Cis-acting mutation affecting GJA5 transcription is underlying the Melanotic within-feather pigmentation pattern in chickens

Significance The molecular mechanisms underlying pigmentation patterns in animals is to a large extent an unresolved mystery in biology. For example, compared with mammals, birds show a stunning diversity in pigmentation patterns. This study advances the knowledge concerning the mechanisms creating periodic pigmentation patterns in individual feathers. We show that a mutation upstream of GJA5 encoding a gap-junction protein is causing the Melanotic phenotype in domestic chickens. Melanotic affects within-feather pigmentation patterns by enhancing the contrast between dark- and light-colored regions in the feather. The result implies that cell–cell communications between melanocytes and other cells in the feather follicle play a critical role for pattern formation.

whereas the remaining 81 samples carried the wild-type allele ml + . All Illumina pairedend FASTQ data were aligned to the red junglefowl genome assembly version GalGal6 using BWA (version: 0.7.12) (3), sorted with SAMtools (version: 1.6) (4), and variants were called with GATK HaplotypeCaller 3.8 (5). Structural variants were called with Lumpy (version: 0.2.13) (6). Sequence data from the backcross population (Pool_pencilling and Pool_double_lacing) were analyzed according to the same procedure, and were used to calculate pairwise FST values with Popoolation2 (7) in a sliding window approach with window size of 50 kb and step size of 10 kb and pool sizes of 60 pencilling individuals: 43 double lacing individuals. The FST values were Ztransformed and plotted against genomic locations.
Linkage mapping: The absRAFdif from SNP-MaP are arbitrary, because of possible microarray-based intensity variation and potential imprecision when construction DNA pools (8). Therefore, individual genotyping was necessary to confirm the allele frequencies and further reduce the candidate region. For the first round of linkage mapping, we selected six SNPs (SI Appendix, Table S1) with relatively high absRAFdif values (when comparing the backcross populations, the maximum absRAFdif is 0.5 in theory) within the peak region on chromosome 1 for further analysis, via the Kompetitive allele-specific PCR assay (KASP), developed by LGC Genomics (Beverly, MA, USA; www.lgcgenomics.com) (9). The six SNPs were genotyped using individual samples of all the F0, F1, and backcross females with pencilling or double lacing phenotypes, from the linkage mapping population. KASP assays were conducted following the manual with minor modifications. Genotyping results were validated by at least two replicates for each sample. Linkage mapping was carried out using the CRIMAP software (10) including the calculation of map distances and log10 odds ratio (LOD) between individual SNPs and Melanotic.
Within the candidate region defined by two of these six SNPs, ten more SNPs, fixed for different alleles in the parental lines and identified by whole genome sequencing, were selected for genotyping (SI Appendix, Table S2). These ten SNPs were used for a second round of linkage mapping, using only five recombinant backcross individuals via standard PCR and Sanger sequencing.
Diagnostic test: After the analysis of the candidate region based on the second round of linkage mapping, six sequence variants were identified as candidate causal mutations, which includes two InDels and four SNPs (SI Appendix, Table S4). A collection of 101 DNA samples from 52 different populations of chickens was studied focusing only on these six mutations. Nineteen of these populations carry Melanotic while the remaining thirty-three are wild-type at the Ml locus (SI Appendix, Table S5). Each of these samples were genotyped by fragment analysis following standard PCR amplifications for the two InDels in SI Appendix, Table S4. Three PCR primers were used for genotyping each of the two InDels (SI Appendix,  Table S6. Each qPCR assay was carried out using three technical replicates. Allelic imbalance test: Total RNA was extracted using Quick-RNA Miniprep Plus Kit (Zymo Research). First-strand cDNA was synthesized using SuperScript TM IV VILO TM Master Mix (Invitrogen). Two PCR primers were designed to quantify the allelic imbalance of GJA5 expression: genomic DNA (gDNA) was amplified using GJA5_AI_F and GJA5_AI_R, while cDNA was amplified using GJA5_A_F or GJA5_B_F (for two different transcripts of GJA5) with the same reverse primer, respectively (SI Appendix, Table S6). PCR products from three cDNA samples which came from the three F1 individuals together with PCR products from gDNA of the same three individuals, were Sanger sequenced using the same reverse primer. The peak heights for three SNPs, rs313638830 and rs731128040 found in transcript NM_205504.2 and rs312762853 found in transcript XM_015295951.2, were quantified using the PeakPicker 2 software (11). Allelic imbalance was estimated as the ratios of the peak height of the Ml allele over the ml + allele in the cDNA or gDNA.
Section immunostaining and in situ hybridization: Fixed embryos were embedded in paraffin and sectioned at 6-7 µm. After de-paraffinization, sections were processed to immunohistochemistry or in situ hybridization. The antibody to MITF was from Abcam (ab122982). The peroxidase staining was used after primary antibody treatment with procedures described (12).
Reporter assay: Previous Cap Analysis of Gene Expression (CAGE) analysis detected the position of one of the transcription start sites (TSSs) of chicken GJA5 using mRNA from whole embryo samples (15). The 1,869 bp sequence (SI Appendix, Fig. S1A) upstream from this TSS in wild-type and mutant GJA5 alleles, including InDel1 and SNP rs316201461, were synthesized in vitro and cloned into the promoterless Firefly Luciferase vector pGL4.15[luc2P/Hygro] by Genscript (Leiden, Netherlands). pRL-TK plasmid monitoring the transfection efficiency was purchased from Promega (E2241). Five ng of each plasmid was used to transform the XL1-Blue competent cells (Agilent, 200236), and plasmid DNA was subsequently prepared from 200 ml overnight culture with an EndoFree plasmid Maxi Kit (Qiagen).
DF-1 chicken fibroblast cells (16) were a gift from Dr. Benjamin Schusser (Technical University of Munich). The DF-1 cells were maintained in DMEM supplemented with 10% FBS, 100 U/ml penicillin, 100 g/ml streptomycin and 292 μg/ml L-Glutamine (ThermoFisher Scientific) at 39℃ with 5% CO2. On day 1, an aliquot of 2.5 X 10 4 cells/well were seeded in a 48-well plate. On the next day, a plasmid mixture containing 247.5 ng pGL4.15[luc2P/Hygro] with or without GJA5 promoters and 2.5 ng pRL-TK was applied to transfect each well of cells using jetPRIME transfection reagent (Polyplus). Four replicates were used for each promoter construct. Medium was replaced by fresh medium 4 h after transfection. On day 3, luminescence of each well was analyzed using a Dual-Luciferase Reporter assay (Promega) on an Infinite M200 Microplate Reader (Tecan Group Ltd., Switzerland), and the luciferase activity was represented as the ratio of firefly (pGL4.15[luc2P/Hygro]) to Renilla (pRL-TK) luminescence.
TRANSFAC analysis: Predictions of putative transcription factor binding sites were done by the MATCH program in the TRANSFAC database (17). The vertebrate database was used and only predicted binding sites with a matrix score of 1.0 were selected for further analysis.