Future Foods 9 (2024) 100323

Contents lists available at ScienceDirect

Future Foods
journal homepage: www.elsevier.com/locate/fufo

Sorghum protein ingredients: Production, compositional variability and

enhancement of aqueous dispersibility through homogenization
Thais Cristina Benatti Gallo a, b, *, Valérie Beaumal a, Bérénice Houinsou-Houssou a, 1,
Michèle Viau a, Lucie Ribourg-Birault a, Hélène Sotin c, Joëlle Bonicel d, Adeline Boire a, Valéria
Aparecida Vieira Queiroz e, Hamza Mameri d, Sylvain Guyot c, Alain Riaublanc a, Vânia
Regina Nicoletti b, Claire Berton-Carabin a, f
a
INRAE, UR1268 Biopolymères Interactions Assemblages (BIA), Interfaces et Systèmes Dispersés (ISD), 44316 Nantes cedex 3, France
b
São Paulo State University (Unesp), Institute of Biosciences, Humanities and Exact Sciences, São José do Rio Preto, Department of Food Engineering and Technology,
15054-000 São José do Rio Preto, São Paulo, Brazil
c
INRAE, UR1268 Biopolymères Interactions Assemblages (BIA), Polyphénols, Réactivité, Procédés (PRP), 35653 Le Rheu, France
d
INRAE, UMR 1208 Ingénierie des Agropolymères et Technologies Emergentes (IATE), F-34060 Montpellier, France
e
Embrapa Milho e Sorgo, 35701-970 Sete Lagoas, Minas Gerais, Brazil
f
Wageningen University & Research, Laboratory of Food Process Engineering, 6700 AA Wageningen, the Netherlands

A R T I C L E I N F O A B S T R A C T

Keywords: Sorghum protein ingredients show promising applications, given that sorghum is a resilient grain with high
Kafirin protein content and bioactive components, such as tannins, in certain genotypes. This study aimed to investigate
Tannins the production and characterization of sorghum protein ingredients extracted from both tannin-free and tannin-
Composition
rich cultivars, along with an examination of their behavior in aqueous suspensions. The use of high-pressure
Sorghum lipids
Particle size
homogenization to improve ingredient dispersibility was also assessed. Protein extracts exhibited protein con­
Solubility tents ranging from 50 to 67 g/100 g (wet basis), with the highest values obtained in tannin-free samples. Total
lipid contents were between 18 and 26 g/100 g (wet basis), with a high contribution of free fatty acids (68–76 g/
100 g lipids), and tocopherol contents ranged from 1080 to 2039 µg/g total lipids. This is substantially higher
than the lipid content in flours, implying an accumulation of lipids during the protein extraction process. While
the dispersibility of the protein extracts in aqueous media was initially limited, high-pressure homogenization
proved effective in reducing the average size of the particles in suspension, from 9–66 µm to 1.8–2.5 µm. This
processing step significantly enhanced protein dispersibility by up to 288 %, especially for the tannin-rich
samples.

1. Introduction 2018). In some countries, such as Japan, the United States and Brazil,
this grain has been studied and recommended as a good alternative for
Sorghum (Sorghum bicolor L.) is a versatile grain commonly culti­ food applications. Sorghum grains are mainly composed of carbohy­
vated in arid regions of Africa, Asia, Australia, and North and South drates (75 g/100 g) and their dietary fiber content is about 6 g/100 g.
America (Girard and Awika, 2018), with a profitable growing owing to Lipids account for approximately 3 g/100 g, with a typical main fatty
its adaptability to various environmental conditions. Sorghum was the acid composition as follows: oleic acid (31.1 - 48.9 %), linoleic acid
sixth most produced cereal in the world in 2020, with a total production (27.6 - 50.7 %) and palmitic acid (11.7 - 20.2 %) (Mehmood et al.,
of almost 59 million tons (FAOSTAT, 2022). In semi-arid zones of Africa, 2008).
South America, and Asia, sorghum is consumed as a staple food (Proietti Sorghum grains serve as an important protein source, with a content
et al., 2015; Queiroz et al., 2011), but its production is mainly intended ranging from 6 to 18 g/100 g (total basis) (de Mesa-Stonestreet et al.,
for industrial conversion into alcohol and for animal feed (Liu et al., 2010). This underscores the potential for using sorghum proteins as a

* Corresponding author at: INRAE, UR1268 Biopolymères Interactions Assemblages (BIA), Interfaces et Systèmes Dispersés (ISD), 44316 Nantes cedex 3, France.
E-mail address: thais.benatti@unesp.br (T.C.B. Gallo).
1
Present address: Centre des Sciences du Goût et de l’Alimentation, CNRS, INRAE, Institut Agro, Université de Bourgogne Franche-Comté, F-21000 Dijon, France.

https://doi.org/10.1016/j.fufo.2024.100323
Received 30 August 2023; Received in revised form 25 February 2024; Accepted 29 February 2024
Available online 9 March 2024
2666-8335/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

