Impact on genetic differences among various chicken breeds on free amino acid contents of egg yolk and albumen

Eggs play important roles as food resources and nutraceuticals, to alleviate malnutrition and to improve health status in the world. Since free amino acids contribute to the nutritional values and food tastes, we investigated a total of 81 eggs from five chicken breeds, which are Australorp, Nagoya (NGY), Rhode Island Red (RIR), Shamo (SHA), Ukokkei, and two F1 hybrids (NGYxRIR and SHAxRIR) to test impact on genetic differences in 10 egg traits, 20 yolk amino acid traits, and 18 albumen amino acid traits. One-way ANOVA revealed significant breed effects on 10 egg traits, 20 yolk amino acid traits, and 15 albumen amino acid traits. Moreover, a significant heterosis effect on yolk aspartic acid was identified. In addition, positive correlations were found broadly among traits within each trait category (egg traits, yolk amino acid traits, and albumen amino acid traits), whereas there were basically no or weak correlations among the trait categories. These results suggest that almost all traits can be dramatically modified by genetic factor, and there will be partially independent production systems of amino acids into yolk and albumen. Since there will be typical quantitative genetic architecture of egg contents, further genetic analyses will be needed.

www.nature.com/scientificreports/ indigenous chicken breeds, which are consisting of approximately 50 breeds. In addition, hybrid vigour based on the crosses between Japanese fancy breeds (e.g., Oh-Shamo, Hinai-dori, and Nagoya) and American breeds (e.g., White Plymouth Rock and Rhode Island Red) has been used frequently to produce more delicious meat as Jidori brand 21,22 . In this study, five breeds (Australorp; AUS, Nagoya; NGY, Rhode Island Red; RIR, Shamo; SHA, and Ukokkei; UKO) 23 and two F 1 hybrids (NGYxRIR and SHAxRIR) of chickens were investigated. Especially, SHA, NGY, and RIR have been used to create three-way crossbred, Shintoku-Jidori chickens in Hokkaido Research Organization, Japan. Large genetic differences among various chicken breeds and heterosis will have a large potential to enhance egg contents at molecular level. A potential for enhancing the nutritional values by using genetic and environmental factors have been reported in egg free amino acids 16,24 . Since free amino acids contribute to the taste of foodstuffs 25,26 , fundamental information of amino acids contents in eggs will lead to increase not only the nutritional values but also food tastes. Until now, there are few reports in phenotypic correlations between egg size-related traits and egg amino acid traits. Therefore, in this study, we aimed to reveal breed and heterosis effects on yolk and albumen amino acids contents and phenotypic correlations between general egg traits and egg amino acid traits using five breeds (AUS, NGY, RIR, SHA, and UKO) and two F 1 hybrids (NGYxRIR and SHAxRIR) of chickens.

Discussion
This study aimed to reveal impact on genetic differences among various chicken breeds in 10 egg traits, 20 yolk amino acid traits, and 18 albumen amino acid traits using 5 breeds and 2 F 1 hybrids. Of them, there were significant breed effects on 10 egg traits, 20 yolk amino acid traits, and 15 albumen amino acid traits. Moreover, a significant heterosis effect on Y_Asp has been identified in the comparisons among NGY, RIR, and NGYxRIR. www.nature.com/scientificreports/ In addition, positive correlations were estimated among traits within each trait category (egg traits, yolk amino acid traits, and albumen amino acid traits), whereas there were basically no or weak phenotypic correlations among the trait categories. However, moderate negative correlations were found between size and weight traits (EW, LLE, LSE, AW, and SW) and almost all albumen amino acid traits, but not in yolk amino acids. Thus, these results suggest that almost all egg traits and yolk and albumen free amino acid traits can be dramatically modified by genetic factor. In general egg traits, five breeds including AUS, NGY, RIR, SHA, and UKO were investigated. Body weights of hens are 1.6 kg in AUS and 0.9 kg in UKO, whereas those of NGY, SHA, and RIR are around 3.0-4.5 kg 16,21 . Our previous report 16 indicated that AUS produce comparable egg weight with RIR, although body weight in AUS is significantly lower than RIR. This study revealed that small size and weight of the egg from UKO will be corresponding to the small body size. This study also indicated that eggshell color of AUS, NGY, SHA, and UKO are tinted, whereas RIR produce brown eggshell color. In addition, F 1 hybrids with the genetic background derived from RIR (NGYxRIR and SHAxRIR) produced brown eggs than the other breeds. Since these general egg traits including size, weight, and color of the egg have been mapped on several chromosomes [10][11][12] , further mapping studies with various breeds will identify more candidate genes for the regulation of the egg.
