Industrial Crops and Products 52 (2014) 136–143

Contents lists available at ScienceDirect

Industrial Crops and Products

journal homepage: www.elsevier.com/locate/indcrop

Valuation of brewer’s spent grain using a fully recyclable integrated

process for extraction of proteins and arabinoxylans
Elsa Vieira a,1 , M. Angélica M. Rocha b,1 , Elisabete Coelho b , Olívia Pinho a,c ,
Jorge A. Saraiva b , Isabel M.P.L.V.O. Ferreira a , Manuel A. Coimbra b,∗
a
REQUIMTE – Deapartamento de Ciências Químicas, Laboratório de Bromatologia e Hidrologia, Faculdade de Farmácia, Universidade do Porto, Rua Jorge
Viterbo Ferreira 228, 450-313 Porto, Portugal
b
QOPNA, Departamento de Química, Universidade de Aveiro, 3810-193 Aveiro, Portugal
c
Faculdade de Ciências da Nutrição e Alimentação, Universidade do Porto, Rua Dr. Roberto Frias, 4200-465 Porto, Portugal

a r t i c l e i n f o a b s t r a c t

Article history: The objective of this work was to design an integrated process to valuate brewer’s spent grain (BSG)
Received 13 May 2013 proteins and arabinoxylans (AX) to be used as food ingredients. For this purpose, a sequential extraction
Received in revised form 30 August 2013 of proteins and AX from BSG with increasing alkali (KOH or NaOH) concentrations of 0.1 M, 0.5 M, and
Accepted 11 October 2013
4 M, was optimized. A ratio of 1:2 (w/v) (weight of BSG by volume of alkali solution) at room temperature
for 24 h was preferred to minimize reagents and energy consumption. To fully integrate the process,
Keywords:
alkaline extracts were acidified to pH 3 with citric acid, to obtain the protein-rich fractions. The AX were
Brewer’s spent grain
recovered by ethanol precipitation and citric acid and ethanol were recycled. This integrated extraction
Proteins
Arabinoxylans
process allowed a yield of 82–85% of the BSG total proteins and 66–73% of total AX with formation of a
Recycling cellulose rich residue almost devoid of nitrogen.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction
Brewer’s spent grain (BSG) is the main by-product gener-
ated during the beer production process. Around 15–20 kg of
BSG is obtained per hectolitre of beer produced. BSG has been
used, almost exclusively, for animal feed, particularly cattle. How-
ever, it is a by-product of great interest, given its nutritional
and functional characteristics for the sectors of biotechnology,
food, and pharmaceutical industry (Mussatto et al., 2006, 2007;
Aliyu and Bala, 2011). BSG proteins and its hydrolysates can be
used as food texture enhancers due to their emulsifying prop-
erties (Celus et al., 2009) and may present immunomodulatory
effects (McCarthy et al., 2013) and antimicrobial activity (Kotlar
et al., 2013). BSG is also considered a major biomass resource
for the production of second generation biofuels, it can be used
for ethanol production using a bioconversion process (Xiros et al.,
2008, 2011).
The spent grains are separated from beer wort by filtration after
the mashing phase. They correspond to the insoluble fraction of the
wort, essentially composed by a lignocellulosic material containing
protein (∼30% on a dry weight basis), lignin (∼28%), hemicelluloses
∗ Corresponding author. Tel.: +351 234 370706; fax: +351 234 370084.
E-mail address: mac@ua.pt (M.A. Coimbra).
1
Both authors contributed equally to the experimental work.
(∼25%), and cellulose (∼17%) (Celus et al., 2006; Treimo et al., 2009).
The major proteins of the BSG are hordeins (A, B, and C), constituting
over 50% of the total amount of proteins, followed by glutenins. The
less abundant protein fraction includes the albumins (about 2%).
The major hemicelluloses of BSG are the arabinoxylans (AX), which
are composed by a backbone of ␤-(1 → 4)-linked xylopyranose, in
part substituted with single units of arabinofuranose at positions 2,
3, or both (Viëtor et al., 1992), which can be extracted under strong
alkali solutions (Mandalari et al., 2005). The AX are considered as
dietary fiber due to their resistance to hydrolysis by digestive tract
enzymes. They can present immunomodulatory activity and the
arabinoxylo-oligosaccharides (AXOS), obtained by partial hydroly-
sis of AX, have been described as prebiotics (Van Craeyveld et al.,
2008; Delcour et al., 2008). Formulations combining AXOS with AX
have been shown to potentiate their prebiotic activity (Broekaert
et al., 2010).
BSG contains 70–80% moisture, thus, a considerable amount of
energy is necessary for its drying. Separation of spent grain com-
ponents can be performed by physical or chemical procedures.
Physical processes, such as pressing and sieving of wet BSG, can
be performed when a suspension in hot water (80 ◦ C) is prepared
and passed through a sieve to obtain two fractions: a proteic frac-
tion (rich in protein and fat and low in fiber) and a fiber fraction
(low in protein and rich in AX (Schwencke, 2006)). Chemical extrac-
tions of BSG proteic fraction in alkaline medium are also described,
namely the alkaline extraction of BSG at pH 11–12 and 104–121 ◦ C

