A cis-regulatory mutation in the EDN3 promoter drives coordinated formation of comb and skin color in Anyi gray chickens

Abstract

Comb and skin pigmentation are key external traits in domestic chickens, closely correlated and serving as important visual cues for consumers. The Anyi Gray chicken, an indigenous breed from Jiangxi Province, China, exhibits two stable phenotypes: red comb with white skin (RAY), and purple-gray comb with purple-gray skin (PAY). Although skin color is completely linked to comb color, the genetic basis of this coordination remains unclear.

We performed whole-genome resequencing (∼ 10 × depth) on RAY and PAY individuals, followed by a case-control genome-wide association study (GWAS). A highly significant variant, g.10818413 T > C, located in the promoter region of EDN3, was identified (P < 4.065343e-09). Transcription factor binding site prediction revealed that this mutation disrupts an NFAT5 binding site. Genotyping of 97 Anyi Gray chickens showed complete concordance between genotype and phenotype: all PAY individuals carried the TC genotype, while all RAY individuals had the TT genotype; no CC genotypes were detected. Validation in an F1 hybrid population (n = 61) and in 140 individuals from other Jiangxi native breeds confirmed the absence of the CC genotype.

Functional assays demonstrated that NFAT5 negatively regulates EDN3 transcription. Dual-luciferase reporter assays confirmed NFAT5 binding to the EDN3 promoter and its suppressive effect. Tissue expression analysis showed significantly higher EDN3 expression in the comb, skin, and chest muscle of PAY individuals compared to RAY individuals, reinforcing the phenotypic linkage.

These findings indicate that the EDN3 g.10818413 T > C mutation reduces NFAT5 binding affinity, cis-regulating EDN3 transcription and driving the coordinated pigmentation of the comb and skin. This study reveals a shared regulatory mechanism underlying the phenotypic correlation between comb and skin pigmentation in Anyi Gray chickens.

Introduction

Comb and skin color are important external phenotypic traits in poultry, playing a critical role in determining the market value of commercial chickens and shaping consumer preferences. In many regions, the visual appeal of poultry products significantly influences purchasing behavior, with pigmentation in the comb and skin serving as key indicators of freshness, health status, and overall product quality (Kennedy et al., 2005; Wideman et al., 2016). For example, the differing consumer preferences for yellow-skinned chickens versus white-skinned chickens in the Chinese market illustrate how regional cultural norms and local dietary traditions shape purchasing behavior (Sirri et al., 2010; Wu et al., 2021a). This consumer-driven demand further elevates the economic importance of these traits, positioning their genetic regulation as a central focus in poultry breeding programs (Huang et al., 2020; Wang et al., 2023). Moreover, vibrant comb coloration is often regarded as a visible marker of flock vitality, enhancing market attractiveness. Therefore, a comprehensive understanding of the genetic basis underlying comb and skin color variation is essential for optimizing poultry production and meeting diverse consumer preferences.

The Anyi Gray chicken is a native breed from Jiangxi Province, China, valued for both meat and egg production. It is distinguished by its distinctive phenotype—gray feathers, gray legs, and a gray beak. Within its core population, two stable morphological types are observed: red comb with white skin (RAY) and purple-gray comb with purple-gray skin (PAY). The two traits are completely correlated (Fig. 1, Supplementary Table 1), yet the molecular mechanisms underlying their coordinated development remain unknown. A systematic investigation into the genetic regulation of these traits is essential for advancing our understanding of breed-specific pigmentation and enabling precise selection in breeding programs.

Fig. 1.

Photographs showing comb and skin coloration in Anyi Gray chickens. (A) Individual with a red comb. (B) Comb of a red-comb individual. (C) Skin of a red-comb individual. (D) Comb of a purple-gray comb individual. (E) Skin of a purple-gray comb individual. (F) Individual with a purple-gray comb.

Avian skin pigmentation is influenced by multiple factors, with genetic mechanisms playing a particularly critical role. Recent studies have identified key genes involved in pigment deposition. For comb color, yellow or orange hues result from carotenoid accumulation in the epidermis (Stettenheim, 2000), while darker combs, such as purple-gray or black, are associated with increased dermal melanocyte activity (Dorshorst et al., 2011). The Endothelin-3 (EDN3) gene has been linked to dermal hyperpigmentation across poultry breeds, where structural variations (e.g., duplications and inversions) upregulate its expression and promote melanocyte proliferation and migration, thereby contributing to fibromelanosis (FM) (Dorshorst et al., 2011). The Bone Morphogenetic Protein 7 (BMP7) gene also exhibits pleiotropic effects, influencing pigmentation and reproductive traits such as egg production performance (Dong et al., 2019). In skin pigmentation, β-carotene oxygenase 2 (BCO2) is a critical gene. Its reduced expression or functional loss to carotenoids accumulation, resulting in yellow skin tones (Eriksson et al., 2008). Further studies have shown that coding sequence polymorphisms (Lobo et al., 2012) and expression level differences (He et al., 2023) in BCO2 are strongly associated with skin color variation. Other genes, including GSTA2 and SCARB1, are also implicated in carotenoid transport and metabolism (Wu et al., 2021b). Despite these advances, the genetic basis of the two distinct pigmentation phenotypes in Anyi Gray chickens—namely, red comb with white skin and purple-gray comb with purple-gray skin—remains poorly understood.

