Laser capture microscopy in a study of expression of structural proteins in the cuticle cells of human hair

Laser capture microscopy in a study of expression ofstructural proteins in the cuticle cells of human hair

George Rogers1 and Kenzo Koike2

1School of Molecular and Biomedical Science, University of Adelaide, Adelaide, SA, Australia;2Beauty Research Center, KAO Corp., Tokyo, Japan

Correspondence: George Rogers, School of Molecular and Biomedical Science, University of Adelaide, Adelaide, SA 5005, Australia,

Tel: 61-8-8303 4624, e-mail: [email protected]

Accepted for publication 31 October 2008

Abstract: The cuticle of human hair consists of several layers of

flat cells that are hardened through their content of cross-linked

proteins and protect the hair structure from the environment.

Known proteins in the cuticle are the sulphur-rich KAP 5 and

KAP10 proteins located in the exocuticle and cross-linked by

disulphide bonds. Isopeptide bonds are also present and led to a

proposal from amino acid analysis that the surface of cuticle cells

also contains keratinocyte cell envelope proteins, loricrin,

involucrin and small proline-rich proteins that contribute to the

stability of the hair cuticle. Confirmation of that proposal by

protein chemical methods is difficult because of the insolubility of

the surface membranes. In the previous studies by other authors,

involucrin was not detected in the cuticle by in situ hybridization

or by immunoelectron microscopy with specific antibodies.

An alternative approach was undertaken to determine whether

mRNAs encoding keratinocyte envelope proteins are expressed in

cuticle cells in the human hair follicle. The study utilized

dissection of the cuticle, cortex and inner root sheath layers from

follicles by laser capture microscopy. RNA was isolated and

subjected to PCR analysis with specific primers to detect

expression of mRNAs encoding cell envelope proteins. Their

presence in the cuticle was not detected, and it was concluded

that the proteins they encode are not produced. The structural

consequences including the possibility that KAPs 5 and 10 are the

prime components cross-linked by both disulphide and isopeptide

bonds are discussed.

Keywords: cross-linking – hair cuticle – involucrin – laser

capture – loricrin

Please cite this paper as: Laser capture microscopy in a study of expression of structural proteins in the cuticle cells of human hair. Experimental

Dermatology 2009; 18: 541–547.

Introduction

The cuticle of a mature human hair consists of several

layers of imbricated flat cells. Within each flattened cell,

three layers can be delineated by transmission electron

microscopy (TEM) – the outermost A-layer, exocuticle and

endocuticle (Fig. 1a). The cells are hardened through their

content of cross-linked proteins within the exocuticle and

afford protection for the hair structure from the environ-

ment. In contrast to the exocuticle, the endocuticle is nei-

ther hardened nor resistant to chemical or proteolytic

attack and is generally considered to be cytoplasmic residue

although the calcium-binding protein S100A3 has been

detected (1). The first proteins to be described in the cuti-

cle of wool fibres were the sulphur-rich KAP 5 and KAP10

located in the exocuticle (2–4) and the outermost A-layer

(Fig. 1b). Although KAP 5 and KAP10 are cross-linked by

disulphide bonds, there is abundant evidence that isopep-

tide (c-glutamyl-e-lysine) bonds are present in the cuticle

and in the keratinized cell membranes of the hair cortex

(5). Analysis has shown that the epicuticle, the chemically-

resistant layer that is released from the outermost surface

of the hair cuticle by treatment with chlorine [the Allwor-

den reaction (6)], is highly resistant to degradation (5,7)

because of isopeptide bonds located in or adjacent to the

exocuticle. Calculations from the amino acid composition

of the epicuticle led Zahn et al. (8) to propose that the

outermost layers of cuticle cells, that are the epicuticle

membrane and part of the underlying A-layer, consist of

keratinocyte cell envelope proteins, including involucrin,

loricrin and small proline-rich proteins (cornifin). It was

proposed that involucrin would be the source of the iso-

peptide cross-linking and loricrin a source of disulphide

bonds. Investigation of the validity of this hypothesis by

protein chemical methods is difficult because of the insolu-

bility of the membranes. An in situ hybridization study by

de Viragh et al. (9) of involucrin in the hair follicle did not

detect the presence of that protein in the cuticle and the

DOI:10.1111/j.1600-0625.2008.00825.x

www.blackwellpublishing.com/EXDOriginal Article

ª 2009 The Authors

Journal compilation ª 2009 Blackwell Munksgaard, Experimental Dermatology, 18, 541–547 541

search for the expression of involucrin and loricrin proteins

by immunoelectron microscopy using specific antibodies

did not reveal their presence (10).

