Analytical Criteria for Predicting Apparent Digestibility of Soybean Protein in Preruminant Calves

Analytical Criteria for Predicting Apparent Digestibility of Soybean Protein in Preruminant Calves J. P. LALLES,l,* H. M. TUKUR,' R. TOULLEC' and

B. G. MlLLERt

'Institut National de la Recherche Agronomique, Laboratoire du Jeune Ruminant, 65, rue de Saint-Brieuc, 33342 Rennes Cedex, France Wniversity of Bristol, Department of Clinical Veterinary Science, Langford House, Langford, Bristol 6S18 7DU, United Kingdom

ABSTRACT

INTRODUCTION

A series of experiments on the use of soybean as a protein source in milk replacers for veal calves was undertaken to determine the relationships between the physicochemical and antinutritional properties and apparent digestibilities of nine soybean products. Soybean provided between 58 and 71% of dietary CP, and skim milk or whey powder provided the remainder. Soybean products were analyzed for CP, native protein, aggregated protein, carbohydratelinked protein, peptides, antitryptic activity, and immunoreactivity of lectin, glycinin, a-conglycinin, and (3-conglycinin. The apparent digestibility of N calculated for soybean protein varied between 59 and 84%. Simple linear correlations were significant between apparent digestibility of soybean N and concentrations of native protein, antitryptic activity, glycinin, a-conglycinin, and (3-conglycinin. However, only variation in antitryptic activity, a-conglycinin, and 0conglycinin contributed significantly to an explanation for the variation in apparent digestibility of soybean N in a multiple linear equation. Under our experimental conditions, (3-conglycinin was the best predictor of digestibility of soybean N. Antitryptic activity became the best predictor when soybean products had no detectable (3-conglycinin. ( Key words: antinutritional factors, calf, protein digestibility, soybean)

Soybean protein is a valuable substitute for milk protein in milk replacers for preruminant calves; however, calf performance and N digestibility are usually lower with soybean diets than with milk protein diets (10). The magnitude of these differences varies widely, depending on amounts of incorporation, antinutritional factors (8), and susceptibility of such products to digestion ( 1 0 ) . Currently, the components of soybean products that affect apparent digestibility of soybean protein in veal calves are unknown. From a study involving 16 experimental products derived from the same batch of soybeans, antitryptic activity, lectin, and aggregated protein, but not antigenic proteins, appeared t o be the most important factors explaining variation in the apparent digestibility of dietary N (28). By contrast, Lallhs et al. ( 1 4 ) observed a negative linear relationship between apparent digestibility of soybean N and concentrations of antigenic glycinin and (3-conglycinin i n 10 commercial products. However, data on other criteria pertaining to soybeans were not available in that study. In this work, we analyzed the physicochemical characteristics of soybean products used previously ( 1 4 ) and determined trypsin inhibitor, lectin, and aconglycinin contents of each. Finally, linear regression analyses were conducted to determine relationships between analytical properties of soybean products and apparent digestibility of soybean N. Some of these results have been presented elsewhere

Abbreviation key: ADSN = apparent digestibility of (25). soybean N, AP = aggregated protein, CLP = carbohydrate-linked protein, KSTI = Kunitz soybean MATERIALS AND METHODS trypsin inhibitor, NP = native protein, R.F = raw flour, TF = toasted flour, "UI = trypsin units inSoybean Products hibited, WEPC = water-extracted and partially proNine commercial defatted soybean products, inteolyzed concentrate, WETC = water-extracted and toasted concentrate. cluding three toasted flours (TF)and six protein concentrates, were evaluated in four separate experiments (16, 17, 24, 26) (Table 1).Soybean flours had CP contents varying between 53 and 56% of DM. Received January 4, 1995, Three concentrates were water-extracted and toasted Accepted October 27, 1995. (WETC) (16, 17, 241, one concentrate was alcohol1Reprint requests. 1996 J Dairy Sci 79:475482

475

476

LALLES ET AL.

