MLPH is located on chicken chromosome 7. The genomic organisation of melanophilin is shown (Figure 2). Sequence analysis carried out on cDNA, and confirmed on genomic DNA, showed that MLPH consists of 17 exons that contain 2052 bp of coding sequence, translating into 683 amino acids. We found that exons 6 and/or 9 are absent in some transcripts, most probably due to alternative splicing events (Figure 3). Exon 1 of chicken MLPH is also homologous to exons 1 and 2 in mammals 51219.

In a similar manner to the expression seen in mice 5, chicken MLPH is expressed predominantly in epithelial-rich tissues (Figure 3). We found no differences in the level of expression with respect to sex or genotype, except in the ovary, where expression was higher in lavender females (data not shown). Expression of the ubiquitously expressed GAPDH was constant across all samples. With respect to age, we detected the highest degree of embryonic expression at embryonic day 14 (E14), consistent with this period having the highest rate of development of feather follicles 28, reducing at E17 and then increasing again at 10 weeks, when the adult plumage is being produced (Figure 3).

We identified 7 SNPs – 5 synonymous and 2 non-synonymous variants (C103T, non-synonymous (R35W); T996C, synonymous; A1056G, synonymous; C1284T, synonymous; A1356G, non-synonymous (I452M); G1701A, synonymous; C1824T, synonymous) – amongst the 94 chicken samples sequenced in their entirety for this study. Of the two non-synonymous SNPs (Table 1), only one was found to be associated with the lavender phenotype (Figure 4A). All the other SNPs were found to be polymorphic in wild type birds. This single point-mutation, found in exon 1, 103 bp from the start codon, results in the amino acid change R35W (where an Argenine in the wild type is replaced by a Tryptophan), which is the same mutation reported in humans with Griscelli syndrome type III 12, and is also included in the deletion found in leaden mice 5 (Figure 4B).

The C103T mutation results in the loss of the single restriction site for the BsrBI restriction enzyme, which cuts the sequence CCGCTC, found only in the wild-type LAV*N allele. This meant that we could design a simple RFLP test in order to identify the presence or absence of the lavender allele. Using a PCR product amplified from genomic DNA, the LAV*N allele is digested into two fragments of sizes 334 bp and 146 bp by BsrBI, while the LAV*L allele remains intact at 480 bp (Figure 5).

We obtained the RFLP genotype for chickens of known lavender genotype in the Nouzilly pedigrees. For the 159 individuals tested in this manner, we found complete association between the lavender allele and the presence of the C103T SNP: all 46 individuals with the lavender phenotype were homozygous for the 480-bp allele and all birds known to be heterozygous for LAV*N/LAV*L (offspring of four lavender dams mated with a wild-type sire) were heterozygous at C103T.

Analysis of cosegregation in a pedigree confirmed this result (Figure 6). Several informative families were produced by mating heterozygous dams to homozygous lavender sires (Figure 6). The segregation of haplotypes was completely in accordance with the observed phenotypes (LOD = 3.91).

The protein sequence of the murine melanophilin gene (NM053015) was used to search for a homologous gene in the chicken genome (The Sequencing Centre, Washington University, St Louis) using TBLASTN 31. Two contigs, accession numbers AADN01050916 and AADN01050915, demonstrated a high degree of similarity to the murine protein sequence. The contigs were analysed in silico in order to predict the chicken MLPH gene sequence using homology and gene prediction programs, primarily GENSCAN 32, to yield a sequence containing the full-length coding region of the chicken melanophilin. The predicted mRNA sequence and genomic structure were used to design primers for DNA and cDNA amplification to verify the sequence (all primer sequences are shown in Table 2). All primer pairs used in these experiments were designed using Primer3 33. The final mRNA sequences were verified and submitted to EMBL (EU007437-40).

Four informative families were produced for pedigree analysis at the experimental facilities of the Institut National de la Recherche Agronomique (INRA), located in Nouzilly, by mating two homozygous lavender (LAV*L/LAV*L) sires to four heterozygous (LAV*L/LAV*N) dams. In total, ten progeny were scored for the homozygous presence of the LAV*L allele, along with seven heterozygote individuals at the LAV locus that displayed undiluted black plumage colour, due to their extended black background. These samples were used in the PCR-RFLP analysis.

A further 75 birds were also produced from the INRA flock's lavender line. The full MLPH coding sequence from 15 of the lavender birds and 14 of the heterozygotes produced from some of these breedings were sequenced in their entirety. The rest, a mixture of LAV*L/LAV*L and LAV*N/LAV*L, were analysed by PCR-RFLP only.