sustainable alternative to animal proteins and as a natural ingredient in Sigma-Aldrich (St Louis, USA). Standards of α, β, γ and δ-tocopherol
foods, which is of interest given the ongoing protein transition and were from Calbiochem Item from Merck and the γ-tocotrienol standard
prevailing clean label trend (Vallons et al., 2010). The production of was from Cayman Chemical (Ann Arbor, USA). Sodium acetate and
sorghum protein-rich ingredients, as well as protein ingredients in β-mercaptoethanol were from Merck (Darmstadt, Germany). Ascorbic
general, requires an extraction process (i.e., separation of the proteins acid was from Fisher Scientific (Loughborough, UK), sodium chloride
from the other components present in the seeds or grains, such as starch, was from VWR International (Radnor, USA) and Alexa 488 was from
fibers and antinutritional compounds). A comprehensive characteriza­ Invitrogen, Thermo Fisher Scientific (Waltham, USA). All the chemicals
tion of the final ingredient is crucial for assessing the impact and effi­ were of analytical grade and ultrapure water was used for all the
ciency of the extraction process. Moreover, this characterization can experiments.
provide insights into how non-protein compounds may influence the
functional properties of plant-based protein-rich ingredients (e.g., 2.2. Production of protein extracts
emulsifying, gelling and moussing properties, and interfacial behavior)
(Keuleyan et al., 2023). Protein-rich fractions were extracted from the sorghum flours
Some sorghum cultivars exhibit a significant polyphenol content, following the methodology described by Taylor et al. (2005). Briefly,
including condensed tannins (procyanidins) (Barros et al., 2014). These each flour was dispersed in an aqueous solution containing 70 % ethanol
tannins have the capacity to bind to proteins and form insoluble com­ (v/v), 0.5 % sodium metabisulfite (w/v) and 0.35 % NaOH (w/v), at 70
plexes. While this interaction is often considered a reason to label tan­ ◦
C for 1 hour, under constant stirring, at a ratio of 1:5 (w/w, flour:
nins as antinutritional factors, their potentially adverse effects may be aqueous phase), followed by centrifugation at 1000 × g for 5 min at 25
offset by other positive functionalities, such as their antioxidant po­ ◦
C (Hitachi CR22N, Tokyo, Japan). The supernatant was collected and
tential. Among the proteins present in sorghum grains, the most abun­ placed overnight in a fume cupboard, at room temperature (25 ◦ C) for
dant is kafirin, belonging to the family of prolamins. Kafirin shares solvent evaporation until a viscous liquid sediment was formed. The
similarities with zein in solubility, molar mass, amino acid composition protein slurry was added to cold distilled water (< 10 ◦ C) and the pH was
and polypeptide structure, although it is more hydrophobic (Xiao et al., adjusted to 5.0 with HCl 1 M to precipitate the proteins. Proteins were
2015). In terms of molar mass, kafirin can be classified into α-kafirin (66 recovered by filtration (14 µm) under vacuum, collecting the material on
- 80 %), with molecular weights of 23 and 25 kDa, β-kafirin (5 - 8 %) and the top of the filter and dispersing it in distilled water to neutralize (7.0)
γ-kafirin (9 - 12 %), with molecular weights of 18 and 28 kDa, respec­ the pH with NaOH 1 M before subsequent freeze-drying. Contrarily to
tively (Espinosa-Ramírez and Serna-Saldívar, 2016; Xiao et al., 2017). the reference procedure described by Taylor et al. (2005), the starting
The low water solubility of sorghum proteins can be a problem for flours were not defatted with hexane in order to evaluate the functional
their application in food products; therefore, dedicated strategies should properties of protein-rich extracts in the presence of lipids, which is
be implemented to facilitate their dispersibility in aqueous media, which relevant given the current incentives for minimal fractionation routes.
implies some structural modifications. For instance, high-pressure ho­
mogenization can be used to modify both the size and structural orga­ 2.3. Preparation of protein suspensions
nization of the protein-based supramolecular assemblies (aggregates,
particles) in suspensions. This mechanical treatment induces an intense Suspensions containing 2 % (w/w) of protein extracts were prepared
shear on the protein suspension by forcing it to pass through a small gap, in ultrapure water. The suspensions were stirred overnight at 4 ◦ C to
which can disrupt protein aggregates (Balny and Masson, 1993; Zamora ensure maximum hydration. The suspensions were then treated by high
and Guamis, 2015). Hence, the application of high pressure homogeni­ pressure homogenization (HPH) at 400 bars for 6 min, corresponding to
zation can enhance the techno-functionalities of proteins (Bader et al., 24 passes through the homogenizer (Panda plus 1000, GEA, Italy) in one
2011), such as their emulsifying properties. This underscores the sig­ continuous process. These conditions were preliminary optimized to
nificance of this processing step as a relevant pre-treatment for have a particle size and distribution fairly comparable among the
plant-based proteins (Burger and Zhang, 2019; Keuleyan et al., 2023; different samples (Supplementary Fig. 1).
Levy et al., 2022; Melchior et al., 2022).
This study therefore aimed to characterize the composition of protein 2.4. Characterization of sorghum flours and freeze-dried protein extracts
ingredients prepared from four sorghum cultivars with contrasted
tannin content. Additionally, the impact of high-pressure homogeniza­ 2.4.1. Proximate composition of the powders
tion on the particle size and protein dispersibility in aqueous suspen­ Flours and protein extracts were characterized according to AOAC
sions prepared from these ingredients was investigated. Thus, this work (2006) official methods for protein, water and ash contents.
represents an important step in facilitating the future utilization of this The protein content was evaluated by the Kjeldahl method, and the
relevant crop in food products. nitrogen-to-protein conversion factor (N:P factor) was computed as the
ratio of total anhydrous mass of amino acids to the total mass of nitrogen
2. Materials and methods (Sosulski and Imafidon, 1990). Aspagarine (Asn) and glutamine (Gln)
were assayed in their acidic form, i.e., as asparagic acid (Asp) and glu­
2.1. Samples and reagents tamic acid (Glu). The relative contents in Asp, Asn, Gln and Glu were
estimated using kafirin sequences (UniProt reference: P14690, P14691
Flours of four different sorghum cultivars (two cultivars without and P14692).
tannins, BR501 and BRS310, and two cultivars rich in tannins, BRS305 Moisture content was analyzed by measuring the mass of water
and SC782) were provided by Embrapa Milho e Sorgo (Sete Lagoas, MG, removed during drying in a vacuum oven (Marconi MA030, Brazil) at 60
Brazil). For clarity, the suffixes (T-) and (T+) are added to the sample ◦
C for 48 h. The ash content was evaluated by incineration in a muffle
codes to recall the absence or presence of endogenous tannins, furnace (Quimis Q318S25T, Brazil) at 550 ◦ C until the residue was white
respectively. or light gray. The total carbohydrate content, including fibers, was
Ethanol, sodium metabisulfite and sodium hydroxide were from calculated by weight difference (FAO, 2003).
Labsynth (Diadema, Brazil). Anhydrous sodium sulfate, toluene and
cyclohexane were from Carlo Erba Reagents (Val-de-Reuil, France). 2.4.2. Quantification of total flavanols
Dichloromethane, methanol and chloroform were from Biosolve The quantification of total flavanols (flavanol monomers (catechin
Chemicals (Dieuze, France). Phloroglucinol, heptadecanoic acid and epicatechin) and procyanidin oligomers and polymers (i.e.,
(C17:0), boron trifluoride-methanol (BF3) and Nile Red were from condensed tannins)) was performed after acidolysis in the presence of

2
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

phloroglucinol and analysis of the reaction media by UHPLC (Waters) evaporated under nitrogen flow (N-Evap 111, Organomation, USA).
coupled to UV–visible diode array detection (Waters) and mass spec­ Then, 1 mL toluene and 1 mL BF3 reagent at 14 % in methanol were
trometry (TQD Quattro Waters) according to a procedure adapted from added and the mixture was incubated at 100 ◦ C for 45 min in a dry bath
Malec et al. (2014). Aliquots of sorghum flours or protein extracts were (Fisher Bioblock Scientific, Ilkirch, France). After cooling, 1 mL cyclo­
weighed (100 mg for tannin-free samples BR501(T-) and BRS310(T-), hexane and 0.5 mL ultrapure water were added, and the mixture was
and 30 and 10 mg for tannin-rich samples BRS305(T+) and SC782(T+), vortexed for 20 s. After centrifugation at 900 × g for 5 min at 20 ◦ C
respectively) and freeze-dried before further analysis. The freeze-dried (Eppendorf 5810 R, Germany), the upper phase containing cyclohexane
samples were mixed with 400 µL methanol/HCl 0.3 N and 800 µL of a and FAMEs was recovered and analyzed by a gas chromatograph (Clarus
mixture containing phloroglucinol (75 g/L) and ascorbic acid (15 g/L). 680, Perkin Elmer, Shelton, USA) on a capillary column (DB 225MS, 30
The mixture was immediately homogenized by sonication and incubated m x 0.32 mm, film thickness 0.25 µm, Agilent Technologies, USA) ac­
for 30 min at 50 ◦ C (Fisher Scientific, Iowa, USA), with intermittent cording to the procedure described by Meynier et al. (2014). The iden­
stirring at 10 and 20 min. The tubes were placed in an ice bath for 5 min, tification of individual fatty acids was made by comparing the retention
stirred again, and added with 1.2 mL sodium acetate (0.2 M) to stop the time of these components with those of a standard mixture (FAME mix;
reaction, following by new stirring and centrifugation at 8000 × g for 5 Supelco, Sigma Saint Quentin Fallavier, France) and the peak areas were
min at 19 ◦ C (Thermo Fisher Scientific Sorvall LYNX 6000, Germany). integrated and corrected by the response factor of each individual fatty
The supernatants were collected, placed in a VivaSpin tube (Reference: acid (Ackman, 2007). The results were expressed in mg fatty acids/g
VS02H01, Membrane: 10.000 MWCO HY) and centrifuged at 8000 × g lipids.
for 30 min at 19 ◦ C (Thermo Fisher Scientific Sorvall LYNX 6000, Ger­ The tocopherol content was measured by HPLC (UHPLC Ultimate
many). Then, the filtrates were analyzed by UHPLC (Waters). This 3000 system, Thermo Fisher Scientific, Germany) as described by
procedure was performed in triplicate. Meynier et al. (2014) with slight modifications. Briefly, 40 µL lipid
extract were injected onto a normal phase column (Acclaim Polar
2.4.3. Lipid extraction and composition analysis of the powder samples Advantage II, Dionex, Si 3 µm, 250 mm × 3 mm) and the separation was
Total lipids were extracted as described by Fogang Mba et al. (2018). achieved in isocratic mode. Hexane/tert‑butyl‑ether (90/10 v/v) was
Briefly, after hydration of the powder in ultrapure water overnight at 4 used as the mobile phase at a flow rate of 0.5 mL/min, and 295 nm and
◦
C under magnetic stirring, dichloromethane/methanol (2:1, v/v) was 330 nm were used as excitation and emission wavelengths for the
added, the system was stirred on a vortex, centrifugated at 1700 × g for fluorescence detector, respectively. The identification and concentration
10 min at 20 ◦ C (Eppendorf 5810 R, Germany) and the lower organic of the tocopherol classes in the samples were determined comparing
phase was recovered in a volumetric flask. This procedure was repeated their retention time with those of standard compounds and using
two more times, and the lower phases were pooled in the same volu­ external calibration curves of tocopherol isomers and γ-tocotrienol from
metric flask. The mixture was put in a separation funnel with the commercial standards, respectively. The results were expressed in μg
addition of 0.73 % (w/v) sodium chloride solution and stored at 4 ◦ C tocopherols/g lipids.
overnight. The lower phase was recovered in another volumetric flask,
previously weighed, using a hopper with glass wool and anhydrous so­ 2.4.4. Fourier transform infrared (FTIR) spectrophotometry
dium sulfate. The solvent was evaporated in a rotative evaporator A FTIR spectrophotometer Thermo Nicolet IS50 (Thermo Scientific,
(R-100, Rotavapor, Büchi, France) at 40 ◦ C and the sample was dried USA) connected to an attenuator of total reflectance (FTIR-ATR) with
under nitrogen flow (N-Evap 111, Organomation, USA). The lipid con­ deuterated triglycine sulfate (DTGS) was used to obtain FTIR spectra for
tent was calculated by weight difference (Eq. (1)) and expressed in g the sorghum flours, the freeze-dried protein extracts and the protein
lipids/g sample. Chloroform was added to the volumetric flask and the suspensions produced thereof. Powdered samples and protein suspen­
lipid extracts were stored at − 80 ◦ C until subsequent analysis of lipid sions were placed in a single-bounce diamond crystal and a ZnSe crystal,
classes, fatty acid composition, and tocopherols. respectively, and 50 scans were collected in the range of 4000 to 650

mass of volumetric flask with lipids − mass of empty volumetric flask

Lipid content = (1)
mass of sample

cm− 1, with 4 cm− 1

resolution.

