台風到来の日曜日、ショー終了後、八重洲南地下街のキリンシティー、帰りのバスの

 予約していた時間までに少し時間があったので遺伝学の専門家とブリティッシュのコートの話になりました。

 ブリティッシュブルーとブリティッシュの他のカラーの場合、コートに変化が生じる、

 その理由を知りたかった。また、ノルウェージャンのコートで、クラシックとマッカレルとでは

 コートの長さが違う、他にもカラーによってコートの質が違う猫種は多い。

 イエネコの長毛という形質は，元々，突然変異によりたまたま生じた長毛の猫を固定したものと考えられている。

という説明から始まって、あの美しいエレガントなペルシャが、元をたどればブリティッシュにあること
そして、それはどのような遺伝子の作用によるものなのか？から興味深い話が始まりました。

この遺伝子の変異の種類は全部で4種類(M1～M4)あり，すべての長毛種は共通にM4の変異をもつことがわかった。

さらに，短毛種のマンクスとブリティッシュショートヘアーと，ブリティッシュショートヘアーとアウトクロスが認められているスコティッシュフォールド・デボンレックスも一定の頻度でM4の変異をもつ。

 外として，M4の変異以外に，メインクーンはM3，ノルウェージャンはM2，ラグドールはM1とM3の変異を一定の頻度でもつ。


ということで、FGF5遺伝子の存在が大きな力を有していることが判明。でも、MC,NF,RDなどの猫種はまた独自の突然変異を遂げたものであるらしい。
 

 コートカラーやパターンによるタイプの違い、気にかかりながら専門家に詳し教えていただく機会も持てずにいまいしたが、幸い今回、詳しく説明していただける専門家のご協力を得て、大事な知識の充実を図っていきたいと思いました。

(専門的な表になりますが、一部公開し、また添付もいたします。もし、詳しくお勉強したい方がいらしたらお申し出に応じてスクールを開催いたします。）  

表3　FGF5遺伝子の変異の頻度

                        　　　　　　　　　　　変異の種類
　　　　　　　　　　　　　　　　　　 M1　　M2　　M3　　M4　　変異なし
長毛種
　アンゴラ                                               1.00
　バリネーズ                                             1.00
　バーマン                                               1.00
　ヒマラヤン                                             1.00
　メインクーン                                    0.50    0.50
　ノルウェージャン                          0.88          0.12
　ペルシャ                                               1.00
　ラグドール                         0.27          0.23    0.50
　サイベリアン                                           1.00
　ソマリ                                                 1.00
　ターキッシュアンゴラ                                   1.00
　ターキッシュバン                                       1.00
　マンクス・LH                                           1.00   0.00
  マンクス・SH                                           0.50   0.50
　スコティッシュフォールド・LH                           1.00   0.00
　スコティッシュフォールド・SH                           0.33   0.67

短毛種
　アビシニアン                                                 1.00
　アメリカンショートヘアー                                     1.00
　ブリティッシュショートヘアー                           0.20   0.80
　バーミーズ                                                   1.00
　シャルトリュー                                               1.00
　コーニッシュレックス                                         1.00
　デボンレックス                                         0.33   0.67
　エジプシャンマウ                                             1.00
　ハバナ                                                       1.00
　オシキャット                                                 1.00
　ロシアンブルー                                               1.00
　サイアミーズ                                                 1.00


なお、詳しい記事は添付した、以下の「アメリカンジェネティックアソシエーション２００７」を参照してください。


The American Genetic Association. 2007. All rights reserved.
For permissions, please email: journals.permissions@oxfordjourna

Journal of Heredity 2007:98(6):555–566 doi:10.1093/jhered/esm072 Advance Access publication September 1, 2007

