Food Research International 39 (2006) 309–317

www.elsevier.com/locate/foodres

Effect of soy protein subunit composition on tofu quality

V. Poysa *, L. Woodrow, K. Yu
Agriculture and Agri-Food Canada, Greenhouse and Processing Crops Research Centre, 2585 County Road 20, Harrow, Ont., Canada N0R 1G0

Received 27 July 2005; accepted 10 August 2005

Abstract

Tofu was made, using two coagulants, from soybean lines which lacked specific glycinin and b-conglycinin protein subunits and
the quality evaluated to determine the effects of specific protein subunits. The group IIb (A3) glycinin subunit played the major role
in contributing to tofu firmness, regardless of coagulant, while the group IIa (A4) subunit had a negative effect on tofu quality in
2002. Soybeans with the group I (A1A2) subunit resulted in tofu with textural properties about one-third higher, expressed as a per-
cent of Harovinton
s values, than tofu prepared from soybeans without the group I subunit. The individual components of group I
had contradictory effects on GDL tofu quality in 2002, with the A1 subunit having a negative effect and A2 having a major positive
effect. Lack of the a 0 subunit of b-conglycinin increased gel hardness relative to the complete 7S protein.

2005 Elsevier Ltd. All rights reserved.

Keywords: Soybean; Tofu; Glycinin; b-Conglycinin; Protein subunits

1. Introduction
Soybeans have long been a staple of the human
diet in Asia, especially as soymilk or tofu, which is
prepared from soymilk (Liu, 1997; Tay & Perera,
2004; Watanabe, 1997). Soybeans are an inexpensive,
high quality protein source. Consumption of soymilk,
tofu, and other soy foods is increasing in North America
due to an increase in Asian immigrants, greater accep-
tance by the general population, and increased recogni-
tion of the health benefits of soy foods, especially by
those who wish to reduce their consumption of animal
products (Murphy, Chen, Hauck, & Wilson, 1997).
In making tofu, soymilk is heated to cause protein
dissociation and a coagulant is added to form a protein
matrix, which gives the tofu its firmness and hardness
(Liu, 1997). The quantity and quality of the protein in
the seed is the major biochemical component influencing
*
Corresponding author. Tel.: +1 519 738 2251x467; fax: +1 519 738
2929.
E-mail address: poysav@agr.gc.ca (V. Poysa).
soybean quality for tofu and other soy food production
(Poysa & Woodrow, 2002). A soybean genotype from
which firmer tofu can be made at a given water:protein
ratio, compared to a second soybean genotype, can be
used for making a greater volume of tofu of a defined
firmness, relative to the second genotype, making it
more valuable.
The principal storage proteins in soybean are glycinin
(11S) and b-conglycinin (7S), which account for about
70% of the total seed protein (Thanh & Shibasaki,
1976). b-conglycinin is a trimeric glycoprotein which
consists of three sub-units, a, a 0 , and b (Thanh & Shiba-
saki, 1978). Glycinin is a hexamer composed of an acidic
(A) polypeptide linked by a disulfide bond to a specific
basic (B) polypeptide (Staswick, Hermodson, & Nielsen,
1984). Glycinin has five subunits, coded for by five
genes, divided into group I (A1aB2; A1bB1b; A2B1a)
[Gy1, Gy2, Gy3], group IIa (A5A4B3) [Gy4] and group
IIb (A3B4) [Gy5] (Beilinson et al., 2002; Nielsen et al.,
1989; Yagasaki, Kaizuma, & Kitamura, 1996). Glycinin
is the predominant soybean seed storage protein,
accounting for over 50% of the seed protein in most

0963-9969/$ - see front matter
2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodres.2005.08.003
310 V. Poysa et al. / Food Research International 39 (2006) 309–317