Lipid classes (neutral lipids – triglycerides, diglycerides, mono­ 2.4.5. Protein composition analysis
glycerides and free fatty acids – and polar lipids – phospholipids) were The protein composition in the flours, protein extracts and protein
analyzed by high performance liquid chromatography (HPLC). suspensions was determined by sodium dodecyl sulfate-polyacrylamide
Approximately 5 μg lipid extract in 10 μL chloroform were analyzed gel electrophoresis (SDS-PAGE). A size exclusion - high performance
with a modular UltiMate 3000 RS system (Dionex, Voisins Le Bre­ liquid chromatography (SE-HPLC) assay was conducted for both flours
tonneux, France) equipped with an Uptisphere Strategy column (150 and their corresponding protein extracts.
mm × 4.6 mm, 2.2 μm, 100 Å; Interchim, Montluçon, France) and For analyzing the powders by SDS-PAGE, total kafirins were
coupled to an evaporative light scattering detector (ELSD) Sedex 85 extracted from raw flour according to D’Almeida et al. (2021) and the
(Sedere S.A., Alfortville, France). Chromatographic conditions, identi­ protein solutions were freeze-dried before being dispersed at 1.5 g/mL
fication and quantification of lipids were as described by Fogang Mba during 15 min at 70 ◦ C in 10 mM borate pH 10.0 buffer containing 2 %
et al. (2017). SDS and 20 mM dithioerythritol (DTE). Equivalents of 12 µg of pro­
For the determination of the fatty acid composition, the lipid extracts tein/sample were loaded on a pre-casted gel and run for 2 h using a
were first derivatized to fatty acid methyl esters (FAMEs) according to Bio-Rad Mini Protean® 3 Cell (USA) system. Pre-stained protein markers
Morrison and Smith (1964). For this, 100 μL of lipid extract (i.e., around with molecular weights ranging from 10 to 250 kDa (Biorad) were
1 mg lipids) were mixed with 100 μL of internal standard solution loaded in a 4–20 % gradient gel (Biorad). Then, the gel was stained and
(C17:0, 1 mg/mL in chloroform) in a glass tube and the solvent was decolored following the method described by Neuhoff et al. (1988).

3
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

For SDS-PAGE analysis of the protein suspensions and of their su­ 2.4.7. Solubility of protein extracts in aqueous media
pernatants after centrifugation (10,000 × g for 30 min at 10 ◦ C), samples Suspensions, both treated and non-treated by HPH, were analyzed
were first denatured and reduced by heating at 95 ◦ C for 10 min in a before and after centrifugation at 10,000 × g for 30 min at 10 ◦ C
buffer solution containing 62.5 mM Tris–HCl buffer (pH 6.8), 2 % (w/v) (Eppendorf 5810 R, Germany) regarding the nitrogen content in the
SDS, 25 % (v/v) glycerol, 0.01 % bromophenol blue and 5 % (v/v) supernatant using the Dumas method (Rapid MAX N exceed, Elementar,
β-mercaptoethanol (β-ME). Pre-stained protein markers (SeeBlue, Germany). The protein solubility (S) was calculated using Eq. (2).
Novex, Life technologies, Thermo Fisher, France), with molecular
Nitrogen contentsupernatant
weights ranging from 3 to 198 kDa, and 5 µL of each sample were loaded S= × 100% (2)
Nitrogen contentsuspension
in an 8–16 % gradient gel. Migration was conducted for 2 h using a Bio-
Rad Mini Protean® 3 Cell (USA) system. Subsequently, the gel was
exposed to a brilliant blue Coomassie G 250 solution under gentle 2.5. Statistical analysis
agitation (Rocker 25, Labnet, Labnet International Inc.) for 2 h.
Following this, it was rinsed with distilled water and subjected to Results obtained from triplicate analytical determinations under­
scanning and analysis using Multi Gauge V3.0 software (Fujifilms). went analysis of variance (ANOVA) and differences between means were
For SE-HPLC analysis of the powders, the procedure was carried out assessed by the Tukey test at 5 % probability level (XLStat).
as described by Morel et al. (2022), with some modifications. In brief, a
protein extraction step was performed to recover soluble proteins, by 3. Results and discussion
dispersing 50 mg powder (flour or protein extract) in 1 mL 10 mM
Na-borate buffer pH 10.0, SDS 2 %. The mixture was stirred for 20 min at 3.1. Characterization of the samples
20 ◦ C, and centrifuged at 20,000 × g for 10 min at 20 ◦ C (Eppendorf
5427 R, Germany). The supernatant was then recovered and added into 3.1.1. Proximate composition of the powder samples
a new tube, repeating this procedure two more times. The pellet The starting flours displayed protein contents ranging from 9.6 to
resulting from the centrifugation was dispersed in 1 mL of the same 10.6 g/100 g (wet basis, w.b.) (Table 1), with no systematic effect of the
buffer complemented with 20 mM of DTE, following the same param­ tannin content. Specifically, samples BR501(T-) and SC782(T+) did not
eters as described above for recovery of the insoluble proteins. The exhibit a significant difference, while a notable distinction was observed
protein analysis was performed at 25 ◦ C, on an Alliance Waters HPLC between samples BRS310(T-) and BRS305(T+)). Martino et al. (2012)
chain equipped with a TSK G4000 (Tosoh bioscience) (7.8 × 300 mm) studied eight Brazilian sorghum cultivars, developed and cultivated by
column and a TSK gel SWXL (4 cm × 6 mm) pre-column, eluted with the Embrapa Milho e Sorgo (Sete Lagoas/MG) and obtained protein contents
phosphate buffer, 0.1 M, pH 6.8, SDS 0.1 %. A volume of 15 μL was ranging from 8.6 to 11.9 g/100 g (w.b.), aligning closely with the
injected and eluted at 0.7 mL/min; the detection of different proteins findings of the present study. The cultivars BR501(T-), BRS310(T-) and
was done at 214 nm. The data were collected with the help of Empower BRS305(T+) were also analyzed by Martino et al. (2012), yielding
software. The mass calibration of the device was carried out by injecting slightly different results for protein content. Antunes et al. (2007)
a standard protein of known molecular weight and the area under the explored the genotype BR501(T-) and reported approximately 11.2
curve was integrated, using a response coefficient determined experi­ g/100 g (dry basis, d.b.) of protein, equivalent to 10.1 g/100 g (w.b.).
mentally, to determine the protein content. This falls within the range observed in the present work and the results
obtained by Martino et al. (2012).
2.4.6. Colloidal morphology of the protein suspensions The protein content in the extracts ranged from 55 to 67 g/100 g (w.
The particle size distribution and mean volume diameter (d4,3) in the b.) (Table 1), with a significantly lower content for the (T+) extracts
aqueous suspensions were determined by static light scattering (Horiba compared to the (T-) extracts. This suggests that the presence of tannins
LA-950, Japan). The refractive indices were set to 1.450 for the tends to decrease the protein content after the extraction process,
dispersed (protein) phase and 1.333 for the continuous (aqueous) phase. possibly due to the ability of tannins to covalently and irreversibly bind
Particle morphology was evaluated by optical microscopy (Zeiss to proteins, hampering their extractability (Duodu et al., 2003). Using a
Axioskop, equipped with Prosilica Digital Camera 5DCAM 1.31, and similar sorghum protein extraction method, Taylor et al. (2005) re­
image software), and confocal laser scanning microscopy (CLSM) ported protein contents in the extracts ranging from 74.6 to 89.3 g/100 g
(Inverted Nikon A1 laser scanning confocal microscope) using Alexa 488 (d.b.), before and after the defatting process, respectively. In a study by
and Nile red to stain the proteins and the lipids, respectively, as Da Silva and Taylor (2004), employing the same reagents with slight
described by Keuleyan et al. (2023). differences in the extraction process, protein contents ranging from 77.9
to 83.1 g/100 g (d.b.) were obtained without the defatting process, and
approximately 88 g/100 g (d.b.) after defatting with hexane.