The study focuses on Anyi Gray chickens, employing whole-genome resequencing to conduct a case-control genome-wide association study (GWAS) aimed at identifying genetic loci associated with comb and skin color variation between RAY and PAY individuals. Functional validation of key variants will be performed through cellular assays, alongside large-scale population genotyping to assess allele frequency and trait association.

Material and methods

Ethical statement

All animal experiments were approved by Jiangxi Agricultural University (Approval No. JXAULL-2024-03-10) and conducted in strict accordance with the current regulations and guidelines on animal welfare and research ethics in China.

Experimental animals and sample preparation

A total of 30 Anyi Gray chickens were selected from the Anyi Gray Chicken Breeding Farm in Jiangxi Province for whole-genome resequencing and genome-wide association analysis. Among these, 9 individuals exhibited the RAY phenotype, while the remaining 21 displayed the PAY phenotype. For population genotyping verification of the EDN3 g.10818413T>C locus, three groups of chickens were analyzed. The first group comprised 97 Anyi Gray chickens from the original breeding population. The second group included 61 F1 individuals derived from RAY and PAY parental lines, generated through three specific mating combination: TC (♂) × TT (♀), TT (♂) × TC (♀), and TC (♂) × TC (♀). The third group consisted of 140 individuals from seven indigenous Chinese chicken breeds including Ningdu Yellow Chicken (ND), Baier Yellow Chicken (BER), Kangle Chicken (KL), Dongxiang Blue-eggshell Chicken (DX), Yugan Black-bone Chicken (YG), Taihe silkies Chicken (TH), and Chongren Partridge Chicken (CR). Approximately 2 mL blood was collected from the wing vein of each chicken using EDTA-coated anticoagulant tubes to prevent clotting. Samples were immediately stored at −20 °C. Genomic DNA was extracted using the standard phenol-chloroform method and dissolved in TE buffer (Green and Sambrook, 2012). Only DNA samples with OD260/280 ratios between 1.8 and 2.0 and showing intact bands on agarose gel electrophoresis were used for subsequent experiments (Green and Sambrook, 2012).

Additionally, five individuals with the PAY phenotype and five with the RAY phenotype were randomly selected from the above F1 population. Samples of comb, breast skin, and breast muscle tissues were collected, immediately frozen in liquid nitrogen, and subsequently stored at −80 °C (Supplementary Fig. 1). Total RNA was extracted from these tissues using the Trizol reagent method (Invitrogen, Carlsbad, CA). Total RNA was extracted from these tissues using the Trizol reagent method (Invitrogen, Carlsbad, CA). Only RNA samples meeting quality standards (OD260/280 ratios between 1.8 and 2.1 and RIN ≥ 7.0; 30 in total) were used for further analyses(Schroeder et al., 2006).

Validation of transcriptional expression across multiple tissues

EDN3 expression levels in comb, skin, and chest muscle tissues of PAY and RAY chickens were quantified by RT-qPCR, with GAPDH serving as the internal reference. The RT-qPCR method was performed on the Applied Biosystems StepOne Real-Time PCR System using the Platinum SYBR Green qPCR SuperMix-UDG kit (Invitrogen Life Technologies, Carlsbad, CA, USA).

Thermal cycling conditions consisted of an initial denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 45 s(annealing/extension). A melting curve analysis was conducted to confirm that a single specific product was amplified in each reaction. Primer sequences are provided in Supplementary Table 2.

Whole-genome resequencing and case-control based genome-wide association analysis

Genomic DNA from 30 qualified Anyi gray chicken was subjected to 10 × whole-genome re-sequencing using the Illumina NovaSeq platform (Shanghai Personalbio Biotechnology Co., Ltd., Shanghai, China). Sequencing reads were aligned to the Gallus_gallus-6.0 reference genome (https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_000002315.6/) using BWA v0.7.17 software (Li and Durbin, 2009). Sorting and indexing, were carried out with Samtools v1.6 (Li et al., 2009), followed by variant calling using GATK v4.2.0.0, generating raw SNP VCF files. Initial SNP filtering was performed using GATK v4.2.0.0 software with the following parameters: QD < 2.0, MQ < 40.0, FS > 60.0, SOR > 3.0, MQRankSum < −12.5, and ReadPosRankSum < −8.0(McKenna et al., 2010), Further filtering was conducted using PLINK v1.90 software (Purcell et al., 2007)with the parameters –geno 0.1 and –maf 0.05, removing SNPs with >10% missing genotypes or minor allele frequency <5%. The final dataset contained 12,299,085 high-quality SNPs. Case-control GWAS was performed using GEMMA software (Zhou and Stephens, 2012)with a mixed linear model to compare the PAY and RAY populations. Results were visualized using Manhattan plots and QQ plots, and significantly associated loci were functionally annotated. Candidate genes were prioritized based on their potential biological relevance.

