Gene 215 (1998) 269–279
Isolation, sequence and expression of the gene encoding
human keratin 131
Ahmad Waseem a,*, Yasmin Alam a, Bilal Dogan b,2, Kenneth N. White c, Irene M. Leigh b,
Naushin H. Waseem d
a Department of Craniofacial Development, Guy’s Dental School, Floor 28, Guy’s Tower, London Bridge, London, SE1 9RT, UK
b ICRF Skin Tumour Biology Unit, The Royal London Hospital, London, UK
c School of Biological and Applied Sciences, University of North London, London, UK
d Paediatrics Research Unit, Division of Molecular and Medical Genetics, UMDS, Guy’s Hospital, London, UK
Received 19 March 1998; accepted 20 May 1998; Received by E. Boncinelli
Keratins are a family of highly homologous proteins expressed as pairs of acidic and basic forms which make intermediate
filaments in epithelial cells. Keratin 13 (K13) is the major acidic keratin, which together with K4, its basic partner, is expressed
in the suprabasal layers of non-cornified stratified epithelia. The mechanism which allows mucosal-specific expression of this
keratin remains unknown. To provide insight into the tissue-specific expression, we have isolated the human K13 gene by screening
a chromosome 17 library with a specific K13 cRNA probe. Sequence analysis of unidirectional deletions produced by transposon
Tn3 has revealed that the gene is 4601 nucleotides long and contains seven introns and eight exons. When driven by the CMV
promoter, the gene produced K13 protein in MCF-7 cells, which normally do not express this protein. Two transcription-start
sites were identified, the major being at 61 and the minor at 63 nucleotides upstream of ATG. The upstream sequence contained
a TATA box and several other putative transcription factor binding sites. A single copy of the K13 gene was detected in the
human genome by Southern hybridisation and polymerase chain reaction. K13 mRNA shows differential expression in cultured
keratinocytes, and in A431 cells the RNA levels remained independent of calcium concentrations in the culture medium.
Characterisation of the human K13 gene will facilitate elucidation of the molecular mechanism regulating K13 expression in
mucosal tissues. © 1998 Elsevier Science B.V. All rights reserved.
Keywords: Chromosome 17; Bacterial transposon; Keratin gene; Retinoid sensitive gene
nails) members. There are two distinct classes of kera-
tins: type I are small (Mr40–56.5 kDa) and acidic,
whereas type II are large (Mr53–67 kDa) and relatively
basic or neutral (reviewed in Fuchs and Weber, 1994).
Expression of keratins in an epithelial tissue is always
as pairs of type I and type II, because two polypeptides
from each sub-family can form a heterotetramer
(Coulombe and Fuchs, 1990; Stewart, 1993) which acts
as a building block for the keratin cytoskeleton. The
genes encoding individual keratins are expressed in
various combinations along different pathways of epithe-
lial differentiation (Quinlan et al., 1985), such that
certain keratin pairs become characteristic of a tissue.
For example, in stratified squamous epithelium the basal
keratinocytes express K5/K14 as the major keratin pair
(Nelson and Sun, 1983), and when the basal keratino-
cytes leave the basement membrane, the K5/K14 pair is
down-regulated, and during their journey through subse-
Keratins are a family of structural proteins which are
expressed mostly in epithelial cells. The family consists
of 21 epithelial and 10 trichocytic (keratins of hair and
* Corresponding author. Tel:+44 171 9554992; Fax:+44 171 9552704;
1 The nucleotide sequence of human K13 has been deposited in
GenBank under accession number AF049259.
2 Present address: GATA Haydarpasa EGt. Hst. Dermatoloji K1.S.
Yrd., Istanbul, Turkey.
Abbreviations: DMEM, Dulbecco’s modified Eagle’s medium; PCR,
polymerase chain reaction; FCS, foetal calf serum; dNTP, deoxy-
nucleotide triphosphate; RNAsin, bovine pancreatic ribonuclease
inhibitor; MAb, monoclonal antibody; nt, nucleotide; DEPC, diethyl
pyrocarbonate; DMBA, 7,12-dimethylbenz[a]anthracene; TPA, 12-O-
tetradecanoyl-phorbol-13-acetate; RA, retinoic acid; HGMP, Human
Genome Mapping Project.
