Analysis of the source of aggressiveness in gamecocks.

Although the fighting behaviour in gamecocks has evolved because of artificial selection, it is unknown whether the selection for aggressiveness affects neurotransmitter levels in the avian central nervous system. We sought to identify the source and origin of this trait. We collected the brain samples from 6 female Shamo gamecocks and 5 Shaver Brown chickens (control; bred for egg production). The midbrain levels of norepinephrine (NE) were significantly higher in Shamo gamecocks (P = 0.0087) than in the controls. Moreover, alleles encoding adrenergic receptors differed between the breeds in terms of response to NE. Gene mutations specific to Shamo and potentially associated with fighting behaviour were in sites T440N of ADRα1D; V296I of ADRα2A; and T44I, Q232R, and T277M of ADRβ2. The evolutionary analysis indicated that the ADRβ2 (T44I and Q232R) mutations were heritable in all Galliformes, whereas the T440N mutation of ADRα1D and V296I mutations of ADRα2A were unique to Shamo and originated by artificial selection. A high NE level may confer a selective advantage by enabling gamecocks to be aggressive and pain tolerant. Therefore, the strong fighting behaviour of Shamo has resulted from a combination of naturally inherited and mutant genes derived by artificial selection.

Seventeen varieties of domesticated chicken (Gallus gallus domesticus) known as Japanese ornamental chickens were developed by artificial selection over a long period for their cultural entertainment value. They are characterised by different body colours and shapes [1][2][3][4][5] . Shamo gamecocks were bred for fighting. The birds were artificially selected under extreme stress to isolate desirable traits such as strength, aggression, and anxiety and pain endurance during cockfighting 3,4 . The shape and colour characteristics of modern-day Japanese Shamo have been depicted in wildlife caricatures approximately 1000 y ago. Shamo have body shapes and fighting styles distinct from those of other chickens 5,6 . After a match, losing Shamo males are euthanised by the breeders, whereas the victorious males are bred with healthy Shamo females. These pedigreed lines have been passed down from generation to generation and are highly valuable to breeders 2 .
The objective of the present study was to elucidate the source of the fighting behaviour in Shamo gamecocks. Our previous molecular evolutionary studies based on the mitochondrial D-loop region revealed that the Shamo gamecock breed originated from red junglefowl 2,6 . As Japanese ornamental chickens phenotypically differ from each other, intensive artificial selection may have been conducted before these varieties diverged from ancestral Shamo in Okinawa 2,6 . We also investigated the degree of genetic differentiation among Shamo chickens by focusing on the genes encoding dopamine receptors D2, D3, and D4 in domesticated chicken populations 7 . The dopamine receptors receive the neurotransmitter when it is released from presynaptic nerve terminals. They trigger crucial physiological responses regulating movement, cognition, reward, and hormone release. Genetic differentiation was evaluated using the nucleotide differentiation (N ST ) index. We found that the N ST of DRD4 (dopamine receptor D4 gene) in Shamo (0.072) was significantly higher than that of the other DRD genes. The genes responsible for aggressiveness, behaviour, and other traits were analysed by array comparative genomic hybridisation (aCGH) in culturally domesticated chickens, gamecocks, and ornamental chickens 1 . The assay revealed 782 gene probe candidates for artificial selection pressure in culturally domesticated chickens.
Here, the aim was to examine the effects of neurotransmitters and their receptors in the brain on Shamo strength and aggression under high-stress conditions used for artificial selection. To the best of our knowledge, no previous research has explored the stress response in Shamo from different perspectives. Moreover, there has been no study on the neurotransmitters in the brains of Shamo gamecocks. This information will increase our understanding of the mechanism underlying the aggressive behaviour of Shamo gamecocks. It may also help Analysis of polymorphisms related to adrenergic receptor genes. There were significant differences between Shamo and Shaver Brown in terms of their brain Epi and NE levels. Therefore, we analysed the genes encoding the receptors for these neurotransmitters. We sampled three Shamo individuals with the highest NE levels and three Shaver Brown birds with the lowest NE levels (Fig. 1A). We then sequenced these samples to localise mutation sites in the ADRα1A, ADRα1B, ADRα1D, ADRα2A, ADRα2B, ADRα2C, ADRβ1, ADRβ2, and ADRβ3 receptor genes (Tables 1 and S4).
