Characterization of the Rabbit Agouti Signaling Protein (ASIP) Gene

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Genomics 95 (2010) 166–175

Contents lists available at ScienceDirect

Genomics

journal homepage: www.elsevier.com/locate/ygeno

Characterization of the rabbit agouti signaling protein (ASIP) gene: Transcripts and phylogenetic analyses and identiﬁcation of the causative mutation of the nonagouti black coat colour☆

Luca Fontanesi a,⁎, Lionel Forestier b, Daniel Allain c, Emilio Scotti a, Francesca Beretti a, Séverine Deretz-Picoulet d, Elena Pecchioli e, Cristiano Vernesi e, Terence J. Robinson f, Jason L. Malaney g, Vincenzo Russo a, Ahmad Oulmouden b,⁎

a DIPROVAL, Sezione di Allevamenti Zootecnici, University of Bologna, Via F.lli Rosselli 107, 42100 Reggio Emilia, Italy b INRA, UMR1061 Unité de Génétique Moléculaire Animale, Université de Limoges, 87060 Limoges Cedex, France c UR631, Station d'Amélioration Génétique des Animaux, INRA, 31326 Castanet Tolosan, France d UE967, Génétique Expérimentale en Productions Animales, INRA, 17700 Surgères, France e Research and Innovation Centre, Fondazione Edmund Mach, Viote del Monte Bondone, 38040 Trento, Italy f Department of Botany and Zoology, Evolutionary Genomics Group, University of Stellenbosch, PB X1, Matieland 7602, South Africa g Museum of Southwestern Biology and Dept. of Biology, University of New Mexico, Albunquerque, NM 87131-0001, USA

article info abstract

Article history: The agouti locus encodes the agouti signalling protein (ASIP) which is involved in determining the switch Received 16 September 2009 from eumelanin to pheomelanin synthesis in melanocytes. In the domestic rabbit (Oryctolagus cuniculus) Accepted 27 November 2009 early studies indicated three alleles at this locus: A, light-bellied agouti (wild type); at, black and tan; a, Available online 11 December 2009 black nonagouti. We characterized the rabbit ASIP gene and identiﬁed the causative mutation (an insertion in exon 2) of the black nonagouti allele whose frequency was evaluated in 31 breeds. Phylogenetic analysis Keywords: Agouti of ASIP sequences from Oryctolagus and 9 other species of the family Leporidae placed Oryctolagus as sister Coat colour species to Pentalagus and Bunolagus. Transcription analysis in wild type agouti rabbits revealed the presence Lagomorpha of two major transcripts with different 5′-untranslated regions having ventral or dorsal skin speciﬁc Mutation expression. ASIP gene transcripts were also detected in all examined rabbit tissues distinguishing the rabbit Oryctolagus cuniculus expression pattern from what was observed in wild type mice. Rabbit © 2009 Elsevier Inc. All rights reserved. Transcription analysis

Introduction mining the wild type (banded-hair) agouti coat colour is caused by the mutually exclusive binding of the MC1R by the ASIP protein or by the α- Coat colour genetics has been the subject of a large number of studies melanocyte-stimulating hormone (α-MSH) [6,7]. In mice, as well as in that, in mice, led to the identiﬁcation of more than 300 loci affecting other species, recessive loss-of-function mutations of the agouti gene, pigmentation [1]. Among these, the agouti and the extension loci interact which impair either the protein function or reduce the level of its mRNA to determine the relative amount of two types of melanin, pheomelanin synthesis, impact the production of eumelanin that, in turn, results in (yellow-red pigment) or eumelanin (black-brown pigment), present in darker coat colour [3,8–17]. Conversely, dominant gain-of-function hair [2].Theagouti locus encodes the agouti signaling protein (ASIP) mutations which display ubiquitous and constitutive expression of this which is a 131-amino acid paracrine peptide that in wild type mice is gene, such as the Ay (lethal yellow) allele in mice, cause yellow produced only in specialized dermal cells at the base of the hair follicle pigmentation [3,4,18]. In mice carrying the Ay allele, the effect on coat [3,4].Theextension locus encodes the melanocortin 1 receptor (MC1R), colour is accompanied by obesity, infertility, tumour susceptibility and which is a seven-transmembrane G protein-coupled receptor located on other defects (i.e. [19]), whereas in mice with the wild type alleles (A the surface of the melanocytes [5]. The biological mechanism deter- and Aw) agouti expression is restricted to the skin with a speciﬁcpattern of expression derived from different gene transcripts [3,20]. Neverthe- less, ubiquitous expression of the ASIP gene has been reported in ☆ Sequence data from this article have been deposited with the DDBJ/EMBL/ human, cattle and white sheep without the detrimental phenotypes GenBank Data Libraries under Accession Nos. AM748787, AM748788, FN547135– observed in the mice models [15,21–23]. FN547150. In contrast to agouti, recessive loss-of-function and dominant ⁎ Corresponding authors. L. Fontanesi is to be contacted at fax: +39 522 290523. A. Oulmouden, fax: +33 5 55457653. gain-of-function mutations of the MC1R gene cause pheomelanic and E-mail addresses: [email protected] (L. Fontanesi), eumelanic phenotypes, respectively [5,8,13,24–32].

Following the rediscovery of the Mendel's laws at the beginning of the last century, the domestic rabbit (Oryctolagus cuniculus) was considered, together with the mouse, a model species to study coat colour genetics due to the availability of several strains with an extreme differentiation of this phenotype [33,34]. Actually, most rabbit breeds and strains, many of which are known under different names in different countries, are identified mainly by characteristic coat colour and patterns leading breeder organizations to establish specific breed-standards. Early studies summarized by Castle [35], Robinson [36] and Searle [2] have suggested the existence of five alleles at the rabbit extension locus. These are listed according to their dominance level: ED (dominant black), ES (steel, weaker version of ED), E (normal grey or normal extension of black), eJ (Japanese, mosaic distribution of black and yellow) and e (non-extension of black, yellow/red with white belly). We recently sequenced the rabbit MC1R gene in different breeds and showed that two in-frame deletions are associated with dominant black (c.280_285del6; alleles ED or ES) and recessive red (c.304_333del30; allele e) coat colours, respectively [31]. Currently, the rabbit agouti locus has been only studied using classical genetic approaches which suggest the presence of three alleles (A, wild type allele; at, black and tan; a, black nonagouti, recessive) that have been assigned to linkage group IV [2,35–38]. The effect of the rabbit wild type A allele (Fig. 1) resembles that of the mouse Aw (light-bellied agouti) allele described to be the true wild type allele in murine species [39,40]. Moreover, the most frequent genotype of some rabbit breeds at this locus has been deduced or established by pedigree studies [35,36,38]. For example, several black/dark breeds such as Alaska, Champagne d'Argent and Blue American or Blue Vienna and other breeds (Supplementary Fig. 1) have been assigned the a/a genotype. Here, we investigated the rabbit ASIP gene at the genomic and transcriptional levels and show that this gene is expressed in all examined tissues thus mimicking observations in cattle and human but differing from the expression patterns observed in wild type mice. Moreover, by analysing rabbits from 31 breeds with different coat colours, we identified the causative mutation of the nonagouti black phenotype and several other polymorphisms that were organized in five haplotypes which we compared to orthologous sequences obtained from other species of hares and rabbits of the family Leporidae. Fig. 1. Castor Rex rabbit with light-bellied agouti phenotype (A) exhibiting dorsal Results agouti banded hairs (B) and ventral white hairs (C), a pattern similar to Aw/Aw mice.