Screening of promoter region mutation sites and prediction of transcription factor binding sites

GWAS results identified a significant SNP, EDN3 g.10818413T>C, located in the promoter region of the EDN3 gene, which was strongly associated with both comb and skin color. This mutation exhibited cosegregation with the pigmentation phenotypes in Anyi Gray chickens, suggesting a regulatory role in gene expression.

To investigate its potential functional impact, both the reference and mutant promoter sequences of EDN3 were retrieved from the NCBI database (https://www.ncbi.nlm.nih.gov/). Transcription factor binding site prediction was performed using the JASPAR database (https://jaspar.elixir.no/). Comparative analysis of the binding motifs before and after the mutation was conducted to assess changes in transcription factor binding affinity or specificity. This analysis aimed to determine whether the mutation could alter transcriptional regulation of EDN3, thereby influencing pigmentation traits.

Confirmation of the complex fibromelanosis (FM) structure

To detect variants in the FM complex structure of the EDN3 gene, we employed a previously reported three-primer multiplex PCR system(Dorshorst et al., 2011) (Accurate Biology, Changsha, China). A total of 53 chicken samples were analyzed, including 13 Anyi Gray chickens (11 RAY and 2 PAY individuals), 20 Yugan Black-bone chickens, and 20 Dongxiang Blue-eggshell chickens. Primers sequences are listed in Supplementary Table 2.

PCR reactions were performed in a 50μl system containing 1μl of DNA template, primers, Multiplex DNA Polymerase (1 U/ul), 25μl of 2 × Multiplex PCR Buffer (Mg2+ and dNTP plus), and 20μl of ddH₂O. Thermal cycling conditions were as follows: initial denaturation at 95°C for 2 min, followed by 30 cycles of 95°C for 30 s, 60°C for 30 s, and 68°C for 30 s, with a final extension at 68°C for 10 min. PCR products were analyzed by agarose gel electrophoresis. The presence of two distinct bands indicated the existence of FM complex structures in the genome.

Population genotyping

Genotype verification at the EDN3 g.10818413 T > C locus was conducted using PCR product sequencing. The study analyzed 97 Anyi Gray chickens, 61 F1 individuals derived from RAY and PAY parental lines, and 140 chickens representing seven other Chinese indigenous chicken breeds. Specific primers targeting the mutation site were designed using the NCBI database (https://www.ncbi.nlm.nih.gov/) and synthesized by Tsingke Biotechnology Co., Ltd. (Hunan, China). Primer sequences are listed in Supplementary Table 2.

PCR amplification was performed on the collected samples, and the products were sequenced using Sanger sequencing (Tsingke, Hunan, China). Sequencing data were processed with SeqMan Pro v11.1.0 software. Genotype frequencies were subsequently calculated, and their distribution patterns across the sampled populations were statistically evaluated.

Functional validation of transcription factor NFAT5 via overexpression and knockdown

To investigate the regulatory role of NFAT5 on EDN3 expression, both overexpression and knockdown vectors were constructed. The coding sequence (CDS) of NFAT5 was amplified and cloned into the pIRES2-EGFP vector to generate the overexpression construct (pIRES2-EGFP-NFAT5). An empty vector (pIRES2-EGFP-Basic) served as the control. All vectors were obtained from Tsingke Biotechnology Co., Ltd. (Hunan, China). In addition, a short hairpin RNA (shRNA) vector targeting NFAT5 (ShRNA-NFAT5) was designed and constructed, along with a corresponding negative control vector.

DF-1 cells were cultured in a medium containing 10% fetal bovine serum (FBS; GIBCO, Waltham, MA) and 90% DMEM HIGH (GIBCO, Waltham, MA) under standard conditions (37°C, 5% CO₂). When cells reached approximately 90% confluence, they were harvested into a single-cell suspension and seeded into 12-well plates. At approximately 70% confluence, the four vectors were transfected using Lipofectamine 3000 (Thermo Fisher, Waltham, MA) and Opti-MEM (GIBCO, Waltham, MA). Transfection complexes were incubated at room temperature for 15 min before being added to the wells. The experiment included four groups: pIRES2-EGFP-NFAT5, pIRES2-EGFP-Basic, ShRNA-NFAT5, and Negative Control, with six replicate wells per group. After 6 h of transfection, the medium was replaced with fresh complete medium. Cells were collected 48 h post-transfection, and total RNA was extracted using the Trizol method (Invitrogen, Carlsbad, CA). NFAT5 and EDN3 expression levels were quantified by RT-qPCR, with primer sequences provided in Supplementary Table 2.

