtered samples were transferred to glass vials and 30 μL were injected in the HPSEC system equipped with a Security Guard Cartridge (GFC 4000, 4 × 3.0 mm ID) and a Yarra SEC-4000 column (300 × 7.8 mm, 3 μm) (Phenomenex, Værløse, Danmark) connected in series and set at 25 °C. The column has a nominal separation range between 15 kDa and 1500 kDa. The mobile phase used was 0.02 μm filtered 50 mM sodium phosphate buffer (pH 7.2) containing 150 mM NaCl and 0.02 % sodium azide at a flow rate of 0.5 mL/min. Sample preparation and injection was performed at least in triplicate and data were analyzed using ASTRA software (Wyatt Technology Corporation, Goleta, CA, USA) selecting a first-order Zimm plot procedure. Bovine serum albumin (BSA) (2 mg/mL) was used for normalization of the MALS signals, correction of band broadening and inter-detector volume delays. The specific refractive index increment $dn/dc$ was assumed to be 0.185 mL/g for lupin proteins and BSA in aqueous solution and the second viral coefficient $A_2$ was assumed to be negligible.

2.5. Reducing gel electrophoresis of lupin flours and protein isolates

Sodium Dodecyl Sulphate-Polyacrylamide Gel Electrophoresis (SDS-PAGE) was performed under reducing conditions. Briefly, 8 mg of lupin flours or 2 mg of protein isolates were dissolved in 2 mL of 100 mM Tris buffer (pH 7.4) containing 1 % SDS and 1 % DTE for a protein concentration of ∼1 mg/mL. The Eppendorf tubes containing the samples were sonicated for 10 min for solubilization of the protein, followed by centrifugation at 10.000×g for 15 min and 4 °C and the supernatants were collected. Then, the reduced samples were prepared by mixing 13 μL of the soluble supernatants with 5 μL NuPAGE LDS sample buffer (4X) (Invitrogen, Thermo Fisher Scientific, CA, USA) and 2 μL of 1 M DTE solution. Samples were incubated at 95 °C for 5 min in a thermomixer (Thermo Fisher Scientific, CA, USA) and centrifuged at 4 °C and 10,000 g for 5 min. Afterwards, 10 μL of the mixes were loaded into 4–12 % precast Bis-Tris gels (Invitrogen, Thermo Fisher Scientific, CA, USA). 5 μl of pre-stained protein ladders were also loaded in the first and last well. Electrophoretic separation was conducted at 200 V for 45 min using NuPAGE MES-SDS running buffer (20x) (Invitrogen, Thermo Fisher Scientific, CA, USA). Bands were resolved using SimplyBlue staining solution (Invitrogen, Thermo Fisher Scientific, CA, USA) for 2 h, de-stained overnight in ultrapure water and the rinsed gels were scanned using ChemiDoc XRS + gel imaging system (Bio-Rad lab., CA, USA). Samples were prepared and run in duplicate.

Furthermore, SDS-PAGE gel electrophoresis experiments were also performed on the acidic supernatants obtained after isoelectric precipitation, to evaluate the amount and composition of protein remaining soluble after the isoelectric precipitation. The procedure for running the gels was the same above described, 13 μL of the acidic supernatants (1:1 dilution) were mixed with the LDS sample buffer and 1 M DTE, and 10 μL of the mix was added into the well.

2.6. Differential Scanning Calorimetry (DSC) of lupin flours and isolates

To evaluate potential differences in the thermal transitions of different lupin isolates, 6 selected lupin protein isolates (LPI-1, 2, 4, 7, 9, 11) were also measured using DSC, and compared with their corresponding initial lupin flours. These 6 samples were selected to include 3 commercial and 3 breeding lines based on their different protein purity (i.e., lowest and highest protein content) and variations in their ratio of α-to total α- and β-conglutins (57–80 %) as determined by HPLC (section 2.4). Endothermic peaks corresponding to protein denaturation were analyzed in a DSC Q-2000 (TA instruments, Crawley, UK) equipped with a refrigerated cooling system (RCS 90). Prior to the measurements, calibration for enthalpy and temperature was carried out using indium. Tzero hermetic aluminum DSC pans (TA Instruments) were employed and an empty pan was used as reference. Dry nitrogen at a flow rate of 50 mL/min was used as the purge gas. Lupin flours (6 mg) and protein isolates (3 mg) were loaded into the aluminum pans and ultrapure water (18 or 12 μL, respectively) was added with a micro-syringe. Samples were hermetically sealed and allowed to equilibrate for 24 h at 25 °C before testing in the DSC instrument. Samples were heated from 25 to 110 °C at 5 °C/min to denature the protein. The peak temperature (T_p) and enthalpy of denaturation, ΔH (expressed as J/g of protein, dry basis) corresponding to the endothermal peak of protein denaturation were calculated. Samples were analyzed in triplicates.