Shamo and Shaver Brown had the same original mutations, namely, V494M in ADRα1B and V58I and D273E in ADRα2a. R342C, S396P, and Q404L of ADRβ3 were observed in both breeds.
Prediction of transmembrane helices in proteins encoded by the ADR genes. We analysed transmembrane helices in the proteins encoded by each mutant allele using TMHMM server. The ADR genes encode membrane proteins. Each of these proteins has seven transmembrane structures ( Fig. 3A-C; Table S5A-C). The TMHMM analysis indicated that the most highly mutated gene was ADRβ2: T277M, T44I ( Fig. 3C; Table S5C). The mutated sectors were located in the TMhelix (transmembrane region), whereas Q232R was intracellular on the cytoplasmic side. A15T was extracellular and not directly related to behaviour.
Several mutations were observed in D273E and V296I of ADRα2A ( Fig. 3A; Table S5A). In ADRα2B, mutations were detected at the R138Q and R210H sites in bird S6 and the V292M site in birds S7 and S9. The mutations at the R138Q and R210H sites were intracellular on the cytoplasmic side and possibly related to behaviour. As the V292M site was extracellular, it was probably not associated with bird behaviour. The T440N site of ADRα1D was also found in Shamo ( Fig. 3B; Table S5B). The differences in the number of mutations in this membrane protein suggest that ADRβ2 is more closely associated with fighting behaviour than ADRα2. T 441 I (TMhelix) and Q232R (inside) of ADRβ2 may regulate fighting behaviour ( Fig. 3C; Table S5C). The TMHMM analysis disclosed that Shaver Brown birds N5, N6, and N7 had mutations on intracellular ADRα1A (S365G) and ADRβ1 (Q403R). These mutations conferred the birds a selective advantage as egg layers. Furthermore, ADRα1B (V494M), ADRα2A (V58I and D273E), and ADRβ3 (R342C, S396P, and Q404L) mutated simultaneously with breeder domestication.

Molecular phylogeny analysis of aggressiveness in Shamo based on ADRα1D, ADRα2A, and
ADRβ2. We conducted an evolutionary analysis to identify the origin of aggressiveness. We constructed phylogenetic trees using ADRα1D(1536 bp): T440N, ADRα2A(1332 bp): D273E, V296I, and ADRβ2(1170 bp): T44I, Q232R. We also extracted sequences of red junglefowl, wild turkey, Guinea fowl, and Japanese quail from the NCBI database 17,18 . These species exhibit strong aggressive behaviours. For example, pheasants can be territorial [19][20][21] . The evolutionary analysis indicated that these birds have similar mutations. The ADRα1D: T440N, ADRα2A: D273E and V296I mutations appeared exclusively in Shamo and not in any other Phasianidae species. In contrast, ADRβ2: T44I and Q232R were confirmed in wild turkey, Guinea fowl, Japanese quail, and Shamo ( Fig. 4). Hence, red junglefowl was the only species that lacked these two mutations.

Discussion
Here, we examined the effects of brain neurotransmitters on Shamo gamecock behaviour. In this breed, artificial selection under extreme stress has improved bird strength and aggression. We analysed 10 brain neurotransmitters and the ADR genes that may either regulate aggression in Shamo used for cockfights or determine calmness of Shaver Brown bred for egg production under high population densities.
We observed significant differences between Shamo and Shaver Brown in terms of Mid NE level. The Mid NE levels were found to be closely related to Shamo aggressiveness. The Mid NE level in S6 was four times higher than that in N5 (Fig. 1A). NE is secreted from the ends of the sympathetic nerves in response to fighting stimulation. It serves as an alarm system in the brain and attenuates stress and tension caused by sudden unpleasant internal and external stimuli. The NE alarm system reduces anxiety and fear [22][23][24][25][26] . NE increases attention, concentration, judgment, and motivation. It is also known as the 'angry hormone' as it induces corporeal tension or excitement in response to stress 23,27,28 . Aggressiveness increases as negative emotions intensify. NE also increases the heart rate and blood flow to the brain and skeletal muscles 29 . Therefore, excess NE increases nerve activity, and thus, the animal becomes restless and aggressive 30 . A portion of the Mid is related to the pathway of sensory and motor neurons [31][32][33][34] , maintains reflex eyeball movements, and adjusts iris contraction 35 .
