Brown coat color phenotypes have been recognized in the domestic cat for over 100 years (CFA 1993). The cat locus for the brown phenotypes was originally designated as Brown, B, which follows nomenclature for the mouse. Domestic cats have a suggested allelic series, B> b > bl (Robinson 1977). The wild-type black (B) allele is dominant with normal, black (eumelanin) coloration. Cats with a brown (b) allele, either bb or bbl genotypes, are chocolate (aka chestnut), and the phenotype resulting from light brown homozygotes, blbl, are cinnamon (aka red). Several of the oldest and founding cat breeds, such as Abyssinians and Siamese, segregate for brown variants but neither the segregation nor the allelic series of the brown color variants has been documented in a peer-reviewed journal for the domestic cat.

Tyrosinase-related protein-1 (TYRP1) is the gene assigned to the B locus (Jackson 1988). The Tyrp1 enzyme has been shown to be involved in the synthesis of the eumelanin pigment (Jimenez–Cervantes et al. 1994; Kobayashi et al. 1994). As shown in mice, brown fur has 30%–40% less eumelanin content than wild-type black fur (Tamate et al. 1989; Hirobe et al. 1998). Brown phenotypes resulting from mutations in tyrosinase-related protein-1 (TYRP1) have been recognized in several species such as mice (Jackson 1988; Javerzat and Jackson 1998), cattle (Berryere et al. 2003), and dogs (Schmutz et al. 2002). However, some brown phenotypes of other cattle breeds (Adalsteinsson et al. 1995) and horses (Rieder et al. 2001) are non-TYRP1-related.

Because of the involvement of TYRP1 in melanin synthesis and since mutations in other species cause similar brown phenotypes, TYRP1 was investigated as a candidate gene for the brown phenotypes in the domestic cat. Presented are both linkage analyses and sequence analyses that associate TYRP1 mutations with feline brown color variants. Breeding and sequencing data also support the historically suggested allelic series.

Materials and methods

Sample collection and DNA preparation

Domestic cat breed samples were collected as previously described (Lyons et al. 2005). Phenotypes were verified by visual inspection, pedigrees, and photographs. DNA was isolated from buccal swabs as previously described (Young et al. 2005). DNA was extracted from white cell preparations from whole blood by standard phenol/chloroform extractions (Sambrook and Russell 2001).

Pedigree and linkage analysis

Whole-blood DNA samples were collected from cats that formed a 60-member multigenerational pedigree that segregates for the chocolate coat color phenotype (Fig. 1). Phenotypes were confirmed by visual inspection. The relationship of the cats was verified by parentage testing using 19 microsatellites markers of a standardized feline parentage identification panel following the procedures of Lyons (in preparation, data not shown). Based on chromosome homology to humans, seven microsatellites (FCA030, FCA119, FCA242, FCA641, FCA742, FCA743, and FCA744) that are within the potential region for TYRP1 on cat Chromosome D4 (Menotti–Raymond et al. 2003a, b) were analyzed in the cat pedigree as previously described (Grahn et al. 2004). Allele frequencies were determined from the unrelated cats of the pedigree by direct counting. The chocolate phenotype was examined as an autosomal recessive locus with complete penetrance using LINKAGE software (Lathrop et al. 1984).

Pedigree records from registered Abyssinian cats were analyzed to confirm the inheritance and segregation of the light brown (aka cinnamon or red) allele.

PCR and exon-based primer design

To generate TYRP1 exon sequence, eight internal exon primer pairs were designed from publicly available sequences in GenBank (http://www.ncbi.nlm.nih.gov/Gen bank/index.html) from human (NM_000550), dog (AY052751), mouse (NM_031202), and goat (AF136926). These sequences were compared for regions of high homology using Megalign software (DNASTAR, Madison, WI). The 3′ region of exon 8 had poor cross-species homology, thus the reverse primer for exon 8 was developed from human sequence only. Primers pairs were designed and optimized for use in the cat as previously described (Lyons et al. 1997). Generated PCR products were amplified and verified by sequence analyses as previously described (Lyons et al. 2005).