Dual-luciferase reporter assay

To assess whether NFAT5 binds to and regulates the EDN3 promoter, both wild-type and mutant promoter sequences surrounding the g.10818413 T > C site were cloned into the pmir-GLO vector (Tsingke, Hunan, China). This generated two constructs: pmir-GLO-T (wild-type) and pmir-GLO—C (mutant), with the empty pmir-GLO vector serving as a control. These reporter vectors were co-transfected with the NFAT5 overexpression vector (pIRES2-EGFP-NFAT5) into HEK293T cells.

HEK293T cells were cultured in DMEM High Glucose supplemented with 10% FBS (GIBCO, Waltham, MA) under standard conditions (37 °C, 5% CO₂). At approximately 90% confluence, cells were trypsinized, resuspended, and seeded into 96-well plates. When cell density reached approximately 70%, plasmid vectors were transfected using Hieff Trans™ (YEASEN, 40802ES01). Transfection complexes were incubated at room temperature for 20 min before being added to the wells. After 6 h, the medium was replaced with fresh complete medium, and cells were harvested 48 h post-transfection.

Luciferase activity was measured using the Dual-Luciferase Reporter Assay Kit (Promega, E2920). Firefly and Renilla luciferase activities were quantified, and promoter activity was expressed as the ratio of firefly to Renilla luciferase activity, reflecting the regulatory influence of NFAT5.

Statistical analysis

All statistical analyses were performed using SPSS v 20.0 (IBM Corp., Armonk, NY). Independent samples t-tests were used to compare gene expression and luciferase activity between groups. Each experimental condition included at least three biological replicates, and results are presented as mean ± standard error (Mean ± SE).

Genotype frequencies were calculated by direct counting and compared across groups in population validation and F1 hybridization experiments. All statistical tests were two-tailed, with significance defined as P < 0.05.

Results

Consistency of comb and skin color phenotypes in Anyi gray chickens

In the original Anyi Gray chicken population, 97 individuals were sampled for both phenotypic and genotypic analyses. Among them, 41 individuals exhibited a red-comb and white-skinned (RAY) phenotype, while 56 displayed a purple-comb and grey-skin (PAY) phenotype. The expression of comb and skin color was highly consistent across all individuals, with no discrepancies observed between phenotypic traits and genotypic data (Supplementary Table 1). In addition, the RT-qPCR results demonstrated that the expression levels of EDN3 in various tissues (including comb, chest skin, and chest muscle) of PAY individuals were significantly elevated compared to those of RAY individuals (Fig. 2D).

Fig. 2.

Genome-wide association analysis highlighting the EDN3 g.10818413T>C variant. (A) GWAS results for comb and skin color phenotypes. (B) Sanger sequencing peak diagram of the EDN3 g.10818413T>C polymorphic site in PAY and RAY individuals (red box). (C) LD block analysis of the EDN3 region (Chr20: 10797698–10818602) based on GWAS results. Significant SNPs (-log10(P) ≥ 8.39) are shown as red dots. (D) EDN3 expression levels in PAY and RAY individuals across comb, skin, and muscle tissues. Statistical significance was assessed using a two-tailed Student’s t-test (*P < 0.05, **P < 0.01, ***P < 0.001).

GWAS identifies a promoter variant in EDN3

To explore the genetic basis of comb color and skin pigmentation in Anyi gray chickens, we analyzed 12,299,085 single nucleotide polymorphism (SNP) sites derived from a filtered Variant Call Format (VCF) file. Genome-wide association analysis was conducted using GEMMA software with a mixed linear model (MLM) in a case-control design. Individuals with purple-comb and gray skin phenotype (PAY) were designated as the case group, while those with the red-comb and white skin phenotype (RAY) served as the control group.

At a significance threshold of -log₁₀(P) > 8.39, 227 SNPs showed significant association with the phenotypes. These SNPs were exclusively clustered within a 900 kb region spanning from 10.58 Mb to 11.48 Mb on chromosome 20 (Fig. 2A). Within this strongly associated interval, several candidate genes were annotated, including: SLC12A5, DDX27, EDN3, ZNF831, PRELID3B, TUBB1, ATP5E, GNAS, APCDD1L, C20H20ORF85, and PMEPA1. Importantly, the EDN3 gene lies at the center of this region, and its promoter variant g.10818413 T > C exhibited a strong correlation with comb and skin color phenotypes in subsequent analyses.

The top 50 significant SNP loci are listed in Table 1. Collectively, these findings highlight chromosome 20 as a key genomic region influencing pigmentation traits in Anyi Gray chickens.

Table 1.

Top 50 single nucleotide polymorphism (SNP) loci showing the most statistically significant associations identified by genome-wide association studies (GWAS).