TABLE 1. Identification of soybean products used in experiments with veal calves and digestibility data. Soybean product2 TF 1 TF 2 TF 3 WETC WETC WETC AEHC WEPC

1 2 3

1

WEPC 2

Soybean protein CP

NSP

( % of DM)

(W of CP)

52.9 52.8 56.3 67.7 66.0 68.4 68.5 76.8 58.6

18.7 17.3 15.0 11.2 14.3 14.3 8.5 56.9 24.7

Dietary protein

Measured ADD"

Soybean

Other

Control

70 60 59 58 65 72 65 58 71

305 406 416 426 355 285 355 426 296

93.8 94.7 94.7 94.7 92.3 93.8 92.3 94.7 94.5

Calculated

Soybean

ADSN4

Reference

68.2 77.6 83.6 75.0 77.8 84.1 84.4 87.7 86.1

59 66 76 61 71 81 81 82 84

(16) (15)

(a)

(24) (15) (26)

'Apparent digestibility of dietary N. 2TF = Toasted flour, WETC = waterextracted and toasted concentrate, AEHC = alcohol-extracted and heated concentrate, and WEPC = water-extracted and partially proteolyzed concentrate. 3Nitrogen solubility index ( 1 ). 4Apparent digestibility of soybean N. 6Whey powder. GSkirn milk powder.

extracted and heated (24), and two concentrates were water-extracted and partially proteolyzed (WEPC)(16, 26). Products TF 3 and WETC 3 were subjected to additional treatments that were not detailed by the manufacturers (16, 24). A partially proteolyzed soybean protein isolate studied previously ( 1 7 ) was excluded in this study because the isolate was a mixture of soybean flour and sweet whey (75 and 25%, percentage of CP, respectively). A sample of raw flour ( RF),containing 47.6% CP (percentage of DM) after defatting at room temperature (20°C), was introduced as a positive control in laboratory analyses. The N solubility index ( 1) varied between 11 and 57% of CP, depending on the product (Table 1).

Analytical Procedures

Distribution of N among peptides and native ( NP), aggregated ( AP), and carbohydrate-linked ( CLP) protein fractions was determined according to methods of Visser and Tolman ( 2 8 1. Soybean samples

were extracted in sodium carbonate buffer (100 mM; pH 10) for 2 h at 37°C. The insoluble fraction was assumed to contain AP and CLP (28). The NP was precipitated from the supernatant at pH 4.6 using 1 M HC1. Peptide was considered to be the fraction soluble at pH 4.6. Separate soybean samples were also extracted in sodium carbonate buffer (100 mM; pH 1 0 ) containing 3% SDS and 2% pmercaptoethanol for 2 h at 37°C. Soluble and insoluble fractions were assumed to contain AP and CLP, respectively ( 2 8). Combining the two fractionation In Vivo Measurements protocols allowed protein N to be partitioned among Milk substitutes were formulated to contain 58 to the fractions described ( 2 8 ) . Nitrogen was deter72% of CP supplied by soybean products; the remainder was provided by skim milk or whey powder mined by the Kjeldahl method. Soybean products were also studied using denatur(Table 1). Concentrations of CP and fat were 20 to ing SDS-PAGE electrophoresis (Figure 1) as 22% and 18 to 20% (percentage of DM), respectively. described earlier (23, 27). Stained gels were analyzed Milk replacers were reconstituted with warm water by densitometry (Bio-Image System; Millipore Corp., and given twice daily to 2- to 4-mo-old calves ( n = 5 to Bedford, MA) following the recommendation of the 7). Intake was fixed at 58 to 60 g of DMkg of BW0.75 manufacturer in an attempt t o evaluate changes obper d. Apparent digestibility of soybean N (ADSN) was calculated with the assumption that digestibility served on gels. Specific soybean proteins either puriof N from skim milk or whey and synthetic AA were fied in the laboratory [glycinin, a-conglycinin, and 0similar for diets based on soybean and for control conglycinin; ( 2 7 )I or purchased (lectin, Bowman-Birk diets (Table 1). Calculated ADSN varied between 59 protease inhibitor; Sigma Chemical Co., La Verpiland 84%. Data for individual experiments have been likre, France) were included in SDS-PAGE analyses. published elsewhere (16, 17, 24, 26). Antitryptic activity was measured according to the Journal of Dairy Science Vol. 79, No. 3, 1996

4i7

SOYBEAN PROTEIN DIGESTIBILITY IN CALVES

methods of Kakade (21, and results were expressed as the number of trypsin units inhibited ( TUI).Concentration of lectin was determined by capture ELISA using rabbit anti-lectin serum as the capture antibody and a mouse monoclonal anti-lectin antibody as the detector antibody (B. G. Miller, 1994, unpublished data). A sheep anti-mouse IgG conjugate was used t o reveal and quantitate the reaction. Standards of puri-

Mrx

id

a

94-

6 7 4

fied lectin and soybean samples were run together on the same plate. Quantitation of glycinin, aconglycinin, and 0-conglycinin was conducted by specific immunometric ELISA tests using rabbit antibodies against corresponding purified NP ( 2 7 ). Statistical Analysis