Dorsal full-length transcript and genomic structure of the rabbit ASIP gene whose pattern of spacing is highly conserved among vertebrates (Supplementary Fig. 2). We used rapid amplification of cDNA ends (RACE) on total RNA The rabbit ASIP gene was amplified by long PCR from genomic DNA extracted from dorsal skin of 1-day-old castor rex rabbit to isolate a resulting in a 4590 -bp fragment (GenBank accession no. AM748788) full-length ASIP cDNA. We sequenced fourteen clones for 5′-RACE and that contained the coding sequence (CDS) of the three coding exons ten clones for 3′-RACE. Clones of each category exhibited the same (exons 2, 3 and 4) of the ASIP gene, in addition to intronic sequences. sequence. The combined sequence obtained by RACE experiments Additional upstream (45 bp, GenBank accession nos. FN547135, allowed us to establish the full-length cDNA. The dorsal skin transcript FN547136, FN547137 and FN547138) and downstream (197 bp, (designed 1C transcript; see below) of Castor Rex rabbit was 757 bp GenBank accession no. AM748787) genomic sequences to the 4590 - long (excluding the poly A tail). It is composed of a 402 -bp open bp region were obtained from the resequencing activity and 3′-UTR reading frame (ORF) (including the stop codon), a 162 bp 5′- transcript isolation, resulting in a total of 4832 bp of ASIP genomic untranslated region (UTR) (152 bp of the 1C non-coding exon and sequence. A schematic representation of the rabbit ASIP gene is 10 bp of the exon 2; see below), and a 193 bp of the 3′-UTR (GenBank reported in Fig. 2. The genomic organization of the rabbit ASIP coding accession no. AM748787). The ORF encodes a 133 amino acid protein, exons 2, 3 and 4 and introns 2 and 3 is similar to that reported for all which is 80%, 78%, and 77% identical to the mouse (131 amino acids; other mammals [3,4,14,22,23,40]. Coding exons 2, 3 and 4 are GenBank accession no. L06941 [3]), human (132 amino acids; separated by 1277 and 2911 bp intronic sequence, respectively. GenBank accession no. L37019 [22]) and bovine (133 amino acids; GenBank accession no. X99692 [23]) proteins, respectively. The coded 5′-UTR of ventral ASIP transcripts amino acid sequence has the characteristic signature for mammalian ASIP protein (i.e. [3,4,14,22,23]). A signal sequence of 22 amino acids It is known that the light-bellied-agouti Aw phenotype in mice is is followed by a basic amino-terminal domain of 63 residues established by differential expression of ventral specific transcripts (containing 15 lysines or arginines), a 3 proline region (6 amino (1A, 1A′ and 1AA′), as well as dorsal specific transcripts (1B and 1C) acids), and a 42-residue cysteine-rich carboxy-terminal domain, that differ only by their 5′ UTR [20]. In order to evaluate if different 168 L. Fontanesi et al. / Genomics 95 (2010) 166–175

Fig. 2. (A) Genomic organization of the rabbit ASIP gene. Exons (2, 3 and 4) are shown as boxes. Solid boxes indicate protein coding sequences. Untranslated regions are shown as striped (horizontal or diagonal) boxes. Polymorphisms identiﬁed in the different sequenced rabbit breeds are marked on the genomic organization of the gene. The c.5_6insA mutation is indicated in red. Nonagouti mutations identiﬁed in other species [3,8–11,13–17] are reported above in the correspondence of the position related to the rabbit gene structure. (B) The four ASIP transcript variants are shown below the genomic structure of the gene. The position of the exons 1A and 1C was based on conservation of this region among species. The position of the exon 1dv is not known.

ASIP transcripts are also produced in rabbit, we performed RACE mouse. Therefore, it seems that the rabbit ASIP gene might have only experiments using ventral skin RNA from a 1-day-old Castor Rex one ventral specific 1A transcript (see below). rabbit of the A/A genotype (based on its light-bellied coat colour pattern which resembles the Aw/Aw mouse phenotype; see Fig. 1). Regional and temporal patterns of expression of different rabbit ASIP isoforms Among twenty cDNA clones analysed, the majority (18 clones) exhibited the same 5′-UTR sequence (GenBank accession no. To confirm that the rabbit 1A and 1C ASIP isoforms were expressed FN547132) whereas two clones each had a different 5′-UTR sequence differentially in different skin regions, RNA extracted from dorsal and (GenBank accession nos. FN547133 and FN547134, respectively). ventral skin of 1-, 2-, 5- and 30-day old Castor Rex rabbits were These three isoforms represented different sites of transcription subjected to reverse transcription (RT) and then PCR-amplified using initiation (Fig. 2). None of these 5′-UTRs contained an ATG predicted primers specific for the dorsal (1C) and ventral (1A) isoforms (Fig. 3). to function as translational initiation site. The most frequent ventral The 1C transcript was detected both dorsally and ventrally in all transcript was designated 1A according to the high identity (BLASTN animals regardless of their age. However, the amount of this isoform e-value: 4e-26; identities=101/120 bp, 84%; gaps, 4/120 3%) shown was higher in dorsal skin than in ventral skin as deduced by the band to its counterpart 1A in mouse (GenBank accession no. L76475) (see intensity comparison between the two skin regions (Fig. 3). The 1A below). Of the other two minor isoforms (designated 1dv, dorsal- transcript was expressed specifically in ventral skin, but not in dorsal ventral; and 2Long), the sequence of the 2Long 5′-UTR was identified skin, at the level of sensitivity allowed by RT-PCR (Fig. 3). Both 1C and just upstream to and contiguous with exon 2 (hence the name we 1A transcripts were not detected in brain, kidney, heart, fat, muscle, attributed to it); it showed a perfect match to the upstream genomic liver and spleen of 1, 3, 6 and 30 days old rabbits (data not shown). sequence of GenBank accession no. AM748788, obtained during the resequencing analysis. Both minor isoforms were expressed in all The rabbit ASIP gene is expressed in every studied tissues tissues analysed in this study as evidenced by nested PCR amplification (data not show), and for this reason were not further investigated. We Tissue distribution of rabbit ASIP mRNA was examined by RT-PCR could not identify a second ventral specific1A′ transcript as reported in using gene specific primers to amplify the entire coding region and the L. Fontanesi et al. / Genomics 95 (2010) 166–175 169

Fig. 3. Expression of different rabbit ASIP exons in the ventrum and dorsum of 1-, 2-, 5-, and 30-day-old Castor Rex rabbits. (A) Amplification obtained with a forward primer designed on the 1C exon. (B) Amplification obtained with a forward primer designed on the 1A exon. (C) Amplification obtained from a primer designed on exon 2. (D) Amplification of a 77 -bp fragment of the GAPDH cDNA used as reference. V=ventrum; D=dorsum; M=Molecular weight marker (1 kb Plus Ladder, Invitrogen).

3′-UTR. A single fragment of 595 bp was amplified from brain, fat, heart, (c.163TNA) causing the p.L55M amino acid change located in a kidney, liver, lung, skeletal muscle, and spleen in 1, 3, 6 and 30 days old conserved position of the basic amino-terminal domain of the protein rabbits (Supplementary Fig. 3). PCR fragments were purified and ASIP (Fig. 2 and Supplementary Fig. 2). We identified 4 SNPs in the sequenced to verify that they contained the expected rabbit ASIP coding region of exon 4 of which two were missense mutations sequence. These results showed that the rabbit ASIP gene is expressed (c.230ANG determining the p.K77R amino acid substitution; and in all tissues examined as already reported in human and cattle [21–23]. c.266TNC causing the p.L89P amino acid change) (Fig. 2 and However these data differentiate the rabbit ASIP gene expression from Supplementary Fig. 2) in the Belgian Hare and Tan (Supplementary the wild type agouti mice gene expression [3,20]. Fig. 1) rabbits sequenced herein. These two amino acid substitutions were located in the basic amino-terminal domain and in the 3 proline Identification of polymorphisms in the rabbit ASIP gene region (for which an addition of one proline changes it to a 4 proline region) of the ASIP protein, respectively (Supplementary Fig. 2). The Resequencing of the three coding exons, including parts of two other exon 4 SNPs were synonymous substitutions (c.234GNT and upstream and downstream sequences, was carried out using a panel c.252GNA). Additionally, three SNPs were identified in the 3′-UTR of genomic DNA from 16 rabbits belonging to 10 different breeds with (c.⁎14ANG, c.⁎23GNC and c.⁎41CNT). different coat colour (see Materials and methods). On the whole we The likelihood that the three non-synonymous (amino-acid identified one insertion of 1 bp and 18 single nucleotide polymorph- changing) coding SNPs could cause a putative functional impact on isms (SNPs) (Fig. 2). The insertion (c.5_6insA) was localized in exon 2 the protein was evaluated using the cSNP analysis tool in PANTHER just 2 bp after the ATG starting codon, causing a frameshift of the (Protein ANalysis THrough Evolutionary Relationships). This classifi- reading frame resulting in a premature stop codon after 57 bp cation system calculates subPSEC (substitution position-specific