Yolk and albumen free amino acids were quantified to see effects on the genetic differences. In this study, completely the same mixed feed was fed to all hens from five breeds and two F 1 hybrids. Rearing environment was also fixed in the conventional individual cages, and age at sampling of eggs was fixed as 62 wk of age. Therefore, major parts of the variation in the present results must be caused by genetic factors. In our previous study 16 , yolk cysteine was only significantly changed by breed difference using AUS and RIR. On the other hand, in this study, all yolk amino acids and almost all albumen amino acids were significantly differed by genetic differences. The ranking of almost all amino acid contents was seen as AUS > UKO, SHAxRIR, SHA > RIR > NGY, NGYxRIR in yolk and UKO, SHA > SHAxRIR, NGY > NGYxRIR > RIR, AUS in albumen. In both yolk and albumen amino acid traits, there are similar tendencies in the order, which indicating that UKO, SHA, and SHAxRIR showed higher contents, whereas RIR, NGY, and NGYxRIR indicated lower contents. However, a unique feature was found that AUS produced higher free amino acid contents in yolk but lower in albumen.
Essential amino acids for humans are known as 10 amino acids (Arg, His, Ile, Leu, Lys, Met, Phe, Thr, Tyr, and Val), while essential amino acids for poultry are these 10 amino acids plus two (Gly and Pro) 27 . In this study, several genetic background (UKO, SHA, and SHAxRIR) can be increased essential amino acid contents in both yolk and albumen, which suggesting that essential amino acids enriched eggs using genetics can be utilized for some specific needs e.g., alleviating malnutrition and supplying children and aged people. In addition, free amino acids are well known to contribute food taste 25 . Glu, Asp, Ala, and Tyr contribute umami taste, while Met, Ala, Gly, Pro, Ser, Val, and Lys affect sweet taste 26,28 . Although Asp, Glu, and Lys show sour taste, Pro, Gly, Val, Leu, Tyr, Phe, His, Lys, Ile, Arg, and Trp indicate bitter taste 26,28 . Asp and Lys affect salty and astringent taste, respectively 26,28 . Since these 16 taste-active amino acids can be altered by breeds, taste of yolk and albumen might be different. Given that egg taste can be changed by altered contents of free amino acids, taste added designer eggs will be provided in the egg industry. Moreover, egg components act as nutraceuticals, such like antioxidant properties, antimicrobial activities, immunomodulatory, anticancer, and antihypertensive activities 2,3 . This study revealed the impact on breed in yolk and albumen amino acids contents, and the previous studies have been reported that genetic factor can modify egg contents in fatty acids and minerals [13][14][15] . Therefore, it will be interesting to investigate the genetic factor in the bioactive fractions of eggs in future.
Phenotypic correlations among yolk and albumen free amino acids were estimated in this study. Although there were moderate or high positive correlations among the most amino acids traits in each yolk and albumen, no or weak correlations of amino acid contents were seen between yolk and albumen. These results suggested www.nature.com/scientificreports/ that there will be partially independent accumulating mechanisms of free amino acids into yolk and albumen. In addition, negative correlations were found among size and mass traits and albumen amino acid traits. This may imply that relatively low concentrations of albumen amino acids are found in larger eggs, since larger eggs tend to have higher volume of albumen. This study revealed a heterosis effect on yolk aspartic acid (Y_Asp), indicating that F 1 hybrid was significantly lower content than those of parental breeds (negative heterosis). In recent poultry industry, all commercial stocks utilize heterosis by crossing different strains 18 to enhance performance in reproduction, growth, meat, and egg production using the benefits of a heterosis effect 23 . Comings and MacMurray 29 summarized that hybrids combine the best traits of both parental strains and thus are more resistant to disease and other factors in several organisms. This study found the negative heterosis in yolk aspartic acid in a hybrid between NGY and RIR, but not in a hybrid between SHA and RIR. In Japan, there are many Jidori brands 23 using hybrids between Japanese indigenous breeds and American breeds 21 . Therefore, several combinations of parental breeds should be tested in the future to find better combinations enhancing egg quality at molecular level.