The American Genetic Association. 2007. All rights reserved. For permissions, please email: journals.permissions@oxfordjournals.org.
Four Independent Mutations in the Feline Fibroblast Growth Factor 5 Gene Determine the Long-Haired Phenotype in Domestic Cats
JAMES S. KEHLER,VICTOR A. DAVID,ALEJANDRO A. SCHA ¨ FFER,KRISTINA BAJEMA,EDUARDO EIZIRIK, DAVID K. RYUGO,STEVEN S. HANNAH,STEPHEN J. O’BRIEN, AND MARILYN MENOTTI-RAYMOND
From the Laboratory of Genomic Diversity, National Cancer Institute-Frederick, Frederick, MD 21702 (Kehler, David, Bajema, O’Brien, and Menotti-Raymond); Center for Animal Transgenesis and Germ Cell Research, Room 102, Myrin Building, School of Veterinary Medicine, University of Pennsylvania, Kennett Square, PA 19348 (Kehler); Department of Health and Human Services, National Center for Biotechnology Information, National Institutes of Health, Bethesda, MD 20894 (Scha ¨ffer); Centro de Biologia Geno ˆmica e Molecular, Faculdade de Biocie ˆncias, Pontifı ´cia Universidade Cato ´lica do Rio Grande do Sul, Porto Alegre, RS 90619-900, Brazil (Eizirik); Department of Otolaryngology and Neuroscience, 720 Rutland Avenue, Johns Hopkins University, Baltimore, MD 21205 (Ryugo); and Nestle ´-Purina PetCare Company, St. Louis, MO 63134 (Hannah).
Address correspondence to J. S. Kehler at the address above, or e-mail: jkehler@ncifcrf.gov.
Abstract
To determine the genetic regulation of "hair length" in the domestic cat, a whole-genome scan was performed in a multigenerational pedigree in which the "long-haired" phenotype was segregating. The 2 markers that demonstrated the greatest linkage to the long-haired trait (log of the odds
6) flanked an estimated 10-Mb region on cat chromosome B1 containing the Fibroblast Growth Factor 5 (FGF5) gene, a candidate gene implicated in regulating hair follicle growth cycle in other species. Sequence analyses of FGF5 in 26 cat breeds and 2 pedigrees of nonbreed cats revealed 4 separate mutations predicted to disrupt the biological activity of the FGF5 protein. Pedigree analyses demonstrated that different combinations of paired mutant FGF5 alleles segregated with the long-haired phenotype in an autosomal recessive manner. Association analyses of more than 380 genotyped breed and nonbreed cats were consistent with mutations in the FGF5 gene causing the long-haired phenotype in an autosomal recessive manner. In combination, these genomic approaches demonstrated that FGF5 is the major genetic determinant of hair length in the domestic cat.
The hair follicle provides a unique to mammalian model in which to study the complex genetic regulation between stem and stromal cells during self-renewal and terminal differentiation of a tissue. Genetic modulation of the hair follicle cycle can affect hair length, providing a rapid means for significant phenotypic change under either artificial or natural selection. The large variety of cat breeds with different hair textures and lengths provides a potential wealth of mammalian models with spontaneous mutations at unknown loci affecting hair follicle structure and function (Vella and Robinson 1999). Discovery and comparison of mutations in orthologous genes between mammals can provide additional understanding about the conserved domains that are required for protein function.
The initial molecular studies of the "long hair" locus were done in mice. Breeding experiments of spontaneously occurring "long-haired" Angora mice demonstrated that the go locus was the major determinant of "hair length" in the mouse and that the long-haired phenotype was inherited in an autosomal recessive manner (Dickie 1963; Pennycuik and Raphael 1984). Subsequent targeted mutation of the fibroblast growth factor 5 (Fgf5) gene and crossbreeding experiments between Fgf5 knockout (Fgf5neo) and Angora mice demonstrated that Fgf5neo and go represented null alleles of the same locus (He´bert et al. 1994). FGF5 was originally identified as a human oncogene (Zhan et al. 1987), belonging to a family of 23 related FGF genes (For a review, see Katoh 2002; Katoh and Katoh
555
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
2005). FGF5 shares a b-trefoil superfold structure composed of 5 b-hairpin folds (that is required for receptor binding) and an N-terminal, signal peptide (that is required for paracrine secretion) with the prototypical members, FGF1 and 2 (Ornitz and Itoh 2001). During embryonic and fetal development of the mouse, Fgf5 is first expressed in the extraembryonic ectoderm of the epiblast and then restricted to differentiating myotomes, skeletal muscles, and neurons (Haub and Goldfarb 1991; He´bert et al. 1991). The orthologous human FGF5 gene product is overexpressed in some mammary, prostatic, and renal carcinomas and can be presented on their major histocompatibility complex (MHC) class I receptors after intracellular processing, providing a potential antigenic target for cancer immunotherapy (Vigneron et al. 2004). However, experimental overexpression of FGF5 has been used to induce angiogenesis in the myocardium and to promote photoreceptor survival in animal models without inducing tumors in situ (Giordano et al. 1996; Green et al. 2001). In the adult mouse, FGF5 is normally expressed in neurons in most regions of the brain, in pancreatic b-islet cells involved in glucose-homeostasis, and in hair follicles in the skin (Haub et al. 1990; Go´mezPinilla and Cotman 1993; Hart et al. 2000). During the normal, cyclic hair growth of adult mice, expression of Fgf5 is restricted to cells in the lower third of the outer root sheath (ORS) and in the inner root sheath at the base of the follicle during the anagen phase just prior to progression into the catagen phase (He´bert et al. 1994). Two major Fgf5 isoforms have been described in the mouse (Suzuki et al. 2000); the full-length Fgf5 mRNA results from transcription of all 3 exons, while splicing of exon 1 to an alternative acceptor site in exon 3 introduces a frameshift producing a shorter transcript (Fgf5S) that lacks exon 2 and most of exon 3. Both the FGF5 and FGF5S gene products bind primarily to the FGF receptors 1 and 2, Isoforms c (FGFR1c and 2c) (Ornitz et al. 1996) found on dermal papilla cells (DPCs) (Clements et al. 1993; Rosenquist and Martin 1996). However, the full-length FGF5 actively inhibits DPC stimulation of ORS cell proliferation and synthesis of hair fibers during anagen (Limat et al. 1993; Suzuki et al. 2000), thereby triggering catagen. In contrast, FGF5S antagonizes the inhibitory effects of FGF5, providing a degree of autoregulatory modulation (Ota et al. 2002). The overall effect of the absence of both isoforms in Fgf5-null mice is that the anagen stage is prolonged, resulting in longer hair (He´bert et al. 1994). However, even in these mice, the hair follicles do eventually proceed into the catagen phase, indicating the presence of additional signaling pathways (reviewed by Paus and Foitzik 2004). Comparative studies of Fgf5 orthologs in other mammalian species exhibiting variation in hair length may reveal novel spontaneous mutations identifying key domains that influence protein function. The inheritance of long hair has been documented as a recessive trait in other mammalian species including rabbits (Fraser 1953), dogs (Burns and Fraser 1963), and cats (Vella and Robinson 1999). While we were pursuing our study of the involvement of the FGF5 locus in the control of hair length in cats, another group demonstrated that