Analytical differences between soybean products could not be evaluated statistically because only one sample of each product was available. Therefore, data presented in Tables 2, 3, and 4 are only descriptive. Relationships among the characteristics of soybean products and ADSN were analyzed using simple and multiple linear regression ( 6 ). RESULTS Physicochemical and Antinutritional Propertles of Soybean Products

20 --.+

In RF, NP amounted t o 60% of total N; AP, CLP, and peptide fractions contributed equally to the remaining N (Table 2).By contrast, most unhydrolyzed products except TF 1 presented a much lower but variable ( 8 to 34%) percentage of NP and a much higher but variable (28 to 74%) percentage of AP. As expected, WEPC 1, which had been proteolyzed, had very low NP (3%) and AP (0%) but high peptides (81%). Curiously, WEPC 2 contained a high percentage of CLP (64%) and a surprisingly low percentage of peptides (9.4%). Excluding WEPC 2 and WETC 1, the CLP fraction was between 12 and 16% among products.

TABLE 2. Distribution of N between native protein (NP), aggregated protein (AP), carbohydrate-linked protein (CLP), and peptides. ~

~

Fraction Soybean product1

NP

AP

RF TF 1 TF 2 TF 3 WETC WETC WETC AEHC WEPC WEPC

59.6 50.9 17.2 15.5 21.6 33.7 7.9 9.4 2.9 4.4

13.4 27.8 59.4 59.4 50.7 47.8 74.2 67.5 0 21.9

CLP

Peptides

(% of total N )

Figure 1. The SDS-PAGE electrophoretic profiles of soybean products and purified proteins. Lane l ( a and b), molecular mass standards; lane 2(a and b), raw flour; lane 3(a), toasted flour ( T F ) 1;lane 4 ( a ) , TF 2; lane 5(a), TF 3; lane 6(a), water-extracted and toasted concentrate (WETC) 1; lane 7 ( a and b), glycinin; lane 8(a and b), a-conglycinin; lane 9(a and b), 8-conglycinin; lane 10 (a and b), lectin; lane l l ( a and b), Bowman-Birk protease inhibitor; lane 3(b), WETC 3; lane 4(b), alcohol-extracted and heated concentrate; lane 5 ( b ) , water-extracted and partially proteolyzed concentrate (WEPC) 1;and lane 6 ( b ) , WEPC 2. Mr = Molecular mass.

1

2 3 1

2

13.3 12.3 16.2 17.3 24.3 14.7 12.2 16.6 16.3 64.1

13.6 9.1 7.2 7.8 3.4 3.8 5.7 6.6 81.2 9.4

1RF = Raw flour, TF = toasted flour, WETC = water-extracted and toasted concentrate, AEHC = alcohol-extracted and heated concentrate, and WEPC = water-extracted and partially proteolyzed concentrate. Journal of Dairy Science Vol. 79,

No. 3, 1996

478

LALLES ET AL.

TABLE 3. Densitometric analysis of soluble protein extracted from soybean products after SDS-PAGE separation. Molecular mass Soybean product'

<20 kDa

20 to 42 kDa

RF TF 1 TF 2 TF 3 WETC WETC WETC AEHC WEPC WEPC

18.2 4.7 2.1 23.5 6.5 16.0 54.1 42.1 100.0 87.1

59.0 56.7 75.8 67.2 57.7 57.8 41.7 45.6 0 3.0

42 to 85 kDa

>85 kDa

(a)

1 2 3 1 2

13.4 18.3 13.2 9.1 35.8 26.3 4.2 12.3 0 9.9

9.4 20.4 8.8 0.2 0 0 0 0 0 0

1RF = Raw.flour, TF = toasted flour, WETC = water-extracted and toasted concentrate, AEHC = alcohol-extracted and heated concentrate, and WEPC = water-extracted and partially proteolyzed concentrate.