Table 1
Composition (g/100 g total basis) of sorghum flours and their protein extracts. BR501(T-) and BRS310(T-) are the cultivars without tannins, whereas BRS305(T+) and
SC782(T+) are the cultivars with tannins.
Sample Proteins1 Total lipids Water Ash Total procyanidins Carbohydrates (including fibers)2
ef d a a 3
Flours BR501(T-) 10.30 ± 0.16 * 3.51 ± 0.03 8.27 ± 0.03 1.40 ± 0.03 n.d. 76.52 ± 0.09a
BRS310(T-) 10.60 ± 0.24e 3.04 ± 0.20de 8.18 ± 0.01a 1.38 ± 0.03ab n.d. 76.80 ± 0.15a
BRS305(T+) 9.64 ± 0.09 g 3.12 ± 0.16de 8.19 ± 0.03a 1.29 ± 0.02c 2.06 ± 0.00 75.70 ± 0.13a
SC782(T+) 10.10 ± 0.18f 2.58 ± 0.29e 8.04 ± 0.02b 1.20 ± 0.04de 1.82 ± 0.00 76.26 ± 0.18a
Protein extracts BR501(T-) 65.22 ± 0.56b 22.19 ± 0.45b 4.30 ± 0.07c 1.41 ± 0.01a n.d. 6.88 ± 0.71d
BRS310(T-) 67.14 ± 0.35a 21.88 ± 0.12b 3.73 ± 0.06d 1.35 ± 0.01b n.d. 5.90 ± 0.64e
BRS305(T+) 54.99 ± 0.37c 25.84 ± 0.13a 3.39 ± 0.03f 1.27 ± 0.04cd 1.27 ± 0.00 13.24 ± 0.51c
SC782(T+) 50.08 ± 1.46d 18.71 ± 0.63c 3.59 ± 0.03e 1.15 ± 0.02e 1.95 ± 0.00 24.52 ± 0.92b
1
Protein contents calculated using N factors values determined for each sample in the present work: 5.88, 5.88, 5.92, and 5.87 for BR501(T-), BRS310(T-), BRS305
(T+) and SC782(T+), respectively.
2
Calculated by difference.
3
Values not detected with the given analysis conditions.
*
Values in the same column with different letters are significantly different (ANOVA with Tukey test: p < 0.05). All experiments were carried out in triplicate.

4
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

Espinosa-Ramírez and Serna-Saldívar (2016), who also used the same protein extracts. Regarding the carbohydrate content (i.e., the sum of
extraction methodology, reported protein contents ranging from 72.7 to the starch and fiber contents), higher values were obtained in the pre­
88.6 g/100 g (d.b.) for extracts from whole sorghum grains. These re­ sent study compared to the previous work of Martino et al. (2012),
sults align with the findings of the present study, although the protein especially for tannin-rich cultivars. These disparities in proximate
contents are lower than previously reported, especially for tannin-rich composition, even for the same cultivars, can be explained by variations
cultivars. It should be pointed out that the protein content in the in the cultivation and storage conditions of the grains.
extract depends not only on the cultivar and the extraction methodol­ In terms of total flavanols, encompassing flavanol monomers (cate­
ogy, but also on the used N conversion factor. A generalized value of chin and epicatechin) and procyanidin oligomers and polymers (i.e.,
6.25 is commonly used by manufacturers/suppliers of plant protein in­ condensed tannins), the tannin-free samples BR501(T-) and BRS310(T-)
gredients, including the works cited above, but this value is usually did not contain any flavanols. In contrast, flavanols were present in the
higher than the real values for plant proteins, leading to an over­ tannin-rich samples BRS305(T+) and SC782(T+). Interestingly, the
estimation of the protein content in such ingredients. For instance, proportions of flavanols in both the flours and protein extracts were of
Sosulski and Imafidon (1990) studied the amino acid composition of the same order of magnitude, ranging from 1.3 to 2.1 %. Consequently,
sorghum grain proteins and determined a N conversion factor of 5.93. In no substantial enrichment of tannins in the extracts was observed
the present work, after analyzing the amino acid composition (Supple­ compared to the initial flours. These findings may hold important im­
mentary Table 1) and correlating the results with the protein sequences, plications for the functionality of the ingredients, as flavonoids such as
specific N conversion factors were determined for each sample: 5.88 for flavanols have been recognized for their health benefits and antioxidant
BR501(T-) and BRS310(T-), 5.92 for BRS305(T+), and 5.87 for SC782 properties (Hackman et al., 2008).
(T+).
The lipid content in the flours ranged from 2.5 to 3.5 g/100 g (w.b.) 3.1.2. Lipid composition of the powder samples
(Table 1), with no significant differences between tannin-free and The accumulation of lipids in the sorghum protein extracts can be of
tannin-rich samples. Martino et al. (2012) reported lipid contents high importance for the properties and functionality of these in­
varying from 2.6 to 3.1 g/100 g (t.b.) for BR501(T-), BRS310(T-), and gredients, and therefore a detailed analysis of the composition of the
BRS305(T+) genotypes, aligning closely with the results in the present lipid extracts was conducted. Examples of the chromatograms obtained
study. In the protein extracts, lipid content was noticeably high (18.7–26 for the analysis of lipid classes, fatty acid composition and tocopherol
g/100 g, t.b.), with the tannin-rich samples BRS305(T+) and SC782(T+) content are shown in Supplementary Fig. 2.
exhibiting the highest and lowest values, respectively. Espinosa-Ramírez No phospholipids (i.e., polar lipids) were detected in any sample. The
and Serna-Saldívar (2016) obtained lipid contents ranging between 17.9 predominant lipid class identified was free fatty acids (FFAs) (Fig. 1.A),
and 20.3 g/100 g (d.b.), whereas Da Silva and Taylor (2004) reported ranging from 51.3 to 75.8 % of the total lipids, followed by tri­
values of 12.8–16.7 g/100 g (d.b.) for protein-rich fractions. It is worth acylglycerols (TAGs), ranging from 21.4 to 44.6 %. An exception was
noting that the latter authors employed a Soxhlet extraction method, observed in the BR501(T-) flour, where TAGs (54.1 %) surpassed FFAs
known as a lengthy process and involving hot solvent, which may have (41.5 %). All the flour samples were significantly different from each
degraded some lipid components. Additionally, the Soxhlet method does other regarding FFA and TAG contents, without a clear effect of the
not allow for a quantitative extraction of polar lipids, as discussed by Li tannins. In the protein extracts, where the lipid fraction is substantial (as
et al. (2014). The high lipid contents observed in the protein extracts discussed in Section 3.1.2.), (T-) samples significantly differed from the
highlight that the applied extraction process not only concentrates (T+) ones in FFA content, with the former exhibiting the highest values.
proteins, but also results in a marked accumulation of the lipids initially This suggests that the presence of tannins could be linked to a lower FFA
present in the flours, which can be related to the fact that the flours were content. Osagie (1987) and Price and Parsons (1975) reported that
not defatted. sorghum grains contained a small quantity of phospholipids and gly­
The water content was comparatively higher in the flours than in the colipids, with neutral lipids being the predominant class (> 85 %),
respective protein extracts, whereas the ash content showed no signifi­ consistent with our study’s findings. However, Osagie (1987) observed a
cant difference between the two (Table 1). This observation can prob­ large contribution of TAGs within neutral lipids (85 %), which was not
ably be attributed to the freeze-drying process used to prepare the found in the present study. The high FFA proportions present in our

Fig. 1. Content in the different lipid classes (A) and tocopherol concentration (B) found in sorghum flours (suffix F) and their protein extracts (suffix PE). BR501(T-)
and BRS310(T-) are the cultivars without endogenous tannins, whereas BRS305(T+) and SC782(T+) are the cultivars with tannins. TAG: triacylglycerol; DAG1:
diacylglycerol-1–3; DAG2: diacylglycerol-1–2; MAG: monoacylglycerol; FFA: free fatty acid. *Different letters for the same lipid class or the same tocopherol isomer
indicate significant differences (ANOVA with Tukey test: p < 0.05).