mutations in the canine FGF5 gene explained the longhaired phenotype observed in multiple dog breeds (Housley and Venta 2006). Similar to dogs, the foundation of cat breed groups committed to the propagation of specific phenotypic traits resulting from spontaneous mutations has provided opportunities to study the genetic basis of several traits. Coat lengths and colors can be specified as part of a breed standard allowing for mutations carried by founders to become fixed within these closed breeding populations. The large variety of breeds containing potential mutations affecting hair follicle structure and function from the hairless Sphynx to the kinked-haired Cornish, Devon, and Selkirk Rexes to the long-haired breeds such as Maine Coon, Persian, and Turkish Angora cats provides a wealth of animal models for understanding the complex gene regulation of the hair growth cycle. In this study, we have performed a genetic survey of "long-" and "short-haired" breeds, as well as genetic linkage and pedigree analyses of nonbreed cats to interrogate the influence of the FGF5 locus on hair length in domestic cats.
Materials and Methods Collection of DNA Samples and Phenotypic Data from Breed and Nonbreed Cats
Whole blood preserved with ethylenediaminetetraacetic acid (EDTA) or cheek swab samples were taken from cats for DNA extraction using a salt-precipitation/column-filtration kit (Qiagen, Valencia, CA). Samples were obtained from 119 unrelated cats with the permission of private owners, from 12 "short-haired" breeds (Abyssinian, American Shorthair, British Shorthair, Burmese, Chartreux, Cornish Rex, Devon Rex, Egyptian Mau, Havana, Ocicat, Russian Blue, and Siamese), 12 long-haired breeds (Angora, Balinese, Birman, Himalayan, Maine Coon, Norwegian Forest Cat, Persian, Ragdoll, Siberian, Somali, Turkish Angora, and Turkish Van), and2 breeds (ManxandScottish Fold)thatcurrentlymaintain separate short- and long-haired registries within the Cat Fanciers Association (CFA). Private owners reported the phenotypes of their cats as either short- or long-haired in accordancewithbreedstandardsandprovidedphotographsin most cases. All breed cats were assigned an anonymous registry (Fc) number, and phenotypic data were recorded in adatabaseattheLaboratoryofGenomicDiversitytopreserve the anonymity of individual cats and their owners. Sampleswereobtainedfrom261relatedcatsthatwereused for nutrition studies at the Nestle´-Purina PetCare Company in St. Louis, MO, and 50 related White Deaf cats that were used for hearing studies at Johns Hopkins University in Baltimore, MD,withtheapprovaloftheirrespectiveInternalAnimalCare and Use Committees. Both colonies were established from random-bred, short- and long-haired cats obtained from commercial vendors and were maintained as closed colonies through planned matings. Because these cats were no longer being randomly mated or out or crossbred, they were defined in this study as "nonbreed" cats to distinguish them from the registered breed cats. As in breed cats, the hair length was
556
Journal of Heredity 2007:98(6)
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
sufficiently different to qualitatively phenotype nonbreed cats as either short or long haired. Photographs were used to confirm the reported phenotypes, but direct measurements of hairlengthwerenottaken.Allnonbreedcatsinthisstudywere maintained in facilities inspected by the US Department of Agriculture, under conditions established by the American AssociationofLaboratoryAnimalCareincompliancewiththe federal Animal Welfare Act.
Whole-Genome Scan
The long-haired trait segregated in the Nestle´-Purina, multigenerational pedigree of nonbreed domestic cats (Eizirik et al. 2003). In conjunction with the development of a third generation, genetic linkage map, we genotyped 261 cats in this pedigree for 483 autosomal and 30 X-linked (513 total) microsatellite markers, using amplification conditions and analyses as previously described (Ishida et al. 2006).
Genetic Linkage Analysis
The long-haired phenotype was modeled as a binary trait with fully penetrant, autosomal recessive inheritance, and log of the odds (LOD) scores were computed using the Superlink software package (Fishelson and Geiger 2002, 2004). For the LOD scores shown in Table 1, the allele associated with the long-haired phenotype was assumed to have a frequency of 0.25 with all other markers of equal allele frequencies. If the trait allele frequency of 0.50 was used, peak LOD scores of all the linked markers were higher by at most 0.6 LOD units, and estimates of the optimal recombination fraction (h) differed by at most 0.01.
PCR Amplification and DNA Sequence Analysis of FGF5
PCR primers were designed in the introns flanking the 3 exons of FGF5 (Supplementary Figure A1 and Table A1). Primers flanking exon 1 were designed from domestic cat
DNA sequence that was detected in the cat 2 whole genome by cross-species MegaBlast (Zhang et al. 2000) (http://www.ncbi.nlm.nih.gov/BLAST/tracemb.shtml). The forward primer for exon 2 was anchored in a region identical in both dog and cow, and the reverse primer was a consensus sequence between dog and cow and contained 2 degenerate oligonucleotide positions. The primers for exon 3 were derived from the dog DNA sequence from a region highly conserved between dog and human. Touchdown polymerase chain reaction (PCR) and DNA sequencing were performed as previously described (Guo et al. 2006; Ishida et al. 2006).
PCR–restriction fragment length polymorphism Genotyping Assay Development
Three fluorescence-based assays were developed to genotype the 4 FGF5 mutations detected (Supplementary Table A2). SeparateassaysweredevelopedtodetectMutation1(c.ins356T) and 2 (c.C.T406), and a third assaywas designed which could detect both the adjacent Mutations 3 (c.del474T) and 4 (c.A.C475),aswellasdeterminethephaseofthe2mutations. In all 3 assays, a fluorescent dye was incorporated into the PCRproduct(Boutin-Ganacheetal.2001)modifiedbyattaching the21M13F primersequence(TGTAAAACGACGGCCAGT) to the 5# end of one of the primers. PCR reactions and product detection were performed as previously described (Guo et al. 2006). For Mutation 1 (c.ins356T), PCR primers were designed flanking the mutation as the 2 alleles could be discriminated by their 1-bp size difference. For the other 2 assays, PCR–restriction fragment length polymorphism (RFLP) approaches were developed. Oligonucleotide substitutions were made in one of the primers such that a restriction site was generated that would recognize one of the alleles at each mutation. Additionally, the restriction enzyme recognition sequence and a pig-tail sequence (gtgtctt) were appended to the 5# end of the primer adjacent to the
Table 1. Linkage analysis of the long-haired trait in the cat to FGF5
Marker
LOD to long haireda hb
Cat RH map positionc
Cat chromosome numberd FCA start
FCA212 3.0 0.07 1188.1 B1 127, 108, 235 FCA074 3.3 0.09 1233.1 B1 133, 650, 025 FCA1144 4.5 0.07 ND B1 135, 390, 194 FCA612 4.6 0.04 1276.7 B1 139, 688, 359 FCA824 9.5 0.03 1356.2 B1 156, 292, 359 FGF5e (11.6) (0.00) ND B1 158, 450, 684 FCA823 6.3 0.06 1476.2 B1 166, 387, 690
a LOD scores were calculated using 513 microsatellite markers typed on 135 potentially informative meioses with the allele associated with the long-haired phenotype assumed to have a frequency of 0.25. LOD scores were higher by at most 0.6 LOD units if an allele frequency of 0.50 was used. b Estimates of the optimal recombination fractions differed at most by 0.01, depending upon the allele frequencies assumed. c Assignment of microsatellite positions within a cat radiation hybrid (RH) map (Murphy et al. ½2007). ND indicates positions not determined in the cat RH map. d Assignment of microsatellite positions to cat chromosome B1 contiguous sequences (contigs) in the 2 feline genome database using the Garfield Cat Genome Browser. Felis catus (FCA) start indicates sequences corresponding to the contig start sequences. e Results of subsequent linkage analyses for the FGF5 locus are presented in parentheses, after genotyping the 3 predicted mutations in the FGF5 CDS found within this pedigree.
557
Kehler et al.  FGF5 Controls Hair Length in Cats
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
Table2.DNAsequenceanalysisofFGF5in50breedcatswithknownhairlengthidentifies