This study clearly indicated that significant impacts on genetic differences among various breeds in yolk and albumen free amino acid contents. Many traits including morphology, behavior, and stress resistance traits have been studies well using animal models. Genetic architecture of quantitative traits is likely to be common feature in other species 30,31 . Amino acid contents in food stuffs have been investigated well in plant. Moderate to high heritability estimates and several QTLs for amino acid contents using genome-wide association study (GWAS) has been reported in Arabidopsis, soybean, and wheat [32][33][34][35][36][37] . These studies always suggested that complex genetic architecture of amino acid contents as well as the animal model 30,31 . Since there will be similar feature in genetic architecture of egg contents in chickens, further genetic analyses with large sample size will be needed.

Animals. Five breeds and two F 1 hybrids of chickens in the Animal Research Center, Agricultural Research
Department, Hokkaido Research Organization, Japan were used. Hokkaido Research Organization has kept all five breeds from generation to generation with more than 100 birds in each breed per generation by artificial insemination. Using the parental information, we carefully selected a wide variety of hens as possible in each group. A total of 81 hens from Australorp (AUS; n = 12), Nagoya (NGY; n = 12), Rhode Island Red (RIR; n = 12), Shamo (SHA; n = 11), Ukokkei (UKO; n = 12), F 1 from a cross between NGY male × RIR female (NGYxRIR; n = 10), and F 1 from a cross between SHA male and RIR female (SHAxRIR; n = 12) were investigated. Hens selected in each group will be representatives to see effect on genetic differences in egg traits. The hens were reared under the 16 h light and 8 h dark photoperiod cycle with free access to feeds and water in the individual cages 23 . Mixed feed for layers (Rankeeper; Marubeni Nisshin Feed Co., Ltd., Japan) was provided from 20 week of age. Detailed ingredients of the mixed feed were shown in Mori et al 16  Egg sampling and egg trait measurement. An egg from each hen (total of 81 eggs) was collected at 62 week of age. A total of 10 egg traits was measured. Egg traits included egg weight (EW), length of the long axis of egg (LLE), length of the short axis of egg (LSE), eggshell weight (SW), yolk weight (YW), albumen weight (AW), eggshell thickness (ST), eggshell color lightness (SCL), eggshell color redness (SCR), and eggshell color yellowness (SCY). Since all the egg traits measured are quantitative traits, it ideally requires much more samples. Weights were measured with an electronic balance (EK-6000H; A&D Company, Ltd., Japan). Two samples from each SHA and RIR lacked data of YW and AW, because yolk membrane was broken when yolk and albumen were divided. Sizes and eggshell thickness were measured using a digital caliper (P01 110-120; ASONE, Japan) and a Peacock dial pipe gauge P-1 (Ozaki MFG Co., Ltd., Japan), respectively. Eggshell colors were measured by a chromameter (CR-10 Plus Color Reader; Konica Minolta Japan, Inc., Japan). After measuring yolk weight, the yolk was diluted two-fold with distilled water (DW) and was mixed. The yolk solution and albumen were separately kept in tubes at − 30 °C until use.