We analysed the receptor genes (ADR) regulated by Epi or NE (Table 1) [36][37][38][39] . In both breeds, 54 amino acid mutation sites were confirmed in eight ADR genes (Shamo, 33 sites; Shaver Brown, 21 sites). Prediction of transmembrane helices in proteins (TMHMM) suggested an association between fighting behaviour and specific mutations, namely, T440N of ADRα1D; V296I and D273E of ADRα2A; and T44 I, Q232R, and T277M of ADRβ2 ( Fig. 3A-C). Several mutations may be related to domestication including ADRα1B (V494M), ADRα2A (V58I), and ADRβ3 (R342C, S396P, and Q404L) in domesticated chickens originally selected from the wild type (G. gallus). The mutation sites in ADRα1A (S365G) may be important for selecting traits associated with calmness and non-aggressiveness. Thus, these mutations may not be related to fighting behaviour, but are advantageous for housed birds raised for human food production. Moreover, T44I and Q232R of ADRβ2 have been detected in other Galliformes members using the molecular phylogeny analysis. We could not confirm these mutations in the wild type (G. gallus) used in our previous study 1,7 .
The wild type still maintains a substantial proportion of ancestral polymorphisms at the genomic level 40 . Our phylogenetic analysis revealed that Shamo are distinctively aggressive and that this trait evolved by combining mutations derived from artificial selection and natural adaptation in Galliformes (Fig. 4).
The vascular smooth muscle relaxes in response to the ADRβ subunit (Gs) and contracts in response to the ADRα subunit (Gi). This mechanism substantially affects the blood vessel responses during sympathetic excitement. In this state, the ADRα subunit (Gi) is expressed mainly in the blood vessels that regulate blood flow to the heart and to the skeletal muscles required for fighting [41][42][43][44] . The blood vessels of the heart, lungs, and skeletal muscles mainly express the ADRβ subunit (Gs) 12 and expand in response to sympathetic excitation [45][46][47][48][49][50][51][52][53][54] . In this way, they maintain blood flow to the organs required for fighting. Therefore, relative differences in NE level and specific receptor mutations were observed in different environments, and they favoured aggression in Shamo and a calm nature in Shaver Brown living communally. These mutations occurred in captive populations and were deemed necessary for domestication 55,56 .
Changes in the NE receptor genes and hormone sensitivity occur as the NE level increases 57 . Excess NE increases the aggressiveness of Shamo, whereas low NE levels result in lethargy but enable Shamo to endure prolonged fighting stress.

Gene name
Shaver Brown (Control chicken) Shamo (Gamecock)   N5  TMHMM  N6  TMHMM  N7  TMHMM  S6  TMHMM  S7  TMHMM  S9  TMHMM   ADRα1A exon2  S365G  inside  S365G  Inside  S365G  inside  ---ADRα1B exon1  -R258Q  Inside  ----ADRα1B exon2  V494M  inside  V494M  Inside  V494M  inside  V494M  inside  V494M  inside  V494M  inside   ADRα1D exon1  ----L58W  Tmhelix  -ADRα1D exon2  ---T440N  inside  T440N  inside  T440N  inside   ADRα2A   -V58I  Tmhelix  V58I  Tmhelix  V58I  Tmhelix  V58I  Tmhelix  V58I  Tmhelix   -D273E  Inside  -D273E  inside  D273E inside  www.nature.com/scientificreports www.nature.com/scientificreports/ Although dopamine is the precursor of NE, we observe no significant differences in the brains of the different breeds in terms of the dopamine levels. However, in our previous study, we found that the N ST of DRD4 in Shamo (0.072) was significantly higher than that of the other genes in domesticated chicken populations 7 . Therefore, the fighting behaviour of Shamo is more closely related to the norepinephrine level than the dopamine level, and mutations in the ADR receptor have the strongest influence on fighting behaviour. Thus, the accumulation of several ADR polymorphisms and elevation in the brain NE levels may have conserved the aggressiveness in Shamo.