5
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

samples most likely result from the hydrolysis of TAGs initially present that, in general, the extraction process had a substantial effect on the
in the seeds, which may happen because of endogenous lipases either in secondary structure of the proteins, especially for the cultivars con­
planta (i.e., during the maturation of the seeds on the sorghum plants) or taining tannins (BRS305 and SC782). This phenomenon could be
after harvesting of the seeds, for instance upon production or storage of attributed to the elevated temperature and reducing conditions (sodium
the flours. metabisulfite) applied during the extraction process (Wang et al., 2009).
The fatty acid composition (Supplementary Table 2) of the sorghum In the aqueous suspensions, the secondary structure predominantly
lipids revealed a high concentration of linoleic acid (45–51 % of all fatty consisted of β-sheets (31–42 %) and α-helices (27–40 %), with some
acids for the protein extracts, and 43–48 % for the flours), oleic acid β-turns (1–15 %) and random coils (10–20 %), except for the SC782(T+)
(27–34 % for the protein extracts, and 29–37 % for the flours), and sample, which exhibited more α-helices (35–37 %) than β-sheets (15–26
palmitic acid (13–15 % for the protein extracts and flours). This in­ %). Comparing the results before and after HPH, a noticeable decrease in
dicates a high content in PUFAs (> 50 % of all fatty acids) and MUFAs (> α-helices and β-sheets was observed following this treatment, indicating
30 %), rather than saturated FAs. A slight increase of linoleic acid and a a modification in the secondary structure of the proteins. This alteration
slight decrease in oleic and palmitic acids were noted when comparing might be attributed to the temperature increase during HPH (typical
the fatty acid composition of the protein extracts with that of the initial temperatures at the beginning and end of HPH were 20 and 45 ◦ C,
flours, suggesting a slight accumulation of PUFAs during the protein
extraction process. Hassan et al. (2017), Osagie (1987) and Price and
Parsons (1975) obtained similar results for sorghum grains, where
linoleic (43–58 %), oleic (21–37 %), and palmitic (13–19 %) acids were
the main components identified in the fatty acid profile.
The total tocopherol content in the lipid extracts obtained from
sorghum flours and their respective protein extracts ranged from 1371
to 2039 µg/g total lipids. Tocopherol analysis (Fig. 1.B) revealed that the
predominant component is γ-tocopherol, representing 85 to 94 % of the
total tocopherol content. A substantial accumulation of tocopherols in
the lipid fraction of the protein extracts was observed compared to the
corresponding starting flour; the tocopherol content in the extracted
lipid phase was multiplied by a factor of 2.1 to 4.6, depending on the
sample, suggesting that the extraction process did not lead to a sub­
stantial chemical damage of these phytochemicals. Instead, it dramati­
cally concentrated them in the final sample. Additionally, a higher
tocopherol content was observed in the sample with a higher tannin
content (SC782), both for flour and protein extract. These findings for
flour samples align with the studies made by Martino et al. (2012), who
reported similar values for tocopherol content, and predominantly
γ-tocopherol. Li et al. (2021) also summarized tocopherol contents
found in sorghum grains and observed the same pattern, having the
highest values for γ-tocopherol.

3.1.3. FTIR spectra of the samples

The overall chemical fingerprint of the powder samples (flours and
extracts), protein suspensions and their respective supernatants after
centrifugation was first assessed by determining their infrared spectra
(Supplementary Fig. 3.A and C). The compositional complexity of these
samples is reflected in the infrared spectra, revealing multiple peaks
corresponding to chemical groups typical of lipids (characterized by
peaks at 2925 and 2854 cm− 1 (methylene (-CH2) and methyl (-CH3)
groups, respectively)), proteins (with peaks at 1652 and 1538 cm− 1
(amide I (C = O) and amide II (N–H, C–N) groups, respectively)), and
carbohydrates (with peaks at 3316.5 and 1151 to 900 cm− 1 (O–H,
C–O, and C–C groups, respectively)) (Lin et al., 2021). All the identi­
fied peaks and functional groups for the flours and protein extracts in
this study align with the results obtained by Lin et al. (2021) and Xiao
et al. (2015), respectively.
The secondary structure of proteins, correlated with the shape of the
amide I band (Xiao et al., 2015) (Supplementary Fig. 3.B and D), was
evaluated by performing a multiple peak fitting using a Gaussian func­
tion. Peaks ranging from 1620 to 1635 cm− 1, 1638 to 1640 cm− 1, 1645
to 1660 cm− 1, and 1670 to 16,790 cm− 1 represent β-sheets, random
coils, α-helices and β-turns, respectively. In the flours, the secondary
structure was predominantly composed of α-helices (30–37 %) and Fig. 2. Electrophoresis (SDS-PAGE) of sorghum proteins present in flours
(suffix F) and their protein extracts (suffix PE) under reducing conditions (A),
β-sheets (22–31 %), with a minor presence of β-turns (17–24 %) and
and protein composition determined by optical density analysis of SDS-PAGE
random coils (10–16 %). Conversely, protein extracts exhibited a greater
gels of total suspensions (B), and their respective supernatants (C) (suffix
contribution of β-sheets (31–34 %) compared to α-helices (27–29 %), SN), before (NP) and after homogenization (HPH). Samples were denatured and
with minor occurrence of β-turns (19–24 %) and random coils (8–9 %). reduced by heating at 95 ◦ C and using β-mercaptoethanol. BR501(T-) and
When comparing a given flour and its respective protein extract, the BRS310(T-) are the cultivars without tannins, whereas BRS305(T+) and SC782
presence of α-helices decreased within a range of 8–25 %. This suggests (T+) are the cultivars with tannins.

6
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

respectively), coupled with enhanced hydration of the powder grains sorghum protein extracts.
facilitated by intense shear and particle size reduction. Upon analyzing the densitometry profiles of the SDS-PAGE gels for
protein suspensions and their corresponding supernatants after centri­
3.1.4. Protein composition of the samples fugation, the distribution of the different protein components was
The protein types present in the sorghum flours from the different determined (Fig. 2.B and C). Minimal differences were observed be­
cultivars and in their respective protein extracts are illustrated in Fig. 2. tween samples before and after HPH, with the protein composition
A. Identifiable proteins include β-kafirin (20 kDa), α-kafirins (23 and 25 mainly featuring β-kafirin (13–17 %), α- and γ-kafirins (57–68 %), di­
kDa) and γ-kafirin (28 kDa) across all samples. Dimers (approx. 50 kDa) mers (10–23 %) and oligomers (3–11 %). In the samples without tannins
were observed in the flours without tannins (BR501 and BRS310) and in (BR501(T-) and BRS310(T-)), some oligomers appeared after HPH,
all protein extracts, suggesting that the protein extraction process suggesting that the treatment may induce covalent aggregation.
potentially induced protein aggregation in the samples with tannins Notably, a distinction in protein components was observed between the
(BRS305 and SC782). Notably, tannin-rich flours exhibited components total suspension (Fig. 2.B) and their respective supernatants (Fig. 2.C)
with molecular weights exceeding 250 kDa, hinting at potential protein- for non-homogenized suspensions of tannin-rich samples. The former
tannin complexation. Large proteins with high molecular weight and exhibited the presence of β-, α- and γ-kafirins, as well as dimers and
flexible, open conformations have been shown to have affinity for tan­ oligomers, whereas only β-, α- and γ-kafirins were seen in the latter.
nins, a tendency that can be enhanced by protein denaturation (Butler,
1982; Rodrigues et al., 2009). The findings for the protein extracts align 3.1.5. Size distribution of the protein components in the different flours and
with studies performed by Belton et al. (2006), Espinosa-Ramírez and extracts
Serna-Saldívar (2016), Taylor et al. (2005), Wang et al. (2009) and Xiao SE-HPLC was employed to analyze the molecular weight distribution
et al. (2015), which demonstrated higher concentrations of α-kafirin and solubility of proteins (Fig. 3). For the flours, the SE-HPLC protein
compared to β- and γ-kafirins, dimers, trimers, and oligomers in various profiles of SDS-soluble proteins (SSP) revealed that the different

Fig. 3. Protein size distribution in sorghum flours and in their respective protein extracts for SDS-soluble (SSP) (A and B) and SDS-insoluble (SIP) (C and D) fractions;
and protein quantification in the flours (E) and protein extracts (F) by size exclusion – high performance liquid chromatography (SE-HPLC). BR501(T-) and BRS310
(T-) are the cultivars without tannins, whereas BRS305(T+) and SC782(T+) are the cultivars with tannins.