0378-1119/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved.
A. Waseem et al. / Gene 215 (1998) 269–279
quent layers the cells express K1/K10 in skin (Leigh
et al., 1993) and K4/K13 in mucosal epithelia (Sawaf
et al., 1991) until they reach the surface. In wound
healing and hyperproliferating conditions, such as psori-
asis and hypertrophic scar, expression of the K6/K16
pair is activated (Machesney et al., 1998).
K13 together with K4, its type II partner, forms the
major keratin network of most internal stratified epithe-
lia (Sawaf et al., 1991) and does not normally express
in epidermis, with the exception of penile foreskin (Stoler
et al., 1988), anal epidermis (Van Muijen et al., 1986)
and in regenerating epidermis (Kallioinen et al., 1995).
K13 is transiently expressed in keratinocytes of foetal
epidermis (Van Muijen et al., 1987), but is later replaced
by K10 in adulthood. In most stratified squamous
epithelia K13 is expressed in suprabasal layers (Van
Muijen et al., 1986; Levy et al., 1988), although in
human urothelium K13 has been identified in basal as
well as in suprabasal layers (Moll et al., 1988). K13 is
heavily O-glycosylated in certain cultured keratinocytes
(King and Hounsell, 1989) but not in mucosal tissues
of human or mouse origin (Kuruc et al., 1989).
Although N-acetylglucosamine is the residue involved
in K13 glycosylation, neither the site nor the function
of this post-translational modification has yet been
Amongst the factors reported to influence K13 expres-
sion, the retinoids and calcium are the most studied
(Jetten et al., 1989; Breitkreutz et al., 1993). Retinoids
suppress differentiation of epidermal keratinocytes and
modulate keratin gene expression. Expression of several
keratin genes, including K1, K5, K6, K10 and K14, is
down-regulated (Tomie-Canie et al., 1996) whereas
expression of K13 and K19 is increased following expo-
sure to retinoids (Eckert and Green, 1984). In epidermal
keratinocytes, K13 expression can be induced by expo-
sure of cells to retinoids both as monolayers (Breitkreutz
et al., 1993) and in skin equivalents (Asselineau and
Darmon, 1995). RA can also induce K13 expression
ectopically in human epidermis (Rosenthal et al., 1992).
High calcium concentrations induce stratification and
K13 expression in immortalised and transformed kerati-
nocytes (Sutter et al., 1991a; Breitkreutz et al., 1993).
In tracheal and laryngeal keratinocytes, however, K13
expression, induced by high cell density and calcium,
can be down-regulated by RA (Jetten et al., 1989;
Mendelsohn et al., 1991).
The normal pattern of keratin expression is severely
altered in chemically and virally induced tumours (Sutter
DMBA/TPA-induced epidermal tumours is suppressed,
with a concomitant induction of K13, which is consid-
ered to be an early marker of carcinoma progression
(Nischt et al., 1988). Keratinocytes derived from these
tumours have characteristics similar to those from
internal stratified epithelia (Sutter et al., 1991a).
Contrary to these observations, the K13 gene is not
induced in human epidermal tumours, including those
undergoing malignant transformation (Kuruc et al.,
1989). In squamous cell carcinoma of the oral cavity
and the female genital tract, K13 expression is severely
reduced (Malecha and Miettinen, 1991) with no signifi-
cant influence on K4 expression. The apparent difference
in K13 expression in human and mouse tumours could
be accounted for by species-specific differences in the
structure of the K13 gene.
To understand the molecular mechanism of K13 regu-
lation, we have isolated and sequenced the human K13
gene by screening a chromosome 17 library. We have
shown that the gene is functional and it is differentially
expressed in cultured keratinocytes. This will provide a
sequences involved in the modulation K13 expression
under various physiological conditions.