Chr	POS	Ref	Alt	-log(P)	gene	Type
20	11445670	C	T	16.13	PMEPA1	intergenic
20	11444816	T	C	16.13	PMEPA1	intergenic
20	11444815	T	C	16.13	PMEPA1	intergenic
20	11431437	G	A	16.13	PMEPA1	intergenic
20	11378378	T	C	16.13	C20H20ORF85	intergenic
20	11403162	C	T	16.12	C20H20ORF85	intergenic
20	11384215	A	G	16.05	C20H20ORF85	intergenic
20	11384187	C	T	16.05	PMEPA1	intergenic
20	11439940	C	G	16.04	PMEPA1	intergenic
20	11414775	G	A	13.64	PMEPA1	intergenic
20	11469759	G	A	13.64	PMEPA1	intergenic
20	11460364	G	T	13.64	PMEPA1	intergenic
20	11460351	C	T	13.64	PMEPA1	intergenic
20	11411671	G	A	13.64	PMEPA1	intergenic
20	11432172	G	T	13.64	PMEPA1	intergenic
20	11452699	G	A	13.55	PMEPA1	intergenic
20	11435651	C	G	13.55	PMEPA1	intergenic
20	11426268	A	G	13.55	PMEPA1	intergenic
20	11426247	C	T	13.55	PMEPA1	intergenic
20	11412219	C	T	13.55	PMEPA1	intergenic
20	11445618	C	T	13.37	C20H20ORF85	intergenic
20	10845460	A	G	12.67	ZNF831	intronic
20	11383651	G	A	12.63	C20H20ORF85	intergenic
20	10837201	T	C	12.63	ZNF831	intergenic
20	10837194	A	C	12.63	ZNF831	intergenic
20	11463744	T	A	12.63	PMEPA1	intergenic
20	11462705	A	G	12.63	PMEPA1	intergenic
20	11459794	C	T	12.63	PMEPA1	intergenic
20	11459780	C	T	12.63	PMEPA1	intergenic
20	11388239	C	G	12.63	PMEPA1	intergenic
20	10851491	A	G	12.63	ZNF831	intronic
20	10839398	G	A	12.63	ZNF831	downstream
20	10819338	T	C	12.63	EDN3	intergenic
20	10793270	A	G	12.63	EDN3	intergenic
20	10779640	C	T	12.63	EDN3	intergenic
20	11448861	G	C	12.63	PMEPA1	intergenic
20	11448732	G	A	12.63	PMEPA1	intergenic
20	11390097	G	T	12.63	PMEPA1	intergenic
20	10834213	C	T	12.63	ZNF831	intergenic
20	10789198	T	A	12.63	EDN3	intergenic
20	10779910	G	A	12.63	EDN3	intergenic
20	10778942	G	A	12.63	DDX27	intergenic
20	10778929	A	G	12.63	DDX27	intergenic
20	10775850	T	C	12.63	DDX27	intergenic
20	11464290	A	G	12.63	PMEPA1	intergenic
20	11458248	A	C	12.63	PMEPA1	intergenic
20	11440322	G	A	12.57	PMEPA1	intergenic
20	10818413	T	C	12.57	EDN3	upstream
20	10789995	G	A	12.57	EDN3	intergenic
20	10805322	A	G	12.48	EDN3	intronic

Prediction of transcription factor binding sites in the EDN3 promoter

GWAS analysis identified a SNP (g.10818413T>C) located 1552 bp upstream of EDN3, which showed a significant association with comb and skin color phenotypes in Anyi Gray chickens. This site lies within the core promoter region of EDN3(Song et al., 2023).Using the JASPAR online platform (https://jaspar.elixir.no/)to predict transcription factor binding sites in the promoter region, we found that the EDN3 g.10818413T>C mutation results in the loss of a binding site for NFAT5(Fig. 3).

Fig. 3.

Schematic diagram illustrating the mechanism by which the EDN3 g.10818413T>C polymorphism modulates EDN3 expression via NFAT5 binding affinity to the promoter.

Genotype validation and absence of CC-type individuals

To further assess the effect of the EDN3 g.10818413T>C variation on comb and skin color phenotypes in Anyi Gray chickens, genotyping was performed on 97 individuals at this locus. All 41 with the RAY phenotype carried the TT genotype, while all 56 chickens with the PAY phenotype carried the TC genotype. No individuals with the CC genotype were detected in the sampled population. Representative Sanger sequencing chromatograms are shown in Fig. 2B, and detailed genotype frequencies are provided in Table 2.

Table 2.

Genotype distribution of the EDN3 g.10818413 T > C polymorphic locus in a large population cohort.