varieties. Various spontaneous or induced mutations and 20 genotypes with different glycinin and b-conglyc-
affect the accumulation of glycinin. Many commercial inin subunit compositions (Table 1). Some combinations
tofu-type cultivars lack group IIa; these are also known of subunit nulls were duplicated in more than one line.
as A4 nulls. A line lacking all 11S protein subunits was These provided an additional check that the effects iden-
developed by combining a c-irradiation induced muta- tified here were due to the subunit composition and not
tion deficient for all group I subunits with a group IIa other non-evaluated factors in a specific genotype. These
null cultivar and a group IIb null wild soybean accession genotypes (subunit null lines) were developed by cross-
(Yagasaki et al., 1996) The Japanese variety Keburi ing and backcrossing Harovinton with selections from
lacks the a 0 subunit of b-conglycinin, while researchers a population segregating for the lack of the glycinin sub-
in Japan developed soybean mutants with b-conglycinin units and the a 0 subunit of b-conglycinin, which was
lacking either or both of the a and a 0 subunits (Takah- kindly provided by Dr. N. Kaizuma, Iwate University,
ashi, Banba, Kikuchi, Ito, & Nakamura, 1994, 1996). Morioka, Iwate, Japan. The line lacking all 11S protein
The soy globulins differ in their functional properties, subunits was developed by crossing a c-irradiation in-
especially in gelation, with gels made from glycinin duced mutation line deficient for all group I and IIa sub-
being harder than gels from b-conglycinin (Renkema, units with a group IIb null wild soybean accession
Knabben, & van Vliet, 2001; Rickert, Johnson, & (Yagasaki et al., 1996; Kaizuma, personal communica-
Murphy, 2004; Saio, Kamiya, & Watanabe, 1969; tion). BC1-F5 lines were grown in 2002 and BC1-F6 lines
Watanabe, 1997; Yagasaki, Kousaka, & Kitamura, were grown in 2003. In the development of the lines used
2000). The specific subunits within glycinin (Mujoo, in this study, selection was for protein subunit composi-
Trinh, & Ng, 2003; Tezuka, Taira, Igarashi, Yagasaki, tion, large seed size and high protein content, to resem-
& Ono, 2000; Yagasaki et al., 2000) and b-conglycinin ble the adapted parent. Harovinton (Buzzell, Anderson,
(Mohamad Ramlan et al., 2004; Mujoo et al., 2003) con- Hamill, & Welacky, 1991), a high-yielding, large-seeded,
tribute differentially to protein gelling properties. Both late maturity group I commercial cultivar, with 44%
Yagasaki et al. (2000) and Tezuka et al. (2000) reported protein content on a dry matter (DM) basis, and yellow
the hardness of gels from glycinin decreased in the order seed coat and hilum, suitable for tofu production, is the
group IIa, IIb, and I. The relative order of hardness for quality standard for Canadian tofu type soybeans.
gels made from the b-conglycinin subunits is a, a 0 and b Protein subunits of the soybean lines were identified
(Mohamad Ramlan et al., 2004). after separation by SDS–PAGE using a PhastSysteme
While the effects of isolated 11S and 7S proteins on (Amersham Biosciences, Piscataway, NJ, USA) appara-
tofu quality have been evaluated, there could be signifi- tus and pre-cast gels. All chemicals were of molecular
cant effects of the subunit composition on other seed
components, particularly other proteins. Tezuka et al. Table 1
(2000) evaluated the textural properties of tofu made Seed protein, oil, and sugar content and seed mass of soybean lines
from soy lines lacking different glycinin subunits, while deficient for specific glycinin and b-conglycinin subunits, along with
Harovinton, grown in 2002 in Harrow
Yagasaki et al. (2000) combined soymilk from low gly-
cinin and low b-conglycinin soybeans for a series of Line Absent Protein Oil Sugar Seed mass
11S:7S ratios to evaluate the effects of different subunits Subunits (%) (%) (%) (mg seed
1)
on tofu gels. Further research is required to test the ef- 1 A2 41.7 20.6 10.9 200
fects of different glycinin and b-conglycinin subunit 2 A3 43.0 20.7 10.1 160
combinations across genotypes, environments, and 3 a 0 A2 45.5 18.6 10.1 183
4 a 0 A2 44.8 19.7 10.0 192
growing seasons. To test the effect of soy protein subunit
5 a 0 A3 45.3 20.6 9.7 205
composition on tofu quality, we have developed a series 6 a 0 A3 42.9 21.5 10.5 167
of lines differing in seed storage protein subunit compo- 7 a 0 A4 45.9 19.0 9.9 175
sition. We studied the effects of soy protein subunit com- 8 A1A2 44.4 19.3 10.4 172
position on tofu quality by making two types of tofu 9 A2A3 43.2 19.6 11.9 187
10 A2A4 41.9 20.9 10.6 189
from soybean lines lacking some or all of the 11S gly-
11 a 0 A1A2 42.8 20.3 10.5 214
cinin subunits and the b-conglycinin subunit a 0 . 12 a 0 A1A3 42.4 20.3 10.4 180
13 a 0 A2A4 44.2 19.8 10.3 209
14 a 0 A3A4 41.0 20.5 10.6 193
2. Materials and methods 15 A1A2A3 41.6 19.5 11.0 153
16 A2A3A4 42.8 19.9 10.3 191
17 a 0 A1A2A3 45.7 18.6 9.8 154
2.1. Materials 18 A1A2A3A4 41.2 19.9 11.5 216
19 A1A2A3A4 42.6 19.7 11.1 190
The soybean lines used in this investigation included 20 A1A2A3A4 42.9 19.8 10.9 199
the tofu-type cultivar, Harovinton, with a complete 21 Harovinton 44.6 20.3 9.7 213
LSD 0.43 0.14 0.68 5.6
complement of glycinin and b-conglycinin subunits,
 V. Poysa et al. / Food Research International 39 (2006) 309–317 311

biology or reagent grade (Sigma–Aldrich, St. Louis,
MO, USA). Soybean samples were ground in a Knife-
tece water-cooled grinder (FOSS, Eden Prairie, MN,
USA) equipped with the manufacturer
s sharp
blade.
Proteins were extracted from the ground powder in a
buffer containing 10 mM Tris–HCl, pH 8.0; 2.5% (w/v)
sodium dodecyl sulphate; 10 mM b-mercaptoethanol;
and 0.01% (w/v) bromophenol blue. Pre-cast PhastGel
Gradiente 10–15% polyacrylamide gradient were
pre-treated with urea to modify protein mobility and
separate the b-conglycinin, a- and a 0 -subunits and to
separate the glycinin A4, A1 and A2 subunits. The proto-
col was adapted for use with pre-cast gels following
Fontes, Moreira, Davies, and Neilson (1984). Prior to
use the gels were soaked in a buffer containing 0.1 M
Tris–acetate, pH 6.4 and 4 M urea for 1 h. The gels were
blotted on both sides with lint-free tissue and air-dried
for 15 min. Aliquots of extract containing 10 lg protein
were applied to each lane. Gels were run using PhastGel
SDS Buffer Stripse in the PhastSystem apparatus using
the following run settings 250 V; 10 mA; 3.0 W; 65 Vh.
The gels were stained with Coomassie Blue R-250 and
digitally photographed using a KODAK DC-120 cam-
era (KODAK, Rochester, NY, USA). The presence or
absence of protein subunit bands was assessed qualita-
tively by visually comparing the band pattern with those
documented by Fontes et al. (1984) for soybeans. This
protocol permitted efficient and highly reproducible
identification of genotypes deficient in glycinin subunits
A1 and A2 (group I); A3 (group IIb); and A4 (group IIa);
as well as the, a, a 0 , and b subunits of b-conglycinin
(Fig. 1). This system does not permit the differentiation
of A1a and A1b subunits. Separate runs showed that the
A4 and A5 glycinin subunits always segregated together
in these lines, as would be expected as they are both con-
trolled by the Gy4 gene (Nielsen et al., 1989).
Seed of each line was produced in 2002 in the field at
Harrow, Ontario, Canada on a Fox sandy loam soil, in
single row plots with three replications. The recurrent
parental check cultivar, Harovinton, was included three
times in each replication for improved comparison. A
subset of 10 of the subunit null lines, along with the
check cultivar Harovinton, were grown in 2003 at Har-
row, Ontario, on a Fox sandy loam soil and at Woodslee,
Ontario on a Brookston clay loam soil in two-replicate,
four row yield trials. For all trials plants were grown in
2.5 m rows with 0.45 m interrow spacing. Normal agro-
nomic practices were followed.
2.2. Seed traits and tofu preparation and evaluation
Seed from each plot was analysed for moisture, pro-
tein, and oil content, on a percent DM basis, using a
Grainspec near infra-red (NIR) whole grain analyser
(Foss Electric Multispec Division, York, UK) calibrated
by the manufacturer for soybeans. In addition, total free
sugars content was determined on the Grainspec using
calibrations developed at GPCRC (Woodrow, unpub-
lished data). Seed mass was determined by weighing
two sub-samples of 100 seed per plot and adjusting to
a DM basis. Samples of soybeans were weighed and
soaked in an excess of water for 22 h in an incubator
set at 13 C, drained, then reweighed. The water absorp-
tion factor was determined by dividing the drained
weight of the soaked beans by the initial weight of the
raw soybeans.

