Tannin content affects negatively nutritive value of pea for monogastrics *

Nine Polish varieties of pea (P. sativum L.) differing in flower colour, from white to purple, were evaluated for chemical and amino acid composition and in vitro for predicted ileal digestibility of protein (pdN) and predicted digestibility of energy (pdE) for pigs. In selected varieties differing in tannin content the apparent metabolizable energy value corrected for zero N balance (AME N ) , apparent digestibility of protein and fat and effect of enzymes reducing viscosity of digesta on this parameters were estimated in chickens; apparent metabolizable energy value (AME), true digestibility and biological value of protein was determined in rats. Tannin content had most pronounced negative effect on protein digestibility in chicken (r= -0.93; PO.05) and rats (r= -0.89; PO.05), pdN for pigs (r= -0.98; PO.001) as well as A M E N for chicken (r= -0.99; PO.001) and AME for rats (r= -0.95; PO.01), while nutrient content in peas had no significant effect on measured parameters. A M E N values of pea did not depend on supplementation of pea-containing diet with xylanase. Seeds of coloured-flowered cultivars, which are rich in tannins are less effectively utilized by monogastric animals than white-flowered ones. The degree of decline in nutrient utilization depends on tannin content in the seeds, which is correlated with the colour of the flowers.


INTRODUCTION
An importance of pea as a component of feed mixtures for poultry and pigs recently increased, as pea protein may substitute part of meat meal protein, withdrawn from animal diets for sanitary reasons.Pea production increased considerably in Europe and many research programmes were established to study the composition of different types and cultivars of pea and relationships between composition criteria and nutritional value of pea (Bastianelli et al., 1998;Grosjean et al., 1998aGrosjean et al., ,b, 1999)).In Poland, among seventeen registered cultivars of pea grown on light soils, twelve (Wiatr, 1999) belongs to a colour-flowered (P.sativum L. ssp.arvense), yielding seeds with higher tannin content than white-flowered varieties (P.sativum L. ssp.sativum).Tannins are recognized as a main antinutrient in colour-flowered field beans and peas (Savage, 1989).Comparison of chemical composition and ileal digestibility of amino acids for pigs of 10 Polish cultivars of pea was done by Gdala et al. (1992), the content of nutrients and antinutritional substances in 15 cultivars of white-flowered pea was determined by Zduhczyk et al. (1997).
The objective of the present study was to compare the composition and nutritional value of peas differing in flowers colour and tannin content in seeds, and to assess effects of feeding them on selected parameters of gastro-intestinal functions in chicken and rats.In experiment with chicken the effects of supplementation of pea diets with xylanase were also determined.

MATERIAL AND METHODS
Seeds (sowing grade) of 9 Polish cultivars of spring pea from 1998 harvest year were obtained from plant breeding stations and comprised the white-flowered: Mazurek, Albatros, Agra, Piast, the red-flowered: Bart, Idol, Dawo, Wiato and the purple-flowered Zuraw.

