Efficiency of lupine seed ( Lupinus angustifolium and Lupinus luteus ) in sow , piglet and fattener feeding

Ewa Hanczakowska1, Jerzy Księżak2, Małgorzata Świątkiewicz3 1 National Research Institute of Animal Production, Department of Animal Nutrition and Feed Science, Krakowska 1 st., 32-083 Balice, Poland 2Institute of Soil Science and Plant Cultivation State Research Institute, Department of Forage Crop Production, 24-100 Puławy, Poland 3 National Research Institute of Animal Production, Department of Animal Genetics and Breeding, Krakowska 1 st., 32-083 Balice, Poland e-mail: ewa.hanczakowska@izoo.krakow.pl


Introduction
Soybean meal is currently the most important protein source in pig feeds.Other legumes cultivated in Europe, such as peas, field beans and lupines, are used on a smaller scale.About 98% of soybean meal present on the feed market is produced from genetically modified (GMO) plants (Sieradzki et al. 2006).According to earlier experi-ments, genetic modification did not influence the nutritive value of feedstuffs (Padgette et al. 1996, Aurlich et al. 2003) and had no effect on animal performance, carcass traits and meat quality (Flachowsky et al. 2005).Also, our experiments on feeding pigs with genetically modified soybean and maize proved that these feeds had no effect on animal performance and there was no transfer of transgenic DNA to animal tissues (Świątkiewicz et al. 2011).However, there is still strong public opposition in some countries against using GMO plants in human and farm animal feeding, thus it is possible that in some cases other protein sources will have to be used.
In European countries this could be legumes such as lupines, peas or field beans.Lupines contain high content of protein, depending on species 28-48% DM, rich in lysine and arginine, as well as they are also a good source of lipids, minerals and vitamins (Sobotka et al. 2016).The presence of antinutritive substances limits the nutritive value of legume seed protein, but the alkaloid content in 'sweet' varieties of lupines was lowered to 0.01% (Ruiz et al. 1977) so that they can be used as supplements, even in food for humans (Lee et al. 2006).In the experiment of Roth-Maier et al. (2004) replacing part of soybean meal by 20% of yellow or blue lupines seeds produced better results in fattener feeding than soybean meal used as a control.In the feeding of young animals, lupine must be used cautiously due to their not fully developed digestive system, but in the experiment of Prandini et al. (2005) white lupine, used at a level of 17%, gave similar results to 16% of soybean meal.Similar results were also obtained by Kim et al. (2008) in their experiment with yellow lupine.On the other hand, according to the earlier experiment of McNiven and Casteli (1995), the amount of lupine (white) in piglet feed should not exceed 10%.Also, Cherriére et al. (2003) recommended to limit the rate of inclusion of lupine to < 10% in weaned piglet diets.

Manuscript received October 2016
There is not much information available about using lupines in sow feeding and that cited by Casper et al. (1991) in their study was not good.Sows fed with lupine meal refused their feed and delivered small litters with a high percentage of stillborn foetuses.In this case, however, not native but dehulled and hexane extracted lupine meal was used.
The aim of this experiment was to examine the possibility of replacing a part of soybean meal in sow, piglet and growing finishing pig feed (i.e. during the whole production cycle) by high and low alkaloid varieties of two species of lupines.