2. Materials and methods
2.1. Cell lines and antibodies
Antibodies used in this project were monoclonal pro-
duced in mice. K13 specific antibody, MAb IC7, was a
gift from Professor F.C.S. Raemaekers, Maastricht,
The Netherlands. The keratinocytes used were human,
derived from non-cornified and cornified epithelia: A431,
a human vulval carcinoma cell line; KB, an oral epider-
mal keratinocyte; Detroit-562; a pharyngeal carcinoma
cell line; SCC4 and SCC27, derived from squamous cell
carcinoma of the oral cavity; and HeLa, a simple
epithelial cell line. The epithelial cells with the exception
of HeLa were grown in keratinocyte culture medium
obtained commercially (Gibco/BRL, UK). HeLa cells
were grown routinely in DMEM supplemented with
2.2. cRNA probes
To make a specific probe for the 5∞ and 3∞ ends of
K13 we amplified the terminal ends of the K13 cDNA
by PCR using a forward and a reverse primer (Table 1).
A recognition site for EcoRI and another site for HindIII
or BamHI were included in the forward and reverse
primers, respectively, to facilitate cloning of the frag-
ments. The PCR products were digested with the respec-
tive enzymes, purified on agarose gel and ligated into
the corresponding sites of pGEM-4 (Promega, USA).
The inserts in the pGEM constructs were confirmed by
DNA sequencing. The K13-5∞ construct contained a
338 bp fragment (nt 1–338) from the 5∞ end, and the
K13-3∞ construct contained a 159 bp fragment (nt
1473–1631) from the 3∞ end of the K13 cDNA (Mischke
A. Waseem et al. / Gene 215 (1998) 269–279
Sequence of primers employed in the PCR amplification
Amplification of intron 1
Amplification of intron 2
Amplification of intron 4
Amplification of introns 6 and 7
a The cloning sites incorporated in the primers are shown in italics.
b The PCR amplified products of K13 cDNA were sub-cloned in pGEM-4 to make constructs used for cRNA probes (Section 2.2).
c The first four primers were used for making the K13 specific probes and the rest were used to amplify specific introns present in the cosmid and
human genomic DNA (Section 3.5).
et al., 1989). The 18S RNA probe was a 209 bp fragment
(nt 1295–1503) derived from the mouse cDNA and
inserted into the PstI site of pGEM-1 vector. The mouse
cDNA for 18S RNA is 98.8% homologous to the
The cRNA probes were prepared by in vitro transcrip-
tion of the pGEM constructs using a commercially
available kit (Promega, USA). The pGEM–K13-5∞ and
pGEM–K13-3∞ constructs were linearised with EcoRI
and used for in vitro transcription with T7 RNA poly-
merase following the instructions supplied by the manu-
facturer. The 18S RNA construct was linearised with
BamHI and used with T3 RNA polymerase for the
preparation of cRNA probe. The probes were freed of
unincorporated radioactivity on spin columns (Bio-Rad,
USA) before use in blotting experiments.
UK) and after crosslinking with UV (120,000 kJ/cm2),
the membrane was hybridised with 3∞- or 5∞-specific K13
probe. When membranes were to be hybridised with the
full-length K13 cDNA, the probe was made by labelling
the K13 cDNA with random hexamers in the presence
of 32P dCTPs using Klenow fragment (Pharmacia
Biotech, UK). The pre-hybridisation was carried out
using either Quick-hyb solution (Strategene, USA) or
Rapid-hyb (Amersham, UK), or using a solution
described elsewhere (Sambrook et al., 1989). The hybrid-
isation using a cRNA or cDNA probe was carried out
at 65°C or 42°C, respectively, for 4–16 h. The membrane
(Sambrook et al., 1989) and exposed to a film at −70°C
for the appropriate length of time. When the radioactive
signal was weak, the blots were developed on a phospho-
photo-image analyser using the Storm 840 system
(Molecular Dynamic, USA).
2.3. Screening of genomic library and Southern
2.4. Cloning and sequencing of genomic clones
A genomic library, prepared by cloning partially
digested human chromosome 17 DNA into laurist-4
cosmid vector, was provided to us by Dr G. Zehetner,
Imperial Cancer Research Fund, Lincolns Inn Fields,
London, UK. Individual clones in this library were
immobilised on a nylon membrane as a grid, such that
the X and Y co-ordinates of a positive clone can be
easily determined. This library has been used previously
for isolating genes located on human chromosome 17
(Ben-Arie et al., 1994). The library was screened with a
3∞-specific and a 5∞-specific K13 probe.