Qualitative differences among soybean products for patterns of soluble proteins were evidenced by SDSPAGE (Figure 1). Molecular mass of P-conglycinin subunits ranged from 42 t o 85 kDa; acidic and basic subunits ranged from 60 to 85 and 42 to 53 kDa, respectively (5,7 ) . Glycinin polypeptides were in the range of 20 to 42 kDa, and acidic and basic polypeptides ranged from 30 t o 42 kDa and from 20 to 25 kDa, respectively (5, 7). Lectin subunits (25 to 30 kDa) were expected t o appear between the acidic and basic polypeptides of glycinin. a-Conglycinin ( 19 to 22 kDa), including Kunitz soybean trypsin inhibitor (IKSTI; 22 kDa), might have migrated close to the

basic polypeptides of glycinin (5,7). Molecules with low molecular mass (c20kDa) included the BowmanBirk protease inhibitor (5,7). In this study, both RF and TF, except TF 3, presented bands of high molecular mass (>85 kDa); but these bands were virtually absent in the other products. Visual examination of gels did not permit a clearer distinction among products, apart from showing the shift of protein bands toward fragments of low molecular mass with proteolyzed products. Densitometric analysis essentially confirmed observations obtained by SDS-PAGE (Table 3 ) , which quantitatively showed that TF 3 presented a higher amount of compounds with low molecular mass (<20 kDa) than did TF 1 and TF 2. Also, WETC 3 had a densitometric profile that was very different from that of the other WETC products studied. As expected, proteolyzed products displayed very high amounts of compounds with low molecular mass that were soluble in borate buffer; however, WEPC 1 was apparently more hydrolyzed than WEPC 2. Indeed, virtually all peptides had a molecular mass c10 kDa in WEPC 1, and those in WEPC 2 had molecular mass between 10 and 20 kDa (data not shown). Antitryptic activity was highest in RF (140 TUV mg of CP) but 20 t o 100 times lower in most soybean products, except in TF 1 (20 TUVmg of CP) (Table 4). Concentration of immunoreactive lectin was the highest in RF (15.4 mg/g of CP), 25 times less in TF 1, and virtually absent from other soybean products tested. Concentration of immunoreactive glycinin was the highest in RF (269 mg/g of CP) and was

TABLE 4. In vitro antitryptic, lectin, glycinin, a-conglycinin, and 0-conglycinin activities of soybean products. Soybean'

Antitrypsin Lectin

Glycinin

a-Conglycinin

RF TF 1 TF 2 TF 3 WETC WETC WETC AEHC WEPC WEPC

(TUIVmg of CP) 140.0 19.9 5.7 6.6 6.5 5.2 2.7 3.4 2.5 1.4

269 39.4 26.8 0.7 20.4 32.9 0.03 0 0 10.5

(mg/g of CP) 155 31.5 15.2 36.1 0.45 13.4 1.13 0 2.93 25.5 3.2 14.7 0.7 0 0 0.45 0 0 0.4 0

1 2 3

1 2

15.4 0.58 0 0.002 0 0.021 0 0 0 0

/3-Conglycinin

Total antigens2

471 90.7 40.7 1.83 48.8 50.8 0.73 0.45 0 10.9

'RF = Raw flour, TF = toasted flour, WETC = water-extracted and toasted concentrate, AEHC = alcohol-extracted and heated concentrate, and WEPC = water-extracted and partially proteolyzed concentrate. 2Total antigens = Lectin + glycinin + a-conglycinin + 0-conglycinin. 3Trypsin units inhibited, Journal of Dairy Science Vol. 79, No. 3, 1996

479

SOYBEAN PROTEIN DIGESTIBILITY IN CALVES

decreased by 10 times in TF 1, TF 2, WETC 1, and WETC 2. Products TF 3 and WETC 3 presented only trace amounts of glycinin. No glycinin was detected in the alcohol-extracted and heated concentrate or in WEPC 1. By contrast, WEPC 2 was still rich in native glycinin or in immunoreactive fragments. @Conglycinin behaved similarly to glycinin among treatments, except for WEPC 2, which had no detectable @-conglycinin.By contrast, a-conglycinin behaved somewhat differently from glycinin o r 0-conglycinin in that a-conglycinin was detectable at substantial concentrations in all soybean products, except WEPC 1. Relationships Among Characteristics of Soybean Products and ADSN

No significant ( P > 0.10) correlations existed among NP, AP, CLP, and peptides, except between AP and peptides, which were negatively correlated ( P < 0.05) (Table 5 ) . As expected, NP was positively correlated ( P < 0.01) with antitryptic activity and with concentrations of lectin, glycinin, a-conglycinin, and 0-conglycinin. Also, all antinutritional activities were positively correlated ( P < 0.05) with one another (Table 5). Simple linear regression between analytical criteria and ADSN showed that it was negatively correlated with NP ( P c 0.01),antitryptic activity ( P < 0.05), and concentrations of glycinin ( P < 0.011, aconglycinin ( P < 0.051, and @-conglycinin ( P < 0.0001) (Table 5). A similar trend ( P = 0.107) occurred between ADSN and lectin. The best predictive analytical parameter for ADSN was (3-conglycinin (Table 6; Figure 2). This prediction could be improved by stepwise introduction of aconglycinin, antitryptic activity, and NP in multiple linear equations. The lowest residual standard error