With these conflicting views in mind, we undertook the

alternative approach of seeking the expression in cuticle

cells of mRNAs that encode the keratinocyte envelope pro-

teins at stages of hair growth, when mRNA and protein

synthesis would be optimal. This approach utilized the dis-

section of the cuticle layer in follicles by laser capture

microscopy (LCM). The follicle layers of cortex and inner

root sheath (IRS) on either side of the cuticle layer were

also dissected to compare the abundance of expression

of the same genes. RNA preparations were subjected to

RT-PCR with specific primers and the products analysed

by agarose gel electrophoresis.

Materials and methods

Collection and preparation of skin for follicledissectionScalp skin from patients undergoing regenerative facial sur-

gery was obtained immediately after removal, the hair cut

close to the skin and the skin cut into pieces 5–10 mm2. The

skin pieces were oriented for transverse sectioning dermal

side down in plastic cryomoulds (Tissue-Tek, Sakura, Tokyo,

Japan) containing OCT (Tissue-Tek, Sakura, Tokyo, Japan)

and rapidly cooled by immersion in isopentane held at liquid

nitrogen temperature. Blocks were stored at )80�C until

used. Transverse sections 8–10 lm were cut in CM1800

cryostat (Leica Instruments GmbH, Nussloch, Germany) set

at )30�C. Sections on glass slides were assessed by light

microscopy during 50–100 lm stages of progression through

follicles from the dermal side of the skin. When levels were

reached where the Henle layer was differentiating and the

cells of the Huxley layer had abundant trichohyalin granules,

the cuticle layer was judged to be developmentally advanced

and suitable for laser dissection. Frozen sections were placed

onto Leica membrane (2.5 lm) covered metal slides and

temporally stored on aluminium foil on a bed of dry ice.

Staining of frozen sections and laser capturedissection of folliclesIt was necessary to dehydrate and lightly stain the sections

on the Leica membranes to distinguish adequately the three

follicle layers: the cortex, cuticle and IRS, for dissection.

The metal slides with sections at )80�C were plunged into

70% ethanol at room temperature for 10 min, rapidly

taken down to water with agitation and stained by a modi-

fication of the SACPIC procedure (11) with Mayer’s haem-

alum for 30 s, water rinse for 1 min, picroindigocarmine

for 15 s, rinsed in water differentiated in 70% ethanol,

dehydrated in two changes of 100% ethanol for 10 min

and air dried (He-PIC staining). Nuclease-free water was

used throughout the staining procedure.

The Leica AS LMD laser capture microscope was used in

these experiments. Stained sections were first scanned with

the 4 · objective to locate follicles that were suitably ori-

ented and at the optimal follicle level for dissection. Three

tissue fractions the cortex, cuticle and IRS, respectively

designated as A, B and C were then dissected from the

selected follicles in that order by the laser using the 40 ·objective. The 40 · objective was obligatory because follicle

layers are relatively small structures. It was mostly necessary

to use a high laser intensity setting greater than 40, the aper-

ture between 1 and 4 and the speed at 3–4. The successive

dissections of the three tissues were each collected into 60 ll

Endo

Exo

A layer

1 µm

Surface membrane

KAPs 5&10 (Envelope proteins?)Disulphide and isopeptide bonds

Exocuticle

Endocuticle

Fatty acid monolayer surface

KAPs 5&10Disulphide bonds

Cytoplasmic protein residuesS100A3 calcium binding protein

EpicuticleA-layer

(a)

(b)