SNP	Ref	Alt	GeneType	PAY	RAY
			TT	0	41
20:10818413	T	C	TC	56	0
			CC	0	0

The F1 generation derived from Anyi Gray chickens were examined for co-segregation of the EDN3 g.10818413T>C locus with phenotype. The F1 generation derived from Anyi Gray chickens were examined for co-segregation of the EDN3 g.10818413T>C locus with phenotype. In hybrid combination 1 (TC ♂ × TT ♀), 17 offspring were produced (10 RAY and 7 PAY); in hybrid combination 2 (TT ♂ × TC ♀), 17 offspring were obtained (8 RAY and 9 PAY); and in hybrid combination 3 (TC ♂ × TC ♀), 27 offspring were generated (5 RAY and 22 PAY). Sanger sequencing confirmed that all PAY individuals in the F1 population carried the TC heterozygous genotype, while all RAY individuals carried the TT homozygous genotype; again, no CC-type individuals were identified. Phenotypic and genotypic frequency data for the offspring are summarized in Table 3.

Table 3.

Genotype distribution of the EDN3 g.10818413T>C polymorphic locus in the F1 generation across three hybrid combinations.

Cross Combination	SNP	Ref	Alt	RAY			PAY
				TT	TC	CC	TT	TC	CC
Combination1	20:10818413	T	C	10	0	0	0	7	0
Combination2				8	0	0	0	9	0
Combination3				5	0	0	0	22	0

Genotype distribution of the EDN3 g.10818413T>C locus was analyzed in 140 individuals from local chicken breeds in Jiangxi Province. Distinct patterns were observed in relation to plumage color. All individuals from red-comb breeds (BER, KL, ND, CR) carried the TT genotype, indicating a uniform genetic profile for this allele. In contrast, all individuals from purple-comb or black-comb breeds (YG, DX, TH) carried the TC genotype, with no CC-type individuals detected across the sampled populations. These distributions are detailed in Table 4.

Table 4.

Genotype distribution of the EDN3 g.10818413T>C polymorphism across seven local chicken breeds in Jiangxi Province.

Breed	SNP	Ref	Alt		GeneType
				TT	TC	CC
BER				20	0	0
KL				20	0	0
ND				20	0	0
CR	20:10818413	T	C	20	0	0
YG				0	20	0
DX				0	20	0
TH				0	20	0

Annotations: BER: Baier Yellow Chicken; KL: KangLe Chicken; ND: Ningdu Yellow Chicken; YG: Yugan Black-bone Chicken; DX: Dongxiang Blue-eggshell Chicken; TH: Taihe silkies Chicken; CR: Chongren Partridge Chicken.

Detection of structural variations in the fibromelanosis (FM) complex

Multiplex PCR typing was performed to examine FM complex structures in 53 individuals, including AY (RAY=11, PAY=2), YG, and DX. The results indicated that 2 PAY individuals, 20 YG individuals, and 20 DX individuals exhibited two distinct bands on agarose gel electrophoresis, whereas all 11 RAY individuals displayed a single band. Representative results are presented in Supplementary Figs. 2 and 3.

Functional role of NFAT5 in regulating EDN3 expression

To examine the regulatory role of NFAT5 on EDN3 expression, DF-1 cells were transfected with either the overexpression vector pIRES2-EGFP-NFAT5 or the knockdown vector ShRNA-NFAT5. Forty-eight hours post-transfection, total RNA was extracted and analyzed by RT-qPCR. Overexpression of NFAT5 significantly increased NFAT5 mRNA levels (P < 0.01), accompanied by a marked reduction in EDN3 expression (P < 0.01). Conversely, NFAT5 knockdown significantly decreased NFAT5 expression (P < 0.05), while EDN3 expression was significantly elevated (P < 0.05) (Fig. 4). These findings indicate that NFAT5 functions as a negative regulator of EDN3 expression in DF-1 cells.

Fig. 4.

Functional validation of NFAT5 regulatory effects on EDN3 expression in DF-1 cells. (A) Fluorescence following transfection with the negative control vector pIRES2-EGFP-Basic. (B) Fluorescence following transfection with the overexpression vector pIRES2-EGFP-NFAT5. (C) Quantitative analysis of NFAT5 and EDN3 expression in DF-1 cells 48 h post-transfection with pIRES2-EGFP-Basic or pIRES2-EGFP-NFAT5. (D) Fluorescence following transfection with the negative control vector. (E) Fluorescence following transfection with the interference vector ShRNA-NFAT5. (F) Quantitative analysis of NFAT5 and EDN3 mRNA levels in DF-1 cells 48 h post-transfection with the negative control or ShRNA-NFAT5 vector. Statistical significance was assessed using a two-tailed Student’s t-test (*P < 0.05, **P < 0.01, ***P < 0.001).