was obtained with these four analytical measures. When five soybean products that were devoid of 0conglycinin were considered, antitryptic activity was the best predictor of ADSN (Table 6; Figure 2). When more criteria were considered, glycinin and lectin provided the most satisfactory equation for predicting ADSN, but that did not decrease the residual standard error. DISCUSSION Antlnutrltlonal Factors and Prediction of ADSN

The present data clearly provide a quantitative basis for predicting ADSN in calves from concentrations of immunoreactive soybean globulins, as determined in vitro ( 2 7 ). This result was anticipated from earlier studies (19, 20, 2 11, which demonstrated that low immunoreactivity of soybean products was a prerequisite for high digestibility of N in calves. Additionally, our results indicate that 0-conglycinin had the best predictive value overall and that antitryptic activity became the best predictor of ADSN when products had undetectable @-conglycinin. However, our observations sharply contrasted with results of the study by Visser and Tolman (28), which concluded that antitryptic activity, lectin, and AP were the most important factors to explain variations in ADSN in young calves. Reasons for such discrepancies are unknown. Although antitryptic activities and concentrations of lectin in RF and TF 1 were similar to those published for similar products (281, large variations were observed for other products. For example, antitryptic activities varied between 1.4 and 6.6 TUI/mg of CP in this study and between 3.1 and

TABLE 5. Matrix of correlation coefficients among variables.'

AP CLP PEP TU1 LEC GLY aCG PCG ADSN

NP

AP

CLP

PEP

TU1

LEC

GLY

aCG

BCG

0.015 -0.369 -0.368 0.893C 0.805c 0.867c 0.895c 0.884C -0.818C

-0.354 -0.730b 0.116 -0.273 -0.121 0.216 0.107 0.075

-0.087 -0.354 -0.219 -0.114 -0.244 -0.248 0.354

-0.212 -0.094 -0.350 -0.189 -0.286 0.351

0.945c 0.682b 0.964d 0.835c -0.764b

0.6108 0.977d 0.734b -0.573

0.703b 0.873c -0.823'

0.832c -0.682"

-0.945e

B P < 0.10. bp < 0.05. C P < 0.01. d P < 0.0001. 'NP = Native protein, AP = aggregated protein, CLP = carbohydrate-linked protein, PEP = peptides, TU1 = trypsin units inhibited, LEC = lectin, GLY = glycinin, aCG = a-conglycinin, PCG = P-conglycinin, and ADSN = apparent digestibility of soybean N. Journal of Dairy Science Vol. 79, No. 3, 1996

LALLES ET AL.

TABLE 6. Equations of prediction of apparent digestibility of soybean N (ADSN; as a percentage) from analytical criteria of soybean products. Equation1

Residual SE

All soybean products ( n = 9 ) ADSN = 80.1 - 0.670dj3CG [l] 3.33 2.94 ADSN = 80.3 - 0.87OCj3CG+ 0.674aCG [2] 0.987 ADSN = 85.0 - 1.67cTuI - 0.796dj3CG + 2.36caCG [31 ADSN = 85.6 - 1.6leTUI - 0.746dj3CG - 0.lOlaNP + 2.47CaCG [41 0.751 Soybean products with no detectable j3CG (n = 5 ) 0.484 ADSN = 85.7 - 1.49 TU1 I51 0.580 ADSN = 81.3 + 0.254aGLY - 2.74bLEC [61

r

P

-0.945 0.964 0.997 0.998

0.0001 0.001 0.0001 0.0001

-0.990 0.990

0.0012 0.019

aP < 0.10. bP < 0.05. CP < 0.001. d P < 0.0001. 1j3CG = j3-Conglycinin (milligram per gram of CP), aCG = a-conglycinin (milligram per gram of CP), TU1 = trypsin units inhibited (per milligram of CP), NP = native protein (percentage of total N ), GLY = glycinin (milligram per gram of CP), and LEC = lectin (microgram per gram of CP).