4putativemutationsassociatedwithlonghair
Catbreed
Cat number
Mutation1Mutation2Mutations3,4a PolymorphismbR51SV61DP65Hins356TR136Xdel474T,T159PN195NQ223QL265L Nucleotide positionc153182194356406474,475585669795
AllelesdG:TT:AC:A:þC:T
T:del474T, A:CC:TG:AG:CGenotypee
Hair length AbyssinianFc640ShortGTCCACAGN,N AbyssinianFc653ShortGTCCACAGN,N AmericanShorthairFc330ShortGTACCACGACGN,N AmericanShorthairFc2387ShortGTCACACACN,N BritishShorthairFc2588ShortGACCACAGN,N BritishShorthairFc2593ShortGTACCACTGGN,4 BritishShorthairFc2596ShortGTACACACTGAGN,N BurmeseFc305ShortGTACACGGN,N BurmeseFc488ShortGTACACGGN,N ChartreuxFc2505ShortGTCCACGCN,N ChartreuxFc2698ShortGTACACGACGN,N CornishrexFc2478ShortGTCCACGCN,N DevonRexFc1455ShortGTCACCACGCN,4 EgyptianMauFc523ShortGTACCACGAGN,N EgyptianMauFc1671ShortGTCCACGCGN,N HavanaFc733ShortTTCCACGCN,N HavanaFc2591ShortGTCACACGACGN,N ManxFc2813ShortGTCACCACGCN,4 OcicatFc2386ShortGTACACACGAGN,N OcicatFc2919ShortGTACACGGN,N RussianBlueFc1143ShortGACCACAGN,N RussianBlueFc2659ShortGTACACACTGCGN,N ScottishFoldFc2536ShortGTACCACGACGN,4 ScottishFoldFc2843ShortGTACCCACGACGN,4 SiameseFc2735ShortGTCCACGAGN,N BirmanFc1916LongGTACCCGC4,4 BirmanFc2349LongGTACCCGC4,4 BirmanFc2569LongGTACCCGC4,4 MaineCoonFc1231LongGTACdel474T/CCTGCG3,4 MaineCoonFc2325LongGTACCCGC4,4 MaineCoonFc2586LongGTACdel474T/CCTGCG3,4 MaineCoonFc2608LongGTACdel474TTGG3,3 NorwegianForestCatFc1914LongGACTACAG2,2 NorwegianForestCatFc2086LongGTACACTC ACGACG2,4 NorwegianForestCatFc2587LongGACTACAG2,2
558
Journal of Heredity 2007:98(6)
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
polymorphic site to introduce a positive control for enzyme digestion (Brownstein et al. 1996). Touchdown PCR cycling conditions were used to amplify Mutation 2 (c.C.T406). For the other 2 assays, the PCR conditions previously described were used (Menotti-Raymond et al. 1999) with the exception thattheinitial93Cdenaturationcyclewasextendedto10min andtheenzymeAmpliTaqGoldwas substitutedforAmpliTaq (Applied Biosystems, Foster City, CA). Fluorescent primer labeling, touchdown PCR amplification, and analysis on an Applied Biosystems Model 3100 DNA sequenher were performed as previously described (Guo et al. 2006). Prior to electrophoresis, digested products were purified by centrifugation through Multiscreen plates (Millipore, Bedford, MA) packedwithSephadexG-50(AmershamBiosciences,Uppsala, Sweden) according to the manufacturer’s instructions.
Results Whole-Genome Scan
We tested for genetic linkage between the long-haired trait and 513 microsatellite markers genotyped in the nonbreed, Nestle´Purina pedigree. A portion of the Nestle´-Purina pedigree included 135 potentially informative meioses from mixed litters of62 progeny produced from F1-by-F1 "short-haired"matings and 11 long-haired cats generated from F1 "short-haired" cats crossed with long-haired cats. The 6 microsatellites that demonstrated single-marker LOD scores at or above 3.0 were all found within contiguous sequences of the assembled cat genome and were assigned to an estimated 39.3-MB region of B1 through the Garfield Cat Genome Browser (http:// lgd.abcc.ncifcrf.gov/cgi-bin/gbrowse/cat) (Pontius J, personal communication) (Table 1). All other markers not on B1 had peaksingle-markerLODscoreslessthan1.75(notshown).The 2 markers (FCA823 and FCA824) that demonstrated the highest LOD scores to thelong-haired locus (6.3, h5 0.06 and 9.5, h 5 0.03, respectively) flanked an estimated 10-Mb region of cat chromosome B1 containing the candidate gene, FGF5 (Table 1). We next sequenced this candidate gene in related nonbreed cats in the Nestle´-Purina pedigree and in an independent research colony at Johns Hopkins University, as well as in unrelated long- and short-haired breed cats.
DNA Sequence Analyses of FGF5 in 50 Short- and Long-Haired Breed Cats
To detect possible mutations in the predicted coding sequence (CDS) of the feline FGF5 gene beginning at position 158, 481, 515 on chromosome B1 of the annotated feline genome (Pontius J, personal communication), primers were designed in flanking regions located in the 5# and 3#untranslated regions and in introns (Supplementary Table A1 and Figure A1) and used to amplify the 3 exons in 50 breed cats with known hair length (Table 2). The size of the assembled predicted CDS was 810 bp, and the inferred fulllength, feline FGF5 protein (Isoform 1 translated from all 3 exons) was 270 amino acids (aas) exhibiting 91% residue identity with the human FGF5 protein (Figure 1A). The
NorwegianForestCatFc2598LongGACTACAG2,2 PersianFc1166LongGTACCCGC4,4 PersianFc1931LongGTACCCGC4,4 PersianFc2064LongGTACCCGC4,4 PersianFc2276LongGTACCCGC4,4 RagdollFc582LongGTC1CACTGC1,1 RagdollFc1621LongGTACdel474T/CCGCG3,4 RagdollFc2271LongGTCAþCCACGC1,4 RagdollFc4013LongGTCAþCCACGC1,4 TurkishAngoraFc2521LongGTACCCGC4,4 TurkishAngoraFc2707LongGTACCCGC4,4 TurkishAngoraFc2742LongGTACCCGC4,4 TurkishVanFc569LongGTACCCGC4,4 TurkishVanFc2517LongGTACCCGC4,4 TurkishVanFc2737LongGTACCCGC4,4 aThehaplotypesofthe2adjacentmutations3and4wereinferredfromtheDNAsequence. del474T:del474T,A/del474T,A;A:NoDel474T,A/NoDel474T,A;C:NoDel474T,C/NoDel474T,