Material and methods
All methods used in this experiment were accepted by the Local Ethics Commission for Experiments on Animals.
Animals and diets 50 sows of Polish Landrace were mated (second parity in all groups) with a Polish Landrace boar and kept and fed individually from mating to the end of 35 d lactation.After mating, the sows were allocated to 5 groups, with 10 animals in each.Group I (control) received standard feed mixture containing soybean meal as the main protein source.The remaining groups received standard feed mixture in which part of the soybean meal was replaced by blue lupine (Lupinus angustifolius) cv.Graf registered 2004, (Group II), blue lupine cv.Karo registered 2001 (Group III), yellow lupine (Lupinus luteus) cv.Mister registered 2003 (Group IV) and yellow lupine cv.Parys registered 1988 (Group V).The Karo and Parys varieties are high-and the Graf and Mister -are low alkaloid ones.All lupines were harvested in 2012.The composition of the diets for the pregnant and lactating sows is given in Table 1.
At 100 d of pregnancy the sows were moved to a farrowing house and also kept in individual pens until the weaning of the piglets.The sows received 2.5 kg of mixture per day from mating to 100 d of pregnancy and 3.5 kg from 100 d of pregnancy to farrowing.During lactation the administered amount of feed depended on the litter size: 1.80 kg per sow and 0.40 kg per piglet.Water was available ad libitum.The sows were weighed at mating, farrowing and weaning of the piglets.
The piglets were kept in group pens, each litter in a separate pen.They were weaned at 35d of life.Before weaning, from the 7 d of life they were fed ad libitum with the same, standard prestarter diet, whose composition is given in Table 2.After weaning the piglets received diets containing lupines corresponding to the feed for the sows.In each group half the animals (5 litters) received the fibrolitic enzyme RONOZYME VP (DSM Nutritional Products Ltd., Mszczonow, Poland) at 200 mg per kg of feed.RONOZYME VP i.e. carbohydrase complex is produced by a submerged fermentation of an Aspergillus aculeatus strain, which contains endo-1,3(4)-beta-glucanase (min.50 FBG/g product) and various hemicellulase and pectic-substance hydrolysing activities.The diet's composition is also given in Table 2.The piglets were weighed at 35 (weaning), 56 and 84 d of life.
The piglets were raised to 84 d of life and then 20 animals (10 gilts and 10 barrows) from each group were randomly chosen for further fattening.They were fed with standard feed mixtures (Table 3) grower (to 60 ± 1.0 kg of body weight) and finisher (from 60 ± 1.0 to the end of the experiment, Table 3).In each group half the animals received the enzyme RONOZYME VP at 200 mg kg -1 feed.The fatteners were slaughtered at an average body mass 107 ± 2.0 kg.The quality of the carcasses was evaluated according to Tyra and Żak (2012).24 h after slaughter the pH of the meat was measured and samples of the longissimus muscle, obtained from the area of the last thoracic and first lumbar vertebra, were taken for analysis.
Apparent digestibility was evaluated in parallel to the fattening experiment, using the same feeds, on 30 barrows weighing around 40 kg (grower) and 80 kg (finisher), not used in the fattening part of the experiment.The plan of the digestibility trial was the same as that of the fattening experiment, except for cv.Karo.Each group consisted of 6 fatterners.The animals were kept individually in balance cages and fed with the same feeds as in the fattening experiment.The preliminary period lasted 10 d and the samples collection lasted 5 d.Faeces from each animal were collected daily, weighed and frozen at -20 °C.At the end of the collection period, faeces samples from each animal were mixed and a representative sample was prepared.

Chemical and physical analyses
Gross composition of the feeds and faeces samples were analyzed according to AOAC (2005) methods.Amino acids were analyzed using the AAA 400 INGOS automatic analyzer and that of alkaloids with the gas chromatographic method according to Carsten and Wink (1992).

Statistics
Statistical analysis of treatment effect was performed by one-way (sow reproductive rates), two-way (weaned piglets rearing indices and apparent digestibility coefficient) and three-way (fattening results) analysis of variance.
Comparison of means was conducted with Duncan's multiple range test at p ≤ 0.05 and p ≤ 0.01 levels of significance.All the analyses were conducted using the STATISTICA 10 package (StatSoft 2011).