In vivo tests
The experiments were complied with obligatory ethical regulations concerning animal experimentation and care of animals under study.
Experiments on chickens.Seeds of 5 cultivars differing in tannin content -Albatros, Piast, Idol, Wiato and Zuraw were used.Experiment was done on 98 fourweek-old broiler chickens (Cobb 500) with a mean initial body weight of 1032 g.The birds were housed individually in balance cages, 8 (test groups) or 9 (control groups) birds per treatment.Two control groups were fed basal diets (A or B) containing (in g/kg): wheat, 590; soyabean oil meal, 334; rape seed oil, 3; mineral-vitamin mixture, 4.4; L-lysine, 0.1 and DL-methionine, 0.1.Diet A was not supplemented, while diet B was supplemented with 1 g of enzyme preparation Bio-Feed Wheat CT (Novo Nordisk) containing 1000 FXU xylanase/g according to producer.Remaining groups were fed test diets composed of basal diet A or B and ground pea combined in proportions as 1:1 on DM basis.To the basal and test diets 3 g Cr 2 0 3 per kg DM was added as a marker.All diets were cold pelleted.
The diets were fed at the level of 100 g/bird/day, in three meals.After 2 days of preliminary feeding the birds were fasted during 17 h then fed on the same diets for 4 days, and fasted for 17 h.Feed intake was recorded and excreta was collected during last 96 h of the experiment.Excreta were immediately frozen and kept at -18°C for further analysis.Apparent metabolizable energy value corrected for zero N balance (AME N ), apparent digestibility of protein, fat and organic matter retention was calculated for basal and test diets.Respective values for evaluated cultivars of pea were calculated by difference method according to Campbell et al. (1983) and Pesti and Ware (1986), the values obtained for basal diets A or B were used as a reference.
At the end of the experiment all birds were killed by decapitation, digestive tract was removed, small intestine was ligated to avoid the post-mortem movement of digesta, and intestinal contents were withdrawn by gently finger stripping.The jejunal digesta were collected to tubes placed on ice, centrifuged at 10000 x g for 10 min, the supernatants were withdrawn and their viscosity determined using a Brookfield digital viscometer (Model DV-II +LV) maintained at 40°C.The ileal digesta from two birds was pooled, mixed with distilled water 1:1 w/w and pH was immediately measured on WTW 340 A digital pH-meter.
Experiments on rats.Seeds of 6 cultivars differing in tannin content: Albatros, Piast, Idol, Dawo, Wiato and Zuraw were used in two experiments.In experiment 1 true digestibility (TD) and biological value (BV) of protein was determined by the Thomas-Mitchell method according to the procedure described by Smulikowska et al. (1997).Semisynthetic diets contained ground pea as the only source of protein at a level corresponding to 95 g crude protein per kg.Each diet was fed to eight four-week old male IF Z :JAZ rats, with mean initial body weight 79 g.Net protein utilization (NPU = BV x TD) and utilizable protein content (NPV = NPU x CP, %) were calculated.
In experiment 2 apparent metabolizable energy value (AME) was measured on six-weeks old male IF Z :JAZ rats, with mean initial body weight of 101 g.Basal diet was composed of (in g/kg): cereals, 618; soyabean oil meal, 150; fish meal, 80; milk powder, 100; soya oil, 20; mineral-vitamin mixture, 32.Test diets contained per kg: 400 g of basal diet and 600 g of ground pea.Basal diet was fed to 14 rats, test diets were fed to 8 rats each in amount 13-15 g per day.After 4 days of preliminary period faeces were collected quantitatively for 6 days.The faeces were kept at -18°C for further analysis.AME of the peas was calculated by the difference method using the values obtained for basal diet as a reference.At the end of experiment rats were killed by C0 2 overdosing, the pancreas and caecum were dissected, caecal digesta were collected, the weight of pancreas, full and empty caecum and pH of caecal contents were measured.

In vitro tests
In all nine cultivars of pea seeds predicted apparent ileal digestibility of protein for pigs (pdN) was evaluated by in vitro method according to Boisen and Fernandez (1995), and predicted total tract digestibility of energy for pigs (pdE) by in vitro method according to Boisen and Fernandez (1997).

Chemical and statistical analysis
Weight of thousand seeds was determined, representative samples of seeds were dehulled and the percentage of hulls was calculated.Chemical composition of the seeds was determined according to AOAC (1990), neutral-detergent fibre (NDF), acid detergent fibre (ADF) and acid detergent lignin (ADL) content according to Van Soest and Wine (1967) and Van Soest (1973) on a Fibertec M (Tecator) apparatus.Amino acids analysis of protein was performed with an automatic Beckman 6300 High Pressure Amino Acid Analyzer after acid hydrolysis.Methionine and cystine were determined after oxidation with performic acid and tryptophane after hydrolysis with BaOH according to Buraczewska and Buraczewski (1981).Trypsin inhibitor activity was analyzed according to Kakade et al. (1974).Tannin content in pea seeds and hulls was evaluated according to the method of Jerumanis (1972) modified by Adams and Novellie (1975).Excreta of chickens prior to analysis were dried in a forced-draft oven at 70°C for 24 h.Chemical composition of diets and excreta was determined according to AOAC (1990), fat in chickens diets and excreta was extracted with diethyl ether after acid hydrolysis.Gross energy was determined by Parr adiabatic oxygen bomb calorimeter KL-10.Chromic oxide in diets and excreta was analyzed spectrophotometrically following wet ashing according to Hinsberg et al. (1953).Faecal N in excreta of chickens was determined according to Ekman et al. (1949).
Statistical evaluation of results was done by Statgraphic plus ver.7 Software.

