J Sci Food Agric 1998, 77, 393È398

Eþect of Temperature and Time during Steam

Treatment on the Protein Quality of Full-fat
Soybeans from Diþerent Origins
G X Qin,1 M W A Verstegen2 and A F B van der Poel2*
1 Department of Animal Science, Jilin Agricultural University, Changchun 130118, PeopleÏs Republic of
China
2 Animal Nutrition Group, Department of Animal Sciences, Wageningen Institute of Animal Sciences
(WIAS), Wageningen Agricultural University, Marijkeweg 40, 6709 PG Wageningen, The Netherlands
(Received 18 December 1996 ; revised version received 14 November 1977 ; accepted 27 November 1997)

Abstract : Soybeans (Glycine max) of Argentinian and Chinese origin were steam-
processed at 102¡C, 118¡C and 136¡C for various durations with the use of a
laboratory-scale steam toaster. Samples of raw and processed soybeans were
analysed for dry matter (DM), crude protein (CP), trypsin inhibitor activity
(TIA), protein dispersibility index (PDI) and reactive lysine content. Chinese raw
soybeans showed higher levels of CP, TIA and PDI, and a lower level of reactive
lysine than Argentinian raw soybeans (366 vs 357 g kg~1, 20.6 vs 15.2 mg g~1,
87.6 vs 85.6% and 17.4 vs 19.4 g kg~1, respectively). Both the TIA and PDI of
the two soybean samples were decreased following a logarithmic pattern with the
lengthening of the heating time when beans were steam heated at various tem-
peratures. The reduction rates for TIA and PDI, however, were di†erent between
the two origins of soybeans. Chinese soybeans required a longer time or a higher
temperature to reduce their TIA to a safe level in comparison with Argentinian
soybeans. In the case that the beans were heated at 136¡C, the di†erence in PDI
between Chinese and Argentinian soybeans was smaller. It is concluded that the
two soybeans origins need di†erent processing conditions to improve their
protein properties. ( 1998 SCI.

J Sci Food Agric 77, 393È398 (1998)

Key words : full-fat soybeans ; antinutritional factors ; soybean origin ; thermal

processing ; protein quality

INTRODUCTION soybeans from di†erent origins respond di†erently to a

given processing technology or to treatment conditions
Raw soybeans have to be processed prior to inclusion in (Friedman et al 1991). A proper processing procedure
diets for monogastric and immature ruminant animal needs to inactivate ANFs to an acceptable level without
due to the presence of certain antinutritional factors reducing the availability of nutrients (Van der Poel et al
(ANFs) (Monari 1993). The contents of trypsin inhibi- 1995 ; Voragen et al 1995). The responses to a treatment,
tors (TI) (Sangle et al 1993) and/or the chemical com- however, may vary depending on the origins of soy-
position (Zarkadas et al 1994) of soybeans are beans with regard to both the inactivation of ANFs and
inÑuenced by genotype (Herkelman et al 1990, 1992 ; the change in the availability of nutrients. Thus, each
Qin and Chen 1995), geographical region of their batch of soybeans requires its own optimum treatment
growing (Zhao et al 1991 ; Fu and Lu 1992) and their procedure in terms of the improvement of nutritional
maturity (McGrain et al 1992). It was found that the value.
In the present experiment, trypsin inhibitor activity
* To whom correspondence should be addressed. (TIA), Ñuorodinitrobenzene reactive lysine (FDNB) and
393
( 1998 SCI. J Sci Food Agric 0022È5142/98/$17.50. Printed in Great Britain
394 G X Qin, M W A V erstegen, A F B van der Poel