Fig. 1. SDS-PAGE electrophoresis of extracts obtained from protein variant soybean lines. The top of the gel corresponds with the origin and the
direction of migration is from top to bottom. LOX indicates lipoxygenase. The electrophoresis protocol used to adequately separate the a 0 and a
subunits results in the A5 subunit running off the gel. From left to right the lanes contain: lane 1, Harovinton with all subunits present; lane 2, lacking
subunits A1, A2; lane 3, lacking subunits A2, A3, A4; lane 4, lacking subunits A1, A2, A3, A4; lane 5, lacking subunits a 0 , A2; lane 6, lacking subunits
a 0 , A3; lane 7, lacking subunits a 0 , A4; lane 8, lacking subunits a 0 , A1, A3..
312 V. Poysa et al. / Food Research International 39 (2006) 309–317
Soymilk and kinugoshi, or silken tofu, were prepared
from each plot following the procedures described by
Mullin et al. (2001). Because tofu is a soy protein gel,
the amount of soy protein used to make the soymilk
can significantly affect tofu yield and quality. To permit
differentiation based on protein quality, rather than
quantity, we evaluated the different genotypes at the
same water to protein ratio (18:1) when making soymilk
and tofu. Two types of tofu were prepared, glucono-d-
lactone (GDL) coagulated tofu and calcium sulphate
dihydrate (CaSO4) coagulated tofu. GDL, an acid coag-
ulant, and CaSO4, a salt coagulant, are the most widely
used coagulants for making kinugoshi-tofu in Japan
(Kohyama, Sano, & Doi, 1995). In brief, duplicate
500 ml portions of the soymilk were heated in double
boilers placed on hotplates. After the temperature of
the milk reached 90 C the top portion of the boiler con-
taining the soymilk was removed and placed directly on
the hotplate. A stirring bar was added and activated
then the soymilk was heated to 98 C and maintained
at that temperature for 4 min with additional intermit-
tent stirring. The coagulating agents, 1.5 g of GDL or
2 g of CaSO4 Æ 2H2O, were each dissolved in 20 ml of dis-
tilled water. The heated soymilk and the GDL coagulant
were poured simultaneously into a 1 l beaker so that
good mixing took place. The mixture was left to stand
for 30 min at ambient temperature after which the gel
was carefully removed into a fine mesh polyester cloth
which was then immersed in cold running water for
1 h. The same procedure was followed for the CaSO4 Æ
2H2O coagulant except that the mixture was stirred
lightly, within 3 s after the addition of the mixture, to
ensure proper dispersion of the coagulant. The tofu
was removed from the cooling water, drained then
weighed to determine tofu yield.
Solids content of the tofu was determined using an
Automatic Volatility Computer, Model AVC 80,
(CEM Corp., Indian Trail, NC), as described by Mullin
et al. (2001). The tofu texture was evaluated for compres-
sion hardness and firmness and penetration hardness and
firmness using an Instrone Texture Measuring System
(Instron Corp., Canton, MA, USA), as described by
Mullin et al. (2001). Compression hardness and firmness
were determined by compressing the sample to 50% of its
original height using a 10 kg load cell with a 57 mm
diameter surface and a speed of 10 mm/min. Penetration
hardness and firmness were measured using an 8 mm
cell. Hardness (fracture break point) was measured in
Newtons (N) and firmness (resistance to compression)
in N/mm (Bourne, 1978).
A small amount of tofu, approximately 200 g, was cut
from the main block and ground to a soft paste with a
spatula for pH and for colour measurement. The tofu
paste was agitated in a 250 ml beaker to remove air bub-
bles. The pH was measured with a Horiba F12 pH meter
fitted with a Horiba S-8721 combination glass electrode
and previously calibrated with standard buffers, allow-
ing the reading to stabilize after 3 min. The paste was
transferred to a Labscan cuvette for colour measure-
ment with a HunterLab Labscan spectrocolourimeter
(Reston, Virginia, US). The paste was agitated to re-
move any air bubbles then placed over the window of
the Labscan and covered. The data relating to colour
and whiteness index were then recorded.
2.3. Statistical analysis
The trial results were analyzed by ANOVA according
to the general linear model procedures of SASe (SAS
Institute Inc., 1997). The 2003 experiment was used to
determine genotype (G), location (L), and GxL interac-
tion effects, while the common genotypes grown at Har-
row in 2002 and 2003 were used to estimate genotype
(G), year (Y), and G · Y interaction effects on seed
and tofu traits. The effects of specific protein subunits
were determined by calculating the average difference
in values of lines with and without the subunit(s). For
example, in 2002 the effect of the presence of the 11S
A1 subunit was determined as the average effect of [Line
2 (A2 null)–Line 8 (A1A2 null)] + [(Lines 3&4 (a 0 A2
null)/2)–Line 11 (a 0 A1A2 null)] + [(Lines 5&6 (a 0 A3
null)/2)–Line 12 (a 0 A1A3 null)]. The significance of these
effects was determined by single degree of freedom
orthogonal contrasts using the GLM procedures of
SAS. Protein subunit effects are presented as a percent
of the values for the check cultivar, Harovinton.
3. Results
3.1. Genotype, location and year effects
In 2002 the soybean lines deficient for one or more
11S and 7S protein subunit had protein contents
between 41.0% and 47.1%, a range which included the
protein content of the check cultivar, Harovinton,
(44.6%). Their oil content (18.3–20.9%), free sugar
content (9.7–11.9%) and seed mass (153–216 mg) also
encompassed the check (20.3%, 9.7%, and 213 mg,
respectively) (Table 1). Similarly, in 2003, the protein
(43.0–49.2%), oil (16.5–19.4%), free sugar (9.7–11.9%),
and seed mass (129–209 mg) encompassed the check
(48.9%, 17.8%, 9.8%, and 208 mg, respectively) (Table 2).
While there were highly significant (p < 0.001) differ-
ences among genotypes for seed traits (Table 3), the lines
evaluated in this study had protein contents similar to
commercial tofu soybeans grown in Ontario (40.5–
47.0% protein), as reported in the Ontario Food Grade
Soybean Performance Trials (OOPSCC, 2005). For the
protein content, seed mass, and water absorption fac-
tors, the location and year effects were generally much
larger than the genotype effects, although there were
 V. Poysa et al. / Food Research International 39 (2006) 309–317 313