Figure 1. (a) Transmission electron micrograph of a transverse section

through five cuticle cells of human hair. In one of the cells the electron-

dense A-layer, the exocuticle (exo) and endocuticle (endo) are indicated

as well as the intercellular membrane (arrow). (b) Line diagram (not to

scale) summarizing what is known about the layers within a cuticle cell

of a hair. The outermost layer is the hydrophobic monolayer of fatty

acid molecules (principally, 18-methyleicosanoic acid). These molecules

are linked to the surface membrane and ⁄ or the A-layer by thioester

bonds. Evidence indicates that these outermost protein layers are cross-

linked by isopeptide bonds and predicted (see text) to consist of the

proteins of keratinocyte cell envelopes. The epicuticle that is liberated

from the surface by chlorine-water (the Allworden reaction) probably

consists of the surface membrane observed by electron microscopy (19)

and at least part of the A-layer that is degraded by the chlorine. As

discussed in the text, the A-layer contains sulphur-rich proteins KAP5

and KAP10 as does the exocuticle layer. The endocuticle is less well

characterized although a specific protein S100A3, a calcium binding

protein has been detected (1).

Rogers and Koike

ª 2009 The Authors

542 Journal compilation ª 2009 Blackwell Munksgaard, Experimental Dermatology, 18, 541–547

of lysis buffer containing 1% v ⁄ v b-mercaptoethanol (Qiagen

RNeasy micro kit, Qiagen, Australia) and if necessary stored

overnight at +4�C or at )20�C for longer periods, before

RNA isolation. Problems with LCM that reduced the capture

of tissue fragments were experienced. Inefficient laser cutting

occurred because of residual OCT and was solved by ensur-

ing that there was adequate washing of the sections. It was

also important to ensure that the stained sections were dry,

and so sections were stored over silica gel at room tempera-

ture. Sometimes, the collection cups did not capture the dis-

sected fragments apparently caused by electrostatic charging.

Electron and light microscopic comparisonsTo establish that the developmental stage of the cuticle

routinely sought in the LCM was appropriate for RNA

analysis sections at several follicle levels were compared by

light and electron microscopy.

Human scalp skin 1 mm · 1 mm containing anagen folli-

cles was fixed for 1.5 h in 1% w ⁄ v glutaraldehyde ⁄ 0.2% w ⁄ vpicric acid in PBS pH 7 fixative. The skin piece in 70% v ⁄ vethanol was further separated into pieces under a dissecting

microscope containing two to three follicles. These pieces

were passed through ethanol changes to 100% v ⁄ v, then 50%

v ⁄ v London White resin ⁄ ethanol 16 h, 100% London resin

8 h and finally embedded in the resin with orientation for

transverse sections in gelatin capsules and cured at 50�C for

48 h. Sections for light microscopy and TEM were collected

at intervals of 100–200 lm along the follicle to compare the

stained light microscope and TEM images. Several sections at

a particular level were cut at 700 nm on a Reichert Ultracut

microtome (Leica, Vienna, Austria) and collected on mem-

brane carbon-coated grids, stained in 4% w ⁄ v aqueous uranyl

acetate for 15 min and examined in a CM100 TEM (Philips,

Eindhoven, Netherlands). This was followed by the cutting

of several sections at 2 lm for light microscopy. The thick

sections were collected on a water drop on a microscope

slide, dried and stained at 70�C for 15 s with 1% w ⁄ v aque-

ous toluidine blue, mounted in DePex and images digitally

recorded in a Zeiss Axiophot microscope (Jena, Germany).

Preparation of RNAs and cDNAsRNA was prepared from the lysis buffer volumes from each

of the dissected fractions, cortex, cuticle and IRS, using

the RNeasy micro kit (Qiagen). On-filter incubation with

DNaseI was used to remove DNA contamination. Each

of the final eluted volumes of 12 ll RNA solution was

stored at )80�C.

The concentrations of the respective RNAs were too low

to be measured specrophotometrically. For the synthesis of

single-stranded cDNAs, 10 ll of each of the RNA prepara-

tions were mixed with 1 ll 0.5 lg ⁄ ll oligoDT 12–18

primer (Invitrogen), heated at 70�C for 5 min and then

incubated with 4 ll first strand buffer (Invitrogen), 1 ll

10 mm d-NTP mix (Sigma, St. Louis, Missouri, USA), 2 ll

0.1 m DTT (Invitrogen) and 1 ll Superscript II (Invitrogen,

Carlsbad, California, USA) for 90 min at 42�C, diluted to

60 ll after incubation with nuclease-free water and stored

at )20�C. It was found that RNA prepared from the cor-

tex ⁄ cuticle ⁄ IRS of 20–30 dissected follicles was necessary

for subsequent RT-PCR analysis.