100 TUYmg of CP ( R F excluded) in the study of Visser and Tolman (28 1, and lectin concentrations were between 0 and 0.02 mg/g of CP in this study and between 0 and 9.7 mg/g of CP in the study of Visser and Tolman (2 8). Total antigen concentrations (Table 4) in untoasted and mildly toasted products used by Visser and Tolman (28) were 30 t o 70% higher than in RF and TF 1 used in the present study. The range of antigen concentrations was also larger ( 0 t o 465 mg/g of C P ) in that study ( 2 8 ) than in this study ( 0 to 91 mg/g of CP). To date, reference values and variations for most antinutritional factors in soybean products, particularly antigens, are lacking. Thus, conclusions may depend on the range of variation of analytical criteria in soybean products and probably on experimental conditions, among which age of calves, duration of experiments, and amount of soybean flour incorporated in milk replacers may be of particular importance. Whether relationships between soybean characteristics and ADSN in the present experiment were causal remains unanswered. Apparent digestibility of protein may decrease as a result of reduced true digestion of dietary protein, increased secretion of endogenous protein, and decreased reabsorption of endogenous protein ( 2 2 ) . Tukur et al. ( 2 7 ) showed that substantial amounts of immunoreactive glycinin and, to a lesser extent, a-conglycinin and pconglycinin escape digestion in the small intestine of preruminant calves. Also, a negative linear relationship existed between ileal flow of glycinin and apparent digestibility of N at the ileum ( 2 7 ) . Ileal loss of active trypsin increased when calves were fed milk replacers containing raw pea flour (15). For pigs, KSTI added to a diet based on casein increased ileal Journal of Dairy Science Vol. 79, No. 3, 1996

flow of both endogenous and undigested dietary protein (3). Low concentrations of KSTI increased ileal loss of endogenous protein, but higher KSTI also reduced digestion of soybean concentrate (91. Lectin appeared t o behave differently from KSTI, increasing ileal flow of endogenous protein, regardless of lectin concentration ( 9 1. Thermostable factors in soybean enhanced fluid secretion in ligated guts of pigs by a reaction apparently unrelated to immune phenomena (18). Because most proteins purified from soybean, particularly 0-conglycinin, induced skin reactions in calves fed milk replacers based on skim milk powder, Lallks et al. ( 1 21 postulated that such soybean constituents might account for gut hypersecretion ( 181. Finally, involvement of local immune reactions cannot be excluded in the explanation of increased fluxes of endogenous protein because young calves are particularly prone to mount immune responses to soybean (10, 19). Recently, Lallks et al. ( 1 3 ) showed that most of the purified soybean proteins studied here, in addition to being highly immunogenic, induced specific skin reactions in calves sensitized to TF 1 and that 0-conglycinin was probably implicated in delayed-type, cell-mediated immune reactions. Disruption of small intestine motility patterns, which is usually associated with maldigestion, malabsorption, and diarrhea ( 191, was observed in sensitized calves only when antigenic TF 1 provided more than 18% (percentage of total CP) in milk replacers (11). Antinutritional Factors and Methods of Processing

Although the aim of the present study was not to determine the influence of method of processing on

481

SOYBEAN PROTEIN DIGESTIBILITY IN CALVES

g

85

tR

4

z

4

79

-

77

.

v1

n

75

f

&-

4

rn

b

z 83

X

v1

P

81 -

W

U

-

85

a

65

i2

55

o

io

75

1

'

1

.

r

'

l

.

l

.

i

30 40 B-Conglycinin (mg/g of CP)

-

20

85

85

25 83

X

8 c

v1

9

75

4E

65

81 79

77

75

55

55

65 Predicted

75

85

ADSN (9b)

75

ai a3 as Predicted ADSN (96)

77

79

Figure 2. Relationships among apparent digestibility of soybean N (ADSN; as a percentage) and a ) 0-conglycinin (all soybean products) or b) antitryptic activity (five soybean products with no detectable 0-conglycinin). Relationships between predicted and measured ADSN using equations c) [4] of Table 6 and d ) [ti] of Table 6. TU1 = Number of trypsin units inhibited.

analytical characteristics and digestive behavior of soybean N, our observations corroborate earlier studies (19, 20, 21). Indeed, Sissons et al. (21) demonstrated that industrially heated soybean flours exhibited high titers of immunoreactive globulins in vitro and that such flours were poorly digested and had a high capacity t o sensitize calves. Sissons et al. ( 2 0 1 also showed that titers of immunoreactive glycinin and /3-conglycinin decreased with increased processing temperature between 40 and 80°C and that the best protein denaturation was obtained when samples were treated with 55 t o 76% ethanol a t a temperature between 70 and 80°C.Unfortunately, no information was provided on variations within treatment in antigen concentrations (20,21). Finally, because of the central role of native soybean globulins in

protein resistance t o digestion and calf sensitization, proteolysis would be expected to reduce antigenic activity of soybean products and largely improve their digestive utilization (16, 17). a-Conglycinin contains some KSTI ( 5 , 7 ) . Our results would suggest that KSTI is somewhat resistant to heat, alcohol denaturation, and proteolysis; however, antibodies raised against native KSTI can recognize denatured molecules (4 ) . CONCLUSIONS