protein dispersibility index (PDI) were determined for TABLE 1

raw and steam toasted Argentinian and Chinese soy- E†ects of steam toasting on trypsin inhibitor activity (TIA)
beans, to evaluate the e†ects of di†erent steam heating protein dispersibility (PDI) and Ñuorodinitrobenzene reactive
temperatures and durations on the protein quality of lysine (FDNB) of Argentinian and Chinese commercial soy-
beansa
the two soybean origins.
T reatment T IA PDI FDNB DM
(¡C/min) (mg g~1) (%) (g kg~1) (g kg~1)
MATERIALS AND METHODS
Argentinian
Soybean origins soybeans :
Raw 15.2 85.6 19.3 895.7
Commercial soybeans (Glycine max) from Argentinian 100/5.0 7.3 55.4 18.4 947.7
and China were purchased from Schouten Industries 100/10.0 4.8 48.3 16.4 948.5
(Giessen, The Netherlands) and Food & Oil Import & 100/20.0 2.5 37.5 18.6 947.0
100/40.0 1.6 25.0 18.9 946.2
Export Company (Jilin Province, China) respectively.
118/2.5 3.7 23.9 18.3 942.8
The dry matter contents of the Argentinian and Chinese 118/5.0 2.0 15.8 18.9 934.8
soybeans were 896 g kg~1 and 924 g kg~1, and their 118/10.0 1.2 12.0 17.8 934.0
crude protein contents were 357 g kg~1 and 366 g kg~1, 118/20.0 0.4 9.8 18.0 931.9
respectively. 136/1.5 1.0 9.7 17.5 938.8
136/2.5 0.8 9.5 18.0 936.7
136/5.0 0.3 9.5 16.7 930.6
Treatment of the soybeans 136/10.0 0.0 10.4 16.2 920.0
Chinese
Soybeans of both origins were steam-toasted at di†erent soybeans :
combinations of temperature and heating duration Raw 20.6 87.6 17.4 923.5
(Table 1). The beans were not crushed before the 100/5.0 13.1 64.8 18.6 949.8
thermal processing. 100/10.0 7.4 52.3 18.8 947.3
A laboratory-scale steam toaster was used for the 100/20.0 4.3 42.5 17.6 946.3
processing of whole soybeans in this experiment as 100/40.0 2.4 29.9 17.8 943.6
described by Van der Poel et al (1990). After heating the 118/2.5 8.9 47.1 18.4 945.7
empty toaster to the designed temperature, batches of 2 118/5.0 4.4 29.5 17.7 947.4
118/10.0 2.2 17.2 15.5 940.9
kg soybeans were metered through the inlet sluice. A
118/20.0 0.7 9.9 18.3 930.8
feed rate was chosen providing a single layer of beans 136/1.5 2.6 13.0 17.9 929.8
on the belt conveyor. The speed of the belt conveyor 136/2.5 1.9 11.3 17.7 937.9
was adjusted for each batch to give the required 136/5.0 0.5 9.4 16.9 920.7
residence time. Temperature tolerances were within 136/10.0 0.0 9.7 16.2 918.8
0.3¡C. Immediately after steam processing, the soybeans
were put into screen bottom trays and spread in one a Data in g kg~1 or mg g~1 dry matter.
layer for cooling down for 0.5 h. After cooling, they
were air-dried in a thermo-controlled drying chamber
(36 h at 40¡C). The dried soybeans were then milled to et al (1974) as described by Van Oort et al (1989). PDI
pass through a 1 mm screen (for proximate and PDI was determined with the standard procedure issued by
analysis) or through a 0.5 mm screen (for TIA and AOCS (1964). The FDNB reactive lysine was measured
FDNB reactive lysine analysis), and stored in a cooling according to the method reported by Carpenter (1960)
chamber (4¡C) until analysis. as modiÐed by Booth et al (1971). All parameters were
analysed in duplicate.
Analytical methods
Data processing
Dry matter (DM) content of the samples was measured
by standard method (103¡C for 4 h). Nitrogen was Curve Ðtting was conducted in a logarithmic model for
determined by the Kjeldahl method (ISO 1979), and the the data of TIA and PDI at various treatment condi-
crude protein content (CP) was calculated by N ] 6.25. tions with Slide Write Version 6.0. The model was as
The protein quality of raw and processed soybeans was follows :
estimated by analysing the trypsin inhibitor activity
Y \ a ] a ] ln(x)
(TIA), protein dispersibility index (PDI) and Ñuoro- 0 1
dinitrobenzene (FDNB) reactive lysine. TIA was deter- where Y is TIA (mg g~1) or PDI (%) ; x is heating dura-
mined according to the modiÐed procedure of Kakade tion (min). A linear correlation analysis was carried out
E†ect of steam treatment on protein quality of soybeans 395

between relevant parameters with the software of SAS

(1990).