0926-6690/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.indcrop.2013.10.012
 E. Vieira et al. / Industrial Crops and Products 52 (2014) 136–143
where the proteic fraction is then obtained by isoelectric pre-
cipitation (Sohtaroh et al., 1992). However, this process involves
decomposition of proteins due to the drastic conditions of the high
temperature used, resulting in a low yield of proteins and deterio-
ration of product quality. Furthermore, the process requires high
amount of energy for the high temperatures of extraction. The
preparation of protein concentrates after alkaline extraction of BSG
(17%, w/v) with 0.1 M NaOH at 60 ◦ C was also proposed (Celus et al.,
2007). After 60 min of extraction, samples were filtered and the
proteins precipitated by acidification at pH 4 using 2 M citric acid.
The AX of BSG are usually extracted by a well established
sequentially procedure using solutions of KOH with increasing
concentration of 0.5 M, 1 M, and 4 M (Mandalari et al., 2005). Before
the sequential extraction, the BSG was pre-treated with a solution
of 80% ethanol (v/v) under reflux, followed by two water extrac-
tions, and a pronase treatment (Mandalari et al., 2005).
The goal of this work is the valuation of BSG by the development
of an integrated process of simultaneous extraction of proteins
and AX through the use of alkaline reagents directly from the BSG,
without any pre-treatment. Selective precipitation of proteins by
acidification of the medium and addition of ethanol for AX recov-
ery was performed, carrying out a total recycling of the reagents.
These compounds (proteins and AX) can be used as ingredients in
the food industry, in dietetic and/or pharmaceuticals products.
2. Experimental
2.1. Materials
Five lots of 1 kg of brewer’s spent grain (BSG) from different
grists of malt/wheat were supplied weekly by a brewery industry
(Unicer Bebidas SA, Leça do Balio, Portugal), coded from 1 to 5. The
spent grains in a wet-form were used for BSG characterization and
for integrated extraction of proteins and arabinoxylans (AX).
All chemicals were from Sigma–Aldrich and were of, at least,
analytical grade. Eluents used for the chromatographic separations
were ultra-pure water (obtained from a Seral – Seralpur Pro 90
CN water purifying system) and LiChrosolv acetonitrile (Merck,
Darmstradt, Germany). Trifluoroacetic acid (TFA) (Fluka, Seelze,
Germany) was added to both eluents.
2.2. Extraction of proteins and AX from BSG
2.2.1. Optimization of the sequential extraction of proteins and
arabinoxylans
BSG without pre-treatment (100 g) was added to 500 mL of a
0.1 M KOH solution (ratio 1:5, w/v). Sodium metabisulfite (5 mM)
was added as antioxidant. Two different time and temperature con-
ditions were tested concerning contact of the sample with alkaline
reagent. Test conditions were 24 h, room temperature, and occa-
sional shaking (coded as Test A) and contact of the sample with
alkaline reagent for 2 h, 40 ◦ C, with continuous shaking (coded as
Test B). As summarized in Fig. 1, Residue 1 and the Extract 1were
obtained by sedimentation and filtration of the 0.1 M KOH extract.
Each Residue 1 was extracted with 500 mL of a 0.5 M KOH with
5 mM sodium metabisulfite, according to the conditions estab-
lished for Tests A and B. The separation of the residues from the
0.5 M KOH extracts was performed by decantation and filtration
that were designated, respectively, as Residue 2 and Extract 2.
The Residue 2 was extracted with 500 mL of a 4 M KOH with
5 mM sodium metabisulfite, using the respective Tests A and B
conditions. Each residue was separated from the respective 4 M
KOH extract by decantation and filtration that were designated,
respectively, as Residue 3 and Extract 3.
137
The Extracts 1, 2, and 3 were separately acidified to pH 3 with
a saturated solution of citric acid. This value of pH allows the pre-
cipitation of BSG proteins (mostly hordeins and glutenins) that are
separated by decantation and filtration, giving proteic fractions PF1,
PF2, and PF3 fractions, respectively (Fig. 1). The fractions soluble in
citric acid, which contain the AX, were further acidified with 37%
HCl until a pH below 2 was achieved. The AX were recovered by
precipitation with 70% (v/v) aqueous ethanol solutions. The citric
acid present, due to its full protonation at a pH below 2, remained
soluble in this solution. The AX were separated by decantation and
washed with ethanol. The AX fractions obtained were designated,
respectively, by AX1, AX2, and AX3 (Fig. 1). A final wash with water
was done to the Residue 3, obtaining the last fraction of AX coded
as AX4.
A Test C was performed using similar conditions of Test A except
the ratio of BSG to KOH aqueous solution that was 1:2 (w/v), e.g.
100 g of BSG and 200 mL of alkali solvent.
2.2.2. Sequential extraction of proteins and arabinoxylans and
reagents recycling
Two tests of four sequential and consecutive extractions using
NaOH or KOH reagents were also performed using Test C condi-
tions, using a ratio of BSG to alkali solution of 1:2 (w/v), at room
temperature, overnight.
Another set of extractions was performed including the
recycling of ethanol and citric acid. The ethanol was recovered
by distillation of the ethanolic solution resultant from the precip-
itation of polysaccharides, leaving an aqueous solution containing
citric acid and NaCl dissolved. The NaCl was partially removed as a
precipitate upon concentration of the solution, allowing to obtain
an aqueous solution saturated in citric acid. An extraction with
ethanol was also performed to the precipitated NaCl to recover
co-precipitated citric acid. The ethanol was recovered by evapora-
tion and the two aqueous solutions containing the citric acid were
combined and reused as a saturated citric acid solution.
2.3. Analytical methods
Dry matter was evaluated using an oven at 105 ◦ C until constant
weight. Nitrogen content was estimated by the Kjeldahl method
(AOAC, 1975). The protein content was calculated from nitrogen
using a conversion factor of 6.25. Fat was extracted from dried
BSG (2 g) by a soxhlet extractor with 150 mL of n-hexane dur-
ing 5 h. The ash was determined gravimetrically by incineration
of BSG at 650 ◦ C during 24 h in a muffle furnace. Carbohydrate
content was determined by sugar analysis derivatized as alditol
acetates. Monosaccharides were released from polysaccharides by
a pre-hydrolysis in 0.2 mL of 72% H2 SO4 (w/w) for 3 h at room
temperature followed by 2.5 h hydrolysis in 1 M H2 SO4 at 100 ◦ C
(Selvendran et al., 1979). Neutral sugars were analyzed as their
alditol acetates by gas-chromatography-flame ionization detection
(Blakeney et al., 1983; Harris et al., 1988). The hydrolysis was per-
formed in duplicate.
Chromatographic analyses of proteins present in BSG and
Residues 1, 2 and 3 require extraction under reducing conditions
(Celus et al., 2006). For this purpose, the procedure described by
Schmitt et al. (1989) was used. FP1, FP2 and FP3 fractions were
dissolved in water containing 40% acetonitrile (0.01 g in 10 mL),
acidic extracts were injected directly. The reversed phase high per-
formance liquid chromatography (RP-HPLC) analyses were carried
out using a HPLC unit (Jasco, Tokyo, Japan) composed of a low
pressure quaternary pump (Jasco PU-1580 intelligent HPLC pump),
a degasification unit (Jasco DG-1580-54 4-line degasser), a type
7981 Jones Chromatography column heater (Jones Chromatogra-
phy, Hesperia, CA, USA), a type 7725i Rheodyne injector (Rheodyne,
Rohnert Park, CA, USA), and a UV/VIS detector (Jasco UV-970
138 E. Vieira et al. / Industrial Crops and Products 52 (2014) 136–143