DNA isolated from cosmid clones or from human
leukocytes (Boehringer Mannheim, Germany) was
digested with restriction enzymes and separated on 0.8%
agarose gel. The DNA fragments were transferred from
the agarose gel to Hybond N+ membrane (Amersham,
The unidirection deletions were prepared using Tn3
mutagenesis followed by m13 helper phage-mediated
rescue of single-strand DNA, as described previously
(Davies and Hutchison III, 1991). Briefly, the method
for mutagenesis was as follows. First the target DNA
was sub-cloned into pUNC-19+ (a kanamycin-resistant
plasmid with an f1 origin of replication) in the sense
(pUNC-K13F) and anti-sense (pUNC-K13R) orienta-
tions and transformed into Escherichia coli strain
RDP146 harbouring pLB101 plasmid which can express
Tn3 transposase (Davies and Hutchison III, 1991).
After selection with kanamycin and chloramphenicol,
the bacteria containing both constructs were identified
and mated with NS2114Sm cells containing the transpo-
son pOX38::mTn3 plasmid. This step brings the three
A. Waseem et al. / Gene 215 (1998) 269–279
plasmids together in the same cell, which induces a
transposition event whereby the transposon integrates
unidirectionally across the target DNA. A further
mating step with F(−) NS2114Sml (cre) cells brings
about the action of resolvase, which resolves the
co-integrates into separate deleted plasmids. Single-
strand DNAs were prepared from these plasmids with
DH5aF∞ and M13CO8 as helper phage. The nucleotide
sequence data were assembled using the software
Autoassembler (ABI, USA).
buffer at 55°C for 30 min. The hybridised primer was
extended with 50 units of Expand reverse transcriptase
(Boehringer Mannheim, Germany) in the presence of
extension buffer at 42°C for 1 h. The reaction mixture
was treated with deoxyribonuclease-free ribonuclease,
precipitated with ethanol and separated on a 6% poly-
acrylamide sequencing gel. To determine the size of the
extended product, the labelled primer was used to
sequence pUNC-K13F and run along side the extended
product. The gel was analysed by phosphophotoimager
on a Storm 840 system (Molecular Dymanics, USA).
2.5. RNA preparation and primer extension
2.6. Northern hybridisation
Primer extension was carried out using cytoplasmic
RNA purified from A431 cells using a commercially
available kit (Qiagen, Germany). Total RNA for north-
ern blotting was purified from growing cells using a
guanidium thiocyanate/phenol cocktail available com-
mercially (Genosys, UK). The ratio of A260/280for the
RNA preparation varied between 1.7 and 1.9.
The method used for primer extension was essentially
that described previously (Waseem et al., 1992) with
several modifications. Briefly, about 30 mg cytoplasmic
RNA from A431 cells was hybridised with 50 ng of
K13-specific 33P-labelled primer (CACCCCCGAAA-
CCACCTCCATAGCTGGCAGAGGA) in annealing
Total RNA isolated from six different cell lines, A431,
Detroit 562, KB, SCC4, HeLa and SCC27, was sepa-
rated on two identical agarose gel in the presence of
formaldehyde as described elsewhere (Sambrook et al.,
1989). The RNA in the gel was transferred onto Hybond
N+ membrane (Amersham, UK) and after blocking in
prehybridisation buffer (50% formamide, 5×Denhardts
reagent, 0.5% SDS, 2×SSC) for 16 h, one of the mem-
branes was hybridised with K13-3∞ cRNA, whereas the
other membrane was hybridised with a cRNA probe for
18S RNA. The membranes were washed according to
the standard protocol and exposed to a film at −70°C
Fig. 1. Southern analysis of the cosmid ICRFc105B05170. Purified DNA from the cosmid ICRFc105B05170 was digested with six different restriction
enzymes and after separation and subsequent transfer, the nylon membrane was probed with the cRNA probes derived from the 5∞ end and 3∞ end
of the K13 cDNA. The cRNA probes were made by in vitro transcription of the linearised constructs containing the fragment from the respective
end of the cDNA. Lane 1, EcoRI; lane 2, HindIII; lane 3, BamHI; lane 4, ApaI; lane 5, HincII; and lane 6, SacI. The molecular size of l HindIII
markers are given on the left.