To estimate the moisture and ash content of the 6 selected protein isolates thermogravimetric analyses were conducted using a thermogravimetric analyzer (TGA, TGA-2 STARe system, Mettler Toledo) equipped with STARe software. This method was selected due to the low amount of isolates available. Protein samples (6 mg) were loaded in Mettler Toledo aluminum crucibles with punctured lids and heated from 25 °C to 600 °C at a heating rate of 10 °C min⁻¹ under a high-purity nitrogen environment, followed by an isotherm at 600 °C for 45 min under an air environment. Moisture content was determined as the weight difference between the initial weight at 25 °C and the weight at 100 °C (g moisture/100 g sample). Ash was calculated by the weight difference between the remaining weight at the end of the isotherm at 600 °C and the initial sample (g of ash/100 g of sample).

2.7. Amino acid composition of lupin flours and protein isolates

Amino acid compositional analyses were performed on a selected subset of samples (the 6 selected protein isolates and their corresponding flours) using an ultra-high-performance liquid chromatography tandem mass spectrometry (UHPLC-MS/MS), with running conditions, internal standards and validation previously reported in detail by Roman et al. (2023). Samples were previously hydrolyzed for 20 h at 120 °C with 6 N HCl, 1 % v/v 2-mercaptoethanol in 0.1 N HCl, and 3 % w/v phenol in 0.1 N HCl. The ratio of essential to total amino acids was also calculated and compared with the initial lupin flours. Amino acid analyses were performed at least in duplicate.

2.8. Alkaloids determination in lupin flours and protein isolates

Alkaloids were determined as reported in Roman et al. (2023) in the 6 selected protein isolates (LPI-1, 2, 4, 7, 9, 11). Protein isolates (125 mg) were extracted with 5 mL of 80 % (v/v) methanol solution under sonication in a water bath for 60 min. Afterwards, samples were centrifuged at 3200 g for 15 min at 20 °C, the supernatants were collected and filtered (0.22 μm nylon filter) for liquid chromatography analyses. An UHPLC system (Agilent 1290) coupled to a quadrupole time-of-flight (QTOF) mass spectrometer (Agilent 6540 UHD Accurate-Mass, Agilent, Germany) was used under the conditions described in Roman et al. (2023). Standards of (+)-13α-hydroxylupanine, (−)-angustifoline and isoangustifoline, (+)-lupanine hydrochloride, and (+)-sparteine stock solution were prepared in methanol to build the calibration curves and quantify the alkaloids’ content.

2.9. Protein solubility of selected protein isolates

The solubility of 6 lupin protein isolates (varieties 1, 2, 4, 7, 9 and 11) was assessed at different pH and compared with the commercial lupin isolate (C-LPI). For solubility assay, triplicate lupin protein dispersions (5 mg protein/mL) were adjusted at different pH values (2, 3, 4.5, 6, 7, 8 and 9) using 1 M HCl or 1 M NaOH. Protein samples were stored overnight at 4 °C, before centrifugation at 5000g for 15 min. The residual protein concentration in the supernatant was determined by bicinchoninic acid assay (BCA) (Thermo Fisher Scientific Inc., USA) using BSA as a standard. The absorbance was measured at 562 nm. Solubility (%) was calculated as the ratio between the soluble protein content in the supernatant to the total protein in the dispersion.

2.10. Drop tensiometer measurements at oil-water interface of selected protein isolates

To evaluate differences in protein ability to adsorb at interfaces, two of the most differing isolates, LPI-4 and LPI-7, showing a significant different globulins ratio and similar solubilities, were measured using drop tensiometry. The commercial C-LPI was also evaluated, for comparison. Briefly, 50 mL of protein solutions were prepared by mixing 10 mM phosphate buffer (pH 7.0) with the lupin protein powder to a concentration of 0.1 % protein. Solutions were mixed on a magnetic stir plate for 60 min and left to hydrate at 4 °C overnight. The insoluble fraction was removed by centrifugation (15000 g, 30 min) and the supernatant was collected. The protein content of the final supernatant was measured with the Bradford assay, and the remaining supernatant was diluted to 0.01 % protein for the drop tensiometer analysis.

The interfacial tension between stripped sunflower oil and the protein solutions as well as the dilatational viscoelasticity at the equilibrium were determined using a drop tensiometer equipment (Surface Analyzer LSA100, LAUDA Scientific GmbH, Germany) as previously reported (Grasberger et al., 2022). An oil droplet of 30 mm² was immersed in a 40 mL glass cuvette filled with the protein solution using a rising drop needle. The interfacial tension was measured at 25 °C as a function of time from 0 until 10800 s. The tension drop profile was monitored up to maximum time of 10800 s to guarantee the molecules adsorption at the interface and a quasi-equilibrium interfacial tension value (γ), assumed when the interfacial tension did not change by more than 0.1 mN/m in 30 min. After reaching equilibrium, dilatational rheological experiments of the interface were performed by oscillating the drop volume in the linear viscoelastic region at an amplitude of 5 % and a frequency of 0.1 Hz, for three repetitions followed by a rest period of 100 s. This cycle was repeated three times and the viscous and elastic moduli were calculated. Two measurements were made for two independently prepared samples.