(Supplementary Fig. 4). Animals homozygous for this insertion should evolutionary conservation) and probability (Pdeleterious) scores based not have a functional ASIP protein determining the nonagouti black on an alignment of evolutionarily related proteins [41,42] (see coat colour phenotype in rabbits (see below). Of the 18 SNPs, two Materials and methods). The subPSEC and Pdeleterious values for the were located 5′ to the ATG starting codon (c.-10-34CNT and c.-10- three missense mutations were: p.L55M, −3.15426 and 0.53849; p. 31CNT), two synonymous substitutions were detected in exon 2 K77R, −2.34498 and 0.34186; p.L89P, −1.97871 and 0.26478. We (c.126GNA and c.147GNA), three SNPs were detected in intron 2 conclude therefore that these amino acid substitutions may not have (c.161-223TNA, c.161-192GNA and c.161-70CNT), and three were functional impacts. detected in intron 3 (c.225+5ANG, c.226-92GNA and c.226-63GNA). The 19 polymorphisms (18 SNPs and 1 indel) were organized into One non-synonymous substitution, identified only in the Tan five haplotypes presented in the Median-Joining network of Supple- sequenced rabbit (Supplementary Fig. 1), was located in exon 3 mentary Fig. 5. Two haplotypes (haplotypes 1 and 2), that differed by 170 L. Fontanesi et al. / Genomics 95 (2010) 166–175 only one SNP (c.234GNT), were identified in the Giant Grey rabbits Association between the c.5_6insA mutation and the nonagouti black (Supplementary Fig. 1), one of which (haplotype 1) was also observed coat colour in several other breeds (Burgundy Fawn, Californian, Checkered Giant, Loop and Rhinelander; Supplementary Fig. 1). Another haplotype Among the 19 polymorphisms identified in our study only the (haplotype 3), differing from haplotype 1 by the presence of the c.5_6insA mutation (which causes a frameshift and a predicted c.5_6insA insertion, was identified in Blue Vienna (homozygous; truncated protein of only 21 amino acids; Supplementary Fig. 4) Supplementary Fig. 1) and in Checkered Giant and Californian seems of functional significance. To confirm that this insertion is the (heterozygous) rabbits. The remaining two haplotypes (haplotypes causative mutation of the recessive nonagouti black coat colour in 4 and 5) were very distant from the first ones (Supplementary Fig. 5) rabbit, we developed a PCR-RFLP protocol (Supplementary Fig. 7) and were identified in Belgian Hare and Tan rabbits (Supplementary allowing the genotyping of this polymorphic site in 407 rabbits (of 31 Fig. 1), respectively. breeds) that have different coat colour (Table 1). All animals of breeds Identification of the putative ancestral haplotype can be deduced with black coat colour or variation of black, that also carried a wild from the Median-Joining network by analysing sequences of exon 2 type MC1R allele [31], for example, blue or silver (Alaska, Blanc de and part of the contiguous intronic regions in different rabbits and in Hotot, Blue Vienna, Champagne d'Argent, English Spot, Havana, Mini several other leporid species (see below; Supplementary Fig. 6). The Silver, Russian and Silver; Supplementary Fig. 1), were homozygous two Giant Grey haplotypes could not be separated from this analysis for the c.5_6insA mutation. For most of these breeds, classical genetic because the polymorphic site that distinguishes these two sequences studies have indicated that their black coat colour was due to the is located in exon 4 (Fig. 2), a region that was not included in this homozygous a/a genotype at the agouti locus [35,36,38]. Other phylogenetic analysis. However, considering the widespread distri- breeds with black coat colour but carrying (fixed or not fixed) the bution of haplotype 1 among several breeds, it seems likely that this putative ED-S allele at the extension locus (MC1R c.280_285del6 allele haplotype is the ancestral (wild type) haplotype in rabbit. [31]), were fixed for or had high frequency of the ASIP c.5_6insA

Table 1 Breeds and animals genotyped for the ASIP c.5_6insA (nonagouti mutation, a allele) and for the MC1R c.280_285del6 (allele ED or ES) and c.304_333del30 (allele e) mutations previously reported [31].

Breeds (no. of animals) Coat colour and [proposed agouti and extension genotype]a MC1R genotypeb ASIP genotypec

a/a A/a A/A

Alaska (10) self black [a/a, E/-] E/E 10 –– Belgian Hare (5) reddish laced with black [A/-, E/-] E/E ––5 Blanc de Hotot (3) white with black markings [a/a, E/-] E/E 3 –– Blue Vienna (56) dark blue [a/a, E/-] E/E 56 –– Burgundy Fawn (15) fawn [A/-, e/e] Δ30/Δ30 ––15 Bristle White (3) bristle white [?, ?] E/E ––3 Californian (56) white with black markings [a/a,?] Δ6/Δ6 38 15 3 Champagne d'Argent (42) silver as surface colour and dark blue as under-colour [a/a, E/-] E/E 42 –– Checkered Giant (41) white with black (36) or blue (5⁎) markings [a/a, ED/-] Δ6/Δ6 33 (4⁎) 8(1⁎) – Checkered Small (7) white with black (6) or blue (1⁎) markings [a/a, ED/-] Δ6/Δ6 6(1⁎)1 – Coloured Dwarf (6) bristle white (2) [?, ?]; hare-grey (2) [A/-, E/-]; Havana (1)⁎ [a/a, E/-]; chinchilla E/E 1⁎–5 (1) [A/-, E/E] Dutch (11) with black (10) or blue (1) markings [a/a, E/- or ED/- or ES/-] E/E (5),E/Δ6 (3), Δ6/Δ6 (3) 11 –– English Spot (16) white with black (15) or blue (1) markings (5) [a/a, E/-] E/E 16 –– Fairy Marburg (3) grey-light blue [?, E/-] E/E 3 –– Fairy Pearly (5) pearling grey [?, E/-] E/E – 14 Giant Chinchilla (20) chinchilla [A/-, E/E] E/E ––20 Giant Grey (16) wild-grey [A/-, E/E] E/E – 412 Giant White (5) white albino [?, ?] Δ6/Δ6 ––5 Havana (4) dark brown [a/a, E/E or ED/E] E/E 4 –– Lop (7) wild-grey (6) [A/-, E/-]; with Madagascar markings (1)⁎ [?, e/e] E/E, Δ30/Δ30⁎ 1⁎ 24 Lop Dwarf (6) wild-grey (2)⁎ [A/-, E/-]; self black (1) [a/a, E/-]; shaded yellow/brown (2)⁎⁎ Δ30/Δ30⁎⁎, Δ6/Δ6, E/E 41⁎ 1⁎ [a/a, e/e]; white with black markings (1) [a/a, ED/-] Mini Silver (2) black with silvering [a/a, E/-] E/E 2 –– New Zealand Red (1) red [A/-, e/e] Δ30/Δ30 ––1 New Zealand White (26) white-albino [?, ?] Δ6/Δ6 3914 Rex (2) black Dalmatian (1) [a/a, E/-]; martens (1) [?, ?] E/E 2 –– Rhinelanderd (8) white with black and yellow markings [a/a, E/-] E/E 242 Russian (2) white with black markings [a/a, E/-] E/E 2 –– Silver (11) black with silvering [a/a, E/-] E/E 11 –– Tan (4) black ﬁre [at/at, E/-] E/E ––4 Thuringian (5) shaded yellow/brown [a/a, e/e] Δ30/Δ30 5 –– White Vienna (9) white-blue eyes [?, E/-] E/E 71– Total (407)