RESULTS
The seeds of white-flowered varieties were on average bigger than those of coloured-flowered ones, however in both types there were cultivars with small and big seeds.The white-flowered cultivars had on average lower percentage of seed coat and lower content of dietary fibre than the coloured-flowered cultivars (Table 1).The content of protein and starch was higher in white-flowered cultivars, but there were no distinct differences in the content of other nutrients between both types of pea (Table 2).Lysine content in 8 cultivars averaged 7.46, only in the Agra cv. was below 7 g/16 g N. Methionine and cystine content was on average 0.96 and 1.49 g/16 g N, respectively.There were no big differences in remaining amino acid content between cultivars (Table 3).Trypsin inhibitor activity was low (from 0.96 to 2.37 TlU/mg DM) and did not differ significantly between whiteand coloured-flowered cultivars.Tannin content in white-flowered cultivars was very low and uniform (0.25 g/kg DM on average), in coloured-flowered significantly higher (7.26-12.1 g/kg DM).Most of tannins were present in seed coat (Table 2).Proportion of seed coat in coloured-flowered cultivars was bigger than in white-flowered and there was significant positive correlation between tannin content and percentage of hulls in seeds (r=0.96;PO.01).
In chickens apparent protein digestibility differed significantly between whiteand coloured-flowered cultivars, but apparent fat digestibility was significantly lower only for Zuraw cv.The AME N value for chicken and metabolizability of energy of white-flowered peas was on average 12.86 MJ/kg DM and 64.5%, and differed from the values of coloured-flowered peas which were significantly   of ileal digesta by 5%; fat digestibility and AME N value of basal diet increased significantly (P<0.05),but apparent protein digestibility and organic matter retention was not affected.Enzyme supplementation of test diets also significantly lowered viscosity of digesta by 24% on average, but did not affect significantly pH of jejunal digesta, AME N value of peas or digestibility of their components (Table 4).
In rats the digestibility of protein of white-flowered peas was significantly higher than coloured-flowered cultivars (on average 85 vs 79.5%; P<0.05).In contrast the biological value of protein in 5 cultivars was similar and averaged 86, and only the B V of protein of the Idol cultivar was significantly higher (Table 5).The NPU value of the purple-flowered Zuraw cv. was significantly lower from white-flowered cv.Albatros and Piast and the red-flowered cv.Idol.The AME value for rats of white-flowered peas was on average 16 MJ/kg DM what means 1 to 2 MJ/kg higher than of coloured-flowered peas (P<0.05).After inclusion of red-flowered peas into diet, pH of caecal digesta was significantly (P<0.05)higher than on diets with white-flowered peas (Table 5).
Predicted values of ileal digestibility of protein and energy in pigs determined in vitro differed, pdN and pdE averaged respectively (in %): 86.2 and 85.4 for white-, 77.8 and 82.1 for red-, 73.5 and 77.4 for purple-flowered pea (Table 6).