C;del474T/C:NoDel474T,C/Del474T,A;CA:NoDel474T,C/NoDel474T,A. bMutation1(ins356T)andMutation3(del474T)resultinthepredictedtruncationofFGF5Isoform1. cNucleotidepositioninCDSofpredictedmRNA. dPutativemutationsaremarkedinbold.Whenonly1baseisshown,thecatishomozygousatthatposition.Forins356T,"þ"and""refertoinsertionandnoinsertion,respectively. eGenotypesofbothallelesarecodedasfollows:N,normal;1–4,Mutations1–4.
559
Kehler et al.  FGF5 Controls Hair Length in Cats
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
Figure 1. (A) Alignment of FGF5 Isoform 1 (full length form translated from 3 exons) for human, short-haired (SH) cats, and long-haired (LH) cats with 4 putative recessive mutations. Residue changes after putative mutations are marked in bold. (B) Alignment of FGF5 Isoform 2 (translated from exon 1 spliced to exon 3) for human, short-haired cats, and long-haired cats with 4 putative recessive mutations.
560
Journal of Heredity 2007:98(6)
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
potential shorter feline FGF5S protein (Isoform 2 resulting from alternative splicing of exons 1 and 3) was predicted to be 125 aas in length and shared 86% residue identity with the human FGF5S (Figure 1B). A recent report confirmed that the transcripts for both predicted FGF5 isoforms are expressed in the skin of domestic cats (Dro ¨gemu ¨ller et al. 2007). In our study, comparison of the feline FGF5 CDS in 50 unrelated breed cats identified 10 single-nucleotide polymorphisms (SNPs) (Table 2). By examining the observed allele distributions between short- and long-haired cats (Table 2), 3 SNPs in exon 1 of FGF5 producing nonsynonymous substitutions in both FGF5 isoforms and 3 SNPs encoding synonymous substitutions in exon 3 were all excluded as possible causative mutations for the long-haired phenotype.
The4remainingindependentchangesdetectedinexons1and 3 appeared to be functionally significant (described below as Mutations 1–4). All 25 long-haired breed cats were either homozygous for 1 or compound heterozygous for 2 of these 4 predicted mutations, while none of the 25 short-haired cats carried 2 predicted mutant FGF5 alleles. Mutation 1: In 3 unrelated long-haired Ragdoll cats (Fc582, Fc2271, and Fc4013), an insertion of a thymine base was detected 356 bp (c.ins356T) downstream of the predicted translation initiation start site of FGF5 CDS (Table 2). This frameshift mutation was predicted to introduce 3 nonsynonymous substitutions at the end of exon 1, 31 substitutions in exon 2, and a stop codon in exon 3, truncating the Mutation 1 FGF5 protein prematurely at 160 aas. In comparison, the size of the full-length, wild-type feline FGF5 protein was predicted
Figure 2. Three independent pedigrees demonstrating autosomal recessive inheritance of mutations1, 2, and 4 in FGF5 with the long-haired trait. (A) Two-generation portion of the nonbreed Nestle´-Purina pedigree. (B) Two-generation portion of the nonbreed John Hopkins University pedigree. (C) Three-generation pedigree of long-haired Ragdoll cats. Square, male; circle, female. Open symbol, short haired; filled, long haired. The identification number of cats is written over top of individual symbols. Coded genotypes are listed in parentheses: N, no mutation; 2, Mutation 2 (c.C.T406); 3, Mutation 3 (c.del474T); 4, Mutation 4 (C.A.C475).
561
Kehler et al.  FGF5 Controls Hair Length in Cats
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
tobe 270 aas inlength (Figure1A).The resultof the c.ins356T mutation in the short FGF5 mRNA (FGF5S), if spliced and translated correctly, would be to extend the predicted protein to 226 aas compared with 125 aas in the wild-type FGF5S (Figure1B).Onelong-hairedRagdollcat(Fc582)wasfoundto be homozygous for the c.ins356T genotype, supporting the conclusion that the substantial changes caused by this frameshift mutation are likely to completely disrupt the biological activity of both Mutation 1 FGF5 isoforms. The other 2 Ragdoll cats (Fc2271 and Fc4013) were compound heterozygous for Mutation 1 and a second mutant allele (Mutation4discussedbelow).Noneofthe25short-hairedcats were found to carry the Mutation 1 allele in this survey. Mutation 2: Among all the sequenced breed cats, only Norwegian Forest Cats possessed a potential nonsense mutation introduced in exon 2 by the substitution of a cytosine to a thymine base at nucleotide position 406 (c.C.T406). While the Mutation 2 FGF5S mRNA lacking exon 2 should be translated normally, the full-length Mutation 2 FGF5 protein was predicted to contain a change at aa position 135 from an arginine to a premature stop codon (p.R136X) (Figure 1A). Three long-haired Norwegian Forest Cats (Fc1914, Fc2587, and Fc2598) were homozygous for Mutation 2 in the FGF5 gene, suggesting a loss of biological activity. One Norwegian Forest Cat (Fc2086) was compound heterozygous for Mutations 2 and 4 (discussed below) (Table 2). Mutation 3: A deletion of a thymine base at nucleotide position 474 (c.del474T) was found to introduce a frameshift mutation near the beginning of exon 3 in 1 Ragdoll (Fc1621) and 3 Maine Coon cats (Fc1231, Fc2586, Fc2608) (Table 2). The predicted Mutation 3 FGF5 protein should contain extensive nonsynonymous substitutions starting at aa position158andtruncateprematurelyat260residues(Figure1A). Whereas the predicted Mutation 3 FGF5S should have extensive changes starting at aa position 144 and extended to 235 aas in comparison to 125 aas for the wild-type FGF5S (Figure 1B). As a result of Mutation 3, both isoforms were unlikely to maintain their normal biological activity. One Maine Coon cat (Fc2608) was found to be homozygous for Mutation 3, while the other 3 long-haired cats were compound heterozygous for Mutations 3 and 4. Recombination spanning the single-nucleotide position between Mutations 3 and 4 was not detected by phase haplotype analyses in any long-haired cats (Table 2). None of the 25 short-haired breed cats genotyped carried the Mutation 3 allele. Mutation 4: A change from an adenine to a cytosine base was detected at nucleotide position 475 (c.A.C475) in exon 3 in at least one individual from all the long-haired cat breeds sequenced (Table 2). This change was predicted to result in the missense substitution of a single threonine with a proline residue at aa position 159 (p.T159P) in FGF5 (Figure 1A) and the replacement of a tyrosine with a serine residue at the penultimate position in the potential FGF5S isoform (Figure 1B). The majority of long-haired cats sequenced in this initial survey were homozygous for Mutation 4 (Table 2). While 5 breed cats from short-haired registries (British Shorthair [Fc2593], Devon Rex [Fc1455], Manx [Fc2813], and Scottish Fold cats [Fc2536 and FC2843]) were hetero
zygous carriers of this allele, no short-haired cats were found to be homozygous for Mutation 4 in the FGF5 gene.
DNA Sequence Analyses of FGF5 in Additional Short- and Long-Haired Breed Cats
To assess the frequency of normal and mutant FGF5 alleles in cat populations of 26 breeds, we genotyped 66 additional breed cats (29 short-haired and 37 long-haired) using RFLP assays designed to detect Mutations 1–4 (Supplementary Table A2) and then sequenced the entire FGF5 CDS to detect any other potential mutations. Their genotypes were combined with those of the original 50 cats (Table 2) to estimate the allele frequencies in the 21 original breeds and 5 additional long-haired breeds (Angora, Balinese, Himalayan, Siberian, and Somali) (Table 3). Mutations 1 and 2 remained
Table 3. FGF5 allele frequencies among 116 cats from 26 registered breeds
CFA registered breeds
Number of cats
Mutation number
None
Coat 1234length1