a In breeds for which animals of different coat colour have been sampled, the number of the rabbits showing the distinctive phenotypes is indicated with circled brackets. Squared brackets contains the proposed genotype at the agouti and extension loci, when reported or deduced from the literature [48,49]. Pictures of rabbits of several breeds are reported in Supplementary Fig. 1. b Genotypes obtained for the MC1R gene [31]. Alleles c.304_333del30 (allele e) and c.280_285del6 (allele ED or ES) have been indicated as Δ30 and Δ6, respectively. E=wild type allele according to the genotyping protocol used. The number of the animals showing the different genotypes is reported. Two asterisks (⁎⁎) have been included to link the genotyped animals to their coat colour description when more phenotypes were sampled for a breed. c Genotypes obtained by PCR-RFLP for the ASIP gene (c.5_6insA mutation): a=nonagouti allele (carriers of the insertion); A=wild type allele (without insertion). The number of the animals showing the different genotypes is reported. An asterisk (⁎) has been included to link the genotyped animals to their coat colour description when more phenotypes were sampled for a breed. d The Rhinelander coat colour phenotype is determined by the English spotting locus and by the eJ allele at the extension locus. This extension allele was not characterized by Fontanesi et al. [31], therefore, here all animals of this breed were considered to carry the wild type E allele. L. Fontanesi et al. / Genomics 95 (2010) 166–175 171 mutation (Californian, Checkered Giant, Checkered Small and Dutch; different intensities of black and black-agouti coat colour phenotypes Supplementary Fig. 1). The ASIP c.5_6insA mutation was present in in these animals. non-black rabbits of other breeds (i.e.: New Zealand White, Thuringian and White Vienna) for which mutations in other loci Phylogenetic relationships among the Leporidae (albino, extension, and viennese white, respectively) determine their coat colour [31,35,36,38,43]. In addition, a few Giant Grey rabbits that In addition to Oryctolagus (see above) partial sequences of the ASIP were heterozygous for the mutated allele had the typical wild type gene (exon 2 and partial intronic regions) were obtained for 9 species grey coat colour indicating dominance of the wild type ASIP allele over of the Leporidae.TheseincludedBunolagus monticularis, Lepus the c.5_6insA allele. This explains the appearance of unexpected black americanus, Lepus europaeus, Lepus othus, Lepus timidus, Pentalagus puppies in some Giant Grey crosses in fancy breeder stocks furnessi, Romerolagus diazi, Sylvilagus floridanus,andSylvilagus (Fontanesi, personal observation). nuttallii. Two synonymous SNPs (c.48TNC and c.121TNC) were Dominance of the wild type grey coat colour over the nonagouti identified in P. furnessi, one SNP in intron 2 was observed in S. black phenotype was confirmed in a cross between a Blue Vienna buck floridanus c.160+29CNG) and one SNP in the same gene region was (wild type at the extension/MC1R locus and homozygous for the identified in L. othus (c.160+19GNA). Sequences obtained in the L. c.5_6insA ASIP insertion) and a Giant Grey doe (wild type at the ex- americanus, L. europaeus, L. othus, and L. timidus were identical if we tension/MC1R and ASIP loci) for which all the 7 F1 rabbits were grey exclude the polymorphic site identified in L. othus. All these sequences (data not shown). Epistatic effects between the extension/MC1R and were deposited in GenBank database (GenBank accession nos. ASIP loci were evident in a cross between a Champagne d'Argent FN547139 to FN547150). buck (Supplementary Fig. 1; wild type at the extension/MC1R locus Phylogenetic relationships among these species are shown in Fig. 4 and homozygous for the c.5_6insA ASIP insertion) and a Thuringian and Supplementary Fig. 6. The only significant difference between our doe (Supplementary Fig. 1; homozygous for the MC1R data and more comprehensive phylogenies [44,45] concerns position c.304_333del30 allele causing the recessive red phenotype [31] and of Lepus which is grouped as a sister lineage to a clade comprising homozygous for the c.5_6insA ASIP insertion). The 6 F1 rabbits had Oryctolagus, Bunolagus and Pentalagus in the published works, but black coat colour phenotype due to the recessive homozygous which is outside of this assemblage in our study. This minor conflict in nonagouti allele since at least one allele at the extension locus was topologies most likely reflects the use of a single gene fragment, and of wild type (data not shown). Crossing a Checkered Giant buck the short sequence analysed in the present study. (Supplementary Fig. 1; homozygous for the MC1R c.280_285del6 allele [31] and heterozygous for the c.5_6insA ASIP insertion) and a Discussion Burgundy Fawn doe (Supplementary Fig. 1; homozygous for the MC1R c.304_333del30 allele [31] and homozygous for haplotype 1 of Since the rediscovery of the Mendel's laws, coat colour segregation the ASIP gene) resulted in 13 F1 rabbits in which there was a analysis in the domestic rabbit was used, together with experiments segregation of the c.5_6insA ASIP insertion. However, the heterozy- carried out in mice and other mammals, to verify theoretically based gous genotype for the c.5_6insA ASIP insertion allele (n.=7) was not models of inheritance [33,34]. However, despite the availability of a associated with different intensities of the black phenotype (evident large number of rabbit breeds and strains with different and specific in some F1 animals) when compared to the non-carriers (n.=6), as coat colours and patterns, few studies have attempted to use was hypothesized previously [31] although without supporting molecular approaches to characterize coat colour loci originally genotyping data for the ASIP gene. Therefore other loci might affect studied by classical crossbreeding experiments in this species [31,43].

Fig. 4. Neighbour-Joining tree based on mid-point rooting using partial sequences of the ASIP gene (exon 2 and part of intronic regions) from several leporid species. Nodal supports are based on 1000 bootstrap pseudo-replicates. Four Oryctolagus haplotypes (1, 3, 4 and 5) were included in the analysis (haplotype 2 was not included because it differs from haplotype 1 for a SNP located in exon 4, see text). Two different sequences were included for L. othus, P. furnessi and S. ﬂoridanus because SNPs determining two haplotypes were identiﬁed in each of these species. 172 L. Fontanesi et al. / Genomics 95 (2010) 166–175