Fig. 1. Scheme of the integrated extraction of proteins and arabinoxylans from BSG, using alkaline reagents.

intelligent UV/VIS detector). The column was a Chrompack P 300 RP
(polystyrenedivinylbenzene copolymer, 8 ␮m, 300 Å, 150 × 4.6 mm
i.d.) (Chrompack, Middleburg, The Netherlands). Data acquisition
was accomplished using the Borwin Controller Software, version
1.50 (JMBS Developments, Le Fontanil, France). Chromatographic
conditions were the same used by Silva et al. (2008) for malt hordein
RP-HPLC separations. Daily, before the chromatographic separation
procedures, a 1 mg/mL BSA solution was injected to detect possible
modifications in the HPLC system.
residues, protein, and AX fractions. Descriptive statistics, analyses
of variance (ANOVA, using Type III Sum of Squares) and pairwise
comparisons of mean values (following Tukey’s test at 5% signifi-
cance level), were all performed with SPSS for Windows, v. 18 (SPSS,
Chicago, IL, USA).
3. Results and discussion
3.1. Variability of BSG chemical composition

The moisture content of the five lots of brewer’s spent grain

2.4. Statistical analyses (BSG) studied was 72–73% for all lots, except lot 5, which was 76%
(Table 1). Protein content ranged from 26.4 to 35.4% dry weight.
The averages and standard deviations were calculated for The most abundant protein fractions were B and C hordeins that
each experimental parameter relating the characterization of BSG ranged from 8.9 to 13.3% and from 3.3 to 11.1%, respectively. These
 E. Vieira et al. / Industrial Crops and Products 52 (2014) 136–143
Table 1
Moisture (%, w/w), protein, carbohydrates, fat and ash (%, w/w, in a dry base) of the five lots of BSG.
BSG lots
1 2
Moisture (%) 72 ± 2a 73 ± 1a
Protein (%) 26.4 ± 0.3a 32.2 ± 0.7b
B Hordeins (%) 9.6a 8.9a
C Hordeins (%) 7.4a 11.1b
Carbohydrates (%) 52.0 ± 2.6a 43.4 ± 1.2b
Arabinoxylans (%) 26.2a 20.7b
Glucans (%) 10.4b 7.2a
Cellulose (%) 9.9a 8.9a
Fat (%) 7.5 ± 0.6a 7.5 ± 0.2a
Ash (%) 4.07 ± 0.03a 3.77 ± 0.03a
Values with different letters within the same row differed significantly (p < 0.05).
results are in agreement with those from literature (Santos et al.,
2003; Celus et al., 2006). BSG carbohydrate content in the five dif-
ferent lots analyzed, ranged from 43.4 to 63.7% (in a dry weight
basis). From those, 20.7–26.9% were AX, 8.9–17.8% were cellulose,
and 4.7–13.4% were ␤-glucans and/or residual starch. The fat con-
tent was in the range 6.9–7.9% and the ash content was 3.58–4.39%,
both in dry weight.
A typical chromatogram for BSG RP-HPLC protein separation
is presented in Fig. 2A. Three major peaks from protein fraction
were numbered according to their relative hydrophobicity, being
peak (c) less hydrophobic and peak (a) more hydrophobic. Similar
qualitative profiles were observed for all BSG lots. The chromato-
graphic profiles obtained by Silva et al. (2008) for malt extracts,
using similar chromatographic conditions, were in agreement with
these results. C hordeins eluted with a proportion of approximately
40% acetonitrile and B hordeins eluted with a proportion of approx-
imately 45% acetonitrile. In the present work, the peaks that eluted
with identical proportions of acetonitrile were identified as c and b.
Consequently, these peaks correspond to C and B hordeins, respec-
tively.
Quantitative differences were obtained for mean areas of peaks
(a), (b) and (c) from lots 1 to 5. In some cases, those differences were
statistically significant, namely in peak (b) for the lots 3, 4 and 5, and
in peak (c) for lot 5.The peak areas obtained from the daily injec-
tion of the external standard (1 mg/mL BSA) were not significantly
different and, consequently, the normalization of the samples peak
areas was not performed. Thus, the variation of results for each lot
was evaluated by calculating the respective coefficients of varia-
tion. Good repeatability was obtained since the mean coefficient of
variation was around 10% (n = 5).
3.2. Optimization of the integrated extraction of proteins and AX
BSG samples were submitted to a sequential alkali extrac-
tion using different conditions of temperature/extraction time and
BSG/alkaline solvent ratio: room temperature during 24 h using
a 1:5 solid/liquid ratio (Test A), 40 ◦ C during 2 h using a 1:5
solid/liquid ratio (Test B), room temperature during 24 h using a