Long haired Angora 2 1.00 Long Balinese 2 1.00 Long Birman 6 1.00 Long Himalayan 4 1.00 Long Maine Coon 4 0.50 0.50 Long Norwegian Forest Cat 4 0.88 0.12 Long Persian 7 1.00 Long Ragdoll 13 0.27 0.23 0.50 Long Siberian 3 1.00 Long Somali 6 1.00 Long Turkish Angora 4 1.00 Long Turkish Van 5 1.00 Long Long or short haired Manxa 1 1.00 0.00 Long 2 0.50d 0.50 Short Scottish Folda,b 1 1.00 0.00 Long 3 0.33d 0.66 short Short haired Abyssinian 8 1.00 Short American Shorthair 7 1.00 Short British Shorthairc 5 0.20d 0.80 Short Burmese 2 1.00 Short Chartreux 2 1.00 Short Cornish Rex 1 1.00 Short Devon Rexb,c 3 0.33d 0.67 Short Egyptian Mau 10 1.00 Short Havana 2 1.00 Short Ocicat 2 1.00 Short Russian Blue 3 1.00 Short Siamese 4 1.00 Short
a Breeds with separate short- and long-haired CFA registries. b Breeds that the CFA allows outcrossing with British Shorthair cats. c Breeds with histories of outcrossing with Persian cats. d Mutation 4 alleles were only present in the heterozygous state in shorthaired cats.
562
Journal of Heredity 2007:98(6)
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
unique to the Ragdoll and Norwegian Forest Cat breeds, respectively, whereas Mutation 3 continued to be detected only in Maine Coon and Ragdoll cats. Mutation 4 was the most prevalent mutant FGF5 allele in this study. It was detected in all 14 long-haired breeds sampled and was the only mutation detected in 11 of these breeds (Table 3). No additional mutations were detected in the FGF5 CDS in these additional cats. Most importantly, no short-haired breed cats (n 5 54) were homozygous for any of the 4 FGF5 mutant alleles, whereas all long-haired breed cats (n 5 62) were either homozygous or compound heterozygous for 2 separate mutant alleles. Only 3 of the 6 possible compound heterozygous allelic combinations (1/4, 3/4, 2/ 4) were observed in breed cats in this study (Table 3). The results of this extended breed analysis spurred us to genotype and test whether the inheritance of the 4 mutations detected in the FGF5 gene segregated in pedigrees of nonbreed cats with the long-haired phenotype in an autosomal recessive manner as previously reported (Vella and Robinson 1999).
FGF5 Sequence and Linkage Analyses in Nonbreed Cats
To test for Mendelian segregation of the 4 mutations identified in FGF5 with the long-haired trait, all 3 exons were successfully sequenced in 225 of the 261 cats in the Nestle´Purina pedigree and 40 of 50 cats in a second, nonbreed pedigree from Johns Hopkins University. Of the 211 shorthaired cats genotyped in the Nestle´-Purina pedigree, all were either homozygous for the wild-type FGF5 allele or heterozygous for Mutations 2, 3, or 4 (Supplementary Table A3). Whereas all the 14 long-haired cats were either homozygous for Mutations 2 or 4 or compound heterozygous for paired combinations of Mutations 2, 3, or 4. When the genotypes of the 225 individuals in the Nestle´-Purina pedigree FGF5 were codedfor the 1 wild-type and 4 mutantalleles(Supplementary Table A3) and included in additional linkage analyses, linkage between the long-haired phenotype and FGF5 locus was establishedwithpeakLODscoresof11.6ath50.00(Table1). Similarly, all the 9 genotyped long-haired cats in the Johns Hopkins University pedigree were either homozygous for Mutation4orcompoundheterozygousforMutations3and4, whereas all the 31 short-haired cats were either heterozygous carriers of Mutations 3 or 4 or homozygous for the wild-type FGF5 allele (Supplementary Table A3). As in the breed survey, the concordant phenotype and genotype data for the total of 242 short-haired and 23 long-haired, nonbreed cats were consistent with mutations in the FGF5 gene segregating with the long-haired phenotype in an autosomal recessive manner.
Inheritance of the Long-Haired Phenotype Segregates with Mutations in FGF5
An informative portion of the Nestle´-Purina pedigree clearly demonstrated autosomal recessive inheritance of both Mutations2and4intheFGF5locussegregatingwiththelong-haired phenotype (Figure 2A). Multiple crosses of short-haired cats that were heterozygous for either FGF5 Mutations 2 or 4 produced mixed litters with all the long-haired kittens pos
sessing either 2/2, 2/4, or 4/4 genotypes. Likewise, an informative portion of the Johns Hopkins University pedigree also demonstrated Mendelian inheritance of Mutation 4 segregating with the long-haired phenotype in an autosomal recessive manner (Figure 2B). Although not exclusive of autosomal dominance, a 3-generation pedigree of 9 longhaired Ragdoll cats produced long-haired offspring with all 3 possible combinations of Mutations 1 and 4 consistent with autosomal recessive inheritance (Figure 2C). In combination, thepedigreeanalysesallsupportanautosomalrecessivemode of inheritance of multiple mutant alleles in the FGF5 locus segregating with the long-haired phenotype in domestic cats.
Discussion
A genomic approach was effective at testing the hypothesis that the FGF5 locus is the major determinant of hair length in the cat. A genome wide scan of the Nestle´-Purina pedigree identified a region on cat chromosome B1 containing a likely candidate gene, FGF5, controlling the long-haired trait in the domestic cat. When individuals were genotyped for changes within the CDS and the FGF5 ‘‘marker’’ was included in subsequent linkage analyses, a peak LOD score of 11.6 (h of 0.00) was obtained. Sequence analyses of the feline FGF5 gene in our survey of 12 short-haired, 12 long-haired, and 2 breeds with separate short- and long-haired registries revealed 4 separate mutations that were predicted to disrupt the biological activity of the FGF5 protein. Association analyses of all breed and nonbreed cats demonstrated uniformly that a combined total of 85 genotyped, long-haired cats were either homozygous for a single or compound heterozygous for 2 of 4 FGF5 mutant alleles, while all 296 shorthaired cats were either heterozygous or homozygous for the wild-type allele (Table 3 and Supplementary Table A3). Pedigree analyses of 2 independent, nonbreed colonies and 1 family of Ragdoll cats demonstrated that multiple combinationsofmutationsintheFGF5genesegregatedwiththelonghaired phenotype in an autosomal recessive manner (Figure 2A–C). All these genetic analyses support our conclusion that the FGF5 gene is the major determinant of hair length in the domestic cat. While this study was under review, another group published their interpretation of SNPs within the feline FGF5 gene associated with hair length (Dro ¨gemu ¨ller et al. 2007). They concluded that the c.194C.A (p.P65H) SNP in exon 1 was a predicted mutation for the long-haired trait, although it was clearly in linkage disequilibrium in their pedigree analyses with the c.475A.C SNP in exon 3 (Mutation 4 in this study). They discounted this change as a causative mutation due to the detection of 1 out of 50 crossbred short-haired cats that was homozygous for c.475A.C. Reevaluating the genotype and phenotype of this animal is clearly critical in light of our results. None of the 296 short-haired cats that we genotyped in our study were homozygous for the c.475 A.C change (the most common mutation found in most long-haired cats in our
563
Kehler et al.  FGF5 Controls Hair Length in Cats
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