We report, for the first time, the expression, structure and mutation like element, interferes with normal transcription and/or splicing of analysis of the rabbit ASIP gene. Its widespread expression in all the agouti mRNA in a/a mice [3].InsheeptheASIP gene is analysed tissues contrasts with that of the wild type mouse agouti characterized by a copy number variation determining the white gene which is limited only to skin [3,20]. A constitutive and ubiquitous and tan (AWt) agouti allele and white coat colour, whereas nonagouti expression of some mutated agouti alleles such as Ay (lethal yellow) black sheep having a single copy ASIP gene could carry a regulatory has been described in mice [3,4]. The Ay allele causes increased yellow mutation in an unidentified regulatory region, or deletions in exon 2 pigmentation of the fur, together with obesity, insulin resistance, or a missense mutation in exon 4 [15–17]. Interestingly, we recently premature infertility, increased body length, and tumour susceptibility identified a copy number variant in the goat ASIP gene where animals (i.e. [19]). Ubiquitous expression of a mutated ASIP transcript driven carrying a duplicated ASIP allele are not completely associated with by an ITCH gene promoter was observed in sheep carrying the white coat colour [49]. Black coat colour in goat was not completely dominant white and tan (AWt) agouti allele without any apparent associated with any other identified missense mutation of this gene negative effects, whereas no results have been reported for the sheep [49]. Similarly as observed in rabbit, loss-of-function mutations wild type allele expression pattern [15]. Widespread expression of causing the nonagouti recessive phenotype were reported in fox (a non-mutated ASIP transcripts has been reported in humans and cattle complete deletion of exon 2 [8]), rat (a 19 bp deletion in exon 2 [9]), without any deleterious effect [21–23]. Therefore, from the available cat (a 2 bp deletion in exon 2 [13]), sheep (a 5 -bp deletion in exon 2 data it seems that the restricted expression of the wild type ASIP gene and a missense mutation in exon 4, p.C126S [15,17]), horse (an 11 bp might be limited to the rodents in which agouti gene regulation deletion in exon 3 [10]) and dog (a missense mutation in exon 4, p. pathways might be involved in their characteristic energy homeostasis R96C [14])(Fig. 2). processes [46]. Based on this, we speculate that the ancestral A large number of rabbits belonging to 31 different breeds was expression pattern might be one of widespread expression as this is genotyped to evaluate breed associations with the c.5_6insA and present in different mammalian orders. This would mean that this nonagouti coat colour (Table 1). Rabbits with black coat colour gene should play important roles (even if at present not completely belonging to breeds that, according to classical crossbreeding studies, clear) in tissues other than skin, and in functions other than should have the a/a genotype at the agouti locus, were fixed (with pigmentation (i.e. [47]). Similarity in gene expression between few exceptions) for the c.5_6insA mutation. The widespread human and rabbit could be important in evaluating the role of ASIP diffusion of the mutated allele (a)acrossdifferentbreedssuggests and better defining its function in a non-rodent model. an early origin of the nonagouti phenotype in the domestic rabbit. It Some other peculiarities seem to differentiate the agouti gene seems that the black nonagouti coat colour in this species appeared expression between mouse and rabbit. The results we obtained before 1700 and, together with other coat colour mutations, tracked suggest that light-bellied agouti (wild type) pattern in rabbit uses the domestication process of the rabbit [50]. As a matter of fact, the only two major agouti isoforms (1A, ventral; and 1C, dorsal) in appearance in this period of coat colour mutants is considered to be contrast to the mouse in which five isoforms (1A, 1A′, and 1A1A′, the true hallmark of the rabbit domestication that occurred quite late ventral; 1B and 1C, dorsal hair growth cycle) have been reported [20]. compared to other livestock species [50]. The early origin (related to This difference could reflect (i) that only two alternative promoters the rabbit domestication time) of the nonagouti mutation is are functional in rabbit or (ii) that exons 1A′ and 1B are not present in confirmed by the fact that haplotype 3 (carrying the c.5_6insA this leporid species. Extensive sequencing of 5′-flanking region of the insertion) derived directly from the wild type haplotype (haplotype ASIP gene in rabbit would help clarifying this issue. Indeed, if we 1) (Supplementary Fig. 5) since only this insertion differentiates the assume structure conservation among species, promoters and two haplotypes. alternative exon 1 forms are placed a few tens of kb upstream to Two other divergent haplotypes were observed in Belgian Hare the starting codon of exon 2 [3,18,23]. It remains necessary to clarify and Tan rabbits. According to classical genetic experiments, Tan the role of the two non-region specific transcripts having different 5′- rabbits should carry the at allele at the agouti locus determining a UTR (1dv and 2Long) that seem characteristic of the rabbit, since reddish belly and black body. None of the identified missense BLASTN analysis did not reveal any significant identity with other mutations seem to have caused this allele, therefore it seems plausible sequences (data not shown). to suppose that the black and tan phenotype is determined by a According to the transcripts we identified the regulation of rabbit ASIP promoter/regulatory mutation (as it has been suggested for mouse gene may occur at the level of transcript production. This could reflect [3]) that has yet to be identified. Its appearance would be more recent different promoters regulating the ventral and dorsal transcripts as it has than that of the nonagouti mutation. been indicated in other species [20,23]. Transcript stability and/or Some of the leporid phylogenetic relationships identified from the transcript translation efficiency could be other regulating mechanisms, as analysis of ASIP sequences, specifically the placement of Oryctolagus each variant has a distinct 5′-UTR. In fact, striking difference in secondary (the focus of the present paper) as a close relative to Pentalagus, are structures (data not shown) and stability were predicted for the four identical to other reports that included full species representation for isolated ASIP transcripts (ΔG=−308.2 kcal, −322.8 kcal, −273.8 kcal, Leporidae, and which were based on sequence information from two and −300.7 kcal for the 1A, 1C, 1dv and 2Long mRNAs, respectively) mitochondrial and five nuclear genes [44,45]. Clearly, in spite of the using RNAstructure software [48]. discordant placement of the Lepus in our study compared to the Mutation analysis revealed one bp insertion (c.5_6insA) a few bp published topologies, the resolving power of the ASIP sequences is from the ATG starting codon of exon 2 in black nonagouti rabbits. This impressive for a single nuclear locus. As this gene is involved in mutation causes a frameshift determining a completely different determining an important trait (coat colour) affecting fitness in the protein of only 21 amino acids due to a premature stop codon in the wild, purifying selection might have played an important function in same exon. The recessive pattern of inheritance and epistatic shaping variation during evolution of the studied species (Fontanesi interactions with the extension/MC1R locus were confirmed in et al., manuscript in preparation). crossbreeding experiments. In conclusion, the characterization of the ASIP gene in O. cuniculus In other mammals, nonagouti mutations have been reported to provides comparative expression data that are useful for understand- completely eliminate the function or expression of the ASIP protein or ing the role of the ASIP gene in affecting coat colour as well as other transcripts. This causes a complete absence of an alternative switch physiological