ORIGINAL ARTICLE

Characterization of Human Keratin-Associated Protein 1Family Members

Yutaka Shimomura, Noriaki Aoki, Michael A. Rogers,w Lutz Langbein,n Jˇrgen Schweizer,w and Masaaki ItoDepartment of Dermatology, Niigata University School of Medicine, Niigata, Japan; nThe Divisions of Cell Biology and wBiochemistry of Tissue Speci¢cRegulation, German Cancer Research Center, Heidelberg, Germany

Keratin-associated proteins are involved in the forma-tion of the cross-linked network of the keratin-inter-mediate ¢lament proteins that support hair ¢bers. Inrecent years, several keratin-associated protein genes havebeen identi¢ed and become an attractive topic in hairresearch. More recently, we isolated two cDNA encod-ing novel members of the human keratin-associatedprotein 1 family (human keratin-associated protein1.6 and human keratin-associated protein 1.7), anddescribed their expression in the hair follicle by RNAin situ hybridization. A comparison of human keratin-associated protein 1.6 and human keratin-associated

protein 1.7 with other human keratin-associated protein1 members revealed that keratin-associated protein 1proteins are fundamentally composed of ¢ve distinctdomains, and that they can be classi¢ed primarily bya striking variation in double cysteine-containingpentapeptide repeats in the repetitive I domain. Thesum of the data analyzed suggests that human keratin-associated protein 1 family genes may have arisenmainly through gene duplication of the cysteine-repeatmotifs during evolution. Key words: human gene/in situ hybridization/keratin/keratin-associated protein. JIDSymposium Proceedings 8:96 ^99, 2003

Human hair ¢bers share a common structural orga-nization, in which a multicellular cortex is encasedin a cuticular layer of £attened cells, often with amedulla layer centrally placed in the cortex. Hair¢bers are composed primarily of keratin proteins

that are derived from several multigene families. These hair-spe-ci¢c keratins have been classi¢ed into the keratin intermediate-¢lament proteins (KIF) and their associated proteins (KAP) (Powellet al, 1991). KAP are a major component of the hair ¢ber matrix andparticipate in the formation of a rigid hair shaft by cross-linkingthe KIF to form a meshwork (Powell and Rogers, 1997). Based ontheir amino acid composition, KAP have been originally classi¢edinto three groups: (i) high-sulfur KAP (16^30 mol% cysteine;KAP1-3, 10^16 and 23 families) (Swart and Haylett, 1973; Swartet al, 1976; Powell et al, 1983; Frenkel et al, 1989; Zhumbaeva et al,1992; Huh et al, 1994; Powell and Rogers, 1997; Aoki et al, 1998; Coleand Reeves, 1998; Mitsui et al, 1998; Takaishi et al, 1998; Kuhn et al,1999; Rogers et al, 2001; Rogers et al, 2002, Shimomura et al, 2002a);(ii) ultra-high-sulfur KAP (430 mol% cysteine; KAP4, 5, 9, and 17families) (McNab et al, 1989; MacKinnon et al, 1990; Fratini et al,1994; Jenkins and Powell, 1994; Powell et al, 1995; Perez et al, 1999;Rogers et al, 2001); and (iii) high-glycine/tyrosine KAP (KAP6^8families) (Kuczek and Rogers, 1987; Fratini et al, 1993; Aoki et al, 1997;Rogers et al, 2002). To date, KAP have been subdivided into a totalof 23 families (KAP1.n^KAP23.n), based on their cysteine/tyrosine-glycine content as well as the degree of amino acid homologywithin each family and the nature of repeat structures oftenfound in these molecules, according to an uni¢ed nomenclature

proposed by Rogers and Powell (1993). Of these KAP families,the KAP1 family has been de¢ned as a group of high-sulfur pro-teins expressed in the cortical layer of the hair follicle (Rogersand Powell, 1993), and so far, eight KAP1members have been re-ported from sheep, rabbits, and humans (Elleman and Dopheide,1972; Powell et al, 1983; Zhumbaeva et al, 1992; Mitsui et al, 1998).In humans, two KAP1 genes, hB2A and hB2B, were initially iso-lated from a single genomic clone (Zhumbaeva et al, 1992), andlater renamed hKAP1.1A and hKAP1.2, respectively (Rogers et al,2001). In a recent study, we isolated two cDNA encoding newhKAP1 members during the analysis of the hKAP1.1A gene inpatients with a congenital fragile hair disorder, which we havenamed hKAP1.6 and hKAP1.7 (Shimomura et al, 2002). hKAP1.6and hKAP1.7 show high nucleotide sequence homology withhKAP1.1A. In particular, the three genes share a completely iden-tical sequence in their 30 -noncoding regions. At the amino acidsequence level, the degree of identity between hKAP1.1A,hKAP1.6, and hKAP1.7 is striking (Fig 1). A comparison of hKA-P1.1A and hKAP1.6 revealed a single amino acid change betweenthe proteins as well as a 46-residue deletion in the amino term-inal region of hKAP1.6. Similarly, comparison of hKAP1.1AwithhKAP1.7 revealed complete identity, except for a 92-residue ami-no acid deletion in hKAP1.7. In addition, the multialignment ofhKAP1.6 and hKAP1.7 with sheep and rat KAP1 members de-monstrated the structural features of the KAP1 family at the ami-no acid sequence level (Shimomura et al, 2002a).Unexpectedly, as we identi¢ed hKAP1.6 and hKAP1.7, a large

cluster of human high/ultra-high-sulfur KAP genes was foundon chromosome 17q12^21, resulting in the identi¢cation offour additional novel hKAP1 genes, termed hKAP1.1B, hKAP1.3,hKAP1.4, and hKAP1.5 (Rogers et al, 2001). In particular,hKAP1.1B is nearly identical to hKAP1.1A in DNA sequence.The multialignment of eight human KAP1 members showedthat these proteins are composed of ¢ve individual domains: anN-terminal domain, a repetitive I domain, a central nonrepetitive

Reprint requests to: Yutaka Shimomura, Department of Dermatology,Niigata University School of Medicine, Asahimachi-dori, Niigata 951-8510,Japan. Email: yshimo@med.niigata-u.ac.jpAbbreviations: KAP, keratin-associated protein; KIF, keratin intermediate

¢lament protein.

Manuscript accepted for publication February 1, 2003

0022-202X/03/$15.00 . Copyright r 2003 by The Society for Investigative Dermatology, Inc.

96

domain, a repetitive II domain, and a C-terminal domain (Fig 1).The N-terminal, central nonrepetitive, and repetitive II domainsappear to be signi¢cantly conserved among the hKAP1members.Likewise, these domains are well conserved in the KAP1 mem-bers of other species (Shimomura et al, 2002a), indicating thatthey may play an essential part in KAP1 family function. The C-terminal domain, which is the smallest region, is composed offour to seven amino acid residues and generally contains a uniqueelement, ‘‘EPTC’’, although this is not found in hKAP1.2, andshows high similarity across the family (Fig 1). An additional re-gion of interest in the KAP1 family members is the repetitive Idomain, which is striking in its high degree of repetitivenessand size variability. The repeat structures in this domain appear

to be composed essentially of amino acid variations on a doublecysteine-containing pentapeptide, CCQ(P/T)S and CCETS. Inaddition, hKAP1members have a curious nonrepetitive sequence‘‘FCGF(P/R)SFST(G/S)GTC(D/S)SS’’or ‘‘FCDFLASQLVDLQLS’’between these pentapeptide repeats in the repetitive I domain(Fig 1). These sequences could not be detected in the repetitive Idomains of the KAP1 members of other species already reported(Shimomura et al, 2002). The comparisons of the hKAP1 mem-bers showed high homologies at the N-terminal, central nonre-petitive, repetitive II, and C-terminal domains, but not in therepetitive I domain. Therefore, the distinction between hKAP1members obviously appears to be dependent on the number ofpentapeptide repeats and the arrangement of the repeat segments

Figure1. Multi-alignment of hKAP1 family members. Residues that are identical are colored light blue. Dashes denote gaps in the sequence to max-imize alignment. The pentapeptide repeats are boxed with black lines. Brackets show the ¢ve KAP subdomains. The unique nonrepetitive sequences in therepetitive I domain are boxed with red lines. The accession numbers for the respective protein sequences are: hKAP1.1A (hB2A), X63337; hKAP1.1B(AC007455, nt 8801^9334); hKAP1.2 (hB2B), X63338; hKAP1.3 (AC007455, nt 15377^15880); hKAP1.4 (AC007455, nt 20120^20485); hKAP1.5 (AC007455,nt 23043^23568); hKAP1.6 (AB052868); hKAP1.7 (AB055057).

CHARACTERIZATION OF HKAP1MEMBERS 97VOL. 8, NO. 1 JUNE 2003

and nonrepetitive sequences in the repetitive I domain, indicatingthat this domain is likely to de¢ne the individual functional char-acteristics of each hKAP1member.At present, eight hKAP1 genes have been isolated from hu-

mans. It is intriguing to consider how these hKAP1 genes havearisen in the human genome. The presence of various patternsof pentapeptide repeat structures in the repetitive I domain raisesthe possibility that the hKAP1 genes arose by gene duplication,especially the duplication or deletion of cysteine-repeat segments,during the course of evolution (Rogers and Powell, 1993; Rogerset al, 2001). Furthermore, additional mutations and/or small dele-tion events occasionally seem to have occurred in other domains,such as the N-terminal domain in hKAP1.4 and the C-terminaldomain in hKAP1.2.In situ hybridization studies of human scalp hair using a

probe against the 30 -noncoding region of hKAP1.6 revealedstrong signals, predominantly in the cortical layer, throughoutthe keratogenous zone (Fig 2A). This ¢nding indicated thathKAP1.1A/B and hKAP1.7 mRNA are also expressed in the sameregion, because hKAP1.1A/B, hKAP1.6, and hKAP1.7 have acompletely identical nucleotide sequence in the 30 -noncodingregion. Similarly, hKAP1.3 (Fig 2B) and hKAP1.5 (Fig 2C)mRNA were detected in the cortical layer as well, usingspeci¢c probes, and their region of expression was consistentwith that of the hKAP1.1A/B, hKAP1.6, and hKAP1.7 mRNA.The relative quantities of these mRNA expressed in this regionremain to be analyzed. Expression studies of all the hKAP1 mem-bers, as well as of other hKAP family members, at both themRNA and protein levels, may give us some insight into thecharacteristics that distinguish the hair of di¡erent individuals,

and provide additional impetus to discover the complete set ofhKAP genes.hKAP gene clusters are located on chromosomes 17q12^21

(Rogers et al, 2001) and 21q22.1 (Hattori et al, 2000; Rogers et al,2002); however, the hKAP1.6 and hKAP1.7 genes are not detect-able at these gene loci, indicating that there may be a further clus-ter of hKAP genes elsewhere. Alternatively, the presence of highsequence identities among hKAP1.1A/B, hKAP1.6, and hKAP1.7in the 50 - and 30 -noncoding regions as well as in the open read-ing frame suggests the possibility that hKAP1.6 and hKAP1.7 arethe polymorphic alleles of either hKAP1.1A or hKAP1.1B createdby intragenic deletions during evolution. So far, we have per-formed further analysis of hKAP1 family members concerningthese problems and have got some interesting results (Shimomuraet al, 2002).Mutations in the hair keratin genes, hHb1 or hHb6, have pre-

viously been shown to cause the congenital hair disease monile-thrix, which is characterized by beaded and fragile hairs (Winteret al, 1997a, b; Korge et al, 1999). Along with the keratins, hKAPare the major component of human hair, in which mutations arelikely to cause hereditary hair genodermatoses. Therefore, theanalysis of hKAP genes in patients with hair abnormalities willprovide meaningful information. In addition, the production oftransgenic mice expressing mutated KAP is likely to provide in-sight into the function of the KAP in hair follicles, and may pro-vide a model for human hair disorders, as well.In this study, we gave only one example out of many hKAP

families. The hKAP1 family, however, shares similar characteris-tics with the other hKAP families.Thus, a detailed study of hKAP1members will be a good guide for understanding all the hKAP.

Figure 2. Transcripts of hKAP1 genes are expressed in the cortex of the hair follicle. (A) hKAP1.1A/B, hKAP1.6, and hKAP1.7. (B) hKAP1.3. (C)hKAP1.5. Note that due to the high DNA sequence homology of the hKAP1.1A/B, hKAP1.6, and hKAP1.7 genes, the in situ probe used for (A) recognizesfour transcripts. The red arrows demarcate the area of major KAP expression. No transcripts were detected in the medulla (data not shown) or in the ORS,IRS, and CU. ORS, outer root sheath; IRS, inner root sheath; CU, cuticle of the hair shaft; CO, cortex; DP, dermal papilla. Scale bar: 150 mm.

98 SHIMOMURA ETAL JID SYMPOSIUM PROCEEDINGS

REFERENCES

Aoki N, Ito K, Ito M: Isolation and characterization of mouse high-glycine/tyrosineproteins. J Biol Chem 272:30512^30518, 1997

Aoki N, Ito K, Ito M: Hair follicle has a novel anagen-speci¢c protein, mKAP13.J Invest Dermatol 111:804^809, 1998

Cole SE, Reeves RH: A cluster of keratin-associated proteins on mouse chromosome10 in the region of conserved linkage with human chromosome 21. Genomics54:437^442, 1998

EllemanTC, DopheideTA:The sequence of SCMK-B2B, a high-sulfur protein fromwool keratin. J Biol Chem 247:3900^3909, 1972

Fratini A, Powell BC, Rogers GE: Sequence, expression, and evolutionary conserva-tion of a gene encoding a glycine/tyrosine-rich keratin-associated protein ofhair. J Biol Chem 268:4511^4518, 1993

Fratini A, Powell BC, Hynd PI, Keough RA, Rogers GE: Dietary cysteine regulatesthe levels of mRNA encoding a family of cysteine-rich proteins of wool.J Invest Dermatol 102:178^185, 1994

Frenkel MJ, Powell BC,Ward KA, Sleigh MJ, Rogers GE: The keratin BIIIB genefamily. Isolation of cDNA clones and structure of a gene and a related pseudo-gene. Genomics 4:182^191, 1989

Hattori M, Fujiyama A,Taylor TD, et al: The DNA sequence of human chromosome21. Nature 405:311^319, 2000

Huh N, Kashiwagi M, Konishi C, Hashimoto Y, Kohno Y, Nomura S, Kuroki T:Isolation and characterization of a novel hair follicle-speci¢c gene, Hacl-1.J Invest Dermatol 102:716^720, 1994

Jenkins BJ, Powell BC: Di¡erential expression of genes encoding a cysteine-rich ker-atin family in the hair cuticle. J Invest Dermatol 103:310^317, 1994

Korge BP, Hamm H, Jury CS, et al: Identi¢cation of novel mutations in basic hairkeratins hHb1 and hHb6 in monilethrix: Implications for protein structure andclinical phenotype. J Invest Dermatol 113:607^612, 1999

Kuczek ES, Rogers GE: Sheep wool (glycineþ tyrosine)-rich keratin genes. A fa-mily of low sequence homology. European J Biochem 166:79^85, 1987