study), whereas 6 short-haired cats were homozygous for the c.194C.A SNP (Table 2). Similarly, a second SNP in exon 1 c.182T.A was detected within 5 Norwegian Forest Cats and concluded to be a causative mutation in their study, but it was likely in linkage disequilibrium with the c.406C.T nonsense mutation in exon 2 (Mutation 2 found in Norwegian Forest Cats in our study). The authors did not detect this c.406.T change because they did not sequence exon 2 in the feline FGF5 gene. However, we detected 2 unrelated short-haired, registered breed cats (Fc1143 and Fc2588) that were homozygous for the c.182T.A SNP (Table 2) and ruled out this change as a causative mutation for long hair. In addition, the other study did not detect the rare c.ins356T frameshift mutation in exon 1 (Mutation 1 found exclusively in Ragdoll cats in this study). Importantly, all 85 of the long-haired cats in our study were either homozygous or compound heterozygous for 2 of the 4 independent, predicted mutations, whereas none of the 296 short-haired cats were homozygous for any of these 4 predicted mutations in the FGF5 gene. Discovery of spontaneous mutations in FGF5 orthologs of other domestic and wild species may implicate key functional protein domains. As in the Angora mice that lack all transcription of the Fgf5 gene, long-haired dogs with a single FGF5 mutation and long-haired cats with multiple combinations of FGF5 mutations show no additional gross phenotypic changes. A SNP detected in exon 1 of the canine FGF5 gene in long-haired dogs (Housley and Venta 2006) was predicted to substitute a phenylalanine for a conserved cysteine residue at aa position 95 within the b-1 loop required for the formation of hydrogen bonds with FGFRs (Yeh et al. 2003). Whereas the profound predicted changes caused by Mutations 1, 2, and 3 in the feline FGF5 protein preclude further functional interpretation, the feline Mutation 4 (T159P) FGF5 protein implicates an additional key domain that is likely to be required for FGF5 inhibitory activity. The threonine residue at aa position 159 falls within b-chain 7 of FGF5 (Katoh and Katoh 2005) and is a part of a conserved motif found in FGF9, 14, and 20 orthologs from multiple mammalian species (Supplementary Figure A2). Substitution with a proline residue may change the orientation of the following b-chains 8 and 9 and interfere with FGF5#s ability to bind to its cognate receptors (Mohammadi et al. 2005, Zhang et al. 2006). Although we have not directly assessed the biological activity of the feline wild-type and 4 mutant FGF5 proteins, singly or in combination, the observation of multiple long-haired cats that were homozygous for each of the 4 FGF5 mutant genotypes supports the conclusion that each allele is dysfunctional. By analogy to FGF5 function in the mouse, all 4 feline mutant FGF5 proteins likely lose their inhibitory activity on DPC proliferation, prolonging the anagen stage and extending hair growth in long-haired domestic cats. The 4 mutations detected in the feline FGF5 locus may have arisen independently in geographically isolated, freeranging populations of domestic cats and increased in frequency through natural selection in cold climates where the long-haired trait might be advantageous. Subsequent
artificial selection for this recessive phenotype by breeders may have fixed the alleles carried by founders within specific long-haired breeds. While reconstructing the origins of the FGF5 mutations within the context of breeds is largely conjecture due to their limited historical documentation, Mutation 4 may represent the oldest mutant allele. It was present in all 14 and was the only mutant allele in 11 longhaired breed registries sampled in this survey, including the Turkish Angora with a documented history older than 400 years (Fogle 2001). In contrast Mutations 1, 2, and 3 were limited to 3 breeds with documented histories less than 200 years old (Vella and Robinson 1999). In this study, Mutations 1 and 2 were found tobe unique to theRagdollandNorwegian Forest Cat breeds, respectively. Although Mutation 3 was found in both Maine Coon and Ragdoll cats, the more recently established Ragdoll breed (1960) may have obtained this allele by interbreeding with Maine Coon cats (first shown in 1860) (Fogle 2001). The inadvertent introduction of heritable congenital cardiomyopathy reported in Maine Coon cats (Kittleson et al. 1999) into some lines of Ragdolls supports these reports of prior crossbreeding with Maine Coon cats (Traas A, personal communication). The distribution of mutant alleles detected in the breeds sampled in this study supports the hypothesis that Mutations 1, 2, and 3 arose independently in nonbreed cats that were used as founders of Ragdoll cats, Norwegian Forest Cats, and Maine Coon cats, respectively, whereas Mutation4wasmostlikelyalsopresentintheseearlyfounders or subsequently introduced through crosses with other longhaired breeds such as Persian cats. The persistence of heterozygous carriers of FGF5 Mutation 4 among 4 short-haired breed registries (Table 2) may also derive from this practice of outcrossing or from the original founders of these breeds originally established in the British Isles. Although long hair is now a disqualification for registry as a British Shorthair cat with the CFA, British Shorthair cats were occasionally crossed with Persian cats beginning after World War II, up until 1978 (Fogle 2001). Despite selection against the long-haired phenotype by breeders, the Mutation 4 allele was still present in the American population of British Shorthair cats sampled in this study. The CFA currently permits both Devon Rex and Scottish Fold cats to be outcrossed with British Shorthair cats, a possible source of Mutation 4 in these breeds. In addition, the founders of the Scottish Fold and Manx breeds who maintain separate registries based on hair length included long-haired individuals (Fogle 2001; Stephens 2001). In support of this historical reconstruction, the long-haired Manx and Scottish Fold cats that we sequenced were homozygous for Mutation 4 (Table 3). Although not detected in our study, it would not be unexpected to find carriers of Mutation 4 in extended populations of shorthaired breeds such as the Siamese from which the longhaired Balinese breed was derived and recognized by the CFA during the 1960s (Vella and Robinson 1999). Although this study included 62 unrelated individuals from 14 long-haired breed registries and 23 long-haired nonbreed cats from the Johns Hopkins University and
564
Journal of Heredity 2007:98(6)
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
Nestle-Purina pedigrees, it is possible that additional mutations in the feline FGF5 gene may be present in unsampled long-haired breed and nonbreed cats. In addition, the reported quantitative and qualitative differences between the coats of long-haired cat breeds indicate that other independent loci may modify the major influence of FGF5 on hair length in the domestic cat (Vella and Robinson 1999). Finally, this study raises the question as to whether wild felid species with long hair possess FGF5 mutations, providing an avenue to test whether free-ranging populations have undergone natural selection for genes that could affect coat length and their current phylogeographic distribution.
Funding
National Institutes of Health/National Center for Research Resources KO1/Special Emphasis Research Career Award (RR019677-01) to J.S.K; National Institutes of Health/ National Institutes on Deafness and Other Communication Disorders (DC00232) to D.K.R.
Supplementary Material
Supplementary material can be found at http://www. jhered.oxfordjournals.org/.
Acknowledgments