RESULTS

The results of TIA, PDI, FDNB reactive lysine and DM

of the samples of the soybeans treated at various condi-
tions are presented in Table 1.
The results indicated that TIA, PDI, FDNB reactive
lysine and CP in raw soybeans are di†erent between the
two origins. The level of TIA, PDI and CP of Chinese
raw soybeans are 5.4 mg g~1, 1.9% unit and 9 g kg~1
higher than that of Argentinian raw soybeans, respec-
tively. The FDNB reactive lysine in Argentinian raw
soybeans is 2 g kg~1 higher than in Chinese raw soy-
beans. Fig. 2. PDI variation of Argentine (Ar) and Chinese (Ch) soy-
beans heated at di†erent temperatures (| Ar 100¡C ; > Ch
When heated at various temperatures, the TIA and 100¡C ; L Ar 118¡C ; … Ch 118¡C ; K Ar 136¡C ; = Ch
PDI in both Argentinian and Chinese soybeans 136¡C)
decreased evidently with an increase of heating dura-
tion. These parameters, however, showed di†erent pat- values of Argentinian soybeans. When the soybeans
terns for the two origins of soybeans. The residual TIA were heated at the same temperature for a longer time
in the Chinese soybeans appeared to be higher com- (eg 118¡C for 20 min and 136¡C 10 min), the di†erence
pared to TIA in Argentinian soybeans, especially for in the PDI value between the two soybean origins dis-
samples heated for short time at various temperatures. appeared (see Figs 1 and 2). As the PDI was reduced to
The di†erences in the relative reduction of TIA between a level of about 10%, it remained unchanged, even with
the soybean origins, however, diminished with the pro- more severe heating treatments.
longation of heating (see Figs 1 and 2). In Table 2, the curve Ðtting equations for TIA and
The di†erences between the soybean origins in PDI are given. The TIA as well as the PDI variations of
responses to steam heating are more pronounced for the the two origins of soybeans after di†erent treatments
PDI value than for any other parameters. With regard
to PDI, Chinese soybeans were found to be less sensi- TABLE 2
tive to the higher temperature, but not to the duration The values of a and a in Ðtting equationa for TIA and PDI
heating than Argentinian soybeans. For example, the 0 1
and the determination coefficients (R2) of the equations
PDI values of Chinese soybeans heated at 118¡C for 2.5,
5.0 and 10 min, and at 136¡C for 1.5 min, were 1.97, a a R2
0 1
1.87, 1.43 and 1.34 times as high as the corresponding
TIA :
1000C :
Argentinian soybeans 11.47 [2.80 0.9816
Chinese soybeans 20.25 [5.08 0.9714
1180C :
Argentinian soybeans 4.85 [1.54 0.9794
Chinese soybeans 11.61 [3.87 0.9685
1360C :
Argentinian soybeans 1.25 [0.55 0.9936
Chinese soybeans 3.13 [1.44 0.9835
PDI :
1000C :
Argentinian soybeans 80.55 [14.73 0.9926
Chinese soybeans 91.09 [16.51 0.9989
1180C :
Argentinian soybeans 28.43 [6.69 0.9592
Chinese soybeans 60.83 [17.86 0.9830
1360C :
Fig. 1. TIA variation of Argentine (Ar) and Chinese (Ch) soy- Argentinian soybeans 9.36 0.33 0.6794
beans heated at di†erent temperatures (| Ar 100¡C ; > Ch Chinese soybeans 13.21 [1.82 0.8999
100¡C ; L Ar 118¡C ; … Ch 118¡C ; K Ar 136¡C ; = Ch
136¡C) a Equation : Y \ a ] a ] ln(x).
0 1
396 G X Qin, M W A V erstegen, A F B van der Poel