Epidermis RNARNA from the neonatal human epidermis was the control

for the experiments on the hair cuticle because all of the

keratinocyte cell envelope proteins, and hence their mRNAs

are present in that tissue. Fresh neonatal foreskin was imme-

diately placed in RNAlater (Ambion-Austin, Texas, USA) for

transfer to the laboratory, and the epidermis was dissected

from the dermis after placing the skin pieces in 10 mm

Tris-HCl at 65�C for 45 s (Dr J. Rothnagel, personal com-

munication). The epidermal tissue was homogenized in an

Eppendorf tube with a pestle, the RNA extracted using the

RNeasy micro kit (Qiagen) including on-filter incubation

with DNaseI to remove DNA contamination and transcribed

into cDNA as described for the follicle fractions.

RT-PCR analysisHuman sequences for primers to detect expressed genes

were synthesized (GeneWorks, Adelaide, Australia) and are

given in Table 1. They were selected using the Primer 3

program (http://primer3.sourceforge.net/) and mRNA

sequences were sourced from the NCBI site.

For RT-PCR reactions, the standard protocol consisted

of 2 ll of cDNA, 0.5 ll of each primer strand

Table 1. Specific primer sequences, melting temperature (Tm) and

product size

Gene Sequence TmAmpliconsize (bp)

Ribosome S27a 5¢-CCAGGATAAGGAAGGAATTCCTCCTG 64 2965¢-CCAGCACCACATTCATCAGAAGG-3¢

Loricrin 5¢-CCAGGGTACCACGGAGGCGAAGGA 68 2045¢-TGAGGCACTGGGGTTGGGAGGTAG 64

Involucrin 5¢-GATGTCCCAGCAACACACAC 58 2315¢-TGCTCACATTCTTGCTCAGG 60

Trichohyalin 5¢-ATGGGTCGGTTTGTTTAATGAC 60 2255¢-TGGGCTGATTTTACAGGAAGTT 60

Transglutaminase1 5¢-CCAGTGGGCAGAATCTGAA 60 1525¢-CCAGGGGTTGAAGAGGATGT 60

Cornifin (SPRR1b) 5¢-CATTCTGTCTCCCCCAAAAA 60 1725¢-ATGGGGGTATAAGGGAGCTG 60

KAP5.5 5¢-CACACCAGTGCTTCCGAAACT-3¢ 57 1535¢-GCTGTCAGGGTCTA AGGGGTCT-3¢

KAP 9.2 5¢-GCAGACAGTCGTGGGGTAGT 60 2245¢-GCCCAACTTGCTGTCAAAAC 60

Sciellin 5¢-TCCCAGGGAATCACTACAGG 60 2295¢-CAGGGCGTTTCTTTATCCAA

S100A3 5¢-ACATTCCCGAAACTCAGTCG 60 2265¢-ACACCCGAACTGGTCAACTC

Structural proteins of the hair cuticle

ª 2009 The Authors


(10 pmol ⁄ ll), 10 ll Sybr Premix ExTaq (Takara, Bio Inc.

Shiga, Japan) and 7 ll water. A 2 ll water control (minus

cDNA) was included. The reactions were performed in a

RotorGene 6 real time RT-PCR platform (Corbett Research

Australia, NSW, Australia). The usual conditions were 35

cycles with each cycle of 30 s at primer Tm and extension

at 72�C for 60 s. The products were analysed by electro-

phoresis in 1.5% w ⁄ v agarose ⁄ TAE gels containing ethi-

dium bromide (Sigma). The levels of PCR product using a

primer for S27a (12) enabled an assessment of the relative

expression of that housekeeping gene in the dissected tis-

sues, cortex, cuticle and IRS (fractions A, B and C).

Results

The follicle level at which laser dissection could capture

cuticle cells close to terminal differentiation was determined

by comparing sections that were cut consecutively by elec-

tron and light microscopy (Fig. 2a,b). It can be seen that

the exocuticle layers in all of the 8 cuticle cells are almost

fully formed although the A-layer is not evident. This can

be accounted for by the fact that only section staining with

uranyl acetate was used, whereas the A-layer is observed

after bulk fixation of the follicle tissue with osmium tetrox-

ide. The levels shown in Fig. 2a,b are the maximum used,

whereas the dissections covered all levels of the follicle from

above the bulb where the cuticle is clearly differentiated

and the Henle layer partly or completely differentiated.