J.S.K. and V.A.D. contributed equally to this work. The authors would like to thank Bethany Buzzell and David Wells for technical assistance and Dr Guo Kei Pei and Lisa Maslan for operating the automated DNA sequencers. We thank the many cat breeders who have contributed in the past to our population genetic database of cat breeds at the Laboratory of Genomic Diversity through cooperation with the CFA and the International Cat Association. We would also like to thank Dr Yasuko Ishida, Dr Solveig Pflueger, and Dr Anne Traas for their useful information about cat breeds and Dr Joan Pontius about the annotation of the cat genome. This research was supported in part by intramural research programs at the National Institutes of Health, National Cancer Institute, and National Library of Medicine. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services or mention of trade names, commercial products, or organizations imply endorsement by the US government.
References
Boutin-Ganache I, Raposo M, Raymond M, Deschepper CF. 2001. M13-tailed primers improve the readability and usability of microsatellite analyses performed with two different allele-sizing methods. Biotechniques. 31:24–26, 28.
Brownstein MJ, Carpten JD, Smith JR. 1996. Modulation of non-templated nucleotide addition by Taq DNA polymerase: primer modifications that facilitate genotyping. Biotechniques. 20:1004–1006, 1008–1010.
Burns M, Fraser MN. 1963. Genetics of the dog: the basis of successful breeding. Edinburgh (UK): Oliver and Boyd.
Clements DA, Wang J-K, Dionne CA, Goldfarb M. 1993. Activation of fibroblast growth factor (FGF) receptors by recombinant human FGF5. Oncogene. 8:1311–1316.
Dickie MN. 1963. New mutations. I. Angora. Mouse News Lett.. 29:39.
Dro ¨gemu ¨ller C, Ru ¨fenacht S, Wichert B, Leeb T. 2007. Mutations within the FGF5 gene are associated with hair length in cats. Anim Genet. 38:218–221.
Eizirik E, Yuhki N, Johnson WJ, Menotti-Raymond M, Hannah SS, O’Brien SJ. 2003. Molecular genetics and evolution of melanism in the cat family. Curr Biol. 13:448–453.
Fishelson M, Geiger D. 2002. Exact genetic linkage computations for general pedigrees. Bioinformatics. 18 (Suppl 1):S189–S198.
Fishelson M, Geiger D. 2004. Optimizing exact genetic linkage computations. J Comput Biol. 11:263–275.
Fogle B. 2001. The new encyclopedia of cats. New York: DK Publishing Inc.
Fraser AS. 1953. A note on the growth of rex and angora coats. J Genet. 51:237–242.
Giordano FJ, Ping P, McKirnan MD, Nozaki S, DeMaria AN, Dillmann WH, Mathieu-Costello O, Hammond HK. 1996. Intracoronary gene transfer of fibroblast growth factor-5 increases blood flow and contractile function in an ischemic region of the heart. Nat Med. 2:534–539.
Go´mez-Pinilla F, Cotman CW. 1993. Distribution of fibroblast growth factor 5 mRNA in the rat brain: an in situ hybridization study. Brain Res. 606:79–86.
Green ES, Rendahl KG, Zhou S, Ladner M, Coyne M, Srivastava R, Manning WC, Flannery JG. 2001. Two animal models of retinal degeneration are rescued by recombinant adeno-associated virus-mediated production of FGF-5 and FGF-18. Mol Ther. 3:507–515.
Guo XC, Scott K, Liu Y, Dean M, David V, Nelson GW, Johnson RC, Dilks HH, Lautenburger J, Kessing B, et al. 2006. General factors leading to chronic Epstein-Barr virus infection and nasopharyngeal carcinoma in South East China: study design, methods and feasibility. Hum Genomics. 2:365–375.
Hart AW, Baeza N, Apelquist A ˚, Edlund H. 2000. Attenuation of FGF signalling in mouse b-cells leads to diabetes. Nature. 408:864–868.
Haub O, Drucker B, Goldfarb M. 1990. Expression of the murine fibroblast growth factor 5 gene in the adult central nervous system. Proc Natl Acad Sci USA. 87:8022–8026.
Haub O, Goldfarb M. 1991. Expression of the fibroblast growth factor-5 gene in the mouse embryo. Development. 112:397–406.
He´bert JM, Boyle M, Martin GR. 1991. mRNA localization studies suggest that murine FGF-5 plays a role in gastrulation. Development. 112:407–415.
He´bert JM, Rosenquist T, Go ¨tz J, Martin GR. 1994. FGF5 as a regulator of the hair growth cycle: evidence from targeted and spontaneous mutations. Cell. 78:1017–1025.
Housley DJE, Venta PJ. 2006. The long and the short of it: evidence that FGF5 is a major determinant of canine ‘hair’-itability. Anim Genet. 37:309–315.
Ishida Y, David VA, Eizirik E, Scha ¨ffer AA, Neelam BA, Roelke ME, O’Brien SJ, Menotti-Raymond M. 2006. A homozygous single-base deletion in MLPH causes the dilute coat color phenotype in the domestic cat. Genomics. 88:698–705.
Katoh M. 2002. WNT and FGF gene clusters. Int J Oncol. 21:1269–1273.
Katoh Y, Katoh M. 2005. Comparative genomics on FGF16 orthologs. Int J Mol Med. 16:959–963.
Kittleson MD, Meurs KM, Munro MJ, Kittleson JA, Liu S-K, Pion PD, Towbin JA. 1999. Familial hypertrophic cardiomyopathy in maine coon cats: an animal model of human disease. Circulation. 99:3172–3180.
Limat A, Hunziker T, Waelti ER, Inaebnit SP, Wiesmann U, Braathen LR. 1993. Soluble factors from human hair papilla cells and dermal fibroblasts dramatically increase the clonal growth of outer root sheath cells. Arch Dermatol Res. 285:205–210.
565
Kehler et al.  FGF5 Controls Hair Length in Cats
 by guest on November 23, 2016http://jhered.oxfordjournals.org/Downloaded from
Menotti-Raymond M, David VA, Lyons LA, Scha ¨ffer AA, Tomlin JF, Hutton MK, O’Brien SJ. 1999. A genetic linkage map of microsatellites in the domestic cat (Felis catus). Genomics. 57:9–23.
Mohammadi M, Olsen SK, Ibrahimi OA. 2005. Structural basis for fibroblast growth factor receptor activation. Cytokine Growth Factor Rev. 16:107–137.
MurphyWJ, Davis B, David VA,Agarwala R, Scha ¨ffer AA,Pearks-Wilkerson AJ, Neelam BA, O’Brien SJ, Menotti-Raymond M. 2007. A 1.5 megabase resolution radiation hybrid map of the cat genome and comparative analysis with the canine and human genomes. Genomics. 89:189–196.
Ornitz DM, Itoh N. 2001. Fibroblast growth factors. Genome Biol. 2:3005. 1–3005.12.
Ornitz DM, Xu J, Colvin JS, McEwen DG, MacArthur CA, Coulier F, Gao G, Goldfarb M. 1996. Receptor specificity of the fibroblast growth factor family. J Biol Chem. 271:15292–15297.
Ota Y, Saitoh Y, Suzuki S, Ozawa K, Kawano M, Imamura T. 2002. Fibroblast growth factor 5 inhibits hair growth by blocking dermal papilla cell activation. Biochem Biophys Res Commun. 290:169–176.
Paus R, Foitzik K. 2004. In search of the "hair cycle clock": a guided tour. Differentiation. 72:489–511.
Pennycuik PR, Raphael KA. 1984. The angora (go) locus in the mouse: hair morphology, duration of growth cycle and site of action. Genet Res. 44:283–291.
Rosenquist TA, Martin GR. 1996. Fibroblast growth factor signaling in the hair growth cycle: expression of the fibroblast growth factor receptor and ligand genes in the murine hair follicle. Dev Dyn. 205:379–386.
Stephens G. 2001. Legacy of the cat. San Francisco (CA): Chronicle Books.
Suzuki S, Ota Y, Ozawa K, Imamura T. 2000. Dual-mode regulation of hair growth cycle by two Fgf-5 gene products. J Invest Dermatol. 114:456–463.
Vella CM, Robinson R. 1999. Robinson’s genetics for cat breeders and veterinarians. Boston: Butterworth-Heinemann.
Vigneron N, Stroobant V, Chapiro J, Ooms A, Degiovanni G, Morel S, van der Bruggen P, Boon T, Van den Eynde BJ. 2004. An antigenic peptide produced by peptide splicing in the proteasome. Science. 304: 587–590.
Yeh BK, Igarashi M, Eliseenkova AV, Plotnikov AN, Sher I, Ron D, Aaronson SA, Mohammadi M. 2003. Structural basis by which alternative splicing confers specificity in fibroblast growth factor receptors. Proc Natl Acad Sci USA. 100:2266–2271.
Zhan X, Culpepper A, Reddy M, Loveless J, Goldfarb M. 1987. Human oncogenes detected by a defined medium culture assay. Oncogene. 1:369–376.
Zhang X, Ibrahimi OA, Olsen SK, Umemori H, Mohammadi M, Ornitz DM. 2006. Receptor specificity of the fibroblast growth factor family. The complete mammalian FGF family. J Biol Chem. 281:15694–15700.
Zhang Z, Schwartz S, Wagner L, Miller W. 2000. A greedy algorithm for aligning DNA sequences. J Comput Biol. 7:203–214.
Received March 19, 2007 Accepted July 11, 2007
Corresponding Editor: Leif Andersson
566