NFAT5 binding to the EDN3 promoter: dual-luciferase reporter assay

To determine whether NFAT5 directly binds to the EDN3 promoter and modulates its transcriptional activity, HEK293 cells were co-transfected with combinations of plasmids, including pmir-GLO-T + pIRES2-EGFP-NFAT5, pmir-GLO—C + pIRES2-EGFP-NFAT5 and pmir-GLO—C + pIRES2-EGFP-Basic. Luciferase activity was measured using a dual-luciferase assay kit. The results showed that promoter activity in the pmir-GLO-T + pIRES2-EGFP-NFAT5 group was significantly reduced compared with the pmir-GLO—C + pIRES2-EGFP-NFAT5 and pmir-GLO—C + pIRES2-EGFP-Basic groups. Moreover, activity in the pmir-GLO—C + pIRES2-EGFP-Basic group was significantly higher than in the pmir-GLO—C + pIRES2-EGFP-NFAT5 group (Fig. 5). These findings suggest that NFAT5 directly interacts with the EDN3 promoter and acts as a negative regulator of transcriptional activity. Mutations at the binding site reduce NFAT5 affinity, thereby altering EDN3 transcriptional regulation.

Fig. 5.

Dual-luciferase reporter assay verifying the targeting relationship between NFAT5 and EDN3. (A) HEK293T cells were co-transfected with pmir-GLO—C or pmir-GLO-T vectors and either pIRES2-EGFP-Basic or pIRES2-EGFP-NFAT5 overexpression vectors. After 48 h, luciferase activity was measured. Statistical significance was assessed using a two-tailed Student’s t-test (*P < 0.05, **P < 0.01, ***P < 0.001).

Discussion

Comb and skin color are important phenotypic traits in chickens that also influence consumer preferences. As a key indigenous breed in Jiangxi Province, China, the genetic mechanisms underlying these traits in Anyi Gray chickens have remained largely unexplored. This study provides comprehensive evidence that the EDN3 g.10818413T>C promoter mutation is a key genetic determinant underlying comb and skin pigmentation in Anyi Gray chickens. Through genome-wide association analysis, population-level genotyping, multi-tissue transcriptomic profiling, and functional validation assays, we demonstrated that this variant modulates EDN3 expression by altering the binding affinity of NFAT5, thereby influencing melanocyte development and pigmentation phenotypes.

Endothelins (EDN) are potent bioactive peptides synthesized primarily by endothelial cells, comprising three isoforms: EDN1, EDN2, and EDN3. Among them, EDN3 plays a pivotal role in the proliferation and differentiation of early neural crest precursors, which later develop into melanocytes(Lahav et al., 1996). EDN3 exerts its biological effects via binding to endothelin receptor proteins (EDNRB), activating the MEK/ERK/MITF signaling cascade, which enhances tyrosinase (TYR) activity and promotes melanin synthesis in melanocyte(Zhang et al., 2013). Meanwhile, elevated EDN3 expression upregulates TYR activity via the EDN/EDNRB signaling pathway, thereby stimulating melanin synthesis(Vachtenheim and Borovanský, 2010). Consistent with this mechanism, our data revealed significantly elevated EDN3 expression in comb, chest skin, and chest muscle tissues of PAY individuals compared with RAY individuals (Fig. 2D).

Diagnostic assays further confirmed structure variation in the fibromelanosis (FM) complex. RAY individuals lacked the FM complex, whereas PAY individuals consistently exhibited it (Supplementary Fig. 2). Additional tests in Jiangxi local breeds showed that heterozygous individuals (TC genotype) consistently exhibited FM structures, while wild-type individuals (TT genotype) did not. The purple-gray comb and skin phenotype in Anyi Gray chickens appears to be influenced by both the FM complex and the EDN3 g.10818413T>C mutation. While it is possible that one or both act merely as genetic markers, our data and functional assays strongly suggest that both are causal variants acting in synergy.

The FM complex, a tandem duplication within the EDN3 locus, is known to increase EDN3 dosage or alter its regulatory landscape, thereby establishing a permissive chromatin environment for hyperpigmentation(Dorshorst et al., 2011). In 2015, Johansson and colleagues found that two Swedish native chicken breeds with dark skin color possessed a similar complex FM structure(Johansson and Nelson, 2015), which also affected the color of the comb, and similar structures have also been reported in Kadaknath chickens in India (Arora et al., 2011). In contrast, the g.10818413T>C mutation directly modifies the NFAT5 binding site, weakening its negative regulatory effect and providing a precise transcriptional regulatory switch. We therefore propose a synergistic causal model: the FM complex provides the broad genetic scaffold for fibromelanosis, while the g.10818413T>C mutation fine-tunes and enhances EDN3 expression within this context, jointly producing the distinct purple-gray phenotype. Future studies, including gene editing approaches, will be required to dissect their independent contributions and confirm whether both variants are causal.

Previous studies have demonstrated that structural variations involving EDN3 are strongly associated with dermal hyperpigmentation and comb color in chickens (Dorshorst et al., 2011; Ma et al., 2024; Shinomiya et al., 2012). In line with these findings, our results suggest that mutations within the EDN3 promoter may play a pivotal role in shaping comb and skin color, underscoring their importance in phenotype determination.