XL PCR and intron-based primer design

The exon-based primers were used in different combinations to amplify the exonic and intronic sequences of TYRP1 by extra-long PCR using DNA from a TYRP1-containing BAC clone and a MasterAmp Extra-Long PCR Kit (Epicentre Technologies, Madison, WI). The TYRP1 BAC clone was isolated from the RPCI–86 BAC library (BACPAC Resources, CHORI, Oakland, CA) using oligo screening following the manufacturer’s recommendations and as previously described (McPherson et al. 2001). The “overgo” primer used to identify the BAC clone was TYRP1-Ova tcaccctcagtttgtcattgccac; TYRP1-Ovb tttcttctgacctc ctggtggcaat. Long PCR reactions were performed according to manufacturer’s suggestions. The XL PCR products of the introns were cloned using the TOPO TA Cloning Kit (Invitrogen, Carlsbad, CA). The plasmids were directly sequenced using the manufacturer’s primers and sequence verified as described (Lyons et al. 2005). New exon-specific primers were designed from generated feline sequence and were placed approximately 100 bp from each exon–intron boundary, within the flanking introns of each exon. For the introns that could not be amplified by XL PCR, introns 1, 3, and 6, primers were developed from dog sequence available through the University of California, Santa Cruz genome browser (http://www.genome.ucsc.edu/). The intron-based primer pairs (Table 1) were then optimized and the sequence confirmed as described above.

Table 1 Primer sequences and accession numbers for domestic cat TYRP1
Full size table

Brown mutation screening

Two cats representing each brown phenotype, wild-type black, chocolate, and cinnamon, were sequenced for the complete coding region, portions of the 5′ untranslated region (UTR), and introns to identify potential SNPs conferring phenotypic changes. The sequences from the different phenotypes were then compared using the Sequencher program (GeneCodes, Ann Arbor, MI). Once putative SNPs were identified, the TYRP1 exons containing the SNPs were amplified from genomic DNA of cats with the appropriate phenotypes using the optimal PCR condition that was consistent for all primers, 1.5 mM Mg2+ at 56°C. DNA samples were analyzed for 11–16 SNPs from 9 random-bred cats and 31 cats representing 11 breeds (Table 2). Additionally, 10 or less of the SNP genotypes for a variety of other cats were determined and are presented in Table 2. The cat breeds represented individuals from the United States, England, Europe, and Australia. Overall, 75 cats representing 19 breeds were genotyped for different combinations of the SNPs.

Table 2 TYRP1 genotypes for cats of wild-type and brown phenotypes
Full size table

Results

Pedigree and linkage analyses

The linkage analyses among seven microsatellites, the chocolate phenotype, and an exon 4 SNP haplotype suggested significant linkage. TYRP1 SNP and microsatellite data were available for all cats; phenotypic data were available for only 49 cats because some kittens had not yet developed their point-restricted coloration. The TYRP1 exon 4 mutations were completely concordant with the chocolate phenotype (Z = 4.05, θ = 0.0). Three markers (FCA742, FCA743, and FCA119) suggested linkage to the brown phenotype (Z = 4.02, 4.02, 3.11, all at θ = 0.0, respectively). Markers FCA742 and FCA743 also suggested linkage to the TYRP1 SNPs (\( \hat Z \)= 3.96 at \( \hat{\theta}\) = 0.14, for both markers), although recombinants could be detected. The other five markers were not statistically significant or did not suggest tight linkage. The linkage data and the identified recombinants between the SNP haplotype and the three linked microsatellite markers are presented in Fig. 1.

Fig. 1
figure 1

Pedigree segregating for TYRP1 “chocolate” and wild-type “black” alleles in Persian cats. Circles represent females, squares represent males, open symbols indicate wild-type black animals, solid symbols indicate brown cats. A diamond is a cat of unknown sex. Symbols with “?” have unknown phenotypes. Numbers in bold represent the laboratory sample number. Genotype haplotypes are drawn for each cat in the order: TYRP1 SNP, FCA742, FCA743, FCA119. The “1” allele represents the wild-type SNPs and the “2” allele represents the brown-associated SNPs. An “R” identifies recombinant individuals between the phenotype and haplotype to the markers.