2.11. Statistical analysis

Statistical differences between the parameters for the different blue lupin varieties were studied by one-way analysis of variance (ANOVA), with Tukey HSD used to describe means with 95 % confidence intervals, using Statgraphics Centurion XVI software (Statpoint Technologies, Inc., The Plains, VA, USA). Coefficient of variation among duplicate or triplicate measurements was generally below 12 %.

3. Results and discussion

3.1. Composition of the lupin protein isolates

Table 1 summarizes the composition of the protein isolates obtained from the 22 blue lupin varieties grown in the greenhouse. The protein content obtained ranged from 70.4 (LPI-1) to 92.1 (LPI-7) (g/100 g). Most of the varieties had purities higher than 80 g/100 g, with no distinction observed between commercial and breeding lines. A protein conversion factor of 6.25 was used in this study, which may result in an overestimation of the purity of the protein isolates. The difference in the protein levels between the varieties is related to the co-precipitation of other components with the proteins, namely lipids, carbohydrates and phytates, which may associate with the globulins during isoelectric precipitation. This would cause the concentration of protein to vary between isolates from different varieties. The ash content ranged, for most varieties between 5 and 7 % (w/w) values lower than those measured for the commercial LPI fraction (8.6 ± 0.1 g/100 g) (Table 2).

Table 1.

Protein content, and weighted average molecular mass of α-conglutin, β-conglutin and the small molecular weight fraction (monomers) and relative ratio of α-conglutin, measured based on refractive index elution areas of α- and β-conglutin areas, for the lupin protein isolates (LPIs) obtained from the 22 commercial and pre-breeding lines (indicated by NS).

Sample	Variety	Protein content (g/100 g, wb)	Mw α-con (kDa)	Mw β-con (kDa)	Mw monomers (kDa)	α-con/(α-con + β-con) (%)
LPI 1	NS '18–25′	70.4 ± 0.1 a	335 ± 18 a	202 ± 22 ab	38.6 ± 0.7 bc	80 ± 5
LPI 2	NS '14-10′	88.7 ± 0.6 jkl	333 ± 13 a	206 ± 8 ab	31.0 ± 2.7 ab	65 ± 5
LPI 3	Rumba	78.1 ± 0.2 bc	315 ± 10 a	189 ± 7 a	29.4 ± 2.7 ab	65 ± 4
LPI 4	Haags blaue	85.1 ± 0.1 fgh	330 ± 13 a	191 ± 9 a	30.6 ± 6.6 ab	65 ± 2
LPI 5	NS '206-5′	80.2 ± 0.8 de	321 ± 1 a	187 ± 1 a	27.2 ± 0.0 ab	72 ± 4
LPI 6	NS '18-18′	86.0 ± 0.7 ghi	344 ± 20 a	200 ± 14 ab	33.1 ± 3.8 abc	75 ± 1
LPI 7	NS '18-14′	92.1 ± 1.5 m	323 ± 15 a	194 ± 23 a	31.9 ± 6.1 abc	80 ± 2
LPI 8	Primadonna	80.9 ± 0.4 e	339 ± 14 a	188 ± 5 a	28.0 ± 1.7 ab	64 ± 2
LPI 9	SM Orion	76.1 ± 0.1 b	340 ± 15 a	194 ± 8 ab	30.9 ± 3.3 ab	72 ± 1
LPI 10	NS '103′	87.4 ± 0.3 ijk	323 ± 3 a	180 ± 2 a	25.7 ± 1.2 ab	72 ± 1
LPI 11	Furman	87.1 ± 0.1 hij	336 ± 4 a	186± 5 a	27.6 ± 3.5 ab	57 ± 1
LPI 12	Samba	87.4 ± 0.2 ijk	321 ± 1 a	183 ± 3 a	25.5 ± 0.1 ab	66 ± 3
LPI 13	NS '14-13′	84.5 ± 0.3 fg	399 ± 33 b	265 ± 60 b	–	75 ± 1
LPI 14	Mirabor	77.9 ± 0.7 bc	349 ± 8 ab	199 ± 4 ab	32.3 ± 2.6 ab	53 ± 1
LPI 15	Bolero	89.7 ± 0.2 l	362 ± 1 ab	207 ± 4 ab	33.8 ± 3.4 abc	61 ± 2
LPI 16	Probor	78.7 ± 0.6 cd	330 ± 2 a	188 ± 5 a	27.7 ± 2.3 ab	60 ± 2
LPI 17	NS '18–23′	76.9 ± 0.1 bc	338 ± 8 a	205 ± 3 ab	36.7 ± 2.3 b	77 ± 2
LPI 18	Boruta	76.7 ± 0.1 bc	400 ± 26 b	253 ± 33 b	–	72 ± 1
LPI 19	Swing	80.6 ± 0.2 de	324 ± 6 a	184 ±1 a	23.3 ± 2.8 a	63 ± 3
LPI 20	Arabella	89.2 ± 0.6 kl	336 ± 2 a	190 ± 6 ab	30.0 ± 2.8 abc	80 ± 2
LPI 21	NS '301-7′	83.6 ± 0.3 f	329 ±7 a	190 ± 6 ab	28.9 ± 1.1 ab	78 ± 0
LPI 22	Szot	80.8 ± 0.2 e	326 ± 7 a	193 ± 1 ab	41.7 ± 6.8 c	62 ± 2

Average values ± standard deviation with the same letter for each column parameter are not significantly different (p < 0.05). “-” not detected signal in the MALS detector for accurate molar mass calculation.