The ADSN, as determined in preruminant calves aged 2 to 4 mo, appeared to be negatively correlated with antitryptic activity and concentrations of glycinin, a-conglycinin, and 0-conglycinin, and NP in comJournal of Dairy Science Vol. 79, No. 3, 1996

482

LALLES ET AL.

mercial soybean products that were incorporated into milk replacers a t 58 to 71% of dietary protein. Residual 6-conglycinin immunoreactivity was the best predictor of ADSN. When soybean products had no detectable P-conglycinin, ADSN was best predicted by antitryptic activity. Conclusions might differ when less refined soybean products and other experimental conditions are used. Further work is required t o better study the influence of method of processing on physicochemical and antinutritional properties of soybean products for calves. ACKNOWLEDGMENTS

The authors thank J. Quillet for collecting the literature, J. Tauran for antitryptic activity determination, and D. Dreau for densitometric analysis. Part of this work was financially supported by the Ministiire de 1’Agriculture et de la PBche (DGER No. 91-1311, the Rdgion de Bretagne, the Dkpartment d’Ille et Vilaine and the District de Rennes. REFERENCES 1 American Oil Chemists’ Society. 1983. Nitrogen Solubility Index. Am. Oil Chem. Soc., Champaign, IL. 2 American Oil Chemists’ Society. 1983. Trypsin Inhibitor Activity. Am. Oil Chem. Soc., Champaign, IL. 3 Barth, C. A., B. Lunding, M. Schmitz, and H. Hagemeister. 1993. Soybean trypsin inhibitor(s) reduce absorption of exogenous and increase loss of endogenous protein in miniature pigs. J. Nutr. 123:2195. 4 Brandon, D.L., S. Haque, and M. Friedman. 1987. Interaction of monoclonal antibodies with soybean trypsin inhibitors. J. Agric. Food Chem. 35:195. 5Brooks, J. R., and C. V. Morr. 1985. Current aspects of soy protein fractionation and nomenclature. J. Assoc. Offic. Anal. Chem. 62:1347. 6 Dagn6lie, P. 1970. Les mkthodes relatives a la regression. Page 265 in Theorie et Methodes Statistiques. Editions J. Duculot, Gembloux, Belgium. 7 Gueguen, J., and J. L. Azanza. 1985. Propriktks biochimiques et physico-chimiques des protkines de lkgumineuses et d‘oleagineux. Page 135 in Les Protkines Vegktales. B. Godon, ed. Tech. Doc. Lavoisier, Paris, France. 8Huisman, J., and A.J.M. Jansman. 1991. Dietary effects and some analytical aspects of antinutritional factors in peas ( P i sum sativum), common beans (Phaseolus vulgaris) and soyabeans ( Glycine max L.) in monogastric farm animals. A literature review. Nutr. Abstr. Rev. Ser. B 61:901. 9 Jansman, A.J.M., H. Schulze, P. van Leeuwen, and M.W.A. Verstegen. 1994. Effects of protease inhibitors and lectins from soya on the true digestibility and endogenous excretion of crude protein in piglets. Page 322 in VIth Int. Symp. Digestive Physiology in Pigs. W. B. Souffrant, and H. Hagemeister, ed. Publ. No. 80. Eur. Assoc. Anim. Prod., Dummerstorf, Bad Doberan, Germany. 10 Lalles, J . P. 1993. Nutritional and antinutritional aspects of soyabean and field pea proteins used in veal calf production: a review. Livest. Prod. Sci. 34:181. 11 Lalles, J. P., D. Benkredda, and R. Toullec. 1995. Influence of soya antigen levels in milk replacers on the disruption of intestinal motility patterns in calves sensitive to soya. J. Vet. Med. Ser. A 42:467. Journal of Dairy Science Vol. 79,No. 3, 1996