Full size image

An informal segregation analysis of 30 Abyssinian cats segregating for the cinnamon mutation was conducted using pedigrees from one Abyssinian cat breeder. The segregation ratios for all mating types of cats with the ruddy (wild-type black) and red (cinnamon, light brown) phenotype are consistent with an autosomal recessive inheritance that is allelic to wild-type black. All red-to-red Abyssinian breedings produced all red cats (N = 10). The ruddy heterozygous × homozygous red mating has produced 65% red cats, but deviation from expected Hardy–Weinberg segregation ratios was not significant as determined with a chi-squared analysis with one degree of freedom (p > 0.15). No gender bias was observed (p > 0.53).

Feline TYRP1

The intron-based primers developed for each TYRP1 exon, product sizes, and GenBank Accession numbers are presented in Table 1. The entire coding sequence of the feline TYRP1 gene and partial sequence of the 5′ UTR and introns II – VII were analyzed. The complete coding sequence for feline TYRP1 and partial intron sequence, as sequenced from genomic DNA, is available as Supplemental Fig. 1. Feline sequence features were based on human TYRP1. Human sequence features was obtained from the Ensembl database (http://www.ensembl.org/). The feline TYRP1 sequence is 1725 bp, of which 1614 bp codes for a 537-amino-acid Tyrp1 protein in the domestic cat. The consensus feline sequence has an 88% nucleotide identity to the human sequence. The dog sequence has a 90% nucleotide identity to cat, while mouse and goat sequences are 84% and 88% identical to cat, respectively.

Figure 2 presents the amino acid translation for the domestic cat and the alignment to other species. At the protein level, the cat TYRP1 has 88% homology in amino acid sequence to human TYRP1, the dog sequence has 92% homology to cat, while the mouse and goat have homologies of 83% and 87%, respectively.

Fig. 2
figure 2

Protein alignment of coding sequence from TYRP1. Alignment of the 537 amino acid translation for TYRP1 between human (Hsap, NM_000550.1), mouse (Mmus, X03687.1), dog (Cfam, AY052751.3), and an orange cat (Fcat wt), a cinnamon cat (Fcat Cinn), and a chocolate cat (Feat Choc). Bold text indicates sites of feline phenotype sequence variants. The cinnamon phenotype has a variant at position 100 that results in a premature stop codon, while the chocolate phenotype has a variant at position 3. For the chocolate cat, exon 6 is underlined, indicating that this exon may be skipped. The possible translation of the known intron 6 sequence is provided should intron inclusion result from the intron 6 mutation. The 3′portion of the sequence is unknown and represented as question marks. The Cfam fragment is missing the terminal 36 amino acids that are indicated by dashes.

Full size image

TYRP1 mutations

Sixteen single nucleotide polymorphism (SNPs) were identified in the feline TYRP1 sequence, eight were within the coding region and the positions are presented in Table 2. The only two SNPs that result in amino acid changes are in exon 2. One polymorphism causes a minor amino acid change in exon 2. The C to G transversion at position 8 causes the incorporation of glycine instead of alanine at amino acid position 3 of the coding sequence, A3G (Fig. 2). This polymorphism is correlated with the chocolate phenotype and is part of the haplotype consistent with chocolate cats (Table 2). One C to T transition is in the 5′ UTR, 50 bp upstream of the translation start codon, and is found in only one cat, a Korat. The locations of the other seven mutations in the noncoding region are presented in Table 2.

Putative brown mutations

Common brown variant phenotypes of the domestic cat are presented in Fig. 3. The C to T transition at position 298 replaces an arginine at amino acid 100 with the opal (UGA) stop codon, R100OPA. This premature stop in exon 2 is concordant in nine cats from four different breeds with the chestnut, cinnamon, fawn, and red phenotypes that are considered light brown, bl. No cats (N = 32) representing nine different breeds with the brown or wild-type black phenotypes had this stop mutation (Table 2). One Ocicat with reported cinnamon silver coloration did not have the stop mutation.