1:2 solid/liquid ratio (Test C). The BSG proteins were extracted
by KOH solution and the solid residues were separated from the
alkaline extracts and were also analyzed by RP-HPLC. The proteins
remaining in the residues highlight extensive protein extraction
throughout the integrated process (Fig. 2B and C). The yield of pro-
tein sequential extraction under Tests A, B, and C conditions was
confirmed by evaluation of total protein content of the Residues
1, 2, and 3 and FP1, FP2, and FP3 fractions as presented in Fig. 3A
and B, respectively, expressing the results as g of protein per 100 g
of BSG. No significant differences were observed between results
obtained from Test A and Test B conditions. BSG protein average
139
3 4 5
73 ± 1a 73 ± 1a 76 ± 1b
28.5 ± 0.4a 32.2 ± 0.7b 35.4 ± 0.4b
13.3b 12.2b 11.1a,b
3.6c 3.3c 10.4b
63.7 ± 2.8c 50.8 ± 1.5a 55.1 ± 2.7a
25.3a 24a 26.9a
13.4b 7.8a 4.7c
16.8b 12.0a,b 17.8b
6.9 ± 0.3a 7.5 ± 0.2a 7.9 ± 0.1a
4.06 ± 0.03a 3.58 ± 0.02b 4.39 ± 0.01c
content was 8.65 g of protein/100 g of BSG. The sum of FP1, FP2,
and FP3 fractions was 7.0–7.1 g of proteins extracted from 100 g of
BSG, e.g. 81–83% of the proteins were extracted after the sequen-
tial alkali extraction. The decrease of protein in the BSG residues
was in agreement with the amount of proteins extracted in each
sequential step. Only small losses of proteins were observed, as
Residue 3 still contains 10–13% of unextracted proteins. Concern-
ing Test C, carried out to decrease the BSG/KOH solution ratio to
1:2, using extraction time of 24 h at room temperature, 79–80%
of the proteins were collected in FP1 + FP2 + FP3 fractions, which
contained 6.8–6.9 g of proteins/100 g of BSG. The yield of the first
alkaline extraction was significantly higher when compared with
results from Tests A and B, whereas the yield of the second alkaline
extraction was significantly lower. However, the Residue 3 contains
similar % of unextracted proteins (about 15%) and no significant
differences were observed between the sum of protein content of
fractions FP1, FP2 and FP3 from Tests A, B, and C. These results
show that the higher ratio of sample to solvent promotes a higher
extraction of proteins in the first alkaline extraction. However, this
does not promote a higher global extraction of BSG proteins when
successive extractions of the same matrix are performed.
The AX extracted throughout the integrated process showed
yields between 16.5 and 22.1% for Test A and between 18.7 and
23.1% for Test B, on a dry weight BSG basis. On average, Test A
allowed a yield of extraction of 19 g AX/100 g of BSG and Test B
allowed 18 g, which represent 63% and 60% of recovery from the
total AX present in BSG. By difference of the AX recovered and
those unextracted, it is possible to infer that in Test A 3.7 g/100 g
of BSG were not precipitated in ethanol and Test B this percent-
age raised to 6.2 g/100 g of BSG, possibly remaining in the ethanol
supernatant. The higher percentage of AX unrecovered in Test B
may result from possible depolymerization that could occur at
the higher temperature used, bringing smaller polymers, as was
observed for ␤-glucans (Ookushi et al., 2006). The low molecu-
lar weight polymers formed tend to be highly soluble in ethanol
solutions (Coimbra et al., 1995). Significant differences were only
observed in the first alkali extraction (AX1), showing that a higher
extraction time should be relevant to the increase of the yield of
AX using 0.1 M KOH solutions (Fig. 4). However, when the mass to
the volume ratio of solvent increased from 1:5 to 1:2 under the
same conditions of time and temperature (Tests A and C), the AX
yield decreased in the first alkali extraction (Fig. 4). Also, this trend
was observed for the second extraction (0.5 M KOH). However, the
amount of AX per volume of solvent for these two extractions was
much higher in Test C, namely for AX2, where 12 g/L of AX were
extracted in Test A and 22 g/L were extracted in Test C. For the third
alkali extraction (AX3) and for the water wash (AX4), a higher yield
of extraction was still observed for Test C. Globally, 58% of AX were
extracted from BSG by Test C, corresponding to 17 g of AX/100 g
140 E. Vieira et al. / Industrial Crops and Products 52 (2014) 136–143