A. Waseem et al. / Gene 215 (1998) 269–279
overnight. Hybridisation with the 18S RNA probe was
used as a control for the quantity of RNA loaded onto
and the fragments were analysed with riboprobes specific
for the 5∞ and 3∞ ends of the K13 cDNA. One of the 11
clones, ICRFc105B05170, when digested separately with
EcoRI, HindIII, ApaI, HincII and SacI, gave a blotting
pattern with the K13-5∞ probe which was different from
that produced by the 3∞ end probe (see Fig. 1). When
the same blot was probed with the full-length K13
cDNA a pattern composite of the 5∞ and 3∞ end probes
was obtained (not shown). However, digestion of the
clone ICRFc105B05170 with BamHI produced a 9.6 kb
fragment which reacted with the 5∞ and the 3∞ end probe
of the K13 (see Fig. 1), suggesting that this fragment
could contain the entire K13 gene. The presence of two
fragments in the ApaI digest detected with the 3∞ probe
(Fig. 1, lane 4), and two fragments each in the HindIII
and SacI digests detected with the 5∞ probe (Fig. 1, lanes
2 and 6), was due to internal restriction sites for
respective enzymes in the cDNA fragments used to
prepare the cRNA probes. The EcoRI digest, when
probed with the 3∞ end cRNA probe, also gave two
fragments (Fig. 1, lane 1), which was perhaps due to
incomplete digestion since the probe did not contain an
internal EcoRI site and, on other occasions, the digest
produced only one band.
2.7. Transfection and immunostaining
The construct to be transfected was prepared by
ligating the target DNA into pCDNA-3, a eukaryotic
expression vector with a CMV promoter. The plasmid
DNA was purified on a caesium chloride gradient in
presence of ethidium bromide. The purified DNA was
used to transfect cells using ClonfectinTM following the
instructions supplied by the manufacturer (ClonTech,
USA). About 48–72 hours post-transfection, the cells
were washed with PBS and fixed in a 1:1 (v/v) mixture
of acetone and methanol, and dried in air before incubat-
ing with primary antibody for 1 h at room temperature
or overnight at 4°C. The cells were washed with PBS
and incubated with an appropriate dilution of fluores-
cein-labelled goat anti-mouse antibody for 1 h at room
temperature. After washing, the cells were mounted in
PBS/glycerol and visualised in a Zeiss fluorescence
2.8. Other methods
3.2. Sequencing strategy
Nucleotide sequencing was routinely carried out man-
ually by the dideoxy method using [35S]ATP as radiola-
belled nucleotide. A few DNA fragments were also
sequenced on an ABI Prism 377 DNA sequencer (ABI,
USA). PCR amplifications were carried out using
the Expand Long Template PCR system (Boehringer
(Machesney et al., 1998). Oligonucleotides were purified
by polyacrylamide gel electrophoresis followed by
reverse phase chromatography on Sep-pak columns
(Sambrook et al., 1989). Purified oligonucleotides were
labelled by T4 polynucleotide kinase (Promega, USA)
as recommended by the manufacturer.
To determine whether the isolated K13 gene was
indeed the gene corresponding to the published K13
cDNA (Kuruc et al., 1989), the 9.6 kb fragment excised
by BamHI was sub-cloned into pUNC-19+ in the
forward, pUNC-K13F, and reverse, pUNC-K13R, ori-
entations. To make unidirectional deletions of the 9.6 kb
fragment, the pUNC-K13 constructs were subjected to
E. coli transposon Tn3-based mutagenesis (Davies and
Hutchison III, 1991) as described in Section 2.4. A total
of more than 300 sequences were generated using this
procedure. When in doubt, both strands were sequenced
or an internal primer was used to confirm the sequence.
On analysis it was noticed that each sequence contained
a small segment of the vector which was identical in all
cases. Once the vector sequence was removed, various
contigs aligned into a single continuous sequence of
about 5.7 kb (see Fig. 2).
The disadvantage of this strategy is that certain
sequences are resistant to transposon integration as has
been reported earlier (Davies and Hutchison III, 1995).