Table 2 of coagulant, than the genotype. As the location and

Seed protein, oil, and sugar content and seed mass of soybean lines genotype by location effects were generally non-signifi-
deficient for specific glycinin and b-conglycinin subunits, along with
Harovinton, grown in 2003 in Harrow and Woodslee
cant (p > 0.05) for most tofu quality parameters in
2003, the results have been reported across locations.
Line Absent Protein Oil Sugar Seed mass
Year effects were more important than genotype effects
Subunits (%) (%) (%) (mg seed
1) for tofu yield, while year and genotype by year effects
1 a 0 A2 49.2 16.5 10.2 176 were highly significant (p < 0.001) for several tofu qual-
2 a 0 A3 45.8 19.4 10.3 191 ity traits, so results are presented for each year
3 a 0 A3 49.2 17.8 9.7 175
4 a 0 A4 48.2 17.5 10.2 179
separately.
5 A1A2 46.3 17.3 11.2 158
6 a 0 A3A4 43.4 19.3 10.9 209 3.2. Subunit effects in 2002
7 A2A3A4 46.3 17.5 10.7 173
8 a 0 A1A2A3 47.5 17.3 11.2 129 Tofu could be made from most subunit null lines in
9 A1A2A3A4 43.5 18.0 11.4 162
10 A1A2A3A4 43.0 17.9 11.9 172
the 2002 crop, although the three lines deficient for all
11 Harovinton 48.9 17.8 9.8 208 11S protein subunits (A1A2A3A4) and the line deficient
LSD 0.82 0.41 0.36 11.5 for A1A2 and A3 subunits (group I and group IIb) failed
to coagulate sufficiently for analysis by the Instrone,
regardless of the coagulant used. Interestingly, tofu
no significant (p > 0.05) genotype by location and geno- made from a line lacking the a 0 subunit of b-conglycinin
type by year interactions for these traits (Table 3). The as well as the group I and group IIb subunits coagulated
genotype effects were highly significant (p < 0.001) for sufficiently for analysis. The two lines deficient for all
all tofu traits evaluated, save pH, regardless of year or 11S protein subunits grown in 2003 also failed to coag-
coagulant, and whiteness index of GDL tofu in 2003. ulate sufficiently for analysis. These lines were excluded
For tofu traits, the location and genotype by location ef- from the analyses of protein subunit effects.
fects were generally insignificant (p > 0.05) and, except In 2002 the effects of the presence of individual 11S
for location effects on colour and percent solids of and 7S subunits evaluated in this study were highly sig-
GDL tofu, much less significant than the genotype ef- nificant (p < 0.01) on the compression hardness and
fects (Table 3). Year effects were more important than firmness of GDL coagulated tofu, but only three of
location effects in this study, but year still had a much the five subunits had significant (p < 0.05) effects when
smaller effect on tofu firmness and hardness, regardless CaSO4 was used as the coagulant (Table 4). Group IIa