Fig. 2. Typical RP-HPLC chromatograms obtained for analyses of protein fraction present in (A) BSG, (B) Residue 1 and (C) Residue 3 using Test C conditions (contact of the
sample with alkali reagent during 24 h, at room temperature and occasional shaking, ratio BSG/KOH solution 1:2).

of BSG, which is highly comparable with the amount extracted by
Tests A and B.
The integrated process presented in this work proposes the use
of a saturated solution of citric acid as the alkaline neutralizing
reagent. This acid has the advantage to be a food grade compound
(E 330), with a higher acidity than most organic acids (pKa1 = 3.15)
and high solubility in water (61.8% w/w at 25 ◦ C), being also solu-
ble in ethanol (38.3% w/w at 25 ◦ C). The selective precipitation of
proteins can be achieved after the acidification of the medium to
pH 3 with citric acid, which remains soluble in this aqueous solu-
tion. In addition, the solubility of citric acid in ethanol solutions
avoids its co-precipitation when ethanol is used to precipitate the
AX. Since citric acid was used to neutralize the alkaline extracts,
part of it is converted into the citrate form, which is much less sol-
uble in ethanol solutions. To convert all citrate into citric acid, the
addition of a strong acid is required. In this process, a concentrated
 E. Vieira et al. / Industrial Crops and Products 52 (2014) 136–143 141

Fig. 3. (A) Mean protein content of BSG and Residues 1, 2, and 3 (n = 5) using Tests A, B, and C conditions. (B) Mean protein content of BSG and fractions FP1, FP2, FP3 (n = 5)
using Tests A, B, and C conditions.

solution of hydrochloric acid (food additive E 507) was used. Addi- recovery of AX without the need of a dialysis step, which signif-
tionally, it should be pointed out that ethanol is a GRAS (generally icantly reduces the duration of the process. In a dialysis process,
recognized as safe) solvent – 184.1293, generally used in the purifi- there is a need for a minimum of three water exchanges, where
cation of polysaccharides by precipitation. This process allows the each dialysis water is in contact with the sample for at least 6 h,
totalizing 18 h. Another disadvantage of this operation is the fact
that the fractions become diluted by the dialysis water, requiring
10 a concentration step that is time consuming and energy costly for
9 Test A the water evaporation.
Test B In general, the use of 40 ◦ C did not increase proteins or AX
8
Test C extraction when compared with at room temperature. Then, the
7 extraction at a lower temperature is preferred due to its lower
g of AX/100 ofg BSG

energy consumption and, consequently, higher economic viability.

6
Moreover, the decrease of the proportion of the alkali extraction
5 solvent is possible without significant loss of protein and AX yield.
The industrial application of this method of extraction using the
4
1:2 ratio consumes 543 kg of KOH, 6 kg of sodium metabisulphite,
3 1114 kg of citric acid, 284 L of 37% HCl, and 20,914 L of ethanol per
2 ton of BSG extracted.

1
3.3. Integrated extraction process and solvent recycling
0
AX1 AX2 AX3 AX4 Final Residue
Three consecutive integrated extractions were performed using
Fig. 4. Mean AX content of AX fractions and residue expressed as g of AX obtained Test C conditions (room temperature, 24 h and 1:2 BSG/alkali solu-
per 100 g of BSG using Tests A, B, and C conditions. tion ratio), using KOH and NaOH solutions. Similar yields were
142 E. Vieira et al. / Industrial Crops and Products 52 (2014) 136–143

Table 2
Mass yield, % of arabinoxylans (AX) and mean of sugar analysis of four extraction cycles (BSG 1–4) obtained using a recycled and integrated extraction procedure with NaOH.