7
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

genotypes, regardless of the presence or absence of tannins, exhibited a SIP fraction (Fig. 3.D). When comparing the flours with their respective
similar protein composition and size distribution. Kafirin monomers (all protein extracts, the SE-HPLC profile of the SIP from the flours mirrored
subclasses) were identified by the peak eluting around 25 kDa. However, that of the SSP from the protein extracts. This alignment is logical,
the tannin-free genotype BR501(T-) displayed some differences, considering the use of sodium metabisulfite (a reducing agent) during
particularly in the peak around 14 kDa, where the intensity was lower the protein extraction process.
(Fig. 3.A). The SE-HPLC profiles of SDS-insoluble proteins (SIP) (Fig. 3. Regarding quantitative aspects (Fig. 3.E and F), the amounts of SSP
C), primarily representing the cross-linked kafirins and glutelin frac­ were higher than the SIP ones. The results unveil the presence of non-
tions, looked fairly similar across different genotypes, although the extractable protein fractions in the flour samples, in varying pro­
kafirin peak (~25 kDa) differed in intensity. Within this fraction, an portions (9 to 17 % of the total proteins) depending on the genotype.
important reduction of polymeric proteins (peaks eluted around 50 and Similar results were reported by D’Almeida et al. (2021), who studied
80 kDa and beyond) occurred due to the reduction of disulfide bonds. two Brazilian sorghum genotypes, one rich in tannins and the other
The absence of the peak eluting around 14 kDa in the SIP for the extracts tannin-free. Additionally, solubility differences were observed between
suggests that they correspond to non-kafirin proteins, probably small (T+) and (T-) genotypes.
albumins and globulins. In conclusion, a robust agreement and complementarity were iden­
In the protein extracts, all proteins were recovered in the SSP fraction tified between the SE-HPLC and SDS-PAGE analyses (Section 3.1.4)
(Fig. 3.B), predominantly as monomers eluting around 25 kDa. How­ concerning protein size distribution and identification.
ever, the presence of some polymeric kafirin forms (peaks at 50 and 80
kDa) was also observed. Only negligible amounts were recovered in the

Fig. 4. Particle size distribution (distribution in volume frequency determined by static light scattering) in the suspensions before (square) and after (triangles) high
pressure homogenization (HPH): (A) BR501(T-), (B) BRS310(T-), (C) BRS305(T+), (D) SC782(T+); and (E) macroscopic images of the suspensions before (NP) and
after (HPH) homogenization. Pictures were taken 24 h after suspension preparation (and homogenization, when applicable).

8
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

Fig. 5. Optical microscopy and CLSM images of two sorghum protein suspensions: (A, B) BR501(T-) before HPH, (C, D) BR501(T-) after HPH, (E, F) SC782(T+)
before HPH, and (G, H) SC782(T+) after HPH. Proteins were stained with fast green and are visualized in green, whereas lipids were stained with Nile red and are
visualized in red. White arrows indicate starch granules.

2.3. Behavior of the protein extracts in aqueous suspensions bottom of the tubes, especially in the samples with tannins (BRS305 and
SC782). This is directly linked to the decrease in particle size, as per the
3.2.1. Colloidal morphology of the protein suspensions Stokes law, where the rate of gravitational separation scales with the
The particle size distributions of all aqueous suspensions of protein square of the particle size in dilute media (McClements, 2015).
extracts, both before and after HPH (Fig. 4), demonstrated that HPH The microstructure of the suspensions was evaluated by optical mi­
decreased the average particle size for all samples (initially ranging from croscopy and CLSM images (Fig. 5), revealing the persistence of some
9 to 66 µm, and decreasing to 1.8 to 2.5 µm after the treatment). This large structures/aggregates (up to tens of μm) with irregular shapes even
effect was particularly pronounced in the suspension of the BR501(T-) after HPH treatment. In the CLSM images, the presence of lipids (stained
protein extract (Fig. 4.A). The shear-based treatment also resulted in in red) was observed, which was expected as the lipid content in the
unimodal particle size distributions, instead of bimodal ones in the non- protein extracts ranged between 18 and 26 % (g/100 g, t.b.). For the
homogenized suspensions. Macroscopic images of the suspensions tannin-rich samples BRS305(T+) and SC782(T+) (Figs. 5.E–H), spher­
(Fig. 4.E) illustrated that HPH mitigated particle sedimentation at the ical, smooth structures were observed (white arrows), especially after

9
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

Fig. 7. Solubility (defined as the fraction of proteins that remained in the su­
pernatant after centrifugation, compared to the total protein content in the
suspension) of sorghum proteins in aqueous suspensions (ultrapure water)
before and after treatment by HPH (suspensions were homogenized for 6 min at
400 bars). NP: non-treated by high pressure homogenization; HPH: treated by
high pressure homogenization.
*Different letters for the same treatment indicate significant differences
(ANOVA with Tukey test: p < 0.05).

3.2.2. Protein solubility in the aqueous suspensions

Conventionally, ‘solubility’ is defined as the fraction of proteins
remaining in the supernatant of a given sample after centrifugation
under specified conditions. However, these supernatants may contain
small particles that do not sediment under these conditions, and that are
often referred to as ‘soluble aggregates’. For simplicity, the terms ‘sol­
ubility’ and ‘soluble’ will be used in the following discussion, but one
should bear in mind that this terminology may deviate slightly from the
strict definition of solubility.
Protein solubility was assessed both before and after HPH treatment
(Fig. 7). Homogenization substantially increased nitrogen concentra­
tion, and thus protein content in the supernatant, resulting in solubilities
rising from 4.8–12.1 % to 12.2–26.5 % (a 1.4- to 3.9-fold increase,
depending on the sample), especially in samples rich in tannins. This
underscores the effectiveness of HPH for this purpose, suggesting
promising technological improvements for handling such suspensions
and unleashing their use for various applications. This observation
aligns with the visual inspection of the samples before and after HPH
(Fig. 4.E), where sedimentation was largely prevented by HPH treat­
ment. Moll et al. (2021) applied Microfluidizer-based HPH on pea pro­
tein suspensions and observed a decrease in particle size and an increase
in solubility (also defined as the ratio between nitrogen concentration in
the supernatant and in the aqueous protein suspension). This confirms
Fig. 6. Polarized light microscopy images for the SC782(T+)-based suspen­ the potential of HPH to improve the processability of poorly soluble
sions: (A) flour, (B) protein suspension before HPH, and (C) protein suspension protein ingredients. A similar conclusion was recently drawn by Keu­
after HPH. leyan et al. (2023), who found that HPH was helpful to improve solu­
bility of various plant protein ingredients from pea and lupin. This was
HPH treatment, identified as starch granules through polarized light especially marked for highly processed ingredients, such a pea protein
microscopy (Fig. 6). The SC782(T+) sample exhibited a substantial isolate, characterized by a high endogenous lipid accumulation and poor
amount of starch granules in the flour (Fig. 6.A), which markedly initial solubility.
decreased in the protein extract suspension (Fig. 6.B and C) This con­
firms that the extraction process effectively removed most of the starch. 4. Conclusions
Additionally, the remaining starch granules were predominantly trap­
ped within large composite structures following the extraction process In this study, a detailed analysis of the composition of sorghum flours
(Fig. 6.B), probably protein- and/or fiber-based aggregates. The HPH from tannin-poor or tannin-rich varieties, along with their derived
treatment facilitated the release of these starch granules into the protein extracts, was carried out. The protein extraction procedure
aqueous medium (Fig. 6.C). proved efficient to yield high protein contents, especially for sorghum
varieties without tannins. Besides achieving its primary objective of
producing protein-rich ingredients, the extraction process also resulted