Table 2.

Ash content and denaturation parameters of selected lupin flours (FL) and their protein isolates (LPI) compared to commercial lupin protein isolate (C-LPI).

	Ash (g/100 g, as is)	Tp1 (°C)	ΔH1 (J/g protein)	Tp2 (°C)	ΔH2 (J/g protein)	ΔH1+2 (J/g protein)	α-con/(α-con + β-con)
FL 1	5.1 ± 0.1 ab	88.4 ± 0.1 a	0.95 ± 0.13 a	105.3 ± 0.2 ab	3.40 ± 0.21 a	4.34 ± 0.08 ab	0.78 ± 0.03 cd
FL 2	4.9 ± 0.1 a	89.8 ± 0.1 b	1.44 ± 0.04 bc	104.7 ± 0.1 a	2.78 ± 0.33 a	4.22 ± 0.29 a	0.66 ± 0.03 b
FL 4	5.8 ± 0.1 ab	89.5 ± 0.2 b	1.71 ± 0.05 c	105.4 ± 0.3 b	3.37 ± 0.09 a	5.08 ± 0.14 bcd	0.66 ± 0.00 b
FL 7	5.8 ± 0.1 ab	88.6 ± 0.3 a	1.06 ± 0.12 a	105.7 ± 0.2 b	4.39 ± 0.22 b	5.45 ± 0.41 d	0.81 ± 0.01 d
FL 9	4.6 ± 0.8 a	89.3 ± 0.0 b	1.20 ± 0.01 ab	105.6 ± 0.0 b	3.25 ± 0.00 a	4.45 ± 0.01 abc	0.73 ± 0.00 bc
FL 11	6.3 ± 0.3 b	89.7 ± 0.1 b	2.33 ± 0.08 d	105.3 ± 0.1 ab	2.85 ± 0.03 a	5.19 ± 0.11 cd	0.55 ± 0.01 a
LPI 1	4.6 ± 0.4 a	88.3 ± 0.1 bc	0.83 ± 0.03 abc	103.7 ± 0.1 d	5.60 ± 0.11 c	9.86 ± 0.77 b	0.87 ± 0.00 bc
LPI 2	11.5 ± 1.1 d	88.5 ± 0.2 c	0.53 ± 0.10 ab	99.9 ± 0.1a	3.51 ± 0.32 ab	11.58 ± 0.68 b	0.87 ± 0.03 b
LPI 4	6.9 ± 0.5 b	88.0 ± 0.2 b	0.88 ± 0.24 bc	101.3 ± 0.6 b	4.29 ± 0.29 b	11.19 ± 1.38 b	0.83 ± 0.04 ab
LPI 7	6.5 ± 0.2 b	88.0 ± 0.1 b	1.01 ± 0.06 cd	101.8 ± 0.2 bc	5.90 ± 0.44 cd	11.62 ± 0.38 b	0.85 ± 0.02 b
LPI 9	5.8 ± 0.3 b	87.5 ± 0.1 a	0.45 ± 0.01 a	102.2 ± 0.2 c	6.55 ± 0.32 d	11.01 ± 0.88 b	0.94 ± 0.00 c
LPI 11	6.6 ± 0.2 b	87.8 ± 0.0 ab	1.30 ± 0.01 d	101.5 ± 0.1 bc	4.06 ± 0.09 b	11.32 ± 0.25 b	0.76 ± 0.00 a
C-LPI	8.6 ± 0.1 c	nd	nd	102.0 ± 0.1 bc	2.71 ± 0.05 a	2.71 ± 0.05 a	–

Average values ± standard deviation for each set of lupin flours and protein isolates with same letter are not significantly different (p < 0.05). T_p, peak temperature of denaturation, ΔH, enthalpy of denaturation. α-con/(α-con + β-con) is calculated as the ratio between the denaturation enthalpy of peak 2 and 1 (ΔH₂/(ΔH₁ + ΔH₂). Nd. Non detected.

The protein composition was analyzed using size exclusion chromatography, coupled with a UV, refractive index and multi-angle light scattering detectors. The protein assemblies elute as a function of time depending on their size, as shown in Fig. 1. The first peak eluting at about 5.5 mL corresponds to the excluded volume of the column. This aggregation peak resulted from the separation of protein aggregates and/or non-protein components. After the aggregation peak, two main peaks were identified (at about 8.4 mL and 9.3 mL), corresponding to α-conglutin and β-conglutin, the major storage proteins of lupin.