Fraction Mass yield (%) AX in extract as % of BSG AX Mol % Total sugar (mg/g)

BSG1 BSG2 BSG3 BSG4 BSG1 BSG2 BSG3 BSG4 Ara Xyl Man Gal Glc

BSG – – – – – – – – 21 36 1 3 38 522
0.1 M NaOH
FP1 14 17 18 13 1 2 2 1 18 20 4 14 44 81
AX1 2 2 1 1 3 4 3 3 40 48 1 3 8 754
0.5 M NaOH
FP2 14 17 28 17 7 11 16 8 15 30 1 2 53 375
AX2 5 5 5 11 17 16 14 32 35 57 0 3 5 886
4 M NaOH
FP3 6 8 8 15 5 12 5 5 13 40 3 2 41 448
AX3 11 20 12 10 24 34 25 23 23 60 1 3 14 703
AX4 4 4 4 3 6 8 5 6 20 44 1 2 32 789
Final residue – – – – 21 18 21 25 15 22 1 2 61 775

obtained concerning protein extraction with KOH or NaOH: Extract
1 removed 37–50% of the total proteins present in the BSG and
Extract 2 removed 23–28%. In the third alkali extraction, 8–12% of
the total proteins present in the BSG were removed leaving a final
residue containing 12–15% of protein.
The Extracts 1, 2 and 3 were acidified to pH 3 with a satu-
rated solution of citric acid to precipitate the proteins, allowing
their separation from AX, which remained soluble. The analyses
of total nitrogen combined with sugars analysis indicated that FP1
was composed by more than 80% protein, whereas FP2 and FP3
were composed by 60% protein. The protein content in these frac-
tions is in agreement with the results presented by Celus et al.
(2007). The RP-HPLC chromatograms of FP1, FP2, and FP3 fractions
dissolved in water and acetonitrile (3:2, v:v) revealed that these
protein fractions included mainly B and C hordeins, whereas the RP-
HPLC chromatograms of the acidic extracts showed small amounts
of proteins that remained in solution (data not shown).
The AX were recovered by acidification to pH below 2 with
a concentrated solution of hydrochloric acid, followed by addi-
tion of ethanol 70% (v/v). The acidification to pH below 2 allows
the complete protonation of citric acid, which was well soluble in
aqueous solutions of ethanol, allowing the separation of the AX
without citrate co-precipitation. As the salt formed (NaCl or KCl)
is more soluble in ethanol solutions than the AX, the possibility of
its co-precipitation with the AX is minimized. Similar yields were
obtained concerning AX extraction with KOH or NaOH. In the first,
second and third alkaline extraction with these alkaline reagents,
3–7%, 14–17%, and 24–27% of the total AX present in the BSG were
extracted, respectively, in AX1, AX2 and AX3 fractions. In order to
maximize the removal of the AX from the BSG matrix, a final wash of
the Residue 4 was carried out, extracting an extra 6–9% of AX (frac-
tion AX4). The final residue accounted for 15–21% of total BSG AX.
Although the yield was similar for the two alkali reagents, the sol-
ubility of NaCl in ethanol is higher than the solubility of KCl (Pinho
and Macedo, 2005). This is relevant in the alkaline extraction with
4 M, as the NaCl tends to remain soluble in the ethanol solutions and
the KCl tends to co-precipitate with the AX. Thus, the use of NaOH
is preferable because it minimizes the possible occurrence of salt
co-precipitation and, in addition, it is less expensive than KOH. For
these reasons, the integrated process was performed using a total
of four BSG batches using recycled solvents and NaOH as alkaline
reagent.
Table 2 shows the mass yield and sugars composition of each
extract (AX1-4 and FP1-3) of four BSG samples (BSG1-4) using
recycled solvents for BSG2-4. The first BSG batch, performed with
fresh solvents, with 0.1 M of NaOH extracted 4% of the total AX
present in the BSG, being 1% present in the FP1 fraction and the
remaining 3% constituting the fraction AX1 (Table 2). The use of
recycled solvents extracted 6, 5 and 4% of the total AX present in
the BSG, in the second, third and fourth batches, respectively. In the
second alkaline extraction, 24% of the total AX present in the BSG
was extracted with 0.5 M NaOH in the first batch, being 7% present
in the FP2 fraction and the remaining 17% constituting the fraction
AX2. With the use of recycled solvents, were extracted 27, 30, and
40% of the total AX present in the BSG. In the third alkali extraction,
with 4 M NaOH, 29% of the total AX of the BSG was extracted in the
first batch, 5% in the FP3 and the remaining 24% in AX3. The other
three batches, using recycled solvents, extracted 46, 30, and 28% of
the total AX present in BSG. In order to maximize the extraction of
BSG AX, a wash with water of the final residue was always carried
out in each batch (AX4), allowing the extraction of 6, 8, 5, and 6% of
AX in first, second, third, and fourth batches, respectively. The final
residue shows 18–25% of AX, showing that these polysaccharides
require a stronger alkali extraction than the 4 M NaOH performed.
Considering the sum of AX recovered in all AX-rich fractions
(AX1 + AX2 + AX3 + AX4, Table 2), it is observed that the amount