12 Lallks, J . P., P. Branco Pardal, and R. Toullec. 1994. Intradermal injection of soya proteins induce specific skin reactions in preruminant calves fed skim milk-based milk replacers. Page 99 in The Developing of Digestive and Metabolic Processes in New Born and Growing Ruminants. W. Barej, R. Zabielski, and P. Ostaszeski, ed. Satellite Symp. 8th Int. Symp. Rum. Physiol. Fundacja Rozwojsggw, Warsaw, Poland. 13 Lalles, J . P., D.Drkau, H. Salmon, and R. Toullec. 1996. Identification of soyabean allergens and mechanisms of dietary sensitivities in preruminant calves. Res. Vet. Sci. 60:111. 14 Lalles, J . P., J. W. Plumb, E.N.C. Mills, M.R.A. Morgan, H. M. Tukur, and R. Toullec. 1993. Antigenic activity of some soyabean products used in veal calf feeding: comparison between in vitro tests (ELISA polyclonal vs monoclonal) and with in vivo data. Page 281 in Recent Advances of Research in Antinutritional Factors in Legume Seeds. A.F.B. van der Poel, J. Huisman, and H. S. Saini, ed. Publ. No. 70, Wageningen Pers, Wageningen, The Netherlands. 15 LallBs, J . P., and R. Toullec. 1994. Trypsin loss a t the ileum of calves fed milk replacers containing legume protein. Ann. Zootech. (Paris) 43:263. 16 Lalks, J. P., R. Toullec, P. Bouchez, and L. Roger. 1995. Antigenicity and digestive utilization of four soyabean products by the preruminant calf. Livest. Prod. Sci. 41:29. 17Lalles, J. P., R. Toullec, P. Branco Pardal, and J . W. Sissons. 1995. Hydrolyzed soy protein isolate sustains high nutritional performance in veal calves. J . Dairy Sci. 78:194. 18 Nabuurs, M.J.A. 1986. Thermostable factorb) in soya producing a net excess of secretion in the ligated gut test in pigs. Vet. Res. Commun. 10:399. 19 Sissons, J. W. 1982. Effects of soya-bean products on digestive processes in the gastrointestinal tract of preruminant calves. Proc. Nutr. Soc. 41:53. 20 Sissons, J. W., A. Nyrup, P. J . Kilshaw, and R. H. Smith. 1982. Ethanol denaturation of soya-bean protein antigens. J . Sci. Food Agric. 33:706. 21 Sissons, J . W., R. H. Smith, and D. Hewitt. 1979. The effect of giving feeds containing soya-bean meal treated or extracted with ethanol on digestive processes in the preruminant calf. Br. J. Nutr. 42:477. 22 Souffrant, W. B. 1991. Endogenous nitrogen losses during digestion in pigs. Page 147 in Digestive Physiology in Pigs. M.W.A. Verstegen, J . Huisman, and L. A. den Hartog, ed. Pudoc, Wageningen, The Netherlands. 23 Stott, D. I. 1989. Immunoblotting and dot blotting. J. Immunol. Methods 119:153. 24 Toullec, R., J . P. Lalles, and P. Bouchez. 1994. Substitution of skim milk protein with soyabean protein concentrates and whey in milk replacers for veal calves. Anim. Feed Sci. Technol. 50: 101. 25 Toullec, R., J. P. LallBs, and H. M. Tukur. 1994. Relationships between some characteristics of soyabean products and nitrogen apparent digestibility in preruminant calves. Page 53 in The Developing of Digestive and Metabolic Processes in New Born and Growing Ruminants. W. Barej, R. Zabielski, and P. Ostaszeski, ed. Satellite Symp. 8th Int. Symp. Rum. Physiol., Fundacja Rozwojsggw, Warsaw, Poland. 26Tukur, H. M., P. Branco Pardal, M. Formal, R. Toullec, J. L. Lallks, and P. Guilloteau. 1995. Digestibility, blood levels of nutrients and skin responses of calves fed soyabean and lupin proteins. Reprod. Nutr. Dev. 35:27. 27Tukur, H. M., J. P. LallBs, C. Mathis, I. Caugant, and R. Toullec. 1993. Digestion of soya-bean globulins, glycinin, aconglycinin and @-conglycinin in the preruminant and the ruminant calf. Can. J. Anim. Sci. 73:891. 28Visser, A., and G. H. Tolman. 1993. The influence of various processing conditions on the level of antinutritional factors in soya protein products and their nutritional value for young calves. Page 447 in Recent Advances of Research in Antinutritional Factors in Legume Seeds. A.F.B. van der Poel, J. Huisman, and H. S. Saini, ed. Publ. No. 70, Wageningen Pers, Wageningen, The Netherlands.