Fig. 1.

Representative chromatograms for 6 selected lupin protein isolates (LPI). The molar mass (g/mol) calculated from the light scattering intensity is also shown in the figure as secondary axis. The commercial lupin protein isolate (C-LPI) is also shown for comparison.

The peak eluting at 8.4 mL had a weight average molar mass ranging between 330 and 430 kDa; this peak contained α-conglutin proteins, principally present in their hexameric form. The differences in weight of the α-conglutins is a result of the mixed presence of different monomers. Immediately after α-conglutins, β-conglutin eluted at about 9.3 mL, and showed a molecular mass between ∼180 and 207 kDa, consistent with the trimeric form of the protein. Interestingly, another peak was also detected at 10.5 mL, and measured by chromatography. The protein present in this peak had a molecular mass below 40 kDa, and may reflect the presence of dissociated monomers of α- and β-conglutins.

The molar mass, calculated for these three main peaks, for the different varieties is shown in Table 1, together with the ratio of α-conglutins. Most of the protein supramolecular assemblies showed same size range, apart from LPI-13 and 18, which showed a higher average mass for both α- and β-conglutins. In addition, the relative abundance of α-conglutins was estimated, for each individual protein isolate, based on the ratio of the elution areas of α- and β-conglutins. The range varied from 0.53 for Mirabor to 0.80 for Arabella and breeding lines LPI-1 and LPI-7. It is not known if such a difference in the ratio between the two main storage proteins will affect the techno-functional properties of the isolates. None of the protein isolates contained a high ratio of β-conglutin. In general, crossbreeding lines denoted a significantly higher proportion of α-conglutin compared to the commercial lines (p < 0.05).

To identify possible differences in protein subunit composition, SDS-PAGE was carried out on the isolates and the original flour samples (Fig. 2). The comparison between flour and isolates should also indicate any potential preferential precipitation of a particular protein subunit. The migration patterns in the reducing-SDS-PAGE gels from all lupin flours were almost identical, with no clear differences between the bands (see also Fig. S1 for all 22 varieties). None of the varieties investigated showed a low concentration of one of the globulins, confirming the size exclusion results shown in Fig. 1.

Fig. 2.

SDS-PAGE of protein in flour (F) and corresponding protein isolate (PI) for selected lupin varieties (LPI 1–5). For α- and β-conglutins the specific bands for acid (A) and basic (B) subunits and high, intermediate and low molecular weight monomer chains (HMW, IMW, and LMW) are highlighted. First and last well in each gel depicts reference ladder (L) with molecular weight in kDa.

All lupin flours and their corresponding protein isolates presented subunits from both α- and β-conglutins. The subunits of the α-conglutin monomers consisted of a basic (20–22 kDa) and acidic (42–52 kDa) subunits, linked through disulphide bonds (Duranti et al., 2008), hence separated by the SDS-PAGE under reducing conditions (Fig. 2). The β-conglutin monomers, assembled non-covalently in a trimeric structure were observed at low (∼17–20 kDa), medium (∼25–46 kDa) and high molecular weight (∼53–70 kDa), and assigned based on prior literature reports (Burgos-Díaz et al., 2016; Duranti et al., 2008). The high molecular weight band (∼100–110 kDa) could be attributed to lipoxygenase (Martínez-Villaluenga et al., 2006; Shi et al., 2020), whose activity has been reported to be reduced after protein isolation (Yoshie-Stark et al., 2004). In the lower portion of the electrophoresis gels, around ∼10 and ∼12 kDa, faint bands were observed, residues of albumin fractions. The similar electrophoretic profiles between flours and isolates demonstrated that all varieties contained a high ratio of the two main globulins (α- and β-conglutins), both precipitating at acidic pH.

To better understand potential differences in the varietal composition, which may be of great interest for lupin selection, the residual supernatant fraction after isoelectric precipitation was also analyzed. Fig. 3 shows the composition of the acidic supernatants. As similar volumes were added to the sample wells, higher band intensities will correspond to a higher amount of residual protein in the supernatant. These results clearly showed dramatic differences in the amount and type of residual protein soluble at acidic pH. Of interest were the bands at < 35 kDa, attributed to the large (∼29 kDa) and small (∼17 kDa) subunits α- β- and γ-conglutin. These subunits did not precipitate at low pH, suggesting differences in post translational modifications, for example, glycosylation, resulting in a different isoelectric point. Their presence would decrease the total protein yield in the isolate, or in turn, increase the opportunity to create a protein side stream from the supernatant fraction. In addition, all supernatant contained albumin fractions, with different level of intensity observed below 15 kDa and more specifically, between ∼9 and 13 kDa. It will be important in the future to evaluate if these low molecular weight proteins are more or less desirable nutritionally or functionally, to ensure an optimal exploitation of lupin protein as ingredient in food.

Fig. 3.

SDS-PAGE gel electrophoresis under reducing conditions of the acidic supernatants (S) obtained from the isoelectric precipitation of the 22 lupin genotypes, with equal volume of supernatants loaded. L in each gel depicts reference ladder with molecular weight in kDa.