Accordingly, gene expression over several follicle levels of

cuticle development was available for analysis.

Typical stages of laser dissection and collection in the

order of cortex, cuticle and IRS (fractions A, B and C)

from stained sections are shown in Fig. 3. Although the

staining method enabled visualization of the layers, cross-

contamination between adjacent layers was inevitable

because the minimal width of the laser beam was 4–5 lm.

The result was that highly enriched but not pure fractions

of the individual layers were obtained. The problem was

least for the removal of the cortex because the melanin

present in the cortex (absent from the cuticle) and the

staining of the distinctive histological structure provided a

clear edge for the laser cut. The second cut was more prob-

lematic because distinguishing the cuticle from the closely

apposed IRS cuticle was impossible and is discussed below.

Gel analyses (Fig. 4) were performed on the RT-PCR

products following isolation of RNA from the dissected

fractions and RT-PCR reactions conducted for major cell

envelope proteins. RNA prepared from the neonatal epider-

mis was used to check that the correct size products for

the cell envelope proteins were obtained. In some instances,

epidermal RNA was included as a control in the RT-PCR

reactions. Expression of ribosomal protein 27a (12) was

chosen as a housekeeping gene control, and primers for

RNAs typical of follicle layers were designed from KAP5.5

(13) for the cuticle, KAP 9.2 (14) for hair cortex and

trichohyalin (15) for the IRS.

As the amounts of RNA isolated from the dissected lay-

ers A, B and C were too low to be measured, the primer

for S27a was used to assess relative cellular content. This

measure of cellular abundance accounted for the area of

IRS CuIRS Cu

Co

5 µm

Hu

(a)

(b)

*

Co

Figure 2. (a) Transmission electron micrograph of the follicle level used

for dissection. The white bracket spans the multicellular cuticle.

Asterisks indicate the outermost and innermost cells of the cuticle. The

dense exocuticle is present in each of the eight cuticle cells. The layers

Hu, Huxley; IRSCu, inner root sheath cuticle; Co, cortex, are indicated.

(b) Light micrograph of a section from the same follicle at a nearby

level stained with toluidine blue. Co, cortex; asterisk, cuticle; bar

10 lm.

Rogers and Koike

ª 2009 The Authors


tissue captured and the level of cytoplasmic activity. Results

from one RNA preparation are presented but several RNA

preparations have been tested during the course of the

study. The cellular content measured in fractions A, B and

C by the S27a primers (Fig. 4a) indicated that cortex was

the highest followed by cuticle and IRS in that order.

The outstanding result was the absence of cuticular

mRNAs for involucrin (Fig. 4b) and loricrin (Fig. 4c). This

result was confirmed by repeating with different primer

sets. The validity of the result was also supported by strong

signals for epidermis where loricrin and involucrin are

major components of the keratinocyte cell envelope (e.g.

loricrin, in E, Fig. 4c).

The signals given for loricrin and involucrin by the cor-

tex suggest that those proteins are present in the mem-

branes of the cortical cells. It is pertinent to note here that

binding of a loricrin antibody to the developing cortex was

observed in an immunochemical study on the wool follicle

(10). The primer for cornifin (SPRR1b) was tested

(Fig. 4d) to see whether it could be a major component of

the cuticle because it is cross linkable by isopeptide bonds.

The RT-PCR result was positive in the epidermal control

but not for cuticle, and gave only a low signal for the IRS,

and was absent from the cortex. The protein sciellin (16),

another candidate protein substrate for transglutaminases,

was sought with a specific primer, but mRNA was not

detected (result not shown).