However, the ease with which mutants are generated
mitigates the disadvantage that one may face with some
nucleotide sequences. In the K13 gene we did find a few
regions resistant to transposon integration and these
were sequenced using internal primers.
3. Results and discussion
3.1. Isolation of a genomic clone containing the human
In situ hybridisation using metaphase chromosome
spreads has determined the precise location of the K13
gene on the long arm of chromosome 17, 17q12–17q21.2
(Romano et al., 1992). A genomic library, prepared by
cloning partially digested human chromosome 17 DNA
into laurist-4 vector, was screened with a riboprobe
derived from the 3∞ end of K13 cDNA. Out of 18,432
independent clones screened, 11 gave a positive reaction
with the K13 riboprobe. The DNAs from the positive
clones were digested with several restriction enzymes
3.3. Structure of the human K13 gene
When we compared our sequence with the K13 cDNA
and the deduced amino-acid sequence, we could identify
A. Waseem et al. / Gene 215 (1998) 269–279
Fig. 2. Nucleotide sequence of the human K13 gene. The consensus sequences of different contigs generated by transposon mutagenesis were aligned
and the positions of introns were identified by comparison with the published sequence for K13 cDNA (Kuruc et al., 1989). The deduced amino-
acid sequence is given in single letter codes. The exon/intron junctions are marked by arrows. The translation termination codon is shown by an
asterisk and the polyadenylation signal is boxed. The nucleotide differences between our sequence and those published K13 cDNA are described
in Section 3.3.
A. Waseem et al. / Gene 215 (1998) 269–279
the coding and non-coding regions of the gene. Our
sequence matched perfectly with the coding and non-
coding regions of the K13 cDNA published by Kuruc
et al. (1989) except for a G?T substitution at nt 4972.
However, there were several differences between our
genomic sequence and the K13 cDNA published inde-
pendently by Mischke et al. (1989). Briefly, there was
an insertion of an extra ‘A’ between nt 494 and 495; a
substitution G?A at 685 which led to an amino-acid
replacement G?D at position 58 in the primary
sequence; four silent substitutions, 3333 C?T, 3470
G?T, 3485 G?T, 3488 G?T, and two deletions, one
between 4970 and 4972 and the other between 5068 and
5070. These differences can be accounted for by reading
error, although it is conceivable that there are polymor-
phic forms of the K13 gene.
The K13 gene consists of a total of 4601 nt divided
into eight exons and seven introns (Fig. 2), which is a
typical feature of type I keratin genes. The first two
nucleotides at the donor splice site were GT in all
exon/intron junctions and the last two nucleotides at
the acceptor site in all intron/exon junctions were AG.
A comparison of the sizes of introns and exons between
K10 (Rieger and Franke, 1988) and K13 shows that
four out of eight exons are identical in size between K10
and K13 from human and mouse origins. Exon 6 in
human and mouse K13s is identical in size, but is larger
in K10 by three nucleotides. When we compared the 5∞
upstream sequence of the human K13 gene with the
corresponding sequence in the murine gene, we noticed
that the strong sequence homology between exons
dropped sharply upstream of the ATG (not shown).
This may indicate that the two genes are regulated by
species-specific mechanisms (see Section 3.7).
Kuruc et al. (1989) have isolated two K13 cDNAs
from A431 cells, and one of the cDNAs upon translation
produced normal K13 whereas the other produced a
K13 protein with an unusually short tail domain due to
a frame shift causing premature termination of the
protein. When we compared the unusual K13 cDNA
with our gene sequence we discovered that exon 7 was
absent, presumably due to an aberrant splicing event.
The splicing of exon 6 to exon 8 results in a change of
reading frame with premature termination. This unusu-
ally spliced K13 mRNA is presumably very rare and we
were unable to detect this species in this study.
Fig. 3. Transfection of K13 in MCF-7 cells. The 9.6 kb fragment con-
taining the K13 gene was sub-cloned in the unique BamHI site present
in pCDNA-3, a eukaryotic expression vector with a CMV promoter.