NE also reduces pain perception [58][59][60][61][62] . ADRα2A, closely associated with pain response, was mutated in Shamo bred for cockfights 63,64 . The Mid NE alleviates pain via the ADRα2 receptor gene 58,63,64 . In combats, Shamo birds either fight or fly away ('fight-or-flight'). They must be able to take immediate action in response to changes in the NE level 55,65 . If Shamo birds were insensitive to pain during fighting, they could either continue fighting or escape. Several receptors may be involved in stress relief. Isolation of mutant alleles in Shamo by aCGH revealed the γ-aminobutyric acid (GABA-a) receptor gene associated with the central nervous system (CNS) 1 . The GABA-a receptor mitigates anxiety and relaxes skeletal muscles [66][67][68] . Artificial selection of Shamo for cockfight enhanced its fighting behaviour by reducing stress related to anxiety and fear.  . Phylogenetic tree of ADRβ2 in Galliformes members. ADRβ2: T44I and Q232R were confirmed in wild turkey, Guinea fowl, and Japanese quail (Fig. 4). Red junglefowl was the only species lacking these three mutations. Diamonds indicate the T44I and Q232R mutations. (2020) 10:7005 | https://doi.org/10.1038/s41598-020-63961-1 www.nature.com/scientificreports www.nature.com/scientificreports/ The CNS responds to stimuli from receptors and induces feelings or sensations. This response may be used to assess the effects of neurotransmitters on the brain. The 'fight-or-flight' response is a conserved behaviour in vertebrates experiencing fear, stress, or intense physical activity. Thus, we can investigate the effects of the neurotransmitters triggering the 'fight-or-flight' response in the brain 69 .
On the basis of the typical brain monoamine levels, we propose that aggressiveness may be related to the ADR genes based on the observed relative differences in brain NE level. Aggressiveness is determined by several genes and neurotransmitters [70][71][72][73][74][75][76] . Identification of these factors may help elucidate their modes of action.
In future research, we will investigate the gene levels of receptors for catecholamines and other substances in various parts of the brains of Shamo and other chickens. We will also measure neurotransmitter levels after cockfights and analyse the exomes of ornamental chickens, Shamo, and other chickens bred for different purposes. We will explore the relationships between chicken domestication and neurotransmitters and the effect of artificial selection on the genes implicated in this process. Here, we identified the adrenergic genes in Shaver Brown crucial for group living and those related to the combative trait in Shamo. Therefore, we will conduct ex vivo assays of the localised NE levels in the Mid of live animals and/or animal models subjected to various stressors.
Stress contributes to numerous human diseases and influences brain neurotransmitter levels. In the future, we will examine the effects of neurotransmitters that govern the 'fight-or-flight' response. Elucidation of the neurochemical mechanisms involved in this process may help improve defences against the deleterious effects of stress.

conclusions
Our results showed that the Mid level of NE was significantly higher in Shamo bred for fighting than in Shaver Brown bred for egg laying. The fighting instinct might be correlated with the type and distribution of ADRβ2 alleles and their responses to the NE level. Gene polymorphisms can change with the NE level. Therefore, we believe that the strong fighting behaviour of Shamo was probably due to the combination of naturally inherited genes with mutant genes derived from artificial selection. Chickens and brain samples. Shamo chicks were randomly selected from various areas and breeders.