Fig. 3
figure 3

Brown phenotypes of the domestic cat. a. Brown (bb) in a chocolate Exotic Shorthair (left), and light brown (bl bl) in a red (aka cinnamon) Abyssinian (right). b. Black (B-) in a wild-type brown tabby domestic shorthair (left) and wild-type black domestic shorthair (right). [Supplemental Fig. 1. Sequence alignment of domestic cat TYRP1 to other species. Messenger RNA, 1611 bp, from human (Hsap, Accession No. NM_000550.1), mouse (Mmus Accession No. X03687.1), and dog (Cfam Accession No. AY052751.3) and genomic sequence data from domestic cats; orange (wt), cinnamon (Cinn), and chocolate (Choc) from domestic cat (Fcat) are aligned. Arrows (↓) and bold text indicate sites of feline phenotype sequence variants. The cinnamon (light brown, bl) phenotype has a single sequence variant at coding position 298 (597 above) that results in a premature stop codon. The brown phenotype has 7 coding sequence variants of which only the C to G transversion at coding position 8 (307 above) results in an amino acid change. The additional seven noncoding sequence variants are also indicated. The Cfam mRNA fragment is missing the terminal 108 bp. The untranslated regions of exons 1 and 2 and flanking intron sequence (lower case) for each exon are provided for the cat.

Full size image

The SNP in intron VI is within 5 bp of the exon 6 boundary, disrupting the highly conserved “G” in position 5 of the exon 6 downstream splice-donor recognition signal. Additionally, all SNPs, except for the exon 2 site that causes the R100OPA, are concordant with the chocolate phenotype and form a consistent haplotype in nine cat breeds (Table 2). These mutations are unique and homozygous in all the chocolate cats analyzed (N = 28). Eighteen chocolate cats were typed for the extended haplotype (Table 2). This haplotype is distinct from the exon 2 stop mutation found for the cinnamon phenotype. The stop mutation found in cinnamon (red) cats has a haplotype consistent with the wild-type black allele. Ten cats with chocolate phenotypes were typed for only the exon 2 or exon 4 mutations and were all consistent with the brown haplotype, which extends the list of breeds to 11 that all have the same chocolate haplotype. Five cats with the dominant black phenotype are heterozygous for several SNPs in the brown haplotype. Two cats, a lilac Munchkin and a chocolate silver Somali, are heterozygous for the stop mutation and the SNPs associated with chocolate.

Discussion

Brown coat color variations have been recognized in a variety of domestic cat breeds and wild felids. Two loci in the cat mimic brown coloration. The wild-type brown tabby cat has only black and yellow pigmentation; the brown phenotype is the result of the optical illusion of agouti fur, and a temperature-sensitive allele of tyrosinase, TYR, termed burmese, cb, produces a brown or sable appearance to the coat for a genetically black cat.

Three alleles are suspected in truly brown-colored cats, normal black pigmentation (B), the brown phenotype (b), and a lighter brown phenotype (bl). The analysis of an extended pedigree segregating for chocolate confirms that the phenotype is allelic to black. Analyses of Abyssinian pedigrees segregating for light brown (aka cinnamon or red in Abyssinians) suggests that light brown is also allelic to black, supporting and documenting the allelic series B > b > bl.

The linkage analyses supported TYRP1 as the candidate gene for the cat brown phenotypes. This analysis suggested that TYRP1 is telomeric on cat Chromosome D4, which is consistent with suspected homologies between humans and cats.

The light brown or cinnamon phenotype in the cat is concordant with a stop mutation in exon 2 that has been identified in all breeds with light brown color variants. The only difference between the cinnamon and wild-type black alleles is the single nucleotide that causes the translation stop site. All other breeds with the same phenotype had the identical stop mutation, suggesting the phenotype is caused by the same mutation that is identical by descent across these breeds. One cat that was designated a cinnamon Ocicat did not have the stop mutation. This suggests that the Ocicat phenotype could caused by a different mutation in TYRP1 or by a mutation in a different gene. A more likely explanation is that the use of the same phenotypic description—cinnamon—for light brown variants in different breeds of cats could be misleading. Since some combinations of color alleles epistatically influence brown coloration, the cinnamon designation for this cat could be erroneous.