are well described by the curve Ðtting equations (Table longation of heating irrespective of the heating tem-
2). perature. This supported earlier observations in
The variation of the FDNB reactive lysine was not soybean heat processing reported by Collins and Beaty
quite regular at the heating temperatures of 100¡C and (1980), Liener and Thomlinson (1981), Qin et al (1996)
118¡C for both soybean origins. But, it was observed and Monari (1993).
that reactive lysine of these two origins of soybeans The results of the TIA determination in the present
decreased with the prolongation of heating time when study indicate that there are substantial di†erences
soybeans were heated at 136¡C. between Argentinian and Chinese soybeans in responses
A correlation analysis, based on the pooled data of to steam toasting. Heated at a low temperature and/or
the two soybean origins, showed that there was a high for a short time, the Argentinian soybeans were more
signiÐcant correlation between TIA and PDI sensitive to a reduction of TIA, but, with the increase of
(R \ 0.9421, P \ 0.01). FDNB active lysine is also cor- heating temperature and/or heating duration, this sensi-
related to TIA and PDI (R \ 0.3301, P \ 0.10 ; tivity decreased. At the most severe heating condition
R \ 0.3707, P \ 0.10, respectively), but, the correlations (136¡C for 10 min), the TIA level in soybeans of both
were only poorly described. origins decreased below detection limit. The di†erences
The parameters and determination coefficients of the in TIA between the two soybean origins varied with the
curve Ðtting equations for the data of TIA and PDI of treatment conditions. The more severe the heating treat-
the soybeans treated at 100¡C, 118¡C and 136¡C are ment employed, the smaller the di†erence between
presented in Table 2. them.
For an adequate processing, the heating conditions
should not be too severe. Chang et al (1984) and
Monari (1993) suggested that the accepted TIA level of
DISCUSSION an adequately processed soya product should be below
4 mg g~1. Based on this standard, proper heating dura-
The TIA levels in the two soybean origins as deter- tions could be calculated with the Ðtted equations for
mined in present experiment are within the normal di†erent heating temperatures (Table 3). Apparently,
range compared with the literature reviewed by Qin and heating time is 10 and 5 min longer for Chinese soy-
Chen (1995). The TIA level of the raw Argentinian soy- beans than for Argentinian soybeans to inactivate TI to
beans, however, is only 65% of that of the Argentinian the safe level at temperature of 100¡C and 118¡C,
soybeans measured in a previous investigation (Qin et respectively. The soybeans containing a lower level of
al 1996). The di†erences also existed between the soy- TIA need less severe heat processing condition. Fried-
beans treated at similar conditions in the present and man et al (1991) also found that low trypsin inhibitor
previous experiments. The PDI values in raw soybeans soybeans needed shorter time than conventional soy-
and the soybeans treated at 118¡C and 136¡C were beans to reduce their TIA to a safe level when heated at
similar to that in similarly treated soybeans before, but, similar temperatures.
the soybeans treated at 100¡C were not. These inconsis- The PDI in both origins of the soybeans decreased
tencies between present and previous data of Argentin- non-linearly with the prolongation of heating duration
ian soybeans may be caused by the origin di†erence and the increase of heating temperature before its level
between batches. reduced to about 10%, then remained stable. At the
The TIA values, for both Argentinian and Chinese most severe heating condition, the PDI even increased a
soybeans, decreased in a nonlinear pattern with the pro- little. This phenomenon was also observed in a previous

TABLE 3
Adequate processing duration (minutes) at di†erent heat temperatures
as calculated from the Ðtted equations based on a residual TIA level
of 4 mg g~1 products

T emperature Soybean Adequate time T IA PDI

(¡C) origin (mg g~1) (%)