DISCUSSION
The protein content in evaluated peas was lower than reported by Gdala et al. (1992) and Zduhczyk et al. (1997), but amino acid profile of protein agrees well with the values reported in both papers and with the mean values published by  W -R -P as in Table 4 WPS A (1992).The protein of evaluated cultivars had high lysine and low sulphur amino acid content.The lysine content in pea protein was about 45% higher than in meat meals, the sulphur-containing amino acids and threonine content was of the same order as in meat meals (WPSA, 1992).It makes pea a good substitute of meat meals in feed mixtures for growing chickens and pigs.In the context of meeting the dietary requirement of indispensable amino acids for poultry and pigs, cereal grains and peas are nutritionally complementary.The significantly higher biological value of protein of the Idol cv. could not be attributed to any difference in amino acid content.Mean trypsin inhibitor activity (TI) in all analyzed cultivars was 1.9 TlU/mg DM.The measured values were much lower and less variable than reported by Gdala et al. (1992), who found 5.1 in the white-and 8.9 TlU/mg DM in the coloured-flowered Polish cultivars from 1985-86 harvest years, and lower than reported by Zduhczyk et al. (1997).The last authors found 4 TlU/mg DM in white-flowered cultivars from 1992-1993 harvest year.Trypsin inhibitor activity in evaluated pea cultivars did not depend on the colour of flowers, similarly as in report of Bastianelli et al. (1998).It seems that Polish plant breeders succeeded in producing cultivars of pea of low activity of protease inhibitors.This means that one of the factors that may interfere with protein digestion in monogastric animals was nearly removed.White-coloured cultivars were low in tannin, as was expected.Tannin content in coloured-flowered cultivars was higher, but within the limits reported by Gdala et al. (1992), Gdala and Buraczewska (1997) and Bastianelli et al. (1998).Statistical evaluation of results indicated, that of all determined components of pea seeds, tannins had the most pronounced negative effect on protein digestibility.Tannin content affected negatively protein digestibility in chicken (r = -0.93;P< 0.05) and rats (r=-0.89;PO.05) and ileal digestibility of protein for pigs determined in vitro (r = -0.98;PO.001).It agrees well with reports of other authors (Lindgren, 1975;Gdala et al., 1992;Grosjean, 1999), who found significant differences in digestibility of protein between white-and coloured-flowered cultivars.The phenolic groups of tannins bind to enzymes and other proteins and form insoluble tanninprotein complexes resistant to digestive enzymes of monogastric animals, hydrogen bonds and hydrophobic interactions appear to be the principal linkages involved (Artz et al., 1987).
The BV of protein of evaluated cultivars was not correlated with tannin content, what indicates that tannins did not bind specifically with any limiting amino acid in pea protein.However NP V of evaluated cultivars depended less on the tannin content and was correlated negatively with trypsin inhibitor activity (r = -0.84;P<0.05).Due to relatively high protein content, the NPV of the Zuraw cv. was the same as NPV of the Idol cv. and higher than Piast cv.despite the differences in protein digestibility.
Metabolizable energy for chicken of white-flowered cultivars averaged 12.86, for coloured-flowered 12.27 MJ/kg DM and were in the range of the values obtained by Grosjean et al. (1999) on adult cockerels (13.18 and 12.72 MJ/kg DM, respectively) and by Grosjean et al. (1998b) on broiler chickens (on average 12.24 MJ/kg DM for white-coloured tannin-free peas).The last authors noted considerable variability in metabolizable energy for broiler chicken of 26 analyzed pea batches (from 11.66 to 12.95 MJ/kg DM), but the correlation between chemical composition criteria and energy values were weak.Similarly in this study, nutrient and dietary fibre content had no significant effect on metabolizable energy value in evaluated cultivars, while tannin content affected negatively AME N value for chicken (r =-0.99;PO.001) and AME for rats (r = -0.95;PO.01).Protein digestibility, AME N values in chickens and AME in rats were also correlated with hulls content in seeds.Coloured-flowered varieties had proportionally more seed coat and higher tannin content, which are mostly present in the hulls.
In chickens due to substitution of half of the basal diet by pea, the viscosity of intestinal digesta increased slightly (on average by 7%).After supplementation of basal and test diets with xylanase the viscosity of digesta decreased significantly (PO.001), but probably the dietary fibre of pea was resistant to the added enzyme, as the difference in viscosity between the basal and pea-containing groups extended (to 15% on average).Igbasan et al. (1997) found, that in the components of non-starch polysaccharides (NSP) of Canadian pea cultivars predominates glucose, followed by uronic acids, arabinose, xylose and galactose (on average 47,22,19,6 and 4% of total NSP,respectively).The relatively high concentrations of uronic acids and arabinose residues indicate that pectic-type substances are important components of the cell wall of peas.The results of the present experiment proves that the effect of pectic substances of pea on viscosity of digesta in chicken may be neglected, as it did not affect the energy and protein values (Table 4).
Significantly higher pH values of caecal digesta in rats fed on colour-flowered pea (Table 5) indicates for an effect of tannins on caecal fermentation.Lower intensity of fermentation may result either from direct inhibitory effect of tannins on microflora or from greater amount of nondigested protein of dietary and endogenous origin reaching large intestine.
Predicted total tract digestibility of energy in pigs (pdE) averaged 85% for 4 white-and 2 cultivars of red-flowered pea, 79% for remaining 2 red-flowered cultivars and 77%) for purple-flowered pea.PdE was not correlated with predicted ileal digestibility of protein (pdN), nor with tannin or nutrient content in seeds.Grosjean et al. (1998a) reported that apparent digestibility of energy in pigs was 89%) for white-and 81%o for coloured-flowered pea.The lower pdE value of the Zuraw cultivar was undoubtedly due to the high tannin content.
The lack of correlation of trypsin inhibitor activity with the relative weight of pancreas in rats confirms the results obtained on chickens, rats and in in vitro evaluation, that at the level present in evaluated cultivars trypsin inhibitors had not affected digestibility of protein.Al-Wesali et al. (1995) have demonstrated that in vitro purified trypsin inhibitor of pea had a far less inhibitory effect on casein digestibility than has soyabean trypsin inhibitor.However recently Hedemann et al. (1999) found, that in lines of pea which have 4-5 times more TI not only digestibility in rats, but also biological value of protein have been lowered in comparison with near-isogenic lines low in TI.It indicates, that trypsin inhibitor activity in new cultivars of pea should be continuously controlled.
The results indicate, that tannins present in colour-flowered pea may lower the digestibility of protein and metabolizability of energy in monogastric animals, while the role of dietary fibre and trypsin inhibitors in modern spring cultivars of pea is of smaller importance.However the amount of protein available to animals depends as well on the protein content in the seed.It seems that high protein content is the most important criterium of usefulness of pea in nutrition of monogastric animals nutrition but both the activity of TI and tannin content in new pea cultivars should be also monitored.