The brown phenotype in the chocolate (aka chestnut) cat is associated with a mutation in intron 6 that likely disrupts the splice-donor downstream of exon 6. This alteration could allow exon skipping and the removal of exon 6 from the mRNA. Although a rarer event than exon skipping (O’Neill et al. 1998), intron inclusion due to the presence of a cryptic splice site in intron 6 may also occur, allowing protein translation to continue into intron 6. The protein translation up to the end of the available sequence does not create a stop codon; however, this could occur downstream. Additionally, if intron inclusion occurs and shifts the reading frame, stop codons would be encountered early in exon 7, considering either reading frame. Similar mutations have been noted in other genes that cause disease (Nicholls et al. 1996; O’Neill et al. 1998). Analysis of the mRNA sequence and expression levels would support and clarify the effects of the intron 6 mutation.

The exon 2 nonsense mutation and several silent mutations found throughout the gene form a consistent haplotype in all chocolate cats. Interestingly, these mutations all occur on the same allele that has not been identified in nonchocolate cats. Unexpectedly the Korat cat breed had the chocolate allele. Korats are accepted in only one color, blue, that is considered a dilution of black. However, there are various shades of blue (aka gray) and the data suggest that Korats may actually be fixed for chocolate and dilution, although the intron 6 mutation still needs to be genotyped in these cats. The analysis of more Korats and other breeds fixed for blue, such as Russian Blues and Chartreux, would support this conclusion. In addition, further analyses in a wider variety of cats may show disruption of the chocolate haplotype.

Five cats that have wild-type phenotypes are heterozygous for the wild-type and chocolate alleles, and two cats with the chocolate phenotype are likely to be heterozygous for the chocolate and red mutations. These data show that wild-type is dominant to chocolate and chocolate is a dominant allele to red, further supporting the allelic series for brown in cats.

None of the currently identified brown mutations are identical between cats, mice, dogs, and humans, including the mutation in humans that causes oculocutaneous albinism 3 (OCA3). The OCA3 is a single-base-pair deletion in exon 6 that leads to a downstream stop codon. The TYRP1 mutations in the cat are associated with different shades of brown, which is also true in the mouse. The three mutations identified in dog TYRP1 are associated with brown phenotypes from different breeds; however, the brown phenotype across the breeds appears to be the same shade of brown.

In the United States, there are approximately 50 cat breeds, some of which are hybrids with wildcat species and long-haired and short-haired varieties of the same breed. Not all breeds or colors are accepted by different cat fancier organizations but basically there are two breeds (Havana Brown and Singapura) that are fixed for brown (chocolate) and six breeds (Abysinnian, Devon Rex, Ocicat, Oriental, Somali, and Sphynx) that segregate for the light brown (cinnamon) allele. The different breeds use different terms for the brown coloration including brown, chestnut, chocolate, champagne, and lilac. Lilac and champagne are brown variants associated with other epistatic color genes such as tyrosinase and dilution. Cinnamon or red is generally recognized as the light brown allele, bl. Fawn and platinum are dilutions of light brown. However, the variations in pigment intensity caused by the environment, dilution factors of other genes, point-restriction, and minor polygenic effects make distinguishing the phenotype of brown present in a cat challenging for breeders.

The fear of undesirable traits, such as particular color variants and diseases, quenches the enthusiasm for opening stud books and allowing the outcrossing of many domestic breeds. Many new breeds have a very limited population size, thus, the eradication of cats that have only a single allele that is undesirable could be very detrimental to the genetic diversity of the population. Genetic testing for some of the undesirable traits can alleviate some of these fears and can promote better genetic diversity, increased heterozygosity, and hence possibly enhancing the health of our companion animal breeds. Thus, if adopted as genetic tests for color, these mutations can improve the efficiency of breeding programs for a variety of breeds.