Table 3
Mean squares for characteristics of seed, and tofu of ten soybean genotypes grown at two locations in 2003 and for two years at Harrow, Ontario
Trait Geno Loc Geno * Loc Residual Geno Year Geno * Yr Residual
Seed
Protein 20.26*** 57.86*** 0.758 0.325 14.57*** 165.7*** 0.817 0.215
Oil 2.571*** 1.688** 0.207 0.074 2.589*** 57.641*** 0.226*** 0.036
Seed mass 23.46*** 234.3*** 1.524 0.636 21.54*** 184.6*** 3.462 1.159
Water absorption factor 0.013*** 0.064*** 0.001 0.002 0.087*** 0.128*** 0.005 0.006
TOFU-GDL
Yield per kg seed DM 0.828*** 0.486 0.051 0.075 0.250* 3.003*** 0.077 0.066
Solids content 0.524*** 1.365** 1.73 0.065 0.349*** 1.871*** 0.132 0.042
Colour (WI) 10.785 144.5*** 3.460 3.322 15.38*** 235.1*** 5.231 2.624
pH 0.005 0.003 0.003 0.006 0.006 0.023 0.005 0.005
Hardness (compression) 0.591*** 0.003 0.061 0.018 0.545*** 0.015 0.061 0.018
Firmness (compression) 0.015*** 0.001 0.002 0.001 0.013*** 0.001 0.001 0.001
Hardness (penetration) 0.033*** 0.001 0.001 0.001 0.029*** 0.001 0.001 0.001
Firmness (penetration) 0.006*** 0.001 0.001 0.001 0.010*** 0.001* 0.001* 0.001
TOFU-CS
Yield per kg seed DM 0.382*** 0.011 0.169* 0.054 0.588*** 2.954*** 0.125 0.043
Solids content 1.517*** 0.003 0.439* 0.087 1.699*** 0.847* 0.304* 0.081
Colour (WI) 41.382*** 27.973* 4.745 2.482 55.92*** 110.2*** 6.050 2.186
pH 0.012 0.117 0.003 0.011 0.013 0.404*** 0.003 0.006
Hardness (compression) 0.275*** 0.079* 0.010 0.006 0.237*** 0.035* 0.051*** 0.004
Firmness (compression) 0.006*** 0.001 0.001 0.001 0.006*** 0.002** 0.001** 0.001
Hardness (penetration) 0.038*** 0.005 0.002 0.001 0.032*** 0.004 0.009*** 0.001
Firmness (penetration) 0.011*** 0.002 0.001 0.001 0.012*** 0.001 0.001** 0.001
*, **, *** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively.
314 V. Poysa et al. / Food Research International 39 (2006) 309–317

Table 4 Table 5
Effect of presence of individual protein subunits on tofu quality for Effect of presence of protein subunit combinations on tofu quality for
lines grown in 2002 in Harrow, as percent of Harovinton values lines grown in 2002 at Harrow, as percent of Harovinton values
Protein Compression Penetration Protein Compression Penetration
Subunit Hardness Firmness Hardness Firmness Subunit Hardness Firmness Hardness Firmness
(N) (N/mm) (N) (N/mm) (N) (N/mm) (N) (N/mm)
GDL Tofu GDL Tofu
Harovinton 1.5093 0.2731 0.3768 0.0720 Harovinton 1.5093 0.2731 0.3768 0.0720
a0
9**
9**
2 ns 0 ns a 0 A1 8* 8* 11** 10 ns
A1
26 ***
19 ***
25 ***
20 ns a 0 A2 16*** 14** 27*** 21 ns
A2 57 *** 46 *** 56 *** 43 ns a 0 A3 40*** 28*** 26*** 23 ns
A3 17 *** 12 ** 2 ns 1 ns a 0 A4 3 ns 2 ns 6*
1 ns
A4
36 ***
24 ***
35 ***
29 ns
A1A2 31*** 27*** 32*** 22 ns
CaSO4 Tofu A1A3 47*** 37*** 27*** 14 ns
Harovinton 0.9402 0.1721 0.3408 0.0579 A2A3 38*** 33*** 37*** 19 ns
a0 1 ns
8 ns
5 ns
5 ns A2A4 21*** 22*** 22*** 14 ns
A1 5 ns
5 ns 0 ns
8 ns A3A4 13** 14** 5 ns 10 ns
A2 18 ** 13 * 37 *** 23 ***
a 0 A1A2 30*** 25*** 29*** 21 ns
A3 0 ns 16** 5 ns 8 ns
a 0 A1A3 58*** 46*** 49*** 41ns
A4
34 ***
30 ***
26 ***
26 ***
a 0 A2A4 34*** 29*** 42*** 42 ns
ns, *, **, *** non-significant, and significant at p < 0.05, 0.01, and a 0 A3A4 26*** 22*** 24*** 14 ns
0.001, respectively. A2A3A4 70*** 60*** 62*** 53 ns
a 0 A1A2A3 63*** 51*** 54*** 34 ns