100 Argentine 15 4 40
China 25 4 38
118 Argentine 2 4 23
China 7 4 26
136 Argentine 0.01 4 8
China 0.5 4 14
E†ect of steam treatment on protein quality of soybeans 397

study (Qin et al 1996). A possible explanation is the REFERENCES

hydrolysis of protein initiated only at the highest tem-
perature (Van der Poel et al 1990). AOCS 1964 American Oil ChemistÏ Society. Protein disper-
For a well-processed soybean product, the PDI has sibility index (PDI) AACC-method 46-24 2.1È2.2.
Booth V H 1971 Problems in the determination of FDNB-
to be kept at certain level. Monari (1993) suggested that available lysine. J Sci Food Agric 22 658È666.
the acceptable PDI level is in the range from 15 to 28. Carpenter K J 1960 The estimation of available lysine in
According to this standard, when the two soybean animal protein foods. Biochem J 77 604È610.
origins were heated at 100¡C and 118¡C for an adequate Chang C L, Tanskley T D Jr, Knabe D A, Zebrowska T,
period to reduce TIA to the safe level (4 mg g~1) for Gregg E J 1984 E†ect of di†erent heat treatments during
processing on nutrient digestibilities of soybean meal by
each of them separately, their PDI values, calculated growing swine. J Anim Sci 59 (suppl 1) 268È269.
with the Ðtted equations, are still kept higher than Collins J L, Beaty B F 1980 Heat inactivation of trypsin
acceptable (see Table 3). However, when heated at inhibitor in fresh green soybeans and physiological
136¡C to inactivate TIA to the safe level, the PDI value responses of rats fed the beans. J Food Sci 45 124È126.
may decrease to a much lower level than acceptable. It Friedman M, Brandon D L, Bates A H, Hymowitz T, Fried-
man M 1991 E†ect of heat on the nutritional quality and
seems that the high temperature (136¡C) denatures total safety of soybean cultivars. Adv Exp Med Biol 289 339È361.
protein more severely than inactivates TIA, which is in Fu C Z, Lu Y C 1992 Determination of trypsin inhibitor
accordance with the Ðndings of Labuza (1973). He con- activity in legumes. Soybean Sci 11(3) 269È272.
verted the data of Hackler et al (1965) into approximate Hackler L R, Van Buren J P, Steinkraus K H, El Rawi I,
kinetic parameters, and found that doubling the rate of Hand D B 1965 E†ect of heat treatment on nutritive value
of soymilk protein fed to weanling rats. J Food Sci 30 723È
trypsin inhibitor destruction by increasing process tem- 728.
perature would increase the destruction of many nutri- Herkelman K L, Cromwell G L, Stakly T S 1990 E†ect of
ents four- to Ðve-fold. heating time on the nutritional value of full-fat soybeans for
The high correlation between TIA and PDI found in chicks. J Anim Sci 68 (suppl 1) 1.
the present study indicates that the inactivation of TIA, Herkelman K L, Cromwell G L, Stakly T S, Pfei†er T W,
Knabe D A 1992 Apparent digestibility of amino acids in
to a certain extent, is accompanied by the denaturation raw and heated conventional and low-trypsin-inhibitor soy-
of the proteins in the soybeans. Qin et al. (1996) also beans for pigs. J Anim Sci 70 818È826.
observed a high correlation between TIA and PDI in ISO 1979 International Standard 5983. Animal feeding stu†s-
soybeans which had been steam toasted under various Determination of nitrogen content and calculation of crude
conditions. However, Tromp et al (1995) reported that protein content.
Kakade M L, Rackis J J, McGhee J E, Puskai G 1974 Deter-
for the roasted whole soybeans, the soluble protein level mination of trypsin inhibitor activity of soy products : a col-
was not directly correlated to the trypsin inhibitor laborative analysis of an improved procedure. Cereal Chem
content. It seems that the treatment e†ect on the corre- 51 376È382.
lation is di†erent between steam heating and roasting. Labuza T P 1973 E†ects of dehydration and storage. Food
The reactive lysine content, as measured by FDNB, T echnol 27 20È27 51.
Liener I, Thomlinson S 1981 Heat inactivation of protease
was not clearly inÑuenced by steam toasting when the inhibition of a soybean line lacking the Kunitz trypsin
soybeans were heated at 100¡C and 118¡C. At 136¡C, inhibitor. J Food Sci 46 1354È1356.
however, the FDNB active lysine content reduced as the McGrain A K, Chen J C, Wilson K A, Tan-Wilson A L 1992
heating time prolonged, and the changes in the FDNB Proteases catalysing processing and degradation of Kunitz
reactive lysine were not paralleled with those of TIA soybean trypsin inhibitor during seed maturation. Phyto-
chemistry 31 421È426.
and PDI. The correlation analysis results showed that Monari S 1993 Fullfat Soya Handbook. American Soybean
FDNB active lysine was not signiÐcantly correlated to Association, Brussels, Belgium, pp 6È9.
TIA and PDI (P [ 0.05). These results indicate that it is Qin G, Chen P 1995 Antinutritional factors in soybeans and
possible to optimize the processing procedure so that the responses of pig to them. In : Fourth Symposium of
the antinutritional factors can be sufficiently inactivated Northeast Pig Production Research Congress Changchun
P.R. China Nov. 20È22
while, simultaneously, the availability of lysine is main- Qin G, Ter Elst E R, Bosch M W, Van der Poel A F B 1996
tained. Thermal processing of whole soya beans : Studies on the
In conclusion, the TIA in the soybeans could be inactivation of antinutritional factors and e†ects on ileal
reduced to an acceptable level by di†erent heating con- digestibility in piglets. Anim Feed Sci T echnol 57 313È324.
ditions (combinations of heating time and temperature). Sangle M, Devi R, Pawar V D 1993 Proximate composition
and anti-nutritional factors of soybeans. J Maharashtra
Both TIA and PDI decreased non-linearly with the Agri Univ 18 391È393.
increase of heating temperature and heating duration. SAS 1990 SAS/STAT UserÏs Guide (Version 6.2). Statistical
These parameters, however, changed in di†erent ways Analysis Systems Institute Inc, Cary, NC, USA.
for soybeans of di†erent origins. It seems that Argentin- Tromp C T, Aghavan G S V, Alikhani Z 1995 Nutritive
ian soybeans need less severe processing conditions quality of whole soybeans heat-treated in a continuous Ñow
processor. J Agric Engng Res 61 49È56.
than Chinese soybeans as evaluated for TIA and PDI. Van Oort M G, Hamer R J, Slager E A 1989 The trypsin
Safe levels of TIA, however, need to be validated by inhibitor assay : improvement of an existing method. In :
animal trials. Recent Advances of Research in Antinutritional Factors in
398 G X Qin, M W A V erstegen, A F B van der Poel