CONCLUSIONS
White flowered peas are digested better than coloured flowered peas.The digestibility is depressed by the presence of tannins.Lowering the tannin content should be included into breeding programs of new cultivars of pea, as the nutrients of high-tannin seeds are less effectively utilized by monogastric animals.
red; p -purple-flowered; 'values of CPD, CFD, OMR and AME N for pea cultivars were calculated by difference method with the use as a reference the values obtained for basal diets A or B x values obtained for basal diet B significantly differs from diet A (P< 0.05); abAB means in columns with different superscripts are significantly different: ab at PO.05; AB at P<0.01

TABLE 1
Seed dimensions, percentage of hulls in seeds and fibre fraction content in seeds, g/kg DM

TABLE 2
Chemical composition of peas and tannin content, g/kg DM, and trypsin inhibitor activity, TlU/mg DM PO.05) lower: 12.27 MJ/kg DM and 61.5% respectively (Table4).Inclusion 50% of pea into diet did not affect significantly the viscosity of jejunal digesta or pH of ileal digesta in chickens (Table4).Enzyme supplementation of basal diet lowered significantly the viscosity of jejunal digesta by about 30% and pH (

TABLE 5
Indices of nutritional value of protein, metabolizable energy value of pea, pH of caecal digesta and weight of pancreas (g/100 g body weight) of rats fed diets with seeds of different pea cultivars W,R ' P as in Table4; a , b -means in columns with different superscripts are significantly different at PO.05

TABLE 6
In vitro ileal digestibility of protein (pdN) and total tract digestibility of energy (pdE) for pigs, %