had a similarly significant negative effect on tofu hard- CaSO4 Tofu

Harovinton 0.9402 0.1721 0.3408 0.0579
ness and firmness, regardless of coagulant. The two
a 0 A1 21*** 4 ns 19**
3 ns
components of group I, A1 and A2, had highly signifi- a 0 A2 14**
1 ns 25*** 11*
cant (p < 0.001) but contradictory effects on GDL tofu a 0 A3 13** 19** 13** 14**
texture. The A1 subunit had a negative effect on GDL a 0 A4
18***
10*
18**
1 ns
tofu hardness and firmness, while the presence of the A1A2 22*** 8 ns 37*** 15**
A2 subunit made the greatest positive contribution to A1A3 47*** 55*** 32*** 15 **
GDL tofu texture, measured as compression firmness A2A3 21*** 9 ns 33*** 10ns
and hardness and penetration firmness. The group IIb A2A4 16 **
18** 11*
4 ns
A3A4 29*** 18 ** 17 ** 10 ns
(A3) subunit also enhanced tofu texture (Table 4).
Among these lines, the presence of 7S subunit a
and a 0 A1A2 25*** 8* 26*** 11*
a 0 A1A3 38*** 21** 42*** 10 ns
11S subunits A1 and A4 had a negative effect on tofu tex-
a 0 A2A4 11** 0 ns 33*** 24***
ture, with lines deficient for these subunits, on average, a 0 A3A4 25*** 16** 22*** 4 ns
forming firmer, harder GDL tofu than the correspond- A2A3A4 47*** 31*** 55*** 33***
ing lines expressing the subunits. The individual sub- a 0 A1A2A3 61*** 54*** 57*** 25***
units had generally less of an effect on the CaSO4 tofu ns, *, **, *** non-significant, and significant at p < 0.05, 0.01, and
quality, although A2 contributed to its penetration hard- 0.001, respectively.
ness and firmness, while A4 had a negative effect on
texture.
As might be anticipated, the removal of two subunits tofu texture (Table 5). Tofu gels made from soybean
had a major negative effect on the hardness and firmness lines lacking two glycinin subunits generally were much
of GDL tofu in 2002 (subunits have a positive effect), less hard and firm than those made from lines lacking
although deficiency for a 0 and group IIb (A4) did not the 7S a 0 subunit in combination with one 11S subunit.
have a significant (p < 0.05) effect on most parameters Especially with the GDL coagulant, tofu made from
measured (Table 5). In contrast to the results with the soybean lacking three or four of the 7S and 11S protein
individual subunits, deficiency for both 7S a 0 and 11S subunits was even less hard and less firm than tofu made
A1 had a modest negative effect on the GDL tofu tex- from seed lacking only two subunits in 2002 (Table 5).
ture. For the CaSO4 tofu, the effects of the subunit com- As might be expected, based on the 11S:7S ratios, tofu
position were generally smaller and less significant than made from the 11S triple null line (A2A3A4 null) was
for the GDL tofu. For pairs of subunit nulls, the 11S softer than tofu made from lines deficient for the 7S a 0
group IIb (A3) subunit generally had the greatest effect subunit in combination with two 11S subunits. Tofu
on firmness and hardness, particularly when combined made from lines deficient for all group I, IIa, and IIb
with 7S a 0 or 11S A1. Deficiency for the paired subunits subunits or for group I and group IIb failed to coagulate
11S group IIa and 7S a 0 had a beneficial effect on CaSO4 sufficiently for textural analysis, regardless of coagulant.
 V. Poysa et al. / Food Research International 39 (2006) 309–317 315