functions that this gene might have given its widespread between eumelanin and pheomelanin synthesis and results only in expression in several tissues. The identification of the ASIP nonagouti eumelanin production with black/dark coat colour. In mice an mutation in rabbit, and its distribution in several breeds, represents an insertion of about 11 kb in intron 1, including a 5.5 kb retrovirus- important step in evaluating the effects of genes and mutations on L. Fontanesi et al. / Genomics 95 (2010) 166–175 173 coat colour in a species for which few studies have identified turer's instructions, including specificprimersRag3andRag4 mutations affecting this phenotypic trait. (Supplementary Table S1) and genomic DNA extracted from 1-day- old Rex Castor rabbit as template. The 4590 -bp-long PCR product was Materials and methods cloned into the Topo XL vector (Invitrogen) and sequenced by primer walking with the sequencing protocol reported below. DNA/RNA extraction and cDNA synthesis Expression pattern of ASIP in rabbit We collected dorsal and ventral skin, brain, kidney, heart, fat, muscle, liver and spleen samples from four Castor Rex rabbits aged 1, 2, Expression analysis of the gene was performed on cDNAs obtained 5 and 30 old-days respectively. The fur of Castor Rex rabbits (Fig. 1) as described above using Rag2 (sense)/Rag3 (anti-sense) primers exhibits the coat colour pattern typically found in agouti Aw/Aw mice (Supplementary Table S1) and amplifying the entire coding region (dorsal agouti-banded hairs and pheomelanic ventral colour). The Aw and 3′-UTR of the ASIP gene. Rag1 and Rag5 sense primers allele has been considered to be the wild type agouti allele [39,40]. (Supplementary Table S1) were designed to investigate dorsal or Genomic DNA and total skin RNA were prepared from blood and from ventral specific transcript, respectively. The anti-sense Rag3 was used different tissue samples of each animal stored at −80 °C as described in both cases. These primer combinations were used to amplify cDNAs previously [22,30]. Total RNA was analysed with the Agilent 2100 obtained from all tissues and animals reported above. PCR was carried Bioanalyzer (Agilent Technologies, Santa Clara, CA) to verify the out in a 25 -μL reaction volume containing 10 pmol of each primer, quality of each extracts (data not shown). cDNA syntheses were 12.5 μL of 2× working concentration of PCR Master Mix (ABgene) and obtained from total RNA extracted from the listed tissues of the four 2 μL of DMSO with the same cycling conditions described above. PCR Castor Rex rabbits. Reverse Transcription was performed from 1 μgof products were cloned into the pCR2.1 vector (Invitrogen) and DNase treated RNA in a total volume of 20 μL, using 0.5 μg oligo (dT) sequenced. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) primer and SuperScript™ II RNase H- Reverse Transcriptase (Invitro- cDNA amplification was used as reference (Supplementary Table S1). gen, Carlsbad, CA), according to the manufacturer's instructions. The PCR conditions were as reported above excluding the use of DMSO. reaction was incubated for 50 min at 42 °C and then for 15 min at 70 °C. Cycling conditions for GAPDH were:92°Cfor2min;35cyclesof95°C for 30 s, 58 °C for 30 s, 72 °C for 40 s; 72 °C for 3 min. RACE experiments and full-length cDNA cloning Resequencing and identification of mutations We performed RACE experiments to isolate the rabbit full-length ASIP cDNA. This was done using 5 μg of RNA obtained from dorsal skin Three PCR primer pairs (Supplementary Table S1) were designed of 1-day-old Castor Rex rabbit and the SMART™ RACE cDNA from intronic or non-coding regions of the rabbit ASIP gene to amplify Amplification kit (Clontech-Takara Bio Europe, Saint-Germain-en- the coding regions of the three coding exons (exon 2, exon 3 and exon 4) Laye, France), following the manufacturer's instructions. Consensus using genomic DNA isolated from 16 rabbits belonging to 10 breeds: one Rag5′.1, Rag5′.2, Rag3′.1, and Rag3′.2 primers (Supplementary Table S1) Belgian Hare, three Blue Vienna, two Burgundy Fawn, two Californian, were designed based on coding sequence homology among human two Champagne d'Argent, one Checkered Giant (with black markings), [21,22], cattle [23] and mouse [3]. Nested PCR with consensus and two Giant Grey, one Loop (grey), one Rhinelander and one Tan. Exon 2 adaptor primers was used for 5′-UTR and 3′-UTR amplification as and parts of intron 1 and intron 2 were also amplified in other 9 leporid follows: Rag5′.1/UPM (Universal Primer Mix) and Rag5′.2/NUP (Nested species (one or two DNA samples for each species): B. monticularis Universal Primer) primer combinations were used for the first and (Riverine rabbit, origin: South Africa), L. americanus (Snowshoes hare, second amplification of 5′-UTR, respectively, and Rag3′.1/UPM and origin: North America), L. europaeus (Brown hare, origin: Alps), L. othus Rag3′.2/NUP primer sets were used for the first and second amplifica- (Alaskan hare, origin: Alaska), L. timidus (Mountain hare, origin: Alps), tion of 3′-UTR respectively (Supplementary Table S1). First and second P. furnessi (Amami rabbit, origin: Amami island, Japan), R. diazi (Volcano PCR amplifications were carried out in a 25 μL reaction volume rabbit, origin: near Mexico City), S. floridanus (Eastern cottontail, origin: containing 10 pmol of each primer and 12.5 μL of 2X working North-East America), and S. nuttallii (Mountain cottontail, origin: concentration PCR Master Mix (ABgene, Epson, UK) with the following North-West America). PCR was carried out using a PT-100 thermal cycling conditions: initial denaturation at 94 °C for 2 min followed by 35 cycler (MJ Research, Watertown, MA, USA) in a 20 -μL reaction volume cycles (96 °C for 30 s, 61 °C for 30 s, 72 °C for 1 min) and one cycle (72 °C containing ∼50 ng genomic DNA, 1 U DNA EuroTaq DNA polymerase for 3 min). PCR products were cloned into the pCR2.1 vector (EuroClone Ltd., Paington, Devon, UK), 1× Euro Taq PCR buffer (exon 2 (Invitrogen) and sequenced as described below. The 5′-and3′-end and exon 3) or 1X enhancer buffer (exon 4), 2.5 mM dNTPs, 10 pmol of sequences were determined and used to design specific primers Rag1 each primer and 2.0 mM of MgCl2 (exon 2 and exon 3) or 2.0 mM of (sense) and Rag2 (anti-sense) (Supplementary Table S1) to amplify full- MgSO4 (exon 4). PCR profilewasasfollows:5minat95°C;35 length cDNAs. The amplification was carried out as described above. In amplification cycles of 30 s at 95 °C, 30 s at 56 (exon 2 and exon 3) or this case nested PCR was not required. PCR products were cloned into 62 °C (exon 4), 30 s at 72 °C; 5 min at 72 °C. pCR2.1 vector (Invitrogen) and sequenced. A second RACE experiment The amplified fragments were sequenced after treatment with 1 μL similar to what described above was performed to obtain 5′-UTR of of ExoSAP-IT® (USB Corporation, Cleveland, Ohio, USA) for 15 min at ventral region from the same animal. This was done using specific 37 °C. Cycle sequencing of the treated PCR products was conducted primers (Supplementary Table S1) deduced from coding regions of full- using the same PCR primers and the Big Dye v3.1 kit (Applied cDNAs obtained as reported above. Rag5′.5/UPM and Rag5′.6/NUP Biosystems). Sequencing reactions, after a few purification steps using primer pairs were used for the first and second amplification of 5′-UTR, EDTA, Ethanol 100% and Ethanol 70%, were loaded on an ABI3100 respectively. Rag3′.7/UPM and Rag3′.8/NUP primer pairs were used for Avant capillary sequencer (Applied Biosystems). the first and second amplification of 3′-UTR, respectively. Sequences were edited and aligned with the help of the CodonCode Aligner software (CodonCode Corporation, Dedham, MA, USA). Long-range PCR Genotyping To obtain the rabbit ASIP genomic sequence (from the ATG start codon to the stop codon), the Expand Long-template PCR (Roche The insertion of 1 bp identified in the ASIP exon 2 (c.5_6insA) was Diagnostics, Basel, Switzerland) was used following the manufac- genotyped by PCR-RFLP in 407 European rabbits of different breeds 174 L. Fontanesi et al. / Genomics 95 (2010) 166–175