The mRNA for transglutaminase-3 gave a high signal in

the cortex with much lower signals in the cuticle and IRS

(data not shown), whereas immunohistochemical studies

(17) showed that transglutaminase-3 is present only in the

cortex and cuticle and not the IRS. Confirmation of a low

level of contamination of the cuticle fraction was demon-

strated by using primers for transglutaminase 1 that is only

expressed in the IRS according to Thibaut et al. (17). Com-

parison of the signals displayed on the gel (Fig. 4e) for

cuticle and IRS with the S27a control indicated that the

signal intensity for transglutaminase 1 was indeed the high-

est in the IRS with a low level of contamination of the

cuticle from the IRS cuticle and no signal in the cortex.

To further determine the extent to which the dissected

follicle tissues were representative of the cortex, cuticle and

IRS layers given the cross-contamination, the mRNAs for

the major proteins of those layers were investigated using

specific primers (Table 1). Thus, trichohyalin, the major

component of the IRS showed the highest abundance in

the IRS by RT-PCR (Fig. 4f). The very low trichohyalin sig-

nal in the cuticle fraction can be ascribed to contamination

from the juxtaposed IRS cuticle as described above. The

faint signal for the cortex could be from the expression of

trichohyalin of medulla cells found in some of the hairs.

The diagnostic primer used for the cuticle was KAP5.5

described by Rogers et al. (4) who showed high specificity

for the cuticle by in situ hybridization. The RT-PCR result

(Fig. 4g) for KAP 5.5 included using the same Tm as used

by Rogers et al. (4) showed the expected greatest abun-

dance in the cuticle, but there was lower expression in the

cortex and the IRS fractions and even in the epidermis

control. A similar result was obtained using a different pri-

mer set. The result for the neonatal epidermis control was

unexpected and not in accord with the findings of Rogers

et al. (4,14). Although the reason is unclear, there is one

report in which a serine-rich ultra-high sulphur protein

was expressed in the epidermis of 8-day-old mice (18), so

it is possible that neonatal but not adult epidermis

expresses KAP5 genes.

The signal from the IRS could have arisen from contam-

ination by the juxtaposed sulphur-rich outermost A-layer

of the cuticle. The KAP9.2 primer is specific for the cortex

according to the in situ hybridization data of Rogers et al.

(4,14). The RT-PCR result (Fig. 4h) was a very strong

signal for the cortex and much lower signals for the cuticle

and IRS. The calcium binding protein S100A3 has been

reported (1) to be present in the endocuticle and is the

only functional protein so far shown to be abundant in

that layer. RT-PCR with a specific primer detected expres-

sion of the mRNA in the cuticle (Fig. 4i), but by compar-

ing the signal with that of the S27a control, the expression

was no better as a marker for the cuticle than KAP5.5.

Discussion

This study appears to be the first to apply the technique

of laser capture microscopy to examine human hair folli-

cles for the expression of genes of structural proteins in

*

(c)

*

(b)

(d)

*

(a)

25 µm

Figure 3. Stages of laser dissection of a hair follicle at a level where

the developing cuticle is wider than in Fig. 2b. (a) Follicle cross-section,

asterisk indicates cuticle. Magnification bar 25 lm. (b) After removal of

cortex the cuticle (asterisk) remains. (c) Cuticle cut before capture.

(d) Cuticle removed and inner root sheath cut before capture.


ª 2009 The Authors


the hair fibre and the IRS of the follicle. The approach

using mRNA to detect the expression of all cuticle pro-

teins should be valid because the cuticle cells selected for

dissection covered the range of follicle levels from where

the Henle layer was partly differentiated to an advanced

stage, where the cuticle cells were almost completely filled

with protein.

The applicability of the laser dissection technique for iso-

lating the three follicle layers examined in this study,

despite the occurrence of cross-contamination, was demon-

strated by the dominant expression of the mRNAs for their

specific proteins namely, KAP9.2 for protein of the cortical

matrix, KAP 5.5 for the cuticle and trichohyalin for the

IRS. The RNA isolated from neonatal epidermis served as

an important positive control, as the mRNAs for the major

keratinocyte cell envelope proteins were detected by the

selected primers used in this study.

The most important conclusion to be drawn from the

study is that mRNAs encoding the keratinocyte cell enve-

lope proteins, involucrin, the major protein cross linkable

by isopeptide bonds and loricrin, were not detected in

the cuticle nor was mRNA encoding SPRR1b and sciellin.

These keratinocyte cell envelope proteins, and others

were postulated to be present in the cuticle cells by Zahn

et al. (8).