For the CaSO4 tofu, the magnitude of the effects of the these subunits, but the differences were not significant
protein quality changes was less than for GDL tofu, but (p > 0.05) for the other textural traits measured. As
the direction was similar (Table 5). was the case in 2002, the effect of removing the three
11S subunits, A2, A3, and A4, was approximately three
3.3. Subunit effects in 2003 times greater than the effect of removing A3 and A4,
combined with the 7S a 0 subunit (Table 6). In general,
In 2003, the presence of the group IIb (A3) subunit the magnitude of the effects of multiple subunit nulls
generally had a highly significant (p < 0.001) positive ef- on the textural properties of tofu, expressed as a percent
fect on tofu hardness and firmness, regardless of coagu- of the check, Harovinton, was similar, regardless of
lant, while the presence of the group IIa (A4) subunit did coagulant, across the two years (Tables 5 and 6).
not have a significant (p > 0.05) effect on the GDL tex-
ture and had a negative effect on CaSO4 tofu firmness
(Table 6). As was seen with the 2002 crop, tofu made 4. Discussion and conclusions
from soybeans lacking both the group IIb and 7S a 0
or both the group IIb and A1 subunits in 2003 was The results presented are based on a limited set of
greatly reduced in hardness and firmness. CaSO4 coagu- genotypes grown at two sites in one year and over two
lated tofu from soybeans lacking both the group IIa and years at one site. As not all subunit compositions were
7S a 0 subunits had significantly (p < 0.05) improved present in the material, the effects of some individual
hardness and firmness, compared to the check cultivar, or paired subunits may have been confounded by addi-
Harovinton, in 2003. This effect was slightly greater in tional effects of complementary deficiencies in the lines
magnitude than in 2002. GDL tofu produced from lines being compared. As has been reported previously, the
lacking the group IIa and 7S a 0 subunits had greater genotype generally had a larger effect on tofu quality
compression hardness, compared to lines possessing than location or year effects, while year effects were more
important than location effects (Poysa & Woodrow,
2002).
Table 6 As Yagasaki et al. (2000) showed, increasing the
Effect of presence of protein subunits on tofu quality for lines grown in 11S:7S ratio, via the a 0 deficiency, resulted in firmer
2003, mean of two locations (Harrow and Woodslee), as percent of GDL tofu (Table 4). The group IIa (A4) deficiency has
Harovinton values been reported to increase firmness in GDL tofu (Nishi-
Protein Compression Penetration nari et al., 1991) and in Nigari tofu (Murasawa, Sakam-
Subunit Hardness Firmness Hardness Firmness oto, Sasaki, & Harada, 1991), although Tezuka et al.
(N) (N/mm) (N) (N/mm) (2000) found it decreased breaking stress when a magne-
GDL Tofu sium chloride solution was used for coagulating. Our re-
Harovinton 1.435 0.2747 0.36225 0.06185 sults in 2002 support the negative effect of the presence
A3 40*** 32*** 31*** 23*** of group IIa on GDL tofu texture, but in 2003 the effect
A4
6 ns
3 ns 4 ns 6 ns
of removing this subunit was non-significant. Our results
a 0 A2 27*** 23*** 31*** 21*** both years also confirmed that presence of group IIb (A3)
a 0 A3 73*** 58*** 47*** 28***
plays a major role in tofu firmness (Nakamura, Utsumi,
a 0 A4
17*
12 ns
8 ns
7 ns
Kitamura, Harada, & Mori, 1984). Tezuka et al. (2000)
A1A2 34*** 32*** 28*** 16** and Tezuka, Yagasaki, and Ono (2004), using a magne-
A1A3 38*** 32*** 31*** 15**
sium chloride coagulant, found group I to have the great-
a 0 A3A4 23** 20** 23*** 16** est effect on tofu firmness, with tofu prepared from
A2A3A4 71*** 60*** 70*** 63*** soybeans with group I having breaking stress values twice
CaSO4 Tofu
as high as those prepared from soybeans without group I
Harovinton 1.00025 0.18392 0.3548 0.05222 (Tezuka et al., 2000). We also found the presence of group
A3 43*** 30*** 41*** 13* I to have a major beneficial effect on tofu quality. Both
A4 0 ns
18** 3 ns
20** types of tofu, prepared from soybeans with group I both
a 0 A2 28*** 15** 30*** 14* years had textural properties about one-third higher, ex-
a 0 A3 80*** 69*** 77*** 60*** pressed as a percent of Harovinton
s values, than tofu pre-
a 0 A4
19***
19***
18***
13* pared from soybeans without the group I subunit. Very
A1A2 35*** 20*** 36*** 6 ns interestingly, we found that the individual components
A1A3 36*** 33*** 37*** 31*** of group I had contradictory effects on GDL tofu quality
a 0 A3A4 25*** 11* 22*** 1 ns in 2002, with the A1 subunit having a negative effect and
A2A3A4 49*** 37*** 57*** 39*** A2 having a major positive effect. The genotypes evalu-
ns, *, **, *** non-significant, and significant at p < 0.05, 0.01, and ated in 2003 did not permit a resolution of group I into
0.001, respectively. its component subunits. The differences in relative
316 V. Poysa et al. / Food Research International 39 (2006) 309–317
magnitude of the effect of group I on tofu quality in this
study and Tezuka et al. (2000) could be due to the type
of coagulant used, the means of measuring textural qual-
ity, or the use of different genotypes. Differences in protein
quantity in seed (35–41% for Tezuka et al. (2000) com-
pared to 41–49% for the current study) also could have
influenced the effect of specific subunits, despite standard-
izing for protein content within each study. Consistent
with the determination by Tezuka et al. (2001, 2004), that
the group IIa subunit had the lowest breaking stress, in
our study soybean lines which possessed only the group
IIa 11S protein, as well as the complete b-conglycinin,
were unable, under our tofu preparation procedures, to
form a tofu cake capable of being measured.
Yagasaki et al. (2000), using a series of glycinin null
lines in the Tamahomare background, reported that
their soymilk storage modulus, which was well corre-
lated with tofu gel breaking stress, decreased in the or-
der all glycinin subunits > I and IIa > I and IIb >
IIa > IIa and IIb > I > IIb. Our results for GDL tofu
hardness differ from this, particularly in suggesting that
group I and IIb tofu can be harder than tofu made with
all glycinin subunits, especially if combined with the 7S
a 0 null, and in the failure of our group IIa line to
coagulate.
Among the b-conglycinin subunits, lack of the a 0
subunit increased gel hardness relative to the complete
7S protein (Mohamad Ramlan et al., 2004), which is
consistent with our 2002 results (Table 4) which
showed a significant negative effect of the a 0 subunit
on GDL tofu compression hardness and firmness.
While our results suggest interactions when the 7S a 0
null is combined with specific 11S nulls, reducing the
effect of the A2 null while increasing the effect of
the A3 null, especially with GDL tofu, experiments de-
signed to specifically examine these interactions would
need to be conducted before a clearer picture of these
interactions can be developed.
In this study two coagulants, GDL, an acid, and
CaSO4, a salt, were used to evaluate the effects of pro-
tein subunit composition on tofu quality. Our results
suggest there is a difference in the effect of specific sub-
units depending on the coagulant used. The 11S IIa
and 7S a 0 null combination consistently had a greater