L egume Seeds, ed Huisman J, van der Poel A F B, Liener I E†ect of some manufacturing technologies on chemical
E, PUDOC, Wageningen, The Netherlands, pp 110È113. physical and nutritional properties of feed. In : Recent
Van der Poel A F B, Blonk J, Van Zuilichem D J, Van Oort Advances in Animal Nutrition, eds Garnsworthy P C and
M G 1990 Thermal inactivation of lectins and trypsin inhib- Cole D Y A, University Press, Nottingham, UK. pp 93È126.
itor activity during steam processing of dry beans Zarkadas C G, Yu Z R, Hope H J, Mineroamador A, Roche-
(Phaseolus vulgaris L) and e†ects on protein quality. J Sci mont J A 1994 Comparison of protein-bound and free
Food Agric 53 215È228. amino acid contents of two Northern adapted soybeans cul-
Van der Poel A F B, Verstegen M W A, Tamminga S 1995 tivars. J Agric Food Chem 42 21È33.
Chemical physical and nutritional e†ects of feed processing Zhao S W, Zhou S H, Hu M X, Meng X X, Li A P, Wang S
technology. In : Proceedings of the 16th W estern Nutrition M 1991 Ti and SP1 locus genotype frequency and distribu-
Conference. Saskatoon, Canada. 70È86. tion of cultivars of soybean from three provinces of North-
Voragen A G J, Gruppen H, Marsman G J P, Mul A J 1995 east. Soybean Sci 10 77È80.