An earlier study of involucrin expression in the hair

follicle by de Viragh et al. (9) using in situ hybridization

indicated no expression in the cuticle and ruled out its

occurrence in the cortex. The present RT-PCR results sug-

gest that both the cortex and IRS express involucrin. The

(a) (b)

(g)

(c)

(f)

(h) (i)

(e)(d)

Figure 4. (a) S27a ribosomal marker (amplicon size 296 bp); (b) involucrin (amplicon size 231 bp); (c) loricrin (amplicon size 204 bp); (d) cornifin

(SPRR1b) (amplicon size 172 bp); (e) transglutaminase-1 (amplicon size 152 bp); (f) trichohyalin (amplicon size 225 bp); (g) KAP5.5 (amplicon size

153 bp); (h) KAP9.2 (amplicon size 224 bp); (i) S100A3 (amplicon size 226 bp). Arrows indicate size of product from 100 bp DNA ladder. W, water;

A, cortex; B, cuticle; C, IRS; E, neonatal epidermis.

Rogers and Koike

ª 2009 The Authors


detection of involucrin, the protein recognized as the sub-

strate for isopeptide crosslinking, in the cortex, is in accord

with an analysis (5) that showed the presence of isopeptide

bonds after extraction of the keratin proteins with protein

denaturing agents that break disulphide bonds. When kera-

tin proteins were extracted from hair and the soluble and

insoluble fractions analysed by mass spectrometer fragmen-

tation by Lee et al. (19), loricrin or involucrin was not

reported in the insoluble fibre fractions. So, the identity of

proteins forming isopeptide bonds was not forthcoming

from that study.

The proposal that keratinocyte cell envelope proteins are

responsible for the rigidity and toughness of the hair cuti-

cle Zahn et al. (8) must now be modified because the

mRNAs for loricrin, involucrin, cornifin and sciellin were

not detected. At this stage of discovering expressed genes

for structural proteins, it seems possible that the KAP5 and

KAP10 proteins are the only cysteine-rich proteins present

in the cuticle layer (exocuticle and A-layer) and therefore

could play a role equivalent to that of loricrin in the

keratinocyte cell envelope. It is theoretically possible for the

KAP5 (and ⁄ or KAP10) proteins to be cross-linked by

isopeptide bonds in addition to disulphide bonds. Mass

spectrometric analysis, of peptide sequences from insoluble

membrane fractions A-layer membranes that had been

enriched from wool by selective protease treatment and

characterized as mainly A-layer, revealed many members of

KAP5 and KAP10 protein families (20). The cross-links,

between these proteins and the possibility of the presence

of some other protein or proteins that could be isopeptide

cross-linked in the A-layer membranes, have yet to be

investigated. In this respect, the incorporation of non-

specific proteins into the keratinocyte cell envelope and

cross-linking by transglutaminase has been reported (21).

An attractive aspect of the A-layer being composed of

KAP5 and KAP10 proteins is that these cysteine-rich pro-

teins could ‘anchor’ one or more of the fatty acids such as

18-methyleicosanoic acid (22) to the fibre surface by

thioester links. Future use of laser dissection of follicles

in conjunction with microarray and proteomic analysis

could be expected to solve these questions and is under

investigation.

Acknowledgements

The authors thank Dr Gwyn Morgan, Dr Chris Kirby and Dr Richard

Hamilton for the supply of human skin; Dr Simon Bawden, Dr Ravinder

Anand-Ivell and Ms Kee Heng for advice. Facilities for LCM and micro-

scopy were provided by colleagues of Adelaide Microscopy (John Terlet,

Director).

References

1 Takizawa T, Takizawa T, Arai S et al. Ultrastructural localization of S100A3, acysteine-rich, calcium binding protein, in human scalp hair shafts revealed byrapid-freezing immunocytochemistry. J Histochem Cytochem 1999: 47: 525–532.

2 MacKinnon P J, Powell B C, Rogers G E. Structure and expression of genes fora class of cysteine-rich proteins of the cuticle layers of differentiating wooland hair follicles. J Cell Biol 1990: 111: 2587–2600.

3 Powell B C, Rogers G E. The role of keratin proteins and their genes in thegrowth, structure and properties of hair. In: Jolles P, Zahn H, Hocker H, eds.Formation and Structure of Hair. Basel: Birkhauser Verlag, 1997: 59–148.