One hundred and fifty progeny, homozygous for the wild type LAV*N allele, were produced from several different breeds within the INRA flock, displaying a range of different, but undiluted, plumage phenotypes. 65 of these were used for sequencing, while the rest were tested for the absence of the lavender allele by PCR-RFLP.

Five chickens with the lavender phenotype were also sourced from two alternative farms in the UK and used for the PCR-RFLP analysis.

Dorsal skin tissue containing developing feather follicles was collected from embryos that had been incubated together until egg day 14 or 17, dependent on the particular breed. In the case of adult tissue samples, feathers were plucked from a patch on the dorsal surface of each bird to induce synchronised follicle development and then the birds were sacrificed 2 weeks later. All tissue samples were immediately submersed in RNAlater and stored at -20°C. In the case of adults, other tissues (liver, muscle, kidney, heart, lung, testis, spleen, brain, and ovary) were also collected at the same time and stored in a similar way to the skin samples.

Up to 100 mg tissue in RNAlater (Ambion, Austin, TX, USA) was homogenised in TRIzol (Invitrogen, Paisley, UK) using a TissueLyser (Qiagen, Crawley, UK). Total RNA was isolated according to the manufacturer's protocol (Invitrogen). RNA quality and quantity were evaluated with the RNA 6000 Nano LabChip kit on the Agilent 2100 Bioanalyzer (Agilent Technologies, South Queensferry, UK) and then stored at -80°C. Subsequent DNA extractions were also carried out using the normal TRIzol protocol.

All RT-PCR reactions were performed on cDNA synthesised from 2 μg of RNA using the RETROscript kit (Ambion), according to the manufacturer's protocol, using oligo-dT in a 20-μl reaction volume. Amplification reactions followed standard PCR protocols in a volume of 50 μl, using a 1.5-μl aliquot of the RT reaction mixture, 1.5 mM MgCl₂, 0.2 mM of each dNTP, 0.04 μM each of the primers RACE4F and ML15R (Table 2), and 1.25 units Taq DNA polymerase (Bioline). PCR was initiated by denaturation for 2 minutes at 94°C, followed by 35 cycles consisting of 30 seconds at 94°C, 1 minute at 64°C and 3 minutes at 72°C, and completed by a final extension of 5 minutes at 72°C. Twenty microlitres of the amplification product was visualised on a 1.5% agarose-TAE gel stained with ethidium bromide.

To determine the 5'- and 3'-regions, 5' and 3' RACE was performed using the SMART RACE cDNA Amplification Kit (Clontech) according to the manufacturer's instructions, with 1 μg of total RNA from chicken using the primers RACE1R, RACE2R, RACE5F and MV8R (Table 2); the products were then separated by agarose gel electrophoresis, gel purified, cloned and sequenced.

PCR products were directly sequenced, and also following cloning, on an Applied Biosystems (ABI) model 3730 DNA sequencer (ABI, Warrington, UK) and the sequencing reactions were carried out using BigDye Terminator v1.1 Cycle Sequencing kit (ABI) under the following conditions: 26 cycles of 94°C for 20 seconds, 52°C for 20 seconds and 60°C for 4 minutes. Sequencing reactions were purified and cleaned by ethanol precipitation. The primers used for initial amplification were subsequently used for sequencing with additional internal primers (Table 2) SJ93, MV2R, MV18F, ML14F, BlotE91F, ML3R, ML16F and MV6F, allowing the whole sequence to be overlapped and confirmed on both strands.

The full coding sequence of the MLPH transcript, including some 5'-UTR and 3'-polyA sequence, was amplified by PCR using primers ML5'-1 and ML3'-1, and was subsequently cloned into the vector pGEM-T Easy (Promega, UK) according to the manufacturer's instructions. The plasmids were completely sequenced in both directions using the universal M13 primers (forward and reverse).

The sequences obtained were edited and aligned using ClustalW software 34 and the Proseq v.2.91 software 35, then re-edited and realigned manually.

A fragment of genomic DNA was amplified for PCR-RFLP analysis using primers ML1F, upstream of the start codon, and RFLP-F1, in intron 1 (Table 2), using standard PCR protocols in a volume of 50 μl that included 1.5 μl of genomic DNA (100 ng/μl), 1.5 mM MgCl₂, 0.2 mM of each dNTP, 0.04 μM each of primer, and 1.25 units Taq DNA polymerase (Bioline). A standard PCR program was used of 40 cycles with a 1-minute annealing step at 58°C and a single extension step of three min at 72°C. Ten microlitres of PCR product was digested in a standard restriction digestion protocol using the BsrBI restriction enzyme (BioLab). Twenty microlitres of the digestion product was visualized on a 1.5% agarose-TAE gel stained with ethidium bromide.