10
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

in a large accumulation of lipids in the extracts. The majority of these Acknowledgements

lipids were identified as free fatty acids (FFAs), mainly composed of
linoleic, oleic, and palmitic acids. Additionally, the lipid fraction The authors would like to gratefully acknowledge Embrapa Milho e
exhibited a remarkably high content in tocopherols. The substantial Sorgo for donating the flours of the sorghum genotypes used in this
presence of lipids, especially FFAs, in sorghum protein ingredients may work, the P2M2 platform (Corsaire, Biogenouest, Le Rheu, France) for
be of importance, particularly for their interfacial properties, which providing equipment and technical support for the LC-UV-MS analyses,
could influence the stability of emulsions produced with such and the contribution of Brigitte Laillet for her help with the lipid
ingredients. extraction.
As plant proteins are not well-soluble in aqueous media, which is a
major limitation to their widespread use in food applications, high Supplementary materials
pressure homogenization (HPH) of the aqueous suspensions prepared
with the protein extracts emerged as a potential means to enhance their Supplementary material associated with this article can be found, in
dispersibility and protein solubility, especially for sorghum varieties the online version, at doi:10.1016/j.fufo.2024.100323.
with tannins. Consequently, HPH proves to be a valuable strategy to
improve the processability of poorly soluble protein ingredients such as References
prolamin-rich fractions, and more broadly, various plant protein
ingredients. Ackman, R.G., 2007. Application of gas–liquid chromatography to lipid separation and
analysis: qualitative and quantitative analysis. In: Chow, C.K. (Ed.), Fatty Acids in
Foods and Their Health Implications. CRC Press, pp. 47–65.
Ethical statement Antunes, R.C., Rodriguez, N.M., Gonçalves, L.C., Rodrigues, J.A.S., Borges, I., Borges, A.
L.C.C., Saliba, E.O.S., 2007. Composição bromatológica e parâmetros físicos de grãos
de sorgo com diferentes texturas do endosperma. Arq. Bras. Med. Veterinária e
This research did not involve human subjects nor animal Zootec. 59, 1351–1354. https://doi.org/10.1590/s0102-09352007000500042.
experiments. AOAC, 2006. Official Methods of Analysis, 18th ed. Association of Official Analytical
Chemists, Gaithersburgs, MD.
Bader, S., Bez, J., Eisner, P., 2011. Can protein functionalities be enhanced by high-
CRediT authorship contribution statement pressure homogenization? – A study on functional properties of lupin proteins.
Procedia Food Sci 1, 1359–1366. https://doi.org/10.1016/j.profoo.2011.09.201.
Thais Cristina Benatti Gallo: Writing – original draft, Visualization, Balny, C., Masson, P., 1993. Effects of high pressure on proteins. Food Rev. Int. 9,
611–628. https://doi.org/10.1080/87559129309540980.
Investigation, Formal analysis. Valérie Beaumal: Writing – review &
Barros, F., Awika, J., Rooney, L.W., 2014. Effect of molecular weight profile of sorghum
editing, Resources, Data curation. Bérénice Houinsou-Houssou: proanthocyanidins on resistant starch formation. J. Sci. Food Agric. 94, 1212–1217.
Writing – review & editing, Resources, Data curation. Michèle Viau: https://doi.org/10.1002/jsfa.6400.
Writing – review & editing, Investigation. Lucie Ribourg-Birault: Belton, P.S., Delgadillo, I., Halford, N.G., Shewry, P.R., 2006. Kafirin structure and
functionality. J. Cereal Sci. 44, 272–286. https://doi.org/10.1016/j.
Writing – review & editing, Investigation. Hélène Sotin: Writing – re­ jcs.2006.05.004.
view & editing, Investigation. Joëlle Bonicel: Investigation. Adeline Burger, T.G., Zhang, Y., 2019. Recent progress in the utilization of pea protein as an
Boire: Writing – review & editing, Resources. Valéria Aparecida Vieira emulsifier for food applications. Trend. Food Sci. Technol. 86, 25–33. https://doi.
org/10.1016/J.TIFS.2019.02.007.
Queiroz: Writing – review & editing, Resources. Hamza Mameri: Butler, L.G., 1982. Polyphenols and their effects on sorghum quality. In: International
Writing – review & editing, Investigation, Formal analysis, Conceptu­ Symposium on Sorghum Grain Quality. Patancheru, India, pp. 291–311.
alization. Sylvain Guyot: Writing – review & editing, Validation, Su­ D’Almeida, C.T.dos S., Mameri, H., Menezes, N.dos S., de Carvalho, C.W.P., Queiroz, V.A.
V., Cameron, L.C., Morel, M.H., Takeiti, C.Y., Ferreira, M.S.L., 2021. Effect of
pervision, Methodology, Investigation, Formal analysis, extrusion and turmeric addition on phenolic compounds and kafirin properties in
Conceptualization. Alain Riaublanc: Writing – review & editing, tannin and tannin-free sorghum. Food Res. Int. 149 https://doi.org/10.1016/j.
Formal analysis. Vânia Regina Nicoletti: Writing – review & editing, foodres.2021.110663.
Da Silva, L.S., Taylor, J.R.N., 2004. Sorghum bran as a potential source of kafirin. Cereal
Validation, Supervision, Project administration, Methodology, Funding Chem 81, 322–327. https://doi.org/10.1094/CCHEM.2004.81.3.322.
acquisition, Conceptualization. Claire Berton-Carabin: Writing – re­ de Mesa-Stonestreet, N.J., Alavi, S., Bean, S.R., 2010. Sorghum proteins: the
view & editing, Validation, Supervision, Project administration, Meth­ concentration, isolation, modification, and food applications of kafirins. J. Food Sci.
75 https://doi.org/10.1111/j.1750-3841.2010.01623.x.
odology, Funding acquisition, Conceptualization.
Duodu, K.G., Taylor, J.R.N., Belton, P.S., Hamaker, B.R., 2003. Factors affecting sorghum
protein digestibility. J. Cereal Sci. 38, 117–131. https://doi.org/10.1016/S0733-
Declaration of competing interest 5210(03)00016-X.
Espinosa-Ramírez, J., Serna-Saldívar, S.O., 2016. Functionality and characterization of
kafirin-rich protein extracts from different whole and decorticated sorghum
The authors confirm that they have no conflicts of interest with genotypes. J. Cereal Sci. 70, 57–65. https://doi.org/10.1016/j.jcs.2016.05.023.
respect to the work described in this manuscript. FAO, 2003. Food Energy – Methods of Analysis and Conversion Factors. FAO Food and
Nutrition Paper 77 [WWW Document]. URL. https://www.fao.org/3/Y5022E/Y50
22E00.htm (accessed 12.20.23).
Data availability FAOSTAT, 2022. Food and Agriculture Organization of the United Nations. FAO.
https://www.fao.org/faostat/en/?#data/QCL (accessed 12.13.22).
Data will be made available on request. Fogang Mba, A.R., Kansci, G., Viau, M., Hafnaoui, N., Meynier, A., Demmano, G.,
Genot, C., 2017. Lipid and amino acid profiles support the potential of
Rhynchophorus phoenicis larvae for human nutrition. J. Food Compos. Anal. 60,
64–73. https://doi.org/10.1016/J.JFCA.2017.03.016.
Funding Fogang Mba, A.R., Kansci, G., Viau, M., Ribourg, L., Muafor, J.F., Hafnaoui, N., Gall, P.L.,
Genot, C., 2018. Growing conditions and morphotypes of African palm weevil
(Rhynchophorus phoenicis) larvae influence their lipophilic nutrient but not their
This research was possible thanks to the scholarship granted from the amino acid compositions. J. Food Compos. Anal. 69, 87–97. https://doi.org/
Brazilian Federal Agency for Support and Evaluation of Graduate Edu­ 10.1016/j.jfca.2018.02.012.
Girard, A.L., Awika, J.M., 2018. Sorghum polyphenols and other bioactive components
cation (CAPES), in the scope of Program CAPES-PrInt, process number
as functional and health promoting food ingredients. J. Cereal Sci. 84, 112–124.
88887.194785/2018-00, mobility number 88887.570753/2020-00. https://doi.org/10.1016/j.jcs.2018.10.009.
Claire Berton-Carabin would like to acknowledge Nantes Métropole and Hackman, R.M., Polagruto, J.A., Zhu, Q.Y., Sun, B., Fujii, H., Keen, C.L., 2008. Flavanols:
Région Pays de la Loire for financial support of this research through her digestion, absorption and bioactivity. Phytochem. Rev. 7, 195–208. https://doi.org/
10.1007/s11101-007-9070-4.
Connect Talent grant. Hassan, S., Imran, M., Ahmad, N., Khan, M.K., 2017. Lipids characterization of
ultrasound and microwave processed germinated sorghum. Lipid. Health Dis. 16,
1–11. https://doi.org/10.1186/s12944-017-0516-4.