expressed in mice carrying the lethal yellow mutation, Genes Dev. 7 (1993) reported in Table 1 and in F1 and parental rabbits obtained from crosses 454–467. between a Vienna Blue buck and a Giant Grey doe (7 F1 rabbits), a [5] L.S. Robbins, J.H. Nadeau, K.R. Johnson, M.A. Kelly, L. Roselli-Rehfuss, E. Baack, K.G. Champagne d'Argent buck and a Thuringian doe (6 F1 rabbits), and a Mountjoy, R.D. Cone, Pigmentation phenotypes of variant Extension locus alleles result from point mutations that alter MSH receptor function, Cell 72 (1993) Checkered Giant buck and a Burgundy Fawn doe (13 F1 rabbits). For the fi 827–834. PCR-RFLP analysis, the ampli ed fragment of 95 or 96 bp was obtained [6] D. Lu, D. Willard, I.R. Patel, S. Kadwell, L. Overton, T. Kost, M. Luther, W. Chen, R.P. using the same exon 2 forward primer used for resequencing and a Woychik, W.O. Wilkison, R.D. Cone, Agouti protein is an antagonist of the mismatched reverse primer that creates an artificial restriction site for melanocyte-stimulating-hormone receptor, Nature 371 (1994) 799–802. EcoRI when the insertion of 1 bp is not present (Supplementary Table [7] M.M. Ollmann, M.L. Lamoreux, B.D. Wilson, G.S. Barsh, Interaction of Agouti protein with the melanocortin 1 receptor in vitro and in vivo, Genes Dev. 12 S1). PCR conditions were as described in the previous paragraph with (1998) 316–330. MgCl2 concentration of 2.0 mM and annealing temperature of 54 °C. [8] D.I. Våge, D. Lu, H. Klungland, S. Lien, S. Adalsteinsson, R.D. Cone, A non-epistatic fi interaction of agouti and extension in the fox, Vulpes vulpes, Nat. Genet. 15 (1997) Digestion of the ampli ed fragment was obtained overnight at 37 °C in a – μ μ 311 315. 25 L reaction volume including 5 L of PCR product, 1X restriction [9] T. Kuramoto, T. Nomato, T. Sugimura, T. Ushijima, Cloning of the rat agouti gene enzyme buffer and 2 U of EcoRI (Fermentas, Vilnius, Lithuania). The and identification of the rat nonagouti mutation, Mamm. Genome 12 (2001) resulting DNA fragments were electrophoresed on 10% 29:1 polyacryl- 469–471. [10] S. Rieder, S. Taourit, D. Mariat, B. Langlois, G. Guerin, Mutations in the agouti amide:bis acrylamide gel and visualized by ethidium bromide staining (ASIP), the extension (MC1R), and the brown (TYRP1) loci and their association to (Supplementary Fig. 7). All rabbits were also genotyped for the coat color phenotypes in horses (Equus caballus), Mamm. Genome 12 (2001) c.280_285del6 (alleles ED or ES) and c.304_333del30 (allele e)ofthe 450–455. [11] P.A. Kanetsky, J. Swoyer, S. Panossian, R. Holmes, D. Guerry, T.R. Rebbeck, A MC1R gene as previously reported [31,51]. polymorphism in the agouti signaling protein gene is associated with human pigmentation, Am. J. Hum. Genet. 70 (2002) 770–775. Sequence and phylogenetic analyses [12] R.J. Miltenberger, K. Wakumatsu, S. Ito, R.P. Woychik, L.B. Russell, E.J. Michaud, Molecular and phenotypic analysis of 25 recessive, homozygous-viable alleles at the mouse agouti locus, Genetics 160 (2002) 659–674. BLASTN algorithm at the NCBI GenBank database (http://www. [13] E. Eizirik, N. Yuhki, W.E. Johnson, M. Menotti-Raymond, S.S. Hannah, S.J. O'Brien, ncbi.nlm.nih.gov) was used to compare the different 5′-UTR regions of Molecular genetics and evolution of melanism in the cat family, Curr. Biol. 13 – the ASIP transcripts. In silico functional analysis of O. cuniculus (2003) 448 453. [14] J.A. Kerns, J. Newton, T.G. Berryere, E.M. Rubin, J.F. Cheng, S.M. Schmutz, G.S. Barsh, missense mutations was obtained using PANTHER [41]. PANTHER Characterization of the dog Agouti gene and a nonagouti mutation in German estimates the likelihood of a particular non-synonymous (amino-acid Shepherd dogs, Mamm. Genome 15 (2004) 798–808. changing) coding SNP to cause a functional impact on the protein. It [15] B.J. Norris, V.A. Whan, A gene duplication affecting expression of the ovine fi ASIP gene is responsible for white and black sheep, Genome Res. 18 (2008) calculates the substitution position-speci c evolutionary conserva- 1282–1293. tion (subPSEC) score based on an alignment of evolutionarily related [16] L.J. Royo, I. Alvarez, J.J. Arranz, I. Fernández, A.L. Rodríguez, Pérez-Pardal, F. proteins [41,42]. The probability that a given variant will cause a Goyache, Differences in the expression of the ASIP gene are involved in the recessive black coat colour pattern in sheep: evidence from the rare Xalda sheep deleterious effect on protein function is estimated by Pdeleterious, such breed, Anim. Genet. 39 (2008) 290–293. that a subPSEC score of −3 corresponds to a Pdeleterious of 0.5. The [17] J. Gratten, J.G. Pilkington, E.A. Brown, D. Beraldi, J.M. Pemberton, J. Slate, The subPSEC score is the negative logarithm of the probability ratio of the genetic basis of recessive self-colour pattern in a wild sheep population, Heredity (2009), doi:10.1038/hdy.2009.105. wild type and mutant amino acids at a particular position. PANTHER [18] M. Girardot, S. Guibert, M.-P. Laforet, Y. Gallard, H. Larroque, A. Oulmouden, The subPSEC scores are continuous values from 0 (neutral) to about −10 insertion of a full-length Bos taurus LINE element is responsible for a (most likely to be deleterious). Phylogenetic analysis of the exon 2 transcriptional deregulation of the Normanne Agouti gene, Pigment Cell Res. 19 (2006) 346–355. sequences obtained for the listed leporid species (Family Leporidae) [19] D.M. Duhl, H. Vrieling, K.A. Miller, G.L. Wolff, G.S. Barsh, Neomorphic agouti including the different O. cuniculus alleles was obtained with the mutations in obese yellow mice, Nat. Genet. 8 (1994) 59–65. Neighbour Joining tree option (with Kimura 2 parameter and 1000 [20] H. Vrieling, D.M. Duhl, S.E. Millar, K.A. Miller, G.S. Barsh, Differences in dorsal and bootstraps) of the MEGA4 software [52]. Median-joining networks ventral pigmentation result from regional expression of the mouse agouti gene, Proc. Natl. Acad. Sci. U.S.A. 91 (1994) 5667–5671. [53] for the rabbit ASIP haplotypes and for the sequences obtained for [21] H.Y. Kwon, S.J. Bultman, C. Löffler, W.-J. Chen, P.J. Furdon, J.G. Powell, A.-L. Usala, exon 2 and parts of the contiguous introns in several leporids were W. Wilkison, I. Hansmann, R.P. Woychik, Molecular structure and chromosomal constructed using Network v. 4.510 [54]. RNAstructure software v. 4.6 mapping of the human homolog of the agouti gene, Proc. Natl. Acad. Sci. U.S.A. 91 (1994) 9760–9764. [48] was used to predict secondary structures and stability of the [22] B.D. Wilson, M.M. Ollmann, L. Kang, M. Stoffel, G.I. Bell, G.S. Barsh, Structure and different ASIP transcripts using default options. function of ASP, the human homolog of the mouse agouti gene, Hum. Mol. Genet. 4 (1995) 223–230. [23] M. Girardot, J. Martin, S. Guibert, H. Leveziel, R. Julien, A. Oulmouden, Widespread Acknowledgments expression of the bovine Agouti gene results from at least three alternative promoters, Pigment Cell Res. 18 (2005) 34–41. We thank several rabbit breeders and ANCI for their help in [24] P. Valverde, E. Healy, I. Jackson, J.L. Rees, A.J. Thody, Variants of the melanocyte- stimulating hormone receptor gene are associated with red hair and fair skin in sampling. This work was supported by the University of Bologna RFO, humans, Nat. Genet. 11 (1995) 328–330. University of Limoges and INRA funds. [25] H. Klungland, D.I. Våge, L. Gomez-Raya, S. Adalsteinsson, S. Lien, The role of melanocyte-stimulating hormone (MSH) receptor in bovine coat color determi- nation, Mamm. Genome 6 (1995) 636–639. Appendix A. Supplementary data [26] L. Marklund, M. Johansson Moller, K. Sandberg, L. Andersson, A missense mutation in the gene for melanocyte-stimulating hormone receptor (MC1R)is Supplementary data associated with this article can be found, in associated with the chestnut coat color in horses, Mamm. Genome 7 (1996) – the online version, at doi:10.1016/j.ygeno.2009.11.003. 895 899. [27] J.M.H. Kijas, R. Wales, A. Törnsten, P. Chardon, M. Moller, L. Andersson, Melanocortin receptor 1 (MC1R) mutations and coat color in pigs, Genetics 150 References (1998) 1177–1185. [28] D.I. Våge, H. Klungland, D. Lu, R.D. Cone, Molecular and pharmacological [1] L. Montoliu, W.S. Oetting, D.C. Bennett, Color Genes. (2009, August). European characterization of dominant black coat color in sheep, Mamm. Genome 10 Society for Pigment Cell Research. World Wide Web (URL: http://www.espcr. (1999) 39–43. org/micemut). [29] J.M. Newton, A.L. Wilkie, L. He, S.A. Jordan, D.L. Metallinos, N.G. Holmes, I.J. [2] A.G. Searle, Comparative Genetics of Coat Colour in Mammals, Logos Press, Jackson, G.S. Barsh, Melanocortin 1 receptor variation in the domestic dog, London, 1968. Mamm. Genome 11 (2000) 24–30. [3] S.J. Bultman, E.J. Michaud, R.P. Woychik, Molecular characterization of the mouse [30] F. Rouzaud, J. Martin, P.F. Gallet, D. Delourme, V. Goulemot-Leger, Y. Amigues, F. Agouti locus, Cell 71 (1992) 1195–1204. Ménissier, H. Levéziel, R. Julien, A. Oulmouden, A first genotyping assay of French [4] M.W. Miller, D.M. Duhl, H. Vrieling, S.P. Cordes, M.M. Ollmann, B.M. Winkes, G.S. cattle breeds based on a new allele of the extension gene encoding the Barsh, Cloning of the mouse agouti gene predicts a secreted protein ubiquitously melanocortin-1 receptor (Mc1r), Genet. Sel. Evol. 32 (2000) 511–520. L. Fontanesi et al. / Genomics 95 (2010) 166–175 175