of AX recovered in the second and fourth batches (62 and 64%,
respectively) was higher than those recovered in the first and third
batches (50 and 47%, respectively). Although variability is expected
in this natural product from the same BSG lot, these results indicate
that the AX usually lost in the first batch because of their solubility
in the supernatant of ethanol, seems to be recovered in the second
batch when the ethanol used in the first batch was recycled. The
same was observed between the third and fourth batches when the
ethanol used in the third batch was recycled and used in the fourth
batch. This effect was not observed from the second to the third
batch. It is possible that the AX present in the ethanol recycled from
the first batch were co-precipitated in the second batch, resulting in
a recycled ethanol with lower amount of soluble AX and a concomi-
tant increase in the second batch AX yield. The ethanol recovered
in second batch and used in the third batch behaves like a fresh
ethanol, solubilizing the AX. These AX are those that co-precipitate
in the fourth batch. From the total AX extracted (62–86%) by this
process, 49–64% which corresponds to 33–59 kg from the 68 kg ini-
tial AX are soluble in water, with the remaining 13–25% being in
the protein fraction, corresponding to 9–17 kg from the 68 kg of the
initial AX.
Sugar analysis of all fractions extracted (Table 2) showed that
AX1 extracts had, on average, 48 mol% of xylose (Xyl) and 40 mol%
of arabinose (Ara), indicative of a high branched AX (Xyl/Ara, 1.2).
With the increase of the NaOH to 0.5 M, AX2 presented, on average,
a Xyl/Ara ratio of 1.6, and with 4 M NaOH the AX3 Xyl/Ara ratio
was 2.6, allowing to observe a decrease in the degree of branch-
ing of the extracted AX, as the less branched AX extracted (AX3)
need a strong alkali strength to disrupt the hydrogen bonds that
tight these polysaccharides within the cell wall cellulosic matrix
 E. Vieira et al. / Industrial Crops and Products 52 (2014) 136–143
(Mandalari et al., 2005). The FP fractions contained from 8 to 45%
of carbohydrates, from FP1 to FP3 respectively. The sugar analy-
sis of FP fractions showed a glucose (Glc) content that, on average,
in a molar percentage, varied between 41 and 53%. These glucose
residues could arise from residual starch and/or ␤-glucans (Roos
et al., 2009). The AX fractions had a carbohydrate content between
70 and 89%, with the minimum observed, on average, for AX3, and
the maximum for AX2 (Table 2). These sugar-rich fractions show
that NaCl co-precipitation was avoided, remaining in the super-
natant of the ethanol solution.
4. Conclusions
This integrated process can extract 79–83% of the proteins and
62–86% of the AX from BSG without pre-treatment. The innovation
of the proposed process is the selective and sequential precipita-
tion of the proteins by acidification of the medium with a solution
of citric acid and the precipitation of the AX by addition of HCl fol-
lowed by ethanol. This innovative environmentally clean process
allows the recycling of citric acid and ethanol, which can be used in
the subsequent steps, saving 93% in costs, as well as the recovery
of the sodium chloride. The intellectual property of this integrated
process was assured (Coimbra et al., 2012). The protein-rich and
the AX extracts may have several uses in food and pharmaceutical
products.
Acknowledgements
This work was supported by QREN – DRECHEVALOR –
Project N◦ 5406. Thanks are also due to Fundação para a
Ciência e a Tecnologia (FCT, Portugal), European Union, QREN,
FEDER, and COMPETE for funding the QOPNA Research Unit
(project PEst-C/QUI/UI0062/2013) and REQUIMTE (project PEst-
C/EQB/LA0006/2011). E. Coelho was supported by the FCT
post-doctoral grant SFRH/BPD/70589/2010. The authors gratefully
acknowledge Unicer Bebidas S.A. for providing the brewers’ spent
grain.
References
Aliyu, S., Bala, M., 2011. Brewer’s spent grain: a review of its potentials and applica-
tions. Afr. J. Biotechnol. 10, 324–331.
AOAC, 1975. Official Methods of Analysis, eleventh ed. AOAC, Washington, DC.
Blakeney, A.B., Harris, P.J., Henry, R.J., Stone, B.A., 1983. A simple and rapid prepa-
ration of alditol acetates for monosaccharide analysis. Carbohydr. Res. 113,
291–299.
Broekaert, W., Courtin, C., Damen, B., Delcour J., 2010. Nutriment containing arabi-
noxylan and oligosaccharides. WO 2010/066012 A2.
Celus, I., Brijs, K., Delcour, J.A., 2009. Fractionation and characterization of brewers’
spent grain protein hydrolysates. J. Agric. Food Chem. 57, 5563–5570.
Celus, I., Brijs, K., Delcour, J.A., 2007. Enzymatic hydrolysis of Brewers’ spent grain
proteins and technofuncional properties of the resulting hydrolysates. J. Agric.
Food Chem. 55, 8703–8710.
Celus, I., Brijs, K., Delcour, J.A., 2006. The effects of malting and mashing on barley