Materials and Methods
These birds would have otherwise been used in cockfights. The Shaver Brown chicks were randomly selected from several farms. Shaver Brown was selected as the docile breed. It has a superior egg production trait. Furthermore, it was previously described as a strong candidate for artificial selection. The brain samples were randomly collected from 6 female Shamo birds and 5 domesticated Shaver Brown chickens. The latter were bred as egg layers and served as the controls 77 (Fig. 5). Shamo females were used because they are invaluable in the genetic improvement of chicken lines 2 . Shamo birds were individually maintained in separate cages. Shaver Brown chickens were maintained in groups outside. Samples were obtained from 24-wk-old chickens, because at this age, the females start ovulating and the males are trained for cockfights. The chickens were decapitated and the backs of their skull were opened with pliers. Then, the brains were rapidly excised and divided into Cx, St, and Mid according to the Chicken Brain Atlas (http://avianbrain.org/nomen/Chicken_Atlas.html) [78][79][80][81] . These tissues were selected because they are easily distinguished from each other and respond to multiple neurotransmitters. The brain samples were frozen on dry ice, weighed, and stored at −80 °C. Images of Shamo brains are shown in Fig. 6.
The skull of Shamo birds was more difficult to remove than that of Shaver Brown chickens as the former birds have been strengthened for cockfighting by selective breeding. Thus, the removal of their skull required www.nature.com/scientificreports www.nature.com/scientificreports/ considerable force and resulted in their skulls shattering into numerous fragments. Moreover, the brain of Shamo birds had a higher tension and gloss than the brain of Shaver Brown birds (Fig. 6). DNA extraction, polymerase chain reaction (PCR) amplification, and sequencing of the adrenaline genes (ADRα1A, ADRα1B, ADRα1D, ADRα2A, ADRα2B, ADRα2C, ADRβ1, ADRβ2, and ADRβ3). Blood was drawn from each chicken and the samples were suspended in 400 μL of TNES-8M urea.

Analysis of monoamines and their metabolites in
Twenty microlitres of proteinase K (20 mg mL −l ) and 20 μL of 1 M dithiothreitol (DTT) were added to the samples, which were then incubated for 1-5 h at 60 °C and mixed with 500 μL of phenol/chloroform/isoamyl alcohol solution (25:24:1) for 3 min. This step was repeated twice. After precipitation with 2-2.5 volumes of ethanol, the pellets were rinsed in 70% v/v cold ethanol and dried. The samples were dissolved in TE buffer [10 mM Tris-HCl (pH 8.0) and 1 mM EDTA]. The PCR was performed to amplify ADRα1A (Ex1 and Ex2), ADRα1B (Ex1 and Ex2), ADRα1D (Ex1 and Ex2), ADRα2A, ADRα2B, ADRα2C, ADRβ1, ADRβ2, and ADRβ3 (Ex1 and Ex2). The PCR primers are listed in Table S6. The PCR enzymes used were KOD-Plus and KOD FX neo (TOYOBO, Osaka, Japan). The PCR was performed under the following conditions: denaturation for 20 s at 98 °C, 30 cycles for 5 s at 98 °C, annealing at 64 °C for 30 s, 68 °C for 1 min, and a final extension at 68 °C for 7 min. The PCR products were purified using the MinElute PCR purification kit (QIAGEN, Duesseldorf, Germany) and ExoSAP-IT (USB Corp., Cleveland, OH, USA) and sequenced using BigDye Terminator v. 3.1 (Thermo Fisher Scientific, Waltham, MA, USA) and the ABI Prism 3730xl DNA sequencer (ABI, Foster City, CA, USA).
Sequence assembly and alignment. The contigs of each ADR gene were assembled from the ABI DNA sequences using ATGC GENETYX v. 13 (GENETYX Corp., Tokyo, Japan). The sequences were then determined and registered in the International Nucleotide Sequence Database Collaboration (INSDC). The accession numbers of the ADR DNA sequences (LC483765-LC483818) were obtained from the DDBJ/EMBL/GenBank. All assembled sequences were aligned using CLUSTALW-MEGA v. 7 82,83 . These analyses confirmed the mutation sites in the ADR genes of each chicken breed. The ADR receptor gene sequences were mapped for both breeds and their mutation sites were determined. The sequence of red junglefowl, the wild ancestor of chickens, was used as the reference (Table S4). Shamo skulls were more difficult to remove than Shaver Brown skulls as the former have been strengthened for cockfighting by selective breeding. Thus, the removal of their skulls required considerable force and resulted in the skulls shattering into numerous fragments. Moreover, the Shamo brains had a higher tension and gloss than the Shaver Brown chicken brains.