Results
Both high alkaloid varieties (Karo and Parys) contained a lower amount of protein when compared to their low alkaloid equivalents within species (Table 4).All lupine seeds contained less protein and more fat and fiber than soybean meal.Lupine protein was relatively poor in methionine and tryptophan.Alkaloids content in the blue lupine was almost a hundred times higher in cv.Karo than in cv.Graf.The yellow lupine cv.Parys contained five times as many alkaloids as the cv.Mister.
In the second part of piglet experiment (Table 6) i.e. between 35 and 84 d of life, both low alkaloid varieties of lupines i.e.Graf and Mister gave significantly (p < 0.01) better body weight gains than the high alkalois varieties, and also better than the soybean meal.There was no difference in feed intake or feed utilization during the whole experiment i.e. between 35 and 84 d of life.In this part of the experiment the enzyme supplement significantly improved (p < 0.05) piglet body weight gains but during its further part (Table 7) the weight gains equaled.In the first half of the fattening period of the experiment i.e. between ca 23 kg and 60 kg of pig body weight, almost all the animals fed with lupine grew faster than the control ones (Table 7) but the differences were not significant.
The exception was blue lupine Karo which produced growth rates significantly (p < 0.01) lower than all the other feeds.The results obtained during the whole fattening period (23-107kg) also showed the lowest body weight gains in pigs fed with blue lupine Karo (p < 0.01), but the best result appeared in control soybean meal group.Pigs fed with blue lupine Karo also characterized by the worst feed mixture utilization.The enzyme addition improved body weight gains (by 3.8%) only during the first fattening period (23-60kg), but the differences were not statistically significant.However, pigs receiving enzymes in the diet utilized feed better than controls (p < 0.01).It was observed that animal sex has not affected the fattening results.
Yellow lupine Parys produced the worst results in carcass evaluation (Table 7).All the important measurements (dressing yield, meat of ham, loin eye area, meat of primal cuts) were significantly (p < 0.01) lower in this group than those of the control.The best of the lupines was blue lupine Graf, which was comparable to the soybean meal.Carcasses of pigs fed with enzymes supplementation characterized of significantly higher loin eye area and thinner backfat (p < 0.01).The most beneficial carcass traits were observed in gilts -their carcasses contained more meat in primal cuts and less backfat.
Acidity of the meat 24h after slaughter (Table 8) ranged from pH 5.51 (control) to 5.66 (cv.Graf and Mister) and these differences, though not big, were statistically significant (p < 0.01).The biggest difference in water holding capacity index (i.e.18.3 and 24.7; p < 0.01) was found within the species L. luteus between varieties Parys and Mister, respectively.The meat of pigs fed with both varieties of blue lupines was more saturated in redness than that of the control's (p < 0.01).In sensory evaluation, the highest values of all the traits were found in the control group and the lowest ones in the group fed with L. luteus cv.Parys.Evaluation of odour was lowered (p < 0.05) by the enzyme supplement.The sex hasn't influenced the meat quality, except some differences in the meat color.
Intramuscular fat (IMF) in the meat of pigs fed with lupines (except that of cv.Parys) contained more saturated fatty acids (SFA) than the control one, though this difference was significant only in the case of cv.Mister (Table 9).There was no significant difference between the groups in the content of monounsaturated (MUFA) or total content of polyunsaturated (PUFA) fatty acids.MUFAs dominated in the intramuscular fat of all the groups and SFAs were second.The meat of all the pigs fed with lupines contained a lower amount of PUFA n-3 acids (p < 0.01) than the control.Animals receiving yellow lupine Parys had the best athero-and thrombogenicity indexes and those fed with cv.Mister the worst.The enzyme addition in pig feed, as well as animals' sex, haven't changed the amount of saturated nor unsaturated fatty acids in IMF of meat.
The animals used in the digestibility tests did not come from the fattening experiment and they were not accustomed to the presence of lupine in feed.They refused consumption of the blue lupine Karo, therefore we could not obtain reliable results.There was no significant difference in protein or fat apparent digestibility of the grower mixtures (Table 10), but in the finisher mixtures the protein digestibility of the control mixture was higher (p < 0.05) than in all the other groups.Fat digestibility was the lowest (p < 0.05) in the finisher period in the mixture containing blue lupine Graf.Significant differences were also found in the fiber digestibility, which in the later part of the experiment was significantly (p < 0.01) higher in all the groups fed with lupines.The enzyme supplement had no significant effect on nutrients digestibility in the grower period, although higher coefficients were noticed in pigs obtaining this additive.Significantly improved apparent digestibility of protein (p < 0.05), fat (p < 0.05) and fiber (p < 0.01) were observed in the finisher period of the experiment, in pigs receiving the enzyme additive.