3.2. Thermal transitions of protein isolates

Selected varieties were also tested to evaluate potential differences in their thermal transitions. Differential scanning calorimetry (DSC) experiments were conducted on the lupin flours and protein isolates on LPI-1, 2, 4, 7, 9 and 11; 3 crossbreeding lines and 3 commercial genotypes based on their differences in the ratio of α-conglutin/total globulins. Both isolates and flours showed two distinct endothermic peaks (Fig. 4), with peaks at around 88 °C for β-conglutin and 101 °C for α-conglutin, in agreement with other literature reports (Devkota et al., 2023; Mazumder et al., 2021; Sirtori et al., 2010). In contrast, the commercial-LPI showed only one endothermic peak (in agreement with the single α-conglutin peak found in native-HPSEC experiments), with a total enthalpy of denaturation (area under the peak) well below that of the experimental, native isolates (Table 2). The presence of two peaks in the gently processed isolates and flours indicates preservation of native protein structures (both α- and β-conglutins). In contrast, the single peak and lower enthalpy in the commercial-LPI suggests that industrial heat processing (e.g., spray-drying) led to partial denaturation, especially of β-conglutin. This could impact functional properties like solubility, emulsification, or gelation.

Fig. 4.

Thermograms of selected lupin protein isolates (LPI) and their corresponding flours (FL). A commercial lupin protein isolate (C-LPI) is also shown for comparison. Heat flow (W/g sample) as a function of temperature. Temperature indicates the peak of thermal denaturation for each denaturation peak.

3.3. Amino acid composition of selected lupin varieties

The essential amino acids threonine, valine, methionine, isoleucine, leucine, phenylalanine, lysine, histidine, in six selected varieties were measured, in both flour and isolates, as reported in Table 3. Tryptophane was not detected (concentration below limit of detection) and, therefore, not quantified due to its degradation during acidic extraction. The most abundant amino acids were leucine and lysine, followed by phenylalanine, regardless of the source, and the limiting amino acid was methionine. In general, for all studied lupin varieties there was a minimal impact of isolation on the amino acids profile, in agreement with previous literature (Burgos-Díaz et al., 2019; Chew et al., 2003); however, there was a slight decrease in lysine and methionine and a slight increase in histidine with isolation, as previously observed (Chew et al., 2003; Rodríguez-Ambriz et al., 2005). The decrease in these amino acids is related to the loss of conglutin δ, an albumin high in lysine and sulfur containing amino acids (Lqari et al., 2002). The increase in histidine should be related to the different composition of the lupin globulin fractions (Duranti et al., 1981).

Table 3.

Essential amino acids (EAA) concentration (g/100 g protein) and ratio (%) of EEA to total amino acids (TAA) of 6 selected lupin varieties, comparing lupin flour (FL) and protein isolate (LPI).

Sample	Valine	Leucine	Isoleucine	Methionine	Phenylalanine	Threonine	Lysine	Histidine	EAA/TAA
FL 1	5.5 ± 0.1	6.1 ± 0.3	2.1 ± 0.0	0.6 ± 0.0 b	5.9 ± 0.3	5.1 ± 0.2	8.2 ± 0.0 b	4.7 ± 1.9	38.2 ± 1.0
LPI 1	4.5 ± 0.7	7.9 ± 0.6	2.8 ± 0.1	0.3 ± 0.1 a	6.0 ± 0.5	3.8 ± 0.6	7.2 ± 0.3 a	5.0 ± 0.5	37.6 ± 0.6
FL 2	4.9 ± 0.2	7.3 ± 2.6	3.0 ± 0.7	0.4 ± 0.1	5.9 ± 0.7	4.1 ± 0.1	8.3 ± 2.0	3.1 ± 0.5	37.0 ± 1.3
LPI 2	4.8 ± 0.1	8.0 ± 0.3	2.8 ± 0.0	0.3 ± 0.1	5.2 ± 0.2	4.0 ± 0.0	6.7 ± 0.2	4.3 ± 0.2	36.0 ± 1.7
FL 4	5.8 ± 0.4	5.8 ± 0.4	2.1 ± 0.1	0.4 ± 0.0 b	5.9 ± 0.4	5.1 ± 0.2 b	8.6 ± 0.6	3.5 ± 0.4	37.1 ± 1.8
LPI 4	5.2 ± 0.2	6.5 ± 0.2	2.6 ± 0.3	0.3 ± 0.0 a	6.6 ± 0.2	4.1 ± 0.1 a	7.3 ± 0.2	4.7 ± 0.3	37.2 ± 0.2
FL 7	6.1 ± 0.2	5.8 ± 0.0	2.3 ± 0.1	0.4 ± 0.0 b	6.4 ± 0.2	5.1 ± 0.3	8.9 ± 0.1 b	2.5 ± 0.1 a	37.5 ± 1.1
LPI 7	5.5 ± 0.1	6.9 ± 0.0	2.3 ± 0.0	0.3 ± 0.0 a	6.5 ± 0.5	3.9 ± 0.2	7.4 ± 0.0 a	5.0 ± 0.0 b	37.8 ± 0.6
FL 9	6.0 ± 0.0 b	6.4 ± 0.1	2.3 ± 0.1	0.6 ± 0.1	6.6 ± 0.5	5.2 ± 0.2 b	9.6 ± 0.1 b	2.1 ± 0.3 a	38.7 ± 1.3
LPI 9	5.4 ± 0.1 a	7.0 ± 0.4	2.4 ± 0.1	0.4 ± 0.0	6.9 ± 0.1	4.0 ± 0.1 a	7.6 ± 0.0 a	4.1 ± 0.2 b	37.7 ± 0.0
FL 11	5.5 ± 0.1	5.9 ± 0.1 a	2.3 ± 0.0	0.5 ± 0.0 b	6.5 ± 0.2	4.9 ± 0.1 b	9.2 ± 0.1 b	2.2 ± 0.2 a	37.1 ± 0.1
LPI 11	5.0 ± 0.2	6.6 ± 0.0 b	2.5 ± 0.1	0.3 ± 0.0 a	6.9 ± 0.1	3.7 ± 0.2 a	7.3 ± 0.0 a	4.6 ± 0.7 b	37.0 ± 1.0