protein extractability. J. Cereal Sci. 44, 203–211.
Coimbra, M.A., Ferreira, I.M.P.L.V.O., Rocha, M.A.M, Vieira, E.M.F., Saraiva, J.M.A.,
Pinho, O.M.C., 2012. Integrated process for extracting proteins and arabinoxy-
lans from brewer’s spent grain, WO/2012/06988.
143
Coimbra, M.A., Rigby, N.M., Selvendran, R.R., Waldron, K.W., 1995. Investigation of
the occurrence of xylan–xyloglucan complexes in the cell walls of olive pulp
(Olea europaea). Carbohydr. Polym. 27, 277–284.
Delcour, J., Courtin, C., Broekaert, W., Swennen, K., Verbeke, K., Rutgeers, P.,
2008. Prebiotic preparation. Japanese Laid-Open Patent Publication No. Sho 51-
129776 (U.S. Pat. no. 2008/102162 Al).
Harris, P.J., Blakeney, A.B., Henry, R.J., Stone, B.A., 1988. Gas-chromatographic deter-
mination of the monosaccharide composition of plant-cell wall preparations. J.
Assoc. Off. Anal. Chem. 71, 272–275.
Kotlar, C.E., Ponce, A.G., Roura, S.I., 2013. Improvement of functional and antimicro-
bial properties of brewery byproduct hydrolysed enzymatically. LWT – Food Sci.
Technol. 50, 378–385.
Mandalari, G., Faulds, C.B., Sancho, A.I., Saija, A., Bisignano, G., LoCurto, R., Waldron,
K.W., 2005. Fractionation and characterization of arabinoxylans from brewers’
spent grain and wheat bran. J. Cereal Sci. 42, 205–212.
McCarthy, A.L., O’Callaghan, Y.C., Connolly, A., Piggott, C.O., FitzGerald, R.J., O’Brien,
N.M., 2013. In vitro antioxidant and anti-inflammatory effects of brewers’ spent
grain protein rich isolate and its associated hydrolysates. Food Res. Int. 50,
205–212.
Mussatto, S.-I., Dragone, G., Roberto, I.C., 2007. Ferulic and p-coumaric acids
extraction by alkaline hydrolysis of brewer’s spent grain. Ind. Crops Prod. 25,
231–237.
Mussatto, S.-I., Dragone, G., Roberto, I.C., 2006. Brewers’ spent grain: generation,
characteristics and potential applications. J. Cereal Sci. 43, 1–14.
Ookushi, Y., Sakamoto, M., Azuma, J.-I., 2006. Optimization of microwave-assisted
extraction of polysaccharides from the fruiting body of mushrooms. J. Appl.
Glycosci. 53, 267–272.
Pinho, S.P., Macedo, E.A., 2005. Solubility of NaCl, NaBr, and KCl in water, methanol
ethanol, and their mixed solvents. J. Chem. Eng. Data 50, 29–32.
Roos, A.A., Persson, T., Krawczyk, H., Zacchi, G., Stålbrand, H., 2009. Extraction
of water-soluble hemicelluloses from barley husks. Bioresour. Technol. 100,
763–769.
Santos, M., Jiménez, J.J., Bartolomé, B., Gómez-Cordové, C., del Nozala, M.J.,
2003. Variability of brewer’s spent grain within a brewery. Food Chem. 80,
17–21.
Schmitt, N., Gille, P., Gaucher, J., Montembault, A.,1989. Storage proteins composi-
tion of barley seeds by RP-HPLC. In: Proceedings of 22nd Congress of European
Brewing Convention. IRL Press, Oxford, pp. 187–194.
Selvendran, R.R., March, J.F., Ring, S.G., 1979. Determination of aldoses and uronic-
acid content of vegetable fiber. Anal. Biochem. 96, 282–292.
Schwencke, K.V., 2006. Sustainable, cost-effective, and feasible solutions for the
treatment of brewers’ spent grains. Tech. Q. Master Brew. Assoc. Am. 43,
199–202.
Silva, F., Nogueira, L.C., Gonçalves, C., Ferreira, A.A., Ferreira, I.M.P.L.V.O., Teixeira, N.,
2008. Electrophoretic and HPLC methods for comparative study of the protein
fractions of malts worts and beers produced from Scarlett and Prestige barley
(Hordeum vulgare L.) varieties. Food Chem. 106, 820–829.
Sohtaroh, K., Takashi, K., Takeshi, M., Kobayashi, H., 1992. Process for producing
protein-rich product, fibrous product and/or vegetable oil from brewer’s spent
grain, US005135765A.
Treimo, J., Westereng, B., Horn, S.J., Forssell, P., Robertson, J.A., Faulds, C.B., Waldron,
K.W., Buchert, J., Eijsink, V.G.H., 2009. Enzymatic solubilization of brewer’s spent
grain by combined action of carbohydrases and peptidases. J. Agric. Food Chem.
57, 3316–3324.
Van Craeyveld, V., Swennen, K., Dornez, E., Van de Wiele, T., Marzorati, M., Ver-
straete, W., Delaedt, Y., Onagbesan, O., Decuypere, E., Buyse, J., De Ketelaere, B.,
Broekaert, W.F., Delcour, J.A., Courtin, C.M., 2008. Structurally different wheat-
derived arabinoxylooligosaccharides have different prebiotic and fermentation
properties in rats. J. Nutr. 138, 2348–2355.
Viëtor, R.J., Angelino, S.A.G.F., Voragen, A.G.J., 1992. Structural features of ara-
binoxylans from barley and malt cell wall material. J. Cereal Sci. 15,
213–222.
Xiros, C., Topakas, E., Katapodis, P., Christakopoulos, P., 2008. Evaluation of Fusa-
rium oxysporum as an enzyme factory for the hydrolysis of brewer’s spent grain
with improved biodegradability for ethanol production. Ind. Crops Prod. 28,
213–224.
Xiros, C., Katapodis, P., Christakopoulos, P., 2011. Factors affecting cellulose and
hemicellulose hydrolysis of alkali treated brewers spent grain by Fusarium oxys-
porum enzyme extract. Bioresour. Technol. 102, 1688–1696.