Discussion
The low alkaloid varieties of both examined species of lupines (Graf and Mister) had a favourable composition when compared to the high alkaloid ones.They contained more protein and, of course, less alkaloids.Results obtained suggest that at the beginning of the experiment pigs at least partially became used to such a feed: piglet weight gains were similar to controls though were worse than those fed with sweet blue lupine.On the other hand, however, when older pigs in the digestibility trial first time received this lupine, they refused consumption.All the lupines contained less protein and amino acids than the soybean meal used as the control but the soybean was extracted (hence the relatively low fat content) and the lupines were unprocessed.This slightly increased the crude fat content in the experimental diets.The amino acid composition of the protein of the tested seeds was similar to that quoted by Schumacher et al. (2011).Differences in amino acid composition of protein of various varieties of legumes are generally small because yield, protein and antinutritive substances content are the main object of interests of plant breeders, and amino acids are of secondary importance (Wang et al. 2003).
There are not many experiments on feeding sows with lupine quoted in the literature.Early experiments performed in 1987 by pork producers using hexane-extracted dehulled lupine meal gave adverse results such as feed refusal, poor growth rate and constipation.Pregnant sows appeared to have an increased incidence of stillborn pigs and reduced litter sizes (Casper et al. 1991).As the lupine used was relatively low alkaloid (< 0.05%) it can be assumed that the poor results were due to the not removed hexane residue.If there was no further processing, hexane extraction alone could reduce the smell and nutritive value of the feed (Honig et al. 1976).In the present experiment reproduction rates of the sows receiving soybean were better than those fed with lupines taking in account the weight of the piglets born.Better results were obtained using lupines and other legumes in sow feeding in the previous experiment of Hanczakowska and Świątkiewicz (2013), in which sows fed L. luteus cv.
Mister mixed with rapeseed press cake born heavier piglets than those fed with soybean.
In the experiment of McNiven and Casteli (1995) weaned (also at 35 d of life) piglets fed with white lupine grew as well as those fed with soybean if the amount of lupine in the feed was lower than 15%.Decreasing body weight gains with higher lupine level in the diet were probably due to lower feed intake.According to these authors the results of all the groups also depended on the environmental conditions (different barn).Prandini et al. (2005), also using white lupine, found no significant difference in body weight gains of piglets between 1 and 42 d of life when compared to those fed with soybean.Extrusion did not improve animal growth.No improvement of protein digestibility of white lupine in piglets after supplementing feed (30% of lupine) with alpha-galactosidases was also found by Pires et al. (2007).Body weight gains cannot be taken into account because the piglets were submitted to an ileo-rectal anastomosis; as such, there was no significant difference in piglet body weight at the end of the experiment.The results of the present and quoted experiments suggest that lupine can be used in piglet diets at the level of 6-15%, and additional treatments such as extrusion at the beginning of the experiment do not improve its nutritive value.Also the supplements of fibrolitic enzymes has only limited effects.
In the first period of fattening i.e. from 23 to 60 kg of body weight the pigs receiving lupines grew slightly better than those fed with soybean but the differences were not significant.In the finisher period (60-107 kg) the control pigs grew faster.Similar results were obtained by Partanen et al. (2006) when blue lupine replaced rapeseed pressed cake at a level of 33%.In the earlier experiment of Hanczakowska and Świątkiewicz (2014) the mean daily body weight gains of fatteners fed with soybean meal (781 g) were intermediate between those of the pigs fed with blue (744 g) or yellow lupine (810 g) mixed with rapeseed press cake and these differences were not significant.In the present experiment the only exception were pigs fed with blue lupine cv.Karo, which grew significantly worse, probably due to its high alkaloid content given that pigs are particularly sensitive to the bitter taste of alkaloids (Godfrey et al. 1985).In this experiment the enzyme supplement had no positive effect on fattening results.The nutritive value of lupine seeds may be improved by dehulling or supplementing feed with fat but the results of different experiments are inconsistent.In any case such treatments complicate and increase the cost of feeding.The nutritive value of lupine for pigs was reviewed by Pisarikova and Zraly ( 2009) and there was no significant difference in the meatiness of the carcass.Generally low alkaloid varieties (Graf and Mister) gave better results than high alkaloid ones.No difference in carcass quality was found also by Sońta et al. (2015) and by Zraly et al. (2007), who used white lupine.
There are only a few publications on the quality of meat of lupine-fed pigs.In the experiment of Leikus et al. (2004), the replacing of soybean meal with white lupine had no effect on carcass and meat quality.This is inconsistent with the results of the presented experiment in which meat quality traits of fatteners fed with the varieties Karo and Parys were significantly lower than those of the control's.Also, the low alkaloid varieties were worse in some organoleptic traits such as taste (cv.Mister) or juiciness (cv.Graf) than the soybean control.These differences could be due to differences in the fatty acids contained in the meat.According to Wood et al. (2008) fatty acid compositions has profound effects on meat quality.In the case of cv.Parys poor odour could be due to the presence of capric acid (C10) which is known for its bad smell (Oprean et al. 2011).Differences in alkaloid content in the seeds seem to have no significant effect on meat taste.
According to Gatlin et al. (2002) the fatty acid profile of fat deposited in pigs is similar to that given in the diet, although according to Lopez-Bote and Rey (2001) this concerns mainly adipose tissue.Also, Froidmont et al. (2005) found big differences between the fatty acid pattern in the meat and backfat of pigs fed with a feed based on wheat and white lupine.In the present experiment there was no significant difference between the groups in the total content of MUFA and PUFA in fat of longissimus thoracis.The biggest difference in SFA content in meat was found between the two varieties of yellow lupine, but because fatty acid patterns of lupines' seeds lipids were not analyzed, it is hard to find reasons for these differences.Anyway, according to data quoted in literature long chain saturated fatty acids predominate in lupine seed fat (Zraly et al. 2007).Andrzejewska et al. (2016) analyzed fatty acid profile in seeds of different lupines, among them variety of blue lupine used also in this experiment (cv.Karo).They found that linoleic acid (C18:2) was most abundant and oleic acid (C18:1) was second.Generally, the fatty acid pattern of the meat found in this experiment (MUFA prevailing, SFA the second) was very similar to that found by Zraly et al. (2007).Despite this similarity PUFA n6/n3 ratio was high in our experiment due to a higher amount of PUFA n6 and a lower amount of PUFA n3 in the lupine groups.Such a ratio cannot be recommended from a health point of view (Kuhnt et al. 2012).According to recommended optimum this latter ratio should be 1/1 to 4/1 (Simopoulos 2004) though in European diets it ranges from 15/1 to about 17/1 (Fernandez et al. 2007).Thus in meat of animals from both experimental groups n6/n3 ratio is considerably above this limit.
There was no difference in the digestibility of the feed protein and fat in the first stage of the experiment.Similar results were obtained by Stanek et al. (2012) but they used older animals (BW about 80 kg), thus their results may be compared rather to the second stage in our experiment in which we found decreasing digestibility of protein and fat but improving digestibility of fiber.Stanek et al. (2012) found no differences in protein and fat digestibility of nutrients of the feed containing soybean meal or blue lupines except for one variety which was significantly worse than the other ones.No difference in protein or fat digestibility in growing pigs was also found in the experiment of Flis et al. (1996) on yellow lupine.Also, Yang et al. (2007) found no improvement in the apparent digestibility of nutrients of feed containing blue lupine, though the apparent ileal digestibility of some amino acids (lysine, methionine and phenylalanine) was better in the expanded seeds.According to these authors this improvement was due to the effects of changes in non-starch poly-and oligosaccharides contained in the seeds.