11
T.C.B. Gallo et al. Future Foods 9 (2024) 100323

Keuleyan, E., Gélébart, P., Beaumal, V., Kermarrec, A., Ribourg-Birault, L., Le Gall, S., Morel, M.H., Pincemaille, J., Lecacheux, L., Menut, P., Ramos, L., Banc, A., 2022.
Meynier, A., Riaublanc, A., Berton-Carabin, C., 2023. Pea and lupin protein Thermodynamic insights on the liquid-liquid fractionation of gluten proteins in
ingredients: new insights into endogenous lipids and the key effect of high-pressure aqueous ethanol. Food Hydrocoll. 123 https://doi.org/10.1016/j.
homogenization on their aqueous suspensions. Food Hydrocoll. 141 https://doi.org/ foodhyd.2021.107142.
10.1016/j.foodhyd.2023.108671. Morrison, W.R., Smith, L.M., 1964. Preparation of fatty acid methyl esters and
Levy, R., Okun, Z., Shpigelman, A., 2022. Utilizing high-pressure homogenization for the dimethylacetals from lipids with boron fluoride–methanol. J. Lipid Res. 5, 600–608.
production of fermented plant-protein yogurt alternatives with low and high oil https://doi.org/10.1016/S0022-2275(20)40190-7.
content using potato protein isolate as a model. Innov. Food Sci. Emerg. Technol. 75, Neuhoff, V., Arold, N., Taube, D., Ehrhardt, W., 1988. Improved staining of proteins in
102909 https://doi.org/10.1016/j.ifset.2021.102909. polyacrylamide gels including isoelectric focusing gels with clear background at
Li, Y., Ghasemi Naghdi, F., Garg, S., Adarme-Vega, T.C., Thurecht, K.J., Ghafor, W.A., nanogram sensitivity using Coomassie Brilliant Blue G-250 and R-250.
Tannock, S., Schenk, P.M., 2014. A comparative study: the impact of different lipid Electrophoresis 9, 255–262. https://doi.org/10.1002/elps.1150090603.
extraction methods on current microalgal lipid research. Microb. Cell Fact. 13, 1–9. Osagie, A.U., 1987. Total lipids of sorghum grain. J. Agric. Food Chem. 35, 601–604.
https://doi.org/10.1186/1475-2859-13-14. https://doi.org/10.1021/jf00076a038.
Li, Z., Zhao, X., Zhang, X., Liu, H., 2021. Bioactive compounds and biological activities of Price, P.B., Parsons, J.G., 1975. Lipids of seven cereal grains. J. Am. Oil Chem. Soc. 52,
sorghum grains. Foods 10, 1–20. https://doi.org/10.3390/foods10112868. 490–493. https://doi.org/10.1007/BF02640738.
Lin, H., Bean, S.R., Tilley, M., Peiris, K.H.S., Brabec, D., 2021. Qualitative and Proietti, I., Frazzoli, C., Mantovani, A., 2015. Exploiting nutritional value of staple foods
quantitative analysis of sorghum grain composition including protein and tannins in the world’s semi-arid areas: risks, benefits, challenges and opportunities of
using ATR-FTIR spectroscopy. Food Anal. Method. 14, 268–279. https://doi.org/ sorghum. Healthcare 3, 172–193. https://doi.org/10.3390/healthcare3020172.
10.1007/s12161-020-01874-5. Queiroz, V.A.V., Schaffert, R.E., Moreira, A.V., Machado, S., Ribeiro, R., Duarte, S., 2011.
Liu, L., Chen, L., Abbasi, A.M., Wang, Z., Li, D., Shen, Y., 2018. Optimization of O sorgo. Rev. Bras. Milho e Sorgo 10, 180–195.
extraction of polyphenols from Sorghum Moench using response surface Rodrigues, W.A., Magalhães, P.C., Paiva, E., dos Santos, F.G., 2009. Padrões
methodology, and determination of their antioxidant activities. Trop. J. Pharm. Res. eletroforéticos das proteínas de reserva em grãos de sorgo com presença e ausência
17, 619–626. https://doi.org/10.4314/tjpr.v17i4.8. de tanino. Cienc. e Agrotecnologia 33, 1972–1977. https://doi.org/10.1590/s1413-
Malec, M., Le Quéré, J.M., Sotin, H., Kolodziejczyk, K., Bauduin, R., Guyot, S., 2014. 70542009000700047.
Polyphenol profiling of a red-fleshed apple cultivar and evaluation of the color Sosulski, F.W., Imafidon, G.I., 1990. Amino acid composition and nitrogen-to-protein
extractability and stability in the juice. J. Agric. Food Chem. 62, 6944–6954. https:// conversion factors for animal and plant foods. J. Agric. Food Chem. 38, 1351–1356.
doi.org/10.1021/jf500336v. https://doi.org/10.1021/jf00096a011.
Martino, H.S.D., Tomaz, P.A., Moraes, É.A., Conceição, L.L.da, Oliveira, D.da S., Taylor, J., Taylor, J.R.N., Dutton, M.F., De Kock, S., 2005. Glacial acetic acid - A novel
Queiroz, V.A.V., Rodrigues, J.A.S., Pirozi, M.R., Pinheiro-Sant’Ana, H.M., Ribeiro, S. food-compatible solvent for kafirin extraction. Cereal Chem. 82, 485–487. https://
M.R., 2012. Chemical characterization and size distribution of sorghum genotypes doi.org/10.1094/CC-82-0485.
for human consumption. Rev. Inst. Adolfo Lutz 71, 337–344. Vallons, K.J.R., Ryan, L.A.M., Koehler, P., Arendt, E.K., 2010. High pressure–treated
McClements, D.J., 2015. Food Emulsions: Principles, Practices, and Techniques, 3rd ed. sorghum flour as a functional ingredient in the production of sorghum bread. Eur.
CRC Press. https://doi.org/10.1201/b18868. ed. Food Res. Technol. 231, 711–717. https://doi.org/10.1007/s00217-010-1316-5.
Mehmood, S., Orhan, I., Ahsan, Z., Aslan, S., Gulfraz, M., 2008. Fatty acid composition of Wang, Y., Tilley, M., Bean, S., Susan Sun, X., Wang, D., 2009. Comparison of methods for
seed oil of different Sorghum bicolor varieties. Food Chem. 109, 855–859. https:// extracting kafirin proteins from sorghum distillers dried grains with solubles.
doi.org/10.1016/j.foodchem.2008.01.014. J. Agric. Food Chem. 57, 8366–8372. https://doi.org/10.1021/jf901713w.
Melchior, S., Moretton, M., Calligaris, S., Manzocco, L., Nicoli, M.C., 2022. High pressure Xiao, J., Chen, Y., Huang, Q., 2017. Physicochemical properties of kafirin protein and its
homogenization shapes the techno-functionalities and digestibility of pea proteins. applications as buildings blocks of funcional delivery systems. Food Funct. 8,
Food Bioprod. Process. 131, 77–85. https://doi.org/10.1016/j.fbp.2021.10.011. 1402–1413. https://doi.org/10.1039/C6FO01217E.
Meynier, A., Leborgne, C., Viau, M., Schuck, P., Guichardant, M., Rannou, C., Anton, M., Xiao, J., Li, Y., Li, J., Gonzalez, A.P., Xia, Q., Huang, Q., 2015. Structure, morphology,
2014. N-3 fatty acid enriched eggs and production of egg yolk powders: an increased and assembly behavior of kafirin. J. Agric. Food Chem. 63, 216–224. https://doi.
risk of lipid oxidation? Food Chem. 153, 94–100. https://doi.org/10.1016/j. org/10.1021/jf504674z.
foodchem.2013.12.028. Zamora, A., Guamis, B., 2015. Opportunities for ultra-high-pressure homogenisation
Moll, P., Salminen, H., Schmitt, C., Weiss, J., 2021. Impact of microfluidization on (UHPH) for the food industry. Food Eng. Rev. 7, 130–142. https://doi.org/10.1007/
colloidal properties of insoluble pea protein fractions. Eur. Food Res. Technol. 247, s12393-014-9097-4.
545–554. https://doi.org/10.1007/s00217-020-03629-2.

12