Average values ± standard deviation. For each essential amino acid, different letters indicate significantly statistical differences for each lupin flour as compared to its corresponding protein isolate (p < 0.05). No letters indicate lack of significant differences between each pair of flour and protein samples (p < 0.05). Dataset does not include tryptophan.

3.4. Quinolizidine alkaloids quantification

Total content of quinolizidine alkaloids in selected protein isolates as compared to their lupin flours is reported in Fig. 5A, together with specific content of each of the alkaloids’ species (Fig. 5B). Differences observed in the alkaloids content among the lupin flours should be explained by intrinsic genotypical differences among the varieties as all the seed samples were cultivated under the same environmental and soil conditions. In all cases, isolation of protein led to a reduction in total alkaloids content, as well as in the three species of alkaloids identified, namely lupanine, 13α-hydroxylupanine and angustifoline (Khedr et al., 2024). Thus, the toxic alkaloids were significantly eliminated during the aqueous extraction procedure performed to produce the isolates, likely because of their solubility at acidic pH.

Fig. 5.

A. Total content (mg/kg) of quinolizidine alkaloids of lupin flours (FL) and their obtained protein isolates (PI) from 6 selected lupin varieties. Total quinolizidine alkaloids (A) are presented in black and grey colored bars for FL and PI samples, respectively. B. Specific content of 13α-hydroxylupanine, lupanine and angustifoline alkaloids of lupin flours (FL) and their obtained protein isolates (PI) from 6 selected lupin varieties.

These results are aligned with the debittering procedures involving acidic or alkaline aqueous conditions that have been successfully performed to eliminate quinolizidine alkaloids in other lupin species (Carvajal-Larenas et al., 2016; Jimenez-Martinez et al., 2010; Resta et al., 2008). All the obtained protein isolates presented a total content of quinolizidine alkaloids below 200 mg/kg, and, hence, considered safe for human consumption according to European and Australian regulations (Schrenk et al., 2019). A 52–72 % reduction in total alkaloids content was achieved with protein isolation, with the highest reduction being observed in LPI-7 and the lowest in LPI-1 and LPI-9. Interestingly, a positive correlation was found between the reduction in alkaloids and the protein content of the isolates (R² = 0.997, p-value <0.05), which may suggest that alkaloids are specially bound to the non-protein components in the seed, as the isolation procedure was the same for all lupin flours. Alkaloids are mainly found within the cotyledon tissue of the seed (Schryvers et al., 2023), and could be associated with the cell walls or other cellular structures, being released during the protein extraction procedure, or eliminated with the fiber fraction. In general, the content of the three alkaloids’ species was reduced to a similar extent, with no selective reduction for any of the three alkaloids, suggesting similar solubility between these compounds.

3.5. Solubility and interfacial properties of lupin protein isolates

Solubility of lupin protein isolates as a function of pH was evaluated for selected varieties, and compared with the solubility of a commercial-LPI. All samples presented the typical U-shape solubility (Fig. 6) with the highest solubility at low and high pH. All the laboratory isolates obtained showed a higher solubility compared to the C-LPI. Regarding pH-solubility differences between the PIs from the 6 studied genotypes, no differences were found for protein solubility at acidic pH, except for LPI-2, which presented the lowest solubility at pH 3 (LPI with the highest ash content). Meanwhile, at basic conditions (pH 7–8), LPI-9 generally exhibited the lowest solubility among the genotypes; this isolate contained a high ratio of α-conglutin. Nonetheless, despite differences in conglutin composition and mineral composition and availability may affect protein solubility, other factors (i.e., exposed polar groups, surface charge, non-protein components), not evaluated in this work could also result in the lower solubility of this isolates at certain pH conditions.