NFAT5 is a transcription factor involved in diverse biological processes, including cell differentiation, migration, and embryogenesis(Aramburu et al., 2006). EDN3, in turn, is essential for neural crest cell proliferation and melanocyte differentiation(Nagy and Goldstein, 2006). We hypothesize that NFAT5 regulates EDN3 by binding directly to its promoter. This study provides the first evidence of such an interaction. Furthermore, it was reported that NFAT5 can modulate the expression of additional genes within the endothelin family(Lakshmipathi et al., 2019). Functional assays and dual-luciferase reporter experiments demonstrated that the g.10818413T>C mutation reduces NFAT5 binding affinity, thereby weakening its negative regulatory effect and resulting in elevated EDN3 expression (Fig. 4, Fig. 5).

Genotypic analysis across seven Jiangxi chicken breeds revealed conservation of the EDN3 g.10818413T>C mutation. Red-comb breeds (KL, ND, GF, CR) uniformly carried the TT genotype, whereas purple- or black-comb breeds (TH, YG, DX) carried the TC genotype, with no CC individuals detected. This suggests strong conservation of the mutation and raises the possibility that the CC genotype is lethal. Beyond pigmentation, EDN3 is critical for enteric nervous system development(Lake and Heuckeroth, 2013). Mutations in EDN3 are associated with congenital megacolon, a potentially fatal condition(Druckenbrod and Epstein, 2009).We hypothesize that individuals with homozygous mutation of EDN3 g.10818413T>C (CC genotype) may fail to develop a functional enteric nervous system, leading to embryonic lethality. Such lethal homozygous mutations are well documented in nature. For example, the Cp gene in chickens causes embryonic lethality in homozygotes(Landauer, 1932), and the ML gene in Manx cats results in embryonic death in homozygous individuals(Deforest and Basrur, 1979).

Beyond its relevance to a single indigenous breed, this study offers insights with broader implications for poultry genetics and pigmentation biology worldwide. EDN3 is a conserved regulator of melanocyte proliferation and differentiation across vertebrates, and regulatory variation affecting its expression has been implicated in pigmentation diversity in multiple species. The identification of a functional promoter mutation that alters NFAT5 binding and modulates EDN3 expression expands current understanding of transcriptional regulation underlying integumentary coloration and provides a useful reference for comparative studies in diverse chicken populations and other avian species.

From a breeding perspective, comb and skin pigmentation are economically important traits in many poultry markets globally. The regulatory variant identified here may serve as a potential molecular marker for marker-assisted selection targeting pigmentation traits across different genetic backgrounds. Furthermore, the integrative analytical framework employed in this study, combining genome-wide association analysis, structural variation detection, multi-tissue transcriptional validation, and functional assays, provides a broadly applicable strategy for identifying causal variants underlying complex traits in poultry and other livestock species. Collectively, These findings therefore contribute valuable genetic and methodological insights for researchers and breeders internationally.

Conclusion

In conclusion, this study identified a functional cis-regulatory mutation, EDN3 g.10818413T>C, located within the promoter region of the chicken EDN3 gene. This variant modulates EDN3 expression by altering the binding affinity of the transcription factor NFAT5, thereby simultaneously influencing the development of comb and skin pigmentation phenotypes. Importantly, our findings provide the first evidence that NFAT5 directly binds to and regulates the EDN3 promoter.

Data availability statement

The sequencing data generated in this study have been deposited in the Genome Sequence Archive (GSA) at the National Genomics Data Center (NGDC), China, under accession number CRA038297. These data are currently under embargo and will be made publicly available upon the publication of this manuscript. During the review process, the data can be accessed upon reasonable request from the corresponding author. The data can be accessed via the following link: https://ngdc.cncb.ac.cn/gsa.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors did not use any Al and Al-assisted technologies

CRediT authorship contribution statement

Jianxi Huang: Writing – original draft, Visualization, Validation, Investigation, Formal analysis, Data curation, Conceptualization. Di Cheng: Methodology, Conceptualization. Suwang Xi: Resources, Methodology, Conceptualization. Zun Wang: Investigation, Formal analysis, Data curation. Yun Tu: Investigation, Formal analysis, Data curation. Sixiang Huang: Investigation, Formal analysis, Data curation. Dongxiao Wu: Investigation. Zengwei Zhang: Investigation. Hongxia Jiang: Methodology, Investigation. Xiaolong Hu: Writing – review & editing, Software, Methodology. Biao Chen: Writing – review & editing, Software, Methodology. Sanfeng Liu: Writing – review & editing, Resources, Methodology. Huirong Mao: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, Formal analysis, Conceptualization.

Disclosures

All authors declare that there are no conflicts of interest.