[31] L. Fontanesi, M. Tazzoli, F. Beretti, V. Russo, Mutations in the melanocortin 1 [43] B. Aigner, U. Besenfelder, M. Müller, G. Brem, Tyrosinase gene variants in different receptor (MC1R) gene are associated with coat colour in domestic rabbit rabbit strains, Mamm. Genome 11 (2000) 700–702. (Oryctolagus cuniculus), Anim. Genet. 37 (2006) 489–493. [44] C.A. Matthee, B.J. van Vuuren, D. Bell, T.J. Robinson, A molecular supermatrix of the [32] L. Fontanesi, F. Beretti, V. Riggio, S. Dall'Olio, E. Gómez González, R. Finocchiaro, rabbits and hares (Leporidae) allows for the identification of five intercontinental R. Davoli, V. Russo, B. Portolano, Missense and nonsense mutations in exchanges during the Miocene, Syst. Biol. 53 (2004) 433–447. melanocortin 1 receptor (MC1R) gene of different goat breeds: association [45] K.M. Halanych, T.J. Robinson, Multiple substitutions affect the phylogenetic utility with red and black coat colour phenotypes but with unexpected evidences, BMC of cytochrome b and 12S rDNA data: examining a rapid radiation in leporid Genet. 10 (2009) 47. (Lagomorpha) evolution, J. Mol. Evol. 48 (1999) 369–379. [33] W.E. Castle, Heredity of coat characters in guinea pigs and rabbits, Carnegie Inst. [46] K.L. Ellacott, R.D. Cone, The role of the central melanocortin system in the Washington Publ. 23 (1905) 1–78. regulation of food intake and energy homeostasis: lessons from mouse models, [34] R.C. Punnett, Inheritance of coat colour in rabbits, J. Genet. 2 (1912) 221–238. Philos. Trans. R. Soc. Lond. B Biol. Sci. 361 (2006) 1265–1274. [35] W.E. Castle, The Genetics of Domestic Rabbit, Cambridge Harvard University Press, [47] S.R. Smith, B. Gawronska-Kozak, L. Janderová, T. Nguyen, A. Murrell, J.M. Stephens, London, 1930. R.L. Mynatt, Agouti expression in human adipose tissue: functional consequences [36] R. Robinson, Genetic studies of the rabbit, Bibl. Genet. 17 (1958) 229–558. and increased expression in type 2 diabetes, Diabetes 52 (2003) 2914–2922. [37] P.B. Sawin, Linkage of “wide band” and agouti genes, J. Hered. 25 (1934) 477–481. [48] D.H. Mathews, M.D. Disney, J.L. Childs, S.J. Schroeder, M. Zuker, D.H. Turner, [38] R.R. Fox, Taxonomy and genetics, in: P.J. Manning, D.H. Ringler, C.E. Newcomer Incorporating chemical modification constraints into a dynamic programming (Eds.), The Biology of the Laboratory Rabbit, Second edition, Academic Press, San algorithm for prediction of RNA secondary structure, Proc. Natl. Acad. Sci. U.S.A. Diego, CA, 1994, pp. 1–26. 101 (2004) 7287–7292. [39] W.K. Silvers, The agouti and extension series of alleles, umbrous, and sable, in: [49] L. Fontanesi, F. Beretti, V. Riggio, E. Gómez Gonzáles, S. Dall'Olio, R. Davoli, V. W.K. Silver (Ed.), The Coat Colors of Mice: A Model for Mammalian Gene Action Russo, B. Portolano, Copy number variation and missense mutations of the agouti and Interaction, Springer-Verlag, New York, 1979, pp. 6–44. signaling protein (ASIP) gene in goat breeds with different coat colours, [40] S.J. Bultman, M.L. Klebig, E.J. Michaud, H.O. Sweet, M.T. Davisson, R.P. Woychik, Cytogenet. Genome Res. 52 (2009), doi:10.1159/000268089. Molecular analysis of reverse mutations from nonagouti (a) to black-and-tan (at) [50] R. Robinson, Rabbit, in: I.L. Mason (Ed.), Evolution of Domesticated Animals, and white-bellied agouti (Aw) reveals alternative forms of agouti transcripts, Longman, London, 1984, pp. 239–246. Genes Dev. 8 (1994) 481–490. [51] L. Fontanesi, M. Tazzoli, V. Russo, Non-invasive and simple methods for sampling [41] P.D. Thomas, M.J. Campbell, A. Kejariwal, H. Mi, B. Karlak, R. Daverman, K. Diemer, rabbit DNA for PCR analysis of melanocortin 1 receptor (MC1R) gene mutations: a A. Muruganujan, A. Narechania, PANTHER: a library of protein families and technical note, World Rabbit Sci 15 (2007) 121–126. subfamilies indexed by function, Genome Res. 13 (2003) 2129–2141. [52] K. Tamura, J. Dudley, M. Nei, S. Kumar, MEGA4: Molecular Evolutionary Genetics [42] P.D. Thomas, A. Kejariwal, N. Guo, H. Mi, M.J. Campbell, A. Muruganujan, B. Analysis (MEGA) software version 4.0, Mol. Biol. Evol. 24 (2007) 1596–1599. Lazareva-Ulitsky, Applications for protein sequence-function evolution data: [53] H.J. Bandelt, P. Forster, A. Röhl, Median-joining networks for inferring intraspecific mRNA/protein expression analysis and coding SNP scoring tools, Nucleic Acids phylogenies, Mol. Biol. Evol. 16 (1999) 37–48. Res. 34 (2006) W645–W650. [54] Fluxus-engineering; Network v. 4.510 [http://www.fluxus-engineering.com].