Fig. 6.

Protein solubility of different lupin protein isolates (LPI) from 6 selected blue lupin genotypes as a function of pH, as compared to commercial LPI (C-LPI).

For studying the interfacial properties and viscoelasticity of the protein layers, LPI-4 and LPI-7 were selected, based on their high protein purity and different relative concentration of α-conglutin: 80 and 65 % for LPI-7 and LPI-4, respectively. The comparison between these two samples may indicate if differences in relative ratio between the proteins affects their adsorption properties at an oil-water interface. Results are shown in Fig. 7. The dynamic interfacial tension measurements denote protein’ ability to diffuse, adsorb and reorientate at the oil-water interface. The dynamic interfacial tension curve showed no significant differences in the adsorption properties at oil-water interface between LPI-4 and LPI-7 protein isolates, with similar drop in the interfacial tension (γ) during the adsorption time measurement until equilibrium conditions. At pseudo equilibrium, the two protein isolates also denoted a similar γ of 16.1–16.4 mN/m. Conversely, the interfacial tension for the commercial sample was significantly lower (12.8 mN/m). This can be attributed to differences in the colloidal state (Grasberger et al., 2024), although the presence of surface-active impurities may not be underestimated. These interfacial results agree with those reported by Geerts et al. (2017), who observed a faster adsorption time at the oil-water interface for commercial pea isolate compared to a more native pea protein fraction, concluding it was associated with the higher aggregated state of the proteins in the commercial and fully denatured protein isolate. After reaching equilibrium conditions, the dilatational viscoelastic properties were studied to characterize the interfacial films formed upon proteins’ adsorption, Again, no differences were seen in the interfacial moduli of the LPI-4 and LPI-7 layers, suggesting differences in their conglutin ratio did not affect their interfacial behavior. Nonetheless, a significantly higher elastic modulus was found for the laboratory made LPI, regardless of the genotype, compared to the commercial sample (values of 31 mN/m and 22 mN/m, respectively). These results would indicate that once adsorbed, the native protein fractions in LPI-4 and LPI-7 may present higher flexibility to re-arrange at the interface layer compared to the aggregated and highly denature protein in the commercial sample reorientating less effectively upon deformation due to their inflexible nature (Grasberger et al., 2024). Furthermore, larger aggregates are less effective in producing homogeneous and dense surface layers (Yang and Sagis, 2021), which could also explain the reported results for their viscoelasticity. Their ability to form a viscoelastic interface suggests that these proteins could be employed in food emulsion applications.

Fig. 7.

Adsorption properties at interfaces of 0.01 % protein suspensions of lupin protein isolates (LPIs) from varieties 4 and 7 compared to a commercial LPI. Dilatational rheological measurements of the interface were performed at equilibrium by oscillating the drop volume, and elastic and viscous modulus reported within the figure.

4. Conclusions

This study demonstrates that genotypical differences between cultivars grown under the same environmental conditions resulted in lupin protein isolates with different protein content and composition, especially in the proportion of the main α-conglutin (legumin-like) and β-conglutin (vicilin-like) fractions. All studied genotypes, and, especially those obtained from experimental breeding lines, had a higher ratio of α-conglutin to β-conglutin. The highest difference amongst the genotypes was found in the amount of acid soluble proteins. This finding may suggest the importance of this trait in selecting varieties based on protein composition, as it would have consequences on protein processing yields, and, potentially, nutritional properties of lupin ingredients. Furthermore, protein isolation was successful in decreasing alkaloids content below 200 mg/kg, producing safe protein ingredients for human consumption. Albeit differences in conglutins’ composition were identified, these differences resulted in similar protein solubility at different pH conditions, which were superior to a commercially available lupin protein isolate. By comparing selected varieties with different α-conglutin relative amounts, it was clearly shown that there were no differences in the viscoelasticity of the protein layers from protein isolates. Significant differences were found between the industrially processed and laboratory made isolates, suggesting that processing is a much more important effect than protein ratios. This research helps identify highly valuable protein-rich genotypes from blue lupins as potential sustainable alternative sources to common protein commodities such as pea and soy, in line with the current interest in increasing the choices of available plant proteins for the food industry.

CRediT authorship contribution statement

Laura Roman: Conceptualization, Visualization, Funding acquisition, Investigation, Methodology, Writing-original draft, Writing-review & editing. Emmanouil Tsochatzis: Investigation, Methodology, Writing-review & editing; Luis Jimenez-Munoz: Investigation, Methodology, Writing-review & editing; Carl-Otto Ottosen: Conceptualization, Visualization, Funding acquisition, Methodology, Writing-review & editing; Milena Corredig: Conceptualization, Visualization, Supervision, Funding acquisition, Methodology, Writing-review & editing.

Disclaimer

This work does not represent an EFSA (European Food Safety Authority) opinion.

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.

Handling Editor: Dr. Xing Chen

Appendix A. Supplementary data

The following is the Supplementary data to this article: