Journal of Dermatological Science 74 (2014) 39–47

Krtap11-1, a hair keratin-associated protein, as a possible crucialelement for the physical properties of hair shafts

Shunsuke Fujimoto, Takahisa Takase, Nanako Kadono, Kenji Maekubo, Yohei Hirai *

Department of Bioscience, Graduate School of Science and Technology, Kwansei Gakuin University, 2-1 Gakuen, Sanda 669-1337, Japan

A R T I C L E I N F O

Article history:

Received 29 October 2013

Received in revised form 17 December 2013

Accepted 18 December 2013

Keywords:

Krtap

Keratin

KAPs

Hacl-1

IF assembly

A B S T R A C T

Background: The physical properties of the hair are predominantly determined by the assembly of

keratin bundles. The keratin-associated proteins (Krtaps) are thought to be involved in keratin bundle

assembly, however, the functional role of the individual member still remains largely unknown.

Objective: The aim of this study is to clarify the role of a unique class of Krtaps, Krtap11-1, in the

development and physical properties of the hair.

Methods: The expression regulation of Krtap11-1 was analyzed and its binding partners in the hair

cortex were determined. Also, the effects of the forcible expression of this protein on the hair follicle

development were analyzed in culture.

Results: The expression pattern of Krtap11-1 was concentrically asymmetric in the faulty hair that

develops in Foxn1nu mice. In cultured keratinocytes, the expression of Krtap11-1 transgene product was

strictly regulated by the keratinization process and proteasome-dependent protein elimination. While

the association with keratin as well as the cohesive self-assembly of Krtap11-1 appeared to be stabilized

by disulfide cross-links, the biotinylated Krtap11-1 probe enabled the adherence to certain type I

keratins in the hair cortex, including K31, 33 and 34, in the absence of disulfide formation. When

embryonic upper lip rudiments were forcibly introduced with Krtap11-1, the hair follicles formed

irregularly arranged globular hair keratin-clumps surrounded by multilayered epithelial cells in culture.

Conclusion: Krtap11-1 may play an important role on keratin-bundle assembly in the hair cortex and

this study provides insight into the physical properties of the hair shaft.

� 2013 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights

reserved.

Contents lists available at ScienceDirect

Journal of Dermatological Science

jo ur n al h o mep ag e: www . jds jo u rn al .c om

1. Introduction

The hair shaft consists of concentric layers of cuticle, cortex andmedulla. The physical properties of the hair shaft, such as thebounce and resilience, are mainly determined by the cortex, whichhas densely packed and longitudinally aligned keratin intermedi-ate filaments (Ifs). The hair and hair follicles repeatedly cyclebetween the anagen (growth), catagen (regression), and telogen(resting) phase, and hair keratin Ifs are actively formed in theanagenic stage [1,2]. The cortical precursor trichocytes in theanagenic hair bulbs radially produce type I and II keratins so as toform heterodimers and assemble them into keratin Ifs, whichfollows the establishment of strong intercellular connections via

numerous intercellular desmosomes [1,3]. During the progressionof hair follicle differentiation, the hair keratin Ifs may be cementedby cysteine-rich IF-associated proteins Krtaps (known also asKAPS) to give rise to the rigid and resistant hair shaft with

* Corresponding author. Tel.: +81 79 565 7234; fax: +81 79 565 7234.

E-mail address: y-hirai@kwansei.ac.jp (Y. Hirai).

0923-1811/$36.00 � 2013 Japanese Society for Investigative Dermatology. Published b

http://dx.doi.org/10.1016/j.jdermsci.2013.12.006

numerous cross-linked networks [4–6]. In humans, more than 100Krtap genes have been identified on certain chromosomes such as11q13.4, 17q21.2 and 21q22.1 and many of them have a singleexon with a short open reading flame [6]. These proteins areclassified into two major groups composed of the high sulfur-type(cysteine content: 11–16%) and the high Glycine/Tyrosine-type(Gly-Tyr content: 13–70%), and are also divided into phylogeneti-cally related subgroups [7,8]. They are reported to be particularlyimportant for hair maturation, since they exhibit strong expressionin a highly defined pattern restricted to the nascent cortical regionof the anagenic hair shaft [7,9]. Among these Krtap proteins,Krtap11-1 (also known as Hacl-1) is a unique member of the highsulfur-type Krtap in terms of lacking the highly conserved primarystructure and in possessing six direct repeats containing aCysteine-Glutamine-Proline (CQP) motif in the C-terminal se-quence [7,10], which is also present in repetitive peptidesequences of certain hair- and epidermal cell-specific proteins[11]. In spite of many investigations on the expression patterns ofthe krtap genes and investigations of their molecular evolution[7,8], the practical role of these proteins still remains to beelucidated, mainly because of their highly insoluble nature.

y Elsevier Ireland Ltd. All rights reserved.

S. Fujimoto et al. / Journal of Dermatological Science 74 (2014) 39–4740

The mouse strain Foxn1nu and HR-1 have been widely used asmodel animals to analyze the molecular mechanism underlyingthe development [12,13] as well as maturation of hair follicles. Themouse strain Foxn1nu has pleiotropic skin disorder phenotype,such as impairment of both the hair and follicles, in which FOXN1, agene for a transcription factor for a number of hair-related genes,was disrupted [14,15]. HR-1 mice, that harbor a mutation in thehairless (hr) gene, shed hair soon after birth and the remaining hairdisplays a feeble structure [16,17].

In the present study, we noticed a unique expression pattern ofthe Krtap11-1 protein in the impaired hair cortex in Foxn1nu mice,which observation was pursued in order to investigate thephysiological role of this protein. Using the previously establishedorgan culture system, the causal effect of the Krtap11-1 protein onthe development and morphogenesis of the hair shaft wasmanifested. The experimental results provide direct evidence forthe crucial function of a Krtap and shed light on the molecularmechanism underlying the formation of rigid and resistant hairshafts.

2. Materials and methods

2.1. Animals and cell culture

ICR, Balb/c, Foxn1nu (Balb/c-background) and Hr-1 mice (5week-old) and Wister rats (7 week-old) were purchased fromJapan SLC. Functionally normal human keratinocyte HaCaT andmouse fibroblast cell line g6 were maintained in DMEM/HamF12medium (DH) supplemented with 10% fetal bovine serum alongwith penicillin and streptomycin (DH10), and the humansquamous carcinoma HSC-1 was maintained in DH mediumsupplemented with 20% fetal bovine serum. In some experiments,20 mg/ml of the calcium ionophore A23187 (Sigma) were added toHaCaT cells to induce keratinization, as described previously[18,19]. In order to trace the normal development of hair follicles inculture, embryonic upper lips in 13-day fetuses of ICR mice weretransfected with an expression plasmid, microsurgically isolatedand placed on porous membranes with the epidermal side up. Theexplant assembly was then floated on DH 10 medium andincubated for five days (for details, see Fig. S1), as the result ofwhich the vibrissae primordia in the skin rudiments underwentnormal morphogenesis into complete hair follicles containingmature hair, as have been shown previously [20]. All of theexperiments using animals were conducted in conformity with thepolicies and procedures of the Institutional Animal Care and UseCommittee of Kwansei Gakuin University.

2.2. Recombinant Krtap11-1

To construct the prokaryotic expression vector for Krtap11-1,the cDNA encoding mouse Krtap11-1 (NCBI-GeneID; 16693)connected with ATGCATCATCATCATCATCAT at the 50 end wasamplified by PCR and subcloned into a Pet3a vector. Therecombinant Krtap11-1 protein with 6X His-tag (r-Krtap11-1)was produced in the E. coli strain BL-21. After induction of theexpression of the transgene product, the bacterial cells werecollected and sonicated in 5 M guanidine hydrochloride. The cellswere then centrifuged at 8000 rpm for 20 min and the resultantpellet was washed twice with 5 M guanidine hydrochloride andPBS. Thus obtained insoluble fraction appeared to contain r-Krtap11-1 with undetectable amount of bacterial proteins asjudged by electrophoresis. To completely reduce the intra- andinter molecular disulfide bonds in r-Krtap11-1, the insolublefraction was re-suspended and sonicated so as to becomedispersed in 1 mg/ml TCEP/HCl (PIERCE). After incubation for2 h, Maleimide-PEG2-biotin (PIERCE) was added to the suspension

(final 200 mg/ml) and incubated for additional 2 h, by which thereactive sulfhydryl groups in cysteine residues were blocked andlabeled with biotin. After washed several time with PBS, thelabeled Krtap11-1 probe was dissolved in 4 M urea and filtrated.The oligomer formation of thus prepared Krtap11-1 probe wasundetectable in western blot analysis under the non-reducingcondition. To prepare the control probe, bacterial cells producinggreen fluorescent protein with 6X His-tag (r-GFP) were similarlytreated. Purified r-GFP, which had been labeled with biotin, wasadded to the control probe solution in order to adjust theconcentration of the probes.

2.3. Antibodies and the immunodetection

Anti-mouse Krtap11-1 polyclonal antibody was prepared byrepetitive immunization of the protein band of the recombinantKrtap11-1 in SDS-PAGE gel into 7 week old-Wister rats. Otherantibodies used for immunostaining and Western blotting includethose against T7-tag (Novagen), HA-tag (Sigma), hair keratins andanagenic inner root sheath (IRS)-specific protein Trichohyalin/AHF[21]. The secondary antibodies labeled with Cye-3, FITC or HRPinclude those against rat, mouse and rabbit IgG (Amersham).Streptavidin labeled with HRP or Cy3 (Amersham) was used tovisualize the biotinylated Krtap11-1 probe that bound to the targetmolecules. Immunohistochemistry, Immunocytochemistry andWestern blotting were carried out as described previously [22].For immunohistochemistry, cryosections (8 mm thick) of mouseupper lip skin or cultured embryonic skin rudiments on slide glasseswere treated with �20 8C methanol for 10 min. For immunocyto-chemistry, cells cultured on a chamber slide (Thermo Scientific,Japan) coated with 0.1% gelatin were fixed with 4% paraformalde-hyde and permeabilized with �20 8C methanol. For some samples,the treatment with 4% paraformaldehyde was omitted to clearlyvisualize the keratin filaments assembled in the cells. The tissuesections on slide glasses, cultured cells on chamber slides or PVDFmembranes onto which protein bands in SDS-PAGE gel had beentransferred, were incubated successively with Tris-buffered saline(TBS) containing 5% skim milk (STBS), primary antibody solution inSTBS and the secondary antibodies in STBS, with an extensivewashing with TBS at each interval. The cell nuclei were counter-stained with DAPI (Sigma–Aldrich) in some of the cell samples.

2.4. Expression plasmids and the transfection

To construct the eukaryotic expression vectors of Krtap11-1,the cDNA for the mouse Krtap11-1 gene, connected with or withoutsynthetic DNA for HA-tag (MYPYDVPDYA) and/or T7-tag(MASMTGGQQMGRI) at the N-terminus, and for the nuclearexport signal (CDLTLRELPPLQL) at the C-terminus was amplified byRT-PCR. These cDNA were subcloned into the EcoRI site in pQCXIN(Invitrogen). To construct a plasmid for the transient expression ofKrtap11-1 in cultured upper lip skins, T7 tagged-Krtap11-1 wasintroduced into the EcoRI site in an EGFP expression construct inwhich the sequence encoding DsRed in pIRES2DsRed (Clontech)had been replaced with that encoding EGFP (a gift from S.Yamamoto of Kyoto University). In this construct, the IRESsequence was sandwiched between the krtap11-1 and egfp genesso that cell populations expressing the Krtap11-1 gene weredetectable by the green fluorescence in culture. The transfectioninto the cultured cells was performed using lipofectamine plus(Invitrogen) following the manufacturer’s protocol and that intoembryonic skin rudiments was performed by electroporation(100 V, pulse width; 50 ms, pulse interval; 100 ms, cycle; 6 times)using an electroporator (Nepa gene; model CUY21pro-vitro,electrode; CUY520), which has been reportedly applied to ex vivo

transfection [23].

S. Fujimoto et al. / Journal of Dermatological Science 74 (2014) 39–47 41

2.5. Analysis of molecules that are bound by Krtap11-1 lacking active

sulfhydryl groups

Several different tissues including hair, ear skin, liver, lung andheart were collected from ICR mice (5 weeks) and treated with 8 Murea/5 M thiourea/5% DTT at 37 8C for 12 h. The tissue extractswere then centrifuged at 15,000 rpm for 3 min to remove anyinsoluble materials. Each well of a 96 well plate was coated withthe prepared tissue extracts (20 ml), dried in air and washedseveral times with TBS. The wells were incubated successively withSTBS for 1 h, a biotinylated probe in STBS (10 mg/ml) for 1 h andstreptavidin–HRP (Amersham) for 1 h, with extensive washingwith TBS at each interval. Then the wells were treated with TMBSoluble Reagents (Scy Tek). After 10–15 min incubation, theabsorbance at 450 nm was measured to quantify the amount ofthe probe bound to the tissue extract. To identify the Krtap11-1binding partners in the hair components, PVDF membranes, onto

Fig. 1. Expression of Krtap11-1 protein. (A) cryosections (8 mm thick) of the dorsal skin of

(red) and IRS-specific protein Trichohyalin/AHF (green). Without primary antibodies, no

the typical shape of the hair. Bar: 100 mm. Lower, the ratio of hair follicles showing the

several mice were analyzed for each category. In Foxn1nu mice, irregular Krtap11-1 expr

normal mice were sonicated in 200 ml of SDS-containing and 2-mercaptethanol-free bu

materials (ISF) were corrected, treated in 200 ml of the same buffer supplemented with 5

portion of Krtap11-1 was extracted in a non-reducing condition and detected as doublet b

with HA-tagged Krtap11-1 expression construct (Krtap11-1) or the empty plasmid (empt

non-reduced condition. b-actin was used for the loading control. Exogenous Krtap11-1d

These analyses were performed three times and the results were the same.

which protein bands from extracts or E. coli expressing Krtap11-1and type I hair keratins had been transferred, were stained with theKrtap11-1 probe and HRP-labeled streptavidin (Amersham).

2.6. Generation of recombinant form of type I hair keratins

The total RNA was extracted from the mouse upper lip skinusing an RNeasyTM mini kit (Qiagen, Germany) and was reverse-transcribed with an RNA-PCR kit (Takara, Japan). Then the cDNAsencoding K31 (NCBI-GeneID; 16660), K33a (NCBI-GeneID;705800), K33b (NCBI-GeneID; 705939), K34 (NCBI-GeneID;16672) and K39 (NCBI-GeneID; 237934) were amplified by PCR.The primer pairs used for this were listed in Supplementary Fig. 2.To generate the cDNA for a K31-K39 chimeric protein, cDNAencoding the N-terminal head domain of K31 with Kpn I restrictionsite, and that encoding the central Rod domain and C-terminal tailwith the Kpn I restriction site were generated independently by

normal (a and b), Foxn1nu (e and f) or HR-1 mice (g and h) was stained for Krtap11-1

signals were detected (c and d, 2nd Abs only). The insets in some of the panels show

concentric and asymmetric localization of Krtap11-1. At least 19 hair follicles from

ession was observed in almost half of the hair follicles. (B) Left, 20 hair follicles from

ffer and the solubilized fraction was applied (non-reduced, a and b). The insoluble

% 2-mercaptethanol and the solubilized fraction was applied (reduced, a’ and b’). A

ands. Insoluble materials also contained Krtap11-1. Right, g6 cells were transfected

y) and the Krtap11-1 protein was analyzed with anti-HA antibody in the reduced or

etected in the non-reduced condition revealed the dimer formation of this protein.

S. Fujimoto et al. / Journal of Dermatological Science 74 (2014) 39–4742

PCR using the primer pairs listed in Supplementary Fig. 2. Aftertreatment with Kpn I, these cDNAs were ligated and the cDNAencoding K31-K39 chimeric molecule was amplified by PCR. Thusprepared cDNAs were subcloned into the prokaryotic expressionvector Pet3a that had an ATG initiation codon introduced followedby a 6 X CAT sequence in order to fuse a His-tag at the N-terminusof the transgene product. The recombinant form of the type Ikeratins was produced in E. coli strain BL-21.

2.7. Statistical analyses

Results are expressed as the mean � SD of three independentexperiments. Data were analyzed using the t-test, and p-values < 0.05 were considered statistically significant.

3. Results

3.1. Expression of the Krtap11-1 protein in the hair follicle

In spite of there being several reports on the mRNA expressionpattern of several Krtap proteins, very few articles have addressed

Fig. 2. Krtap11-1 can specifically bind to extracts from hair in the absence of active su

Recombinant Krtap11-1 was treated with TCEP-HCl and Maleimide-PEG2-biotin to yield

(GFP) was used as the control. (B) Left, the amount of biotinylated probes bound to surfa

fine hair), liver, lung or heart of normal mice was quantified. The binding of the Krtap11-

*p < 0.05. Right, the Krtap11-1 probe, but not the GFP probe bound to a cross section of h

hair extract. ~, type II keratins; ~, type I keratins; *, Krtaps. The Krtap11-1, but not the G

The faint band corresponding to type II keratins (D) in the left panel might come from a

type I hair keratin K39 that has similar Mw to type II keratins, since none of type II hair

keratins (see Fig. 3). CBB, stained with Coomassie Brilliant Blue. These analyses were p

the exact localization of these proteins in the hair follicle. Thus, westarted with the generation of antibodies against mouse Krtap11-1protein. Although we could not completely exclude the possibilitythat the generated antibodies cross-react also with other Krtaps,we concluded the antibodies are useful to specifically detect theKrtap11-1 protein, provided that the antibodies recognized singleprotein band in the skin, the molecular size of which is the same asof the gene product of Krtap11-1 (Fig. 1B) that is unique in terms ofmolecular size and amino acid sequence as the family members[7]. Also, the antibodies strongly reacted with the nascent corticalregion in the anagenic hair shaft (Fig. 1A), which is consistent withthe expression pattern of Krtap11-1 mRNA in the hair follicles thatis reported previously [10]. In Foxn1nu mice, which have fine andfrizzy hair as a result of the impaired transcription of the Foxn1targets, we found that the expression pattern of the Krtap11-1protein was very often extremely irregular: this protein displayeda concentrically asymmetric localization in the middle and lowerpart of many hair shafts (Fig. 1A). Concomitant with Krtap11-1’sabnormal localization, Trichohyalin/AHF, a marker of the anagenic/catagenic inner root sheath [24] also showed the somehowaberrant distribution in this mouse strain (Fig. 1A). In contrast, in

lfhydryl groups. (A) Schema for the generation procedure of the Krtap11-1 probe.

the biotinylated Krtap11-1 probe. A recombinant form of green fluorescent protein

ces coated with extract of hair, skin (from the ear, which contains a small amount of

1 probe to hairs was only statistically significant compared to the GFP probe. N = 5.

air cortex. Bar: 100 mm. (C) The detection of Krtap11-1-bound protein bands in the

FP probe, bound to the protein band corresponding to type I hair keratins and Krtaps.

non-specific background or a very weak binding to an exceptionally high molecular

keratins have a domain similar to the Krtap11-1-binding domain in ceratin type I

erformed three times and the results were the same.

S. Fujimoto et al. / Journal of Dermatological Science 74 (2014) 39–47 43

the hair follicles of HR-1 mice, in which hairless gene disruptionleads to a defective cyclic progression of hairs [16], the Krtap11-1protein exhibited a normal distribution pattern, in despite ofhaving a hairless phenotype similar to that of Foxn1nu mice(Fig. 1A).

3.2. Disulfide cross-link of Krtap11-1 in the skin

The Krtap11-1 protein in the hair-containing skin that was notcompletely released in 2% SDS buffer was further released whenthe same buffer contained the reducing agent 2-mercaptethanol(Fig. 1B). When fibroblastic g6 cells, that are free of any keratins,were transiently transfected with T7-tagged Krtap11-1, most ofthe transgene product was extracted into 2-mercaptethanol-freebuffer in monomeric and dimeric form; the latter disappeared inthe reducing condition, suggesting that disulfide cross-linksparticipate not only in the incorporation of Krtap11-1 into hairkeratin filaments but also in the covalent self-assembly (Fig. 1B).

3.3. The cohesive self-assembly of Krtap11-1 and its association with

type I hair keratins

While Krtap11-1 and keratin are probably cross-linkedcovalently, the Krtap11-1 probe, in which most of the sulfhydrylgroups were inactivated (Fig. 2A), specifically bound to the surfacescoated with the extract from hair (Fig. 2B). In addition, thisKrtap11-1 probe appeared to specifically recognize the hair cortex

Fig. 3. Identification of certain type I keratins as the Krtap11-1 target. (A) Upper, the typ

overall structure and are composed of an N-terminal head, coiled-coil Rod and C-termin

revealed that the homology of the N-terminal domain is exceptionally low in K39 (right pa

(B) Left, Krtap11-1 probe was able to specifically bind to the recombinant form of K31, 33

Krtap11-1 probe to K39 was dramatically increased if K39 had the N-terminal head do

proteins. These analyses were repeated three times and the results were reproducible.

compartment in tissue sections and the major protein bandscorresponding to type I-hair keratin (�40 kDa) and Krtaps(�20 kDa) on a blot (Fig. 2B), suggesting that the sulfhydrylgroups in the Krtap11-1 protein are not required for therecognition of and association with type I keratins and Krtaps,respectively. It is noteworthy that the Krtap11-1 probe enabled aconcentric labeling of a transverse section of hair cortex fromFoxn1nu mice, whereas endogenous Krtap11-1 exhibited anasymmetric distribution pattern, implying that abnormal hairmorphology was caused by the miss-expression of Krtap11-1, butnot of hair keratins. Indeed, the expression of K31, one of Krtap11-1-binding partner (see below), was reported to be normal in thismouse strain [25].

3.4. Krtap11-1 binds to certain type-I hair keratins

We then sought to identify the binding partners of Krtap11-1 byusing the recombinant forms of several mouse type I hair keratinsincluding K31, 33a, 33b, 34, and 39. All the type I hair keratinsshare essentially the same overall structure, which is composed ofa highly conserved N-terminal head, a central Rod domaincontaining the coiled-coil motifs and an unstructured C-terminaltail (Fig. 3A) [26]. We found that a Krtap11-1 probe strongly boundto the recombinant forms of K31, K33a, K33b and K34, but not K39,revealing the existence of a selective association between Krtap11-1 and certain type I hair keratins (Fig. 3B). The characteristicfeature of K39 was found in the N-terminal head domain: the

e I hair keratins used in this study, which include K31, 33a, 33b, 34 and 39, share an

al tail domain (left panel). Comparison of the amino acid sequences in each domain

nel). Lower, the amino acid sequences of N-terminal head domains of these keratins.

a, 33b, 34, but not of K39, produced in the bacterial cells. Right, the reactivity of the

main of K31 (K31/39). CBB, stained with Coomassie Brilliant Blue. *, recombinant

S. Fujimoto et al. / Journal of Dermatological Science 74 (2014) 39–4744

central rod domain, for the filamentous assembly, was highlyconserved with other members, but for the N-terminal head thiswas not the case (Fig. 3A). We found that the reactivity of K39 tothe Krtap11-1 probe was dramatically increased if this moleculehad N-terminal domain (Met1 to Gly50) of another type I hairkeratin K31, demonstrating that the N-terminal head serves as thecritical domain for Krtap11-1 association. In addition, Krtap11-1probe appeared to bind to recombinant form of Krtap11-1 proteinas well, demonstrating the self-assembly potential of Krtap11-1without any disulfide bridges (Fig. 3B).

3.5. The strict control of the expression of Krtap11-1 protein

While a previous report demonstrated that transcription ofthe krtap11-1 gene is severely restricted in the trichocytes at thelower and middle part of the hair cortex [10], we found thatKrtap11-1 may also be post-transcriptionally regulated: whenphenotypically normal human keratinocyte HaCaT cells had theKrtap11-1 transgene stably introduced, the expression of

Fig. 4. Strict control of Krtap11-1 expression at the protein level. (A) HaCaT keratinocytes w

protein in the absence of proteasome inhibitor MG132 (left) or keratinization induction (a a

transfection with the expression construct resulted in the expression of the HA-tagged Krtap

the cell nucleus (c). The nuclear localization was confirmed by confocal microscopy (XZ an

nuclear exporting signal (NES; CDLTLRELPPLQL), the transgene product became detectabl

tagged Krtap11-1 (f) perturbed the filamentous assembly of exogenous T7-tagged hair k

exogenous Krtap11-1 protein was severely suppressed(Fig. 4A). Upon the abrogation of proteasome-dependent proteinelimination or keratinization induction, however, the exogenousKrtap11-1 protein became detectable in the cytoplasm (Fig. 4A).However, the Krtap11-1-positive cell population was still verysmall in this clonal cell mass (Fig. 4A). While the transienttransfection of Krtap11-1 resulted in successful expression inthe form of relatively large granules, the subcellular localizationwas restricted to the cell nucleus, which was confirmed by thefact that addition of the nuclear export signal (NES) to Krtap11-1 altered its localization to the cytoplasm (Fig. 4B). The forciblecytoplasmic expression of NES-tagged Krtap11-1 and a Type Ihair keratin K31 [27] in HSC-1 cells resulted in a severesuppression of keratin If assembly (Fig. 4C). These results notonly implicated a strict control exercised over the expression ofthe Krtap11-1 protein, in addition to its well-recognizedtranscriptional regulation [10], but also suggested a criticalrole of the Krtap11-1 protein in the filamentous assembly ofhair keratin.

ith exogenous HA-tagged Krtap11-1 stably introduced did not express the Krtap11-1

nd b in right panels). Nuclei were counterstained with DAPI. Bar: 20 mm. (B) Transient

11-1 protein (left), but Krtap11-1 formed granules and its expression was restricted to

d YZ images in d). Also, when the Krtap11-1 gene was fused with DNA for an artificial

e in the cytoplasm (arrows in e). Bar: 20 mm. (C) Forcible expression of HA- and NES-

eratin K31 in HSC-1 cells (arrows in g and h). Bar: 20 mm.

S. Fujimoto et al. / Journal of Dermatological Science 74 (2014) 39–47 45

3.6. Forced expression of Krtap11-1 in the hair follicle severely impairs

keratin filament assembly

Finally, we sought to define the physiological role of Krtap11-1in the developmental process and physical properties of hair. Tothis aim, we used an organ culture of the mouse embryonic upperlip rudiments, as described previously [20]. In this culture system,the development and extensive outgrowth of vibrissa hair folliclesencapsulating firm hair shafts took place, along with an increasedexpression of the Krtap11-1 protein (Fig. 5A). When the embryonicskin rudiments were ectopically introduced with an expressionconstruct that possesses the Krtap11-1/IRES/EGFP cassette prior tocultivation, a dramatic alteration of the hair follicle structureensued in the explants. Some of the transgene-expressing cells thatwere surrounded by regularly arranged keratinocyte layers grewas aberrantly large clumps. That Krtap11-1-expressing fluorescentcells were rarely detectable in the dermal and non-trichocyteepidermal compartments in the explant was reminiscent of thestrict expressional control of the Krtap11-1 protein (Fig. 4).Immunostaining of keratin and Krtap11-1 revealed that this clumpstructure was filled with a multi-layering of hair keratin IFs andaccumulation of the Krtap11-1 protein (Fig. 5B). The frequency ofthe appearance of such abnormal hair-like structure was clearlyhigher in tissues transfected with Krtap11-1/IRES/EGFP than EGFPalone (Fig. 5B), suggesting that Krtap11-1 plays a critical role on

Fig. 5. Forcible expression of Krtap11-1 led to abnormal hair keratin assembly in the hair

day 0 and day 5. Extensive growth and development of hair and the follicle were evident.

3 or 4. The analyses were performed three times and a typical expression profile is sho

cultured rudiments on day 5. The embryonic upper lip rudiments were introduced with

cultivation, by which cell populations expressing transgene in the rudiments became rou

cell populations (green) (b and d) in the rudiments. The hair follicles that developed in

abnormal structures composed of hair keratins (f and g) were instructed. (e) Stained with

tagged Krtap11-1 proteins in day 5-rudiments. Lower, the ratio of the abnormal hair fol

exhibited an abnormal phenotype. All the hair follicles from 20 independent samples w

hair IF assembly and/or bundling so as to form the appropriate hairstructure. Taking these results together, Krtap11-1, whichassociates with the N-terminal domain of certain type I hairkeratins and Krtaps without any disulfide bonds may contributenot only to the arrangement of the keratin molecules but also thewell-regulated parallel assembly of the micro/macro fibrils withdisulfide intermolecular bridges, with a consequent impact on thefundamental features of the hair (Fig. 6).

4. Discussion

An irregular expression pattern was detected in the Foxn1nu butnot in Hr-1 mice, both of which produce apparently similar hairlessphenotypes. The main cause of the impaired hair in Hr-1 mice is thetranscriptional defect of a set of genes for the cyclic progression ofthe hair follicle [16,17] and the formation of proper hair structuremay be beyond the control of Krtap11-1 in this mouse strain, giventhe severely impaired hair structure with a nevertheless normaldistribution of this protein. On the other hand, a strikingperturbation of the expression pattern of Krtap11-1 was evidentin the case of the Foxn1nu mice. Although the hair follicles inFoxn1nu mice are apparently normal, they produce and deliverscant and faulty hairs, indicating that the impaired keratinfilamentous assembly occurs in the developmental process ofthe hair. In such an environment, the normalized localization of the

follicles. (A) Left, phenotypic appearances of the upper lip skin (13 g.d.) in culture on

Bar: 150 mm. Right, the Krtap11-1 protein became reproducibly detectable from day

wn. (B) Left, expression constructs for GFP and Krtap11-1. Middle, analyses of the

an expression construct of GFP (a and b) or Krtap11-1-IRES-GFP (c–g) before the

ghly detectable. The phase contrast images (a and c) and GFP-expressing transfected

culture are outlined by dotted lines. (e–g) When Krtap11-1 was forcibly expressed,

hematoxylin/eosin (HE). Bar: 150 mm. Right, upper, the expression of exogenous T7-

licles. In response to the expression of the Krtap11-1 protein, 18 out of 19 follicles

ere analyzed.

Fig. 6. Schematic model of the role of Krtap11-1. Upper, Krtap11-1 interacts with N-

terminal head domain of certain type I hair keratins, which allows densely packed and

longitudinally aligned assembly of keratin fibrils. By cross-linking to keratin molecules

or self-assembly via disulfide bonds, keratin bundles form stable and rigid hair shafts

(upper). Thus, the expression profile of Krtap11-1 may be an important element to

determine the characteristic physical properties of the hair (lower).

S. Fujimoto et al. / Journal of Dermatological Science 74 (2014) 39–4746

Krtap11-1 possibly remedies the hair structure, considering thenormal expression of its target K31 [25]. Provided the Foxn1’s roleas a nexus coordinating the epidermal cell behaviors [25], thistranscription factor may regulate certain mediators responsible fora strict post-transcriptional control of Krtap11-1. These candidatemediators may include Notch, which activates signaling elementsfor hair matrix differentiation [28], and PLCd1, which isindispensable for hair shaft differentiation [29].

As previously shown for other high-sulfur type Krtaps [30–32],Krtap11-1 contains a large number of cysteine residues in its shortbody (12%) that are used for covalent cross-linking with insolublekeratin fibers/bundles and cohesive self-assembly. However, thisstudy also revealed that the association of Krtap11-1 protein withtype I keratins and Krtaps does not require the intermoleculardisulfide cross-links. Presumably, the keratin/Krtap11-1 andKrtap11-1/Krtap11-1 complexes are stabilized by the disulfidecross-links that follow their initial sulfide-independent associa-tion, thus rearranging and reinforcing the hair keratin Ifs.

To confer rigidity upon the hair shaft, the keratin Ifs shouldsubsequently be connected intercellularly via cell-cell adhesionapparatus called desmosomes in the hair cortex [33,34]. A recentstudy revealed that another Krtap member, KAP8.1, appeared toassociate with desmoplakin a scaffold protein for the desmosomalcadherins [35]. Also, a previous study demonstrated that thefunctional absence of desmoglein4, an abundant desmosomalcadherin at the hair cortex, led to hypotrichosis and a beaded hairphenotype [3], which resembles cultured embryonic hair follicles,that have an ectopic over-expression of Krtap11-1. Thus, it ispossible that Krtap11-1 may also play a role in the intercellularconnectivity of the keratin Ifs.

A previous study revealed that human high sulfur KRTAP2 bindsdirectly to the N-terminal head domain in keratin 86 (K86) [30] andwe have shown that the well-conserved N-terminal domain in K31,33 and 34 plays a key role for Krtap11-1 association. Computeranalyses did not detect any distinctive molecular characteristics in

this domain except for its relatively high isoelectric point (PI): type Ihair keratins are the acidic proteins (PI = 4.9–5.4) while this domainis positively charged (PI = 7.47). However, ionic bond may notcontribute to this intermolecular association, since Krtap11-1 is abasic protein (PI = 8.32). In addition, the Krtap11-1 probe stronglyreacted to K31 on a blot even in the presence of 0.5 M urea (data notshown), indicating that hydrogen bond may not be a key element.Although it is thus unclear how this association occurs, hydrophobicinteraction is possibly involved, given that Krtap11-1 exhibits ratherhydrophobic characteristics.

The Krtap11-1 probe, in which most of the SH groups wereblocked, did not bind to the extracts of several different types oftissues, except for hairs, might be connected to the fact that the N-terminal domain of certain type I hair keratins is not conserved inepithelial keratins. On the other hand, Krtap11-1 might cross-linkwith cysteine-rich proteins via disulfide bonding, which wouldaccount for the failure of stable expression in culture. That theartificial cytoplasmic expression of Krtap11-1 and type I hairkeratin in non-trichocyte keratinocytes resulted in the irregularorganization of hair keratin Ifs was pursued in order to investigatethe physiological role of this protein. In a combined application oforgan culture technique with an ex vivo electroporation method, itwas shown that the over-expression of Krtap11-1 severelyperturbs the structure of both hair and hair follicles, such thatrounded clumps composed of piled-up hair keratin Ifs weresurrounded by hair matrix-like epidermal cell layers. AlthoughKrtap proteins are believed to be related with features ofmonilethrix [3], the severely disordered hair shaft in the explantwith the ectopic expression of Krtap11-1 is reminiscent of otherhair diseases, such as Pilitorti, which are accompanied by a defectin the keratin filamentous structure [36–38].

In the organ culture experiments the exogenous expression ofKrtap11-1 was observed almost exclusively in the massive hairclumps and not in the other skin compartments, including bothepidermal and dermal tissues. Consistent with this, the expressionof the Krtap11-1 transgene product in fibroblasts and non-trichocyte keratinocytes was inefficient, even with inhibition ofproteasome-protein elimination machinery and after keratiniza-tion induction. We hypothesize that this translation inefficiencyand/or the protein instability may be due to the internal CQP motifsin the Krtap11-1 sequence, since the transfection of non-keratocyte keratinocytes with an expression plasmid havingDNA for CQP motifs yielded the undetectable amount of thetransgene product under the above described culture conditions(data not shown). Such cell type-dependent translation andstabilization have been reported for other structural proteinspossessing certain repetitive motif in other cell types [11,39–41].

In conclusion, this study shows, for the first time, thephysiological relevance of the role and the control over theexpression of the Krtap11-1 protein, one of Krtaps. The identifica-tions of the molecular elements for regulation of the expression ofKrtap11-1 may afford an approach to pathological defects of hairattributed to the miss-assembly of hair keratins IFs.

Acknowledgements

We are grateful to Dr. Manabe M for HaCaT keratinocyte,Yamamoto S for the expression vector and all members of the Hirailaboratory for helpful discussions. Part of this work was supportedby Grant-in Aid for Scientific Research (KAKENHI 24590365).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the

online version, at http://dx.doi.org/10.1016/j.jdermsci.2013.12.006.

S. Fujimoto et al. / Journal of Dermatological Science 74 (2014) 39–47 47

References

[1] Popescu C, Hocker H. Hair – the most sophisticated biological compositematerial. Chem Soc Rev 2007;36:1282–91.

[2] Popescu C, Hocker H. Chapter 4. Cytomechanics of hair basics of the mechani-cal stability. Int Rev Cell Mol Biol 2009;277:137–56.

[3] Bazzi H, Demehri S, Potter CS, Barber AG, Awgulewitsch A, Kopan R, et al.Desmoglein 4 is regulated by transcription factors implicated in hair shaftdifferentiation. Differentiation 2009;78:292–300.

[4] Gong H, Zhou H, McKenzie GW, Yu Z, Clerens S, Dyer JM, et al. An updatednomenclature for keratin-associated proteins (KAPs). Int J Biol Sci2012;8:258–64.

[5] Rogers GE, Powell BC. Organization and expression of hair follicle genes. JInvest Dermatol 1993;101:50S–5S.

[6] Rogers MA, Langbein L, Praetzel-Wunder S, Winter H, Schweizer J. Human hairkeratin-associated proteins (KAPs). Int Rev Cytol 2006;251:209–63.

[7] Pruett ND, Tkatchenko TV, Jave-Suarez L, Jacobs DF, Potter CS, Tkatchenko AV,et al. Krtap16, characterization of a new hair keratin-associated protein (KAP)gene complex on mouse chromosome 16 and evidence for regulation byHoxc13. J Biol Chem 2004;279:51524–33.

[8] Wu DD, Irwin DM, Zhang YP. Molecular evolution of the keratin associatedprotein gene family in mammals, role in the evolution of mammalian hair.BMC Evol Biol 2008;8:241.

[9] Nam Y, Kim JK, Cha DS, Cho JW, Cho KH, Yoon S, et al. A novel missensemutation in the mouse hairless gene causes irreversible hair loss: genetic andmolecular analyses of Hr m1Enu. Genomics 2006;87:520–6.

[10] Huh N, Kashiwagi M, Konishi C, Hashimoto Y, Kohno Y, Nomura S, et al.Isolation and characterization of a novel hair follicle-specific gene, Hacl-1. JInvest Dermatol 1994;102:716–20.

[11] Marvin KW, George MD, Fujimoto W, Saunders NA, Bernacki SH, Jetten AM.Cornifin, a cross-linked envelope precursor in keratinocytes that is down-regulated by retinoids. Proc Natl Acad Sci USA 1992;89:11026–30.

[12] Benavides F, Oberyszyn TM, VanBuskirk AM, Reeve VE, Kusewitt DF. Thehairless mouse in skin research. J Dermatol Sci 2009;53:10–8.

[13] Mecklenburg L, Tychsen B, Paus R. Learning from nudity: lessons from thenude phenotype. Exp Dermatol 2005;14:797–810.

[14] Mecklenburg L, Nakamura M, Sundberg JP, Paus R. The nude mouse skinphenotype: the role of Foxn1 in hair follicle development and cycling. Exp MolPathol 2001;71:171–8.

[15] Schlake T, Schorpp M, Maul-Pavicic A, Malashenko AM, Boehm T. Forkhead/winged-helix transcription factor Whn regulates hair keratin gene expres-sion: molecular analysis of the nude skin phenotype. Dev Dyn2000;217:368–76.

[16] Panteleyev AA, Paus R, Christiano AM. Patterns of hairless (hr) gene expressionin mouse hair follicle morphogenesis and cycling. Am J Pathol2000;157:1071–9.

[17] Panteleyev AA, van der Veen C, Rosenbach T, Muller-Rover S, Sokolov VE, PausR. Towards defining the pathogenesis of the hairless phenotype. J InvestDermatol 1998;110:902–7.

[18] Kadono N, Miyazaki T, Okugawa Y, Nakajima K, Hirai Y. The impact ofextracellular syntaxin4 on HaCaT keratinocyte behavior. Biochem BiophysRes Commun 2012;417:1200–5.

[19] Okugawa Y, Bascom JJ, Hirai Y. Epimorphin-derived peptide antagonistsremedy epidermal parakeratosis triggered by unsaturated fatty acid. J Der-matol Sci 2010;59:176–83.

[20] Hirai Y, Nose A, Kobayashi S, Takeichi M. Expression and role of E- and P-cadherin adhesion molecules in embryonic histogenesis. II. Skin morphogen-esis. Development 1989;105:271–7.

[21] Takebe K, Oka Y, Radisky D, Tsuda H, Tochigui K, Koshida S, et al. Epimorphinacts to induce hair follicle anagen in C57BL/6 mice. FASEB J 2003;17:2037–47.

[22] Hirai Y, Takebe K, Takashina M, Kobayashi S, Takeichi M. Epimorphin: amesenchymal protein essential for epithelial morphogenesis. Cell1992;69:471–81.

[23] Osumi N, Inoue T. Gene transfer into cultured mammalian embryos byelectroporation. Methods 2001;24:35–42.

[24] Yamamoto S, Hirai K, Hasegawa-Oka Y, Hirai Y. Molecular elements of theregulatory control of keratin filament modulator AHF/trichohyalin in the hairfollicle. Exp Dermatol 2009;18:152–9.

[25] Weiner L, Han R, Scicchitano BM, Li J, Hasegawa K, Grossi M, et al. Dedicatedepithelial recipient cells determine pigmentation patterns. Cell 2007;130:932–42.

[26] Coulombe PA, Omary MB. ‘Hard’ and ‘soft’ principles defining the structure,function and regulation of keratin intermediate filaments. Curr Opin Cell Biol2002;14:110–22.

[27] Takase T, Hirai Y. Identification of the C-terminal tail domain of AHF/tricho-hyalin as the critical site for modulation of the keratin filamentous meshworkin the keratinocyte. J Dermatol Sci 2012;65:141–8.

[28] Potter GB, Beaudoin 3rd GM, DeRenzo CL, Zarach JM, Chen SH, Thompson CC.The hairless gene mutated in congenital hair loss disorders encodes a novelnuclear receptor corepressor. Genes Dev 2001;15:2687–701.

[29] Potter CS, Pruett ND, Kern MJ, Baybo MA, Godwin AR, Potter KA, et al. The nudemutant gene Foxn1 is a HOXC13 regulatory target during hair follicle and naildifferentiation. J Invest Dermatol 2011;131:828–37.

[30] Fujikawa H, Fujimoto A, Farooq M, Ito M, Shimomura Y. Characterization of thehuman hair keratin-associated protein 2 (KRTAP2) gene family. J InvestDermatol 2012;132:1806–13.

[31] Shimomura Y, Ito M. Human hair keratin-associated proteins. J InvestigDermatol Symp Proc 2005;10:230–3.

[32] Thibaut S, Cavusoglu N, de Becker E, Zerbib F, Bednarczyk A, Schaeffer C, et al.Transglutaminase-3 enzyme: a putative actor in human hair shaft scaffolding?J Invest Dermatol 2009;129:449–59.

[33] Bazzi H, Christiano AM. Broken hearts, woolly hair, and tattered skin: whendesmosomal adhesion goes awry. Curr Opin Cell Biol 2007;19:515–20.

[34] McGrath JA, Wessagowit V. Human hair abnormalities resulting from inher-ited desmosome gene mutations. Keio J Med 2005;54:72–9.

[35] Matsunaga R, Abe R, Ishii D, Watanabe SI, Kiyoshi M, Nocker B, et al. Bidirec-tional binding property of high glycine-tyrosine keratin-associated proteincontributes to the mechanical strength and shape of hair. J Struct Biol 2013.

[36] Rakowska A, Slowinska M, Kowalska-Oledzka E, Rudnicka L. Trichoscopy ingenetic hair shaft abnormalities. J Dermatol Case Rep 2008;2:14–20.

[37] Rogers M. Hair shaft abnormalities, Part I. Aust J Dermatol 1995;36:179–84.quiz 185–6.

[38] Rudnicka L, Olszewska M, Rakowska A, Slowinska M. Trichoscopy update2011. J Dermatol Case Rep 2011;5:82–8.

[39] Ku NO, Omary MB. Keratins turn over by ubiquitination in a phosphorylation-modulated fashion. J Cell Biol 2000;149:547–52.

[40] Schneider MR. MicroRNAs as novel players in skin development, homeostasisand disease. Br J Dermatol 2012;166:22–8.

[41] Zhou P, Bogacki R, McReynolds L, Howley PM. Harnessing the ubiquitinationmachinery to target the degradation of specific cellular proteins. Mol Cell2000;6:751–6.