Yolk free amino acid analysis. Amino acid analysis for yolk was performed following the previous method 16 . Five milliliter of the yolk solution was mixed with 5 mL of 16% trichloroacetic acid solution (FUJI-FILM Wako Chemicals, Japan) and vortexed. The samples were centrifuged at 1400×g for 15 min using a tabletop centrifuge (model 2410; KUBOTA Corporation Co., Ltd., Japan) and a centrifuge (RX II series; HITACHI Ltd., Japan). After collecting the supernatant using a 1 mL syringe (NIPRO Corporation, Japan), the solution was filtered out to a microtube through a disposable cellulose acetate membrane filter unit with a 0.45 μm pore size (DISMIC-25CS; Advantec Toyo Kaisha, Ltd., Japan). The filtered solution was heated at 40 °C for 60 min in a vacuum oven (VOS-201SD, Eyela, Japan) until dry. After adding 20 μL of mixing solution (ethanol: DW: triethylamine = 2:2:1), the microtube was vortexed for 20 min using a Micro Mixer E-36 (TAITEC Corporation, Japan). The sample was heated at 40 °C for 60 min in a vacuum oven. After drying, 20 μL of mixing solution (Ethanol: DW: triethylamine: phenylisothiocyanate = 7:1:1:1) was added and mixing for 20 min using a micro mixer. The sample was re-heated at 40 °C for 60 min in a vacuum oven for drying. After preprocessing, the sample tube was placed at − 30 °C until the sample was analyzed.
Yolk amino acids were analyzed by HPLC (LC-2010CHT; Shimadzu Co. Ltd., Japan). Following the same method for sample preprocessing, amino acid standards (types H and B), l-asparagine, and l-glutamine www.nature.com/scientificreports/ (FUJIFILM Wako Chemicals, Japan) were prepared. The standards were analyzed before every 30 samples. The absolute concentration of amino acids in yolk was calculated from the peaks of sample and standard.
Albumen free amino acid analysis. Albumen (100 mg) was vortexed for 1 min with 800 µL of 0.1 M HCl.
Hexan (800 µL) was added and vortex mixed for 1 min. After centrifuging at 22,000×g for 5 min, the lower layer was collected. The lower layer (400 µL) was mixed with 1200 µL of water-acetonitrile (1:2 v/v) by vortexing for 1 min. After centrifuging at 22,000×g for 5 min, the supernatant was filtered through a 0.2 µm pore size filter. Albumen amino acid was analyzed by the UHPLC system (NexeraX2, Shimadzu Co., Ltd., Japan) using a Kinetex 2.6 µm column (EVO C18; 100 × 3.0 mm). Gradient elution program was applied using two mobile phases, which are mobile phase A (17 mM of potassium dihydrogenphosphate and 3 mM of dipotassium hydrogen phosphate) and mobile phase B (mixing solution of DW: acetonitrile: methanol = 15:45:40). Precolumn derivatization was performed by the UHPLC system with 45 µL of mercaptopropionic acid, 22 µL of orthophthalaldehyde, and 7.5 µL of samples. After placing the mixture for 2 min, 5 µL of fluorenylmethyl chloroformate was mixed and placed for 2 min. Then, the sample (1 µL) was injected into the column. Flow rate was set 0.85 mL/ min and column temperature was set 35 °C. An RF-20Axs high-sensitivity fluorescence detector was used at Ch1 (Ex 350 nm and Em 450 nm) and Ch2 (Ex 266 nm and Em 305 nm).
Standard solution of amino acids mixture type H, l-asparagin, l-glutamine, l-cystine, and l-tryptophan (FUJIFILM Wako Chemicals, Japan) were preprocessed following the same method for albumen samples. Calibration curve was calculated by serial dilution of the standards. The absolute concentration of albumen amino acids was calculated from the curve.

Statistical methods.
In order to test the effect on genetic differences in egg traits and yolk and albumen amino acid traits, data were analyzed using one-way analysis of variance (ANOVA) with significant threshold (p < 0.05). For testing heterosis effect, two sets of two parental breeds and their F 1 hybrids (NGY, RIR, and NGYxRIR; SHA, RIR, and SHAxRIR) were used. Tukey's honestly significant difference (HSD) test was conducted to see inter-breed differences (p < 0.05). Data are indicated as the mean and standard deviation. Statistical analyses were performed using R software 38 . Pearson's correlations among all traits were calculated and plotted using 'corrplot' package of R. Significant threshold (p < 0.05) was used to see whether or not correlation coefficients equal zero.

Data availability
The datasets generated during the current study are available from the corresponding author on reasonable request.