Conclusions
The results obtained suggest that blue and yellow lupine seeds may be, in moderate amounts, good replacements of soybean meal in the weaned piglet and growing pig nutrition.Apparent differences were found between the high alkaloid and low alkaloid varieties in the seeds' chemical composition, especially in their protein and alkaloid content.Low alkaloid varieties of both blue and yellow lupine given at a level of 6 or 10% gave better results in piglet feeding than soybean meal.Feed enzyme supplement has only limited effect on piglet and growing pig performance.They also gave results comparable to soybean meal in growing pig feeding but the meat quality was lower, especially in the case of yellow lupines.
mean values in the same row with different letters differ significantly at p ≤ 0.05; A, B = mean values in the same row with different letters differ significantly at p ≤ 0.01; 1 = interaction; NS = not significant; SEM = standard error of the mean a, b = mean values in the same row with different letters differ significantly at p ≤ 0.05; A, B = mean values in the same row with different letters differ significantly at p ≤ 0.01; SFA = sum of saturated fatty acids; UFA = sum of unsaturated fatty acids; MUFA = sum of monounsaturated fatty acids; PUFA = sum of polyunsaturated fatty acids; 1 = interaction; NS = not significant; SEM = standard error of the mean Table 9. Fatty acid composition (g 100 g -1 detected fatty acids) and partial sums of fatty acids and related indices of Longissimus thoracis

Table 1 .
Composition and nutrient contents of the diets for experimental sows (g kg -1 )

Table 10 .
Apparent digestibility coefficients of grower and finisher feed mixtures a, b = mean values in the same row with different letters differ significantly at p ≤ 0.05; A, B = mean values in the same row with different letters differ significantly at p ≤ 0.01; 1 = interaction; NS = not significant; SEM = standard error of the mean