Kuhn F, Lassing C, Range A, Mueller M, HunzikerT, Ziemiecki A, Andres A-C: Pmg-1 and Pmg-2 constitute a novel family of KAP genes di¡erentially expressed dur-ing skin and mammary gland development. Mechanisms Dev 86:193^196, 1999

MacKinnon PJ, Powell BC, Rogers GE: Structure and expression of genes for a classof cysteine-rich proteins of the cuticule layers of di¡erentiating wool and hairfollicles. J Cell Biol 111:2587^2600, 1990

McNab AR,Wood L,Theriault N, GiermanT,Vogeli G: An ultra-high sulfur keratingene is expressed speci¢cally during hair growth. J Invest Dermatol 92:263^266,1989

Mitsui S, Ohuchi A, Adachi-YamadaT, Hotta M,Tsuboi R, Ogawa H: Structure andhair follicle-speci¢c expression of genes encoding the rat high sulfur proteinB2 family. Gene 208:123^129, 1998

Perez C, Auriol J, Gerst C, Bernard BA, Egly JM: Genomic organization and pro-moter characterization of two human UHS keratin genes. Gene 227:137^148,1999

Powell BC, Rogers GE: The role of keratin proteins and their genes in thegrowth, structure and properties of hair. In: Jolles P, Zahn H, H˛cker H(eds). Formation and Structure of Human Hair. Basel: Birkh�user Verlag, 1997:pp 59^148

Powell BC, Sleigh MJ, Ward KA, Rogers GE: Mammalian keratin gene families.Organisation of genes coding for the B2 high-sulphur proteins of sheep wool.Nucleic Acids Res 11:5327^5346, 1983

Powell BC, Nesci A, Rogers GE: Regulation of keratin gene expression in hairfollicle di¡erentiation. Ann NYAcad Sci 642:1^20, 1991

Powell BC, Arthur J, Nesci A: Characterization of a gene encoding a cysteine-richkeratin associated protein synthesized late in rabbit hair follicle di¡erentiation.Di¡erentiation 58:227^232, 1995

Rogers GE, Powell BC: Organization and expression of hair follicle genes. J InvestDermatol 101:50s^55s, 1993

Rogers MA, Langbein L,Winter H, Ehmann C, Praetzel S, Schweizer J: Character-ization of a cluster of human high/ultrahigh sulfur keratin associated protein(KAP) genes imbedded in the type I keratin gene domain on chromosome17q12^21. J Biol Chem 276:19440^19451, 2001

Swart LS, Haylett T: Studies on the high-sulphur proteins of reduced Merinowool. Amino acid sequence of protein SCMKB-3A3. BiochemJ 133:641^654, 1973

Swart LS, Joubert FJ, Parris D: Homology in the amino acid sequences of the high-sulphur proteins from keratins. In: Ziegler K (ed.). Proceedings of the 5th Interna-tionalWoolTextile Research Conference. Aachen: GermanWool Research Institute,1976:pp 254^264

Takaishi M,TakataY, Kuroki T, Huh N-H: Isolation and characterization of a puta-tive keratin-associated protein gene expressed in embryonic skin of mice.J Invest Dermatol 111:128^132, 1998

Winter H, Rogers MA, Langbein L, et al: Mutations in the hair cortex keratinhHb6 cause the inherited hair disease monilethrix. Nature Genet 16:372^374,1997a

Winter H, Rogers MA, Gebhardt M, et al: A new mutation in the type II hair cortexkeratin hHb1 involved in the inherited hair disorder monilethrix. Hum Genet101:165^169, 1997b

Zhumbaeva BD, Gening LV, Gazaryan KG: Cloning and structural characterizationof human hair sulfur-rich keratin genes. Mol Biol 26:550^555, 1992

CHARACTERIZATION OF HKAP1MEMBERS 99VOL. 8, NO. 1 JUNE 2003