beneficial effect on quality of tofu formed with CaSO4
than with GDL. While this might be related to the faster
gelation rate of CaSO4 compared with GDL (Kohyama
et al., 1995), further rheological studies would be re-
quired to determine the cause of this differential re-
sponse. The effects of most glycinin nulls tended to be
greater with GDL tofu than with CaSO4 tofu, although
the relative importance of the subunit composition re-
mained similar across coagulants.
Based on these results, selection for soybean geno-
types lacking both the 7S a 0 and the 11S group IIa
subunits and with an increased level of 11S group I
and IIb subunits should result in soy cultivars which
produce firmer tofu. This would permit the industry
to use less soy, relative to the use of other soy culti-
vars, to make the same quantity of tofu of a desired
firmness. This should provide economic benefits to
the tofu industry.
Changing the protein composition of soy by eliminat-
ing the 11S protein clearly had a very major effect on the
protein gel forming ability, as the 11S null lines did not
coagulate sufficiently to form a tofu cake. Our results
also show that changing the subunit composition of
the 11S protein can have a major effect on the gel form-
ing ability of soy protein. These altered protein compo-
sitions might also result in changed protein functionality
for other soy protein uses.
Acknowledgement
The authors would like to thank the Ontario Soybean
Growers for financial support and Bob Armstrong,
Doug Jessop, and Dale Anderson for excellent technical
assistance.
Appendix A. Supplementary data
Supplementary data associated with this article can
be found, in the online version, at doi:10.1016/
j.foodres.2005.08.003.
References
Beilinson, V., Chen, Z., Shoemaker, R. C., Fischer, R. L., Goldberg,
R. B., & Nielsen, N. C. (2002). Genome organization of glycinin
genes in soybean. TAG, 104, 1132–1140.
Bourne, M. C. (1978). Texture profile analysis. Food Technology, 42,
62–72.
Buzzell, R. I., Anderson, T. R., Hamill, A. S., & Welacky, T. W.
(1991). Harovinton soybean. Canadian Journal of Plant Science, 71,
525–526.
Fontes, E. P. B., Moreira, M. A., Davies, C. S., & Neilson, N. C. (1984).
Urea-elicited changes in relative electrophoretic mobility of certain
glycinin and b-conglycinin subunits. Plant Physiology, 76, 840–842.
Kohyama, K., Sano, Y., & Doi, E. (1995). Rheological characteristics
and gelation mechanism of tofu (soybean curd). Journal of
Agricultural and Food Chemistry, 43, 1808–1812.
Liu, K. (1997). Soybeans: Chemistry, technology, and utilization. New
York: Chapman & Hall, 532 pp.
Mohamad Ramlan, B. M. S., Maruyama, N., Takahashi, K., Yagasaki,
K., Higasa, T., Matsumura, Y., et al. (2004). Gelling properties of
soybean b-conglycinin having different subunit compositions. Bio-
science Biotechnology and Biochemistry, 68, 1091–1096.
Mujoo, R., Trinh, D. T., & Ng, P. (2003). Characterization of storage
proteins in different soybean varieties and their relationship to tofu
yield and texture. Food Chemistry, 82, 265–273.
Mullin, W. J., Fregeau-Reid, J. A., Butler, G. M., Poysa, V.,
Woodrow, L., Jessop, D. B., et al. (2001). An interlaboratory test
 V. Poysa et al. / Food Research International 39 (2006) 309–317
of a procedure to assess soybean quality for soymilk and tofu
production. Food Research International, 34, 669–677.
Murasawa, H., Sakamoto, A., Sasaki, H., & Harada, K. (1991). The
effect of a glycinin subunit on tofu-making. In Proceedings of the
International Conference on Soybean Processing and Utilization (pp.
53–57).
Murphy, P. A., Chen, H. P., Hauck, C. C., & Wilson, L. A. (1997).
Soybean protein composition and tofu quality. Food Technology,
51(3), 86–88, 110.
Nakamura, T., Utsumi, S., Kitamura, K., Harada, K., & Mori, T.
(1984). Cultivar differences in gelling characteristics of soybean
glycinin. Journal of Agricultural and Food Chemistry, 32, 647–651.
Nielsen, N. C., Dickinson, C. D., Cho, T., Thanh, V. H., Scallon, B. J.,
Fischer, R. L., et al. (1989). Characterization of the glycinin gene
family is soybean. Plant Cell, 1, 313–328.
Nishinari, K., Kohyama, K., Zhang, Y., Kitamura, K., Sugimoto, T.,
Saio, K., et al. (1991). Rheological study of the effect of A5 subunit
on the gelation characteristics of soybean proteins. Agricultural and
Biological Chemistry, 55, 351–355.
Ontario Oil and Protein Seed Crops Committee (OOPSCC) (2005).
Ontario food grade soybean performance trials. <http://
www.oopscc.org/fg2005t3.html>. Accessed 23.06.05.
Poysa, V., & Woodrow, L. (2002). Stability of soybean seed compo-
sition and its effect on soymilk and tofu yield and quality. Food
Research International, 35, 337–345.
Renkema, J. M. S., Knabben, J. H. M., & van Vliet, T. (2001). Gel
formation by b-conglycinin and glycinin and their mixtures. Food
Hydrocolloids, 15(4-6), 407–414.
Rickert, D. A., Johnson, L. A., & Murphy, P. A. (2004). Functional
properties of improved glycinin and b-conglycinin fractions.
Journal Food Science, 69(4), 303–311.
Saio, K., Kamiya, M., & Watanabe, T. (1969). Food processing
characteristics of soybean 11S and 7S proteins. Agricultural and
Biological Chemistry, 33, 1301–1308.
SAS Institute Inc. (1997). SAS/STAT Software: changes and enhance-
ments through release 6.12, Cary, NC. SAS Institute Inc.
317
Staswick, P. E., Hermodson, M. A., & Nielsen, N. C. (1984).
Identification of the cystines which link the acidic and basic
components of glycinin subunits. Journal of Biological Chemistry,
259, 13431–13435.
Takahashi, K., Banba, H., Kikuchi, A., Ito, M., & Nakamura, S.
(1994). An induced mutant line lacking the a-subunit of b-
conglycinin in soybean (Glycine max (L.) Merrill). Breeding
Science, 46, 65–66.
Takahashi, K., Mizuno, Y., Yumoto, S., Kitamura, K., & Nakamura,
S. (1996). Inheritence of a-subunit deficiency of b-conglycinin in
soybean (Glycine max (L.) Merrill) line induced by (c)-ray
irradiation. Breeding Science, 46, 251–255.
Tay, S. L., & Perera, C. O. (2004). Physiochemical properties of 7S and
11S protein mixtures coagulated by glucono-d-lactone. Journal of
Food Science, 69(4), 139–143.
Tezuka, M., Taira, H., Igarashi, Y., Yagasaki, K., & Ono, T. (2000).
Properties of tofus and soy milks prepared from soybeans having
different subunits of glycinin. Journal of Agricultural and Food
Chemistry, 48, 1111–1117.
Tezuka, M., Yagasaki, K., & Ono, T. (2004). Changes in characters of
soybean glycinin groups I, IIa, and IIb caused by heating. Journal
of Agricultural and Food Chemistry, 52, 1693–1699.
Thanh, V. H., & Shibasaki, K. (1976). Heterogeneity of beta-
conglycinin. Biochimica et Biophysica Acta, 439, 326–338.
Thanh, V. H., & Shibasaki, K. (1978). Major proteins of soybean
seeds: subunit structure of b-conglycinin. Journal of Agricultural
and Food Chemistry, 26, 692–695.
Watanabe, T. (1997). Science of tofu. Kyoto, Japan: Food Journal Co.
Ltd.
Yagasaki, K., Kaizuma, N., & Kitamura, K. (1996). Inheritance of
glycinin subunits and the characterization of glycinin molecules
lacking the subunits in soybean (Glycine max (L) Merr.). Breeding
Science, 46, 11–15.
Yagasaki, K., Kousaka, F., & Kitamura, K. (2000). Potential improve-
ment of soymilk gelation properties using soybean with modified
protein subunit compositions. Breeding Science, 50, 101–107.