4 Rogers M A, Langbein L, Praetzel-Wunder S, Winter H, Schweizer J. Humanhair keratin-associated proteins (KAPs). Int Rev Cytol 2006: 251: 209–263.

5 Rice R H, Wong V J, Pinkerton K E. Ultrastructural visualization of cross-linked protein features in epidermal appendages. J Cell Sci 1994: 107: 1985–1992.

6 vonAllworden K. The properties of wool and a new chemical method fordetecting damaged wool. Zeitschrift Angewandte Chemie 1916: 29: 77.

7 Allen CF, Dobrowski SA, Speakman PS, Truter EV. Evidence for lipid and fila-mentous protein in Allworden membrane. Proc 7th Int Wool Text Res Conf1985: 1: 143–151.

8 Zahn H, Wortmann F-J, Hoecker H. Considerations on the occurrence ofloricrin and involucrin in the cell envelope of wool cuticle cells. Int J SheepWool Sci. 2005: 53: 1–13.

9 de Viragh P A, Huber M, Hohl D. Involucrin mRNA is more abundant in humanhair follicles than in normal epidermis. J Invest Dermatol 1994: 103: 815–819.

10 Jones L N, Rogers G E. Protein expression in developing wool fibre cuticle cells.Proceedings 11th International Wool Research Conference. Leeds: Fundamen-tal Wool Science, 2005, p. 108.

11 Auber L. The anatomy of follicles producing wool fibres with special referenceto keratinization. Trans R Soc Edinb 1951; 62 (Part I): 191–254.

12 Anand-Ivell R J K, Relan V, Balvers M et al. Expression of the insulin-like pep-tide 3 (INSL3) hormone receptor (LGBR8) system in the testis. Biol Reprod2006: 74: 945–953.

13 Rogers M A, Winter H, Langbein L, Wolf C, Schweizer J. Characterization of a300 kbp region of human DNA containing the type II hair keratin genedomain. J Invest Dermatol 2000: 114: 464–472.

14 Rogers M A, Langbein L, Winter H et al. Characterization of a cluster ofhuman high ⁄ ultrahigh sulfur keratin-associated protein genes embedded inthe type I keratin gene domain on chromosome 17q12-21. J Biol Chem 2001:276: 19440–19451.

15 Lee S-C, Wang M, McBride O W, O’Keefe E J, Kim I-G, Steinert P M. Humantrichohyalin gene is clustered with the genes for other epidermal structuralproteins and calcium-binding proteins at chromosomal locus 1q21. J InvestDermatol 1993: 100: 65–68.

16 Kvedar J C, Manabe M, Phillips S B, Ross B S, Baden H P. Characterization ofsciellin, a precursor to the cornified envelope of human keratinocytes. Differ-entiation 1992: 49: 195–204.

17 Thibaut S, Candi E, Pietroni V, Meloni G, Schmidt R, Bernard B. Transgluta-minase-5 expression in human hair follicle. J Invest Dermatol 2005: 125: 581–585.

18 Wood L, Mills M, Hatzenbuhler N, Vogeli G. Serine-rich ultra high sulphur pro-tein gene expression in murine hair and skin during the hair cycle. J Biol Chem1990: 265: 21375–21380.

19 Lee Y J, Rice R H, Lee Y M. Proteome analysis of human hair shaft. Mol CellProteomics 2006: 5: 789–800.

20 Bringans S D, Plowman J E, Dyer J M, Clerens S, Vernon J A, Bryson W G.Characterization of the exocuticle a-layer proteins of wool. Exp Dermatol2007: 16: 951–960.

21 Michel S, Schmidt R, Robinson S M, Shroot B, Reichert U. Identification andsubcellular distribution of cornified envelope precursor proteins in the trans-formed human keratinocyte line SV-K14. J Invest Dermatol 1987: 88: 301–305.

22 Jones L N, Rivett D E. The role of 18-methyleicosanoic acid in the structureand formation of mammalian hair fibres. Micron 1997: 28: 469–485.


ª 2009 The Authors


Documents

Laser capture microscopy in a study of expression of structural proteins in the cuticle cells of human hair