The construct DNA was purified and transfected into MCF-7 cells
using Clonfectin@ and 48–72 h later the cells were immunostained with
the K13-specific MAb IC-7 and visualised under a fluorescence
into the BamHI site of pCDNA-3, a eukaryotic expres-
sion vector with CMV promoter, and transfected the
construct into MCF7, a human breast carcinoma cell
line which normally does not express K13. As shown in
Fig. 3, MCF-7 cells expressing the human K13 could be
detected in a background of unstained MCF-7 cells,
suggesting that the 9.6 kb fragment can produce
IC7-positive cells. The staining pattern of IC7 in
transfected cells was that of a typical intermediate
filament network, suggesting that K13 expressed from
the gene was capable of integrating into the existing
MCF-7 keratin network. A similar transfection experi-
ment into fibroblasts did not produce positive staining
(not shown), perhaps because a keratin polypeptide
cannot integrate into vimentin filaments and single
unpolymerised keratins are proteolysed rapidly.
3.5. K13 is present as a single copy in the human genome
In order to demonstrate that the gene we have isolated
is present in the human genome, we followed two
different approaches. First, we amplified five individual
introns, introns 1, 2, 4 and 6 and 7 together, by the
PCR using primer sets flanking the individual introns
(Table 1). The PCR products were identical in size with
the products obtained when the same set of primers was
used on human genomic DNA (not shown). This experi-
ment suggested that the human genome contains the
same intronic sequences as those in the K13 gene isolated
from the chromosome 17-specific library, and that there
3.4. Functional nature of the K13 gene
In order to show that the 9.6 kb fragment containing
the K13 gene can produce a functional protein, we
expressed this fragment in tissue culture cells and
detected the protein using the MAb IC7. Our attempts
to use mouse keratinocytes were not successful because
of the cross reaction of IC7 with mouse K13 (not
shown). We therefore sub-cloned the 9.6 kb fragment
A. Waseem et al. / Gene 215 (1998) 269–279
is unlikely to be another copy of the gene unless it is an
almost identical duplicate.
The second approach that we used was to analyse the
restriction fragments obtained from human genomic
DNA by Southern hybridisation. Human leukocyte
DNA was digested using six different restriction enzymes
including BamHI, HindIII, ApaI, SacI, HincII and KpnI
and the fragments were blotted on a nylon membrane
and reacted with 32P-labelled K13-3∞ cRNA probe. The
hybridisation is shown in Fig. 4. The Southern blot was
ICRFc105B05170 (Fig. 1), except that larger fragment
with ApaI digestion was
ICRc105B05170 this faint band was present because of
an internal ApaI site located towards one end of the
cDNA fragment used to prepare the 3∞ cRNA probe. In
the genomic Southern the signal for this ApaI fragment
was perhaps too weak to be detected. All the restriction
enzymes produced only one fragment with the K13
probe and similar results were obtained when the blot
was reacted with the K13-5∞ probe (not shown). These
data taken together with our PCR results clearly suggest
that the human genome does not contain pseudogenes
or isoforms of K13 gene. Non-functional copies for
other keratin genes, such as K8 (Waseem et al., 1990)
and K18 (Kulesh and Oshima, 1988) and distinct iso-
forms of K6 (Takahashi et al., 1995), have been iden-
tified in the human genome. Although the existence of
another K13 gene with a different intronic structure is
unlikely, the presence of subtle allelic differences or
polymorphic forms of K13 cannot be ruled out from
3.6. Transcription-start site and possible regulatory
absent. In the cosmid
The 5∞ upstream sequence of the human K13 gene is
shown in Fig. 5. The transcription start-site was deter-
mined by primer extension as described in Section 2.5.
On extending the DNA:RNA hybrid with reverse tran-
scriptase, four different products were observed when
the annealing of the primer to the RNA was allowed to
proceed at room temperature (not shown). However,
when the RNA and the primer mixture was heated to
85°C, slowly cooled to 42°C and immediately extended
with the transcriptase, two cDNA products, a major
and a minor band, were observed on a sequencing gel
(arrows in Fig. 6). This may indicate that the 5∞ end of
the K13 gene forms loops induced when the RNA:DNA
hybrid is annealed at lower temperatures. An analysis
of the 5∞ end using the StemLoop program (from the
HGMP Resource Centre, Hinxton, UK) confirmed a
tendency of this region to form folds and loops. The
minor cDNA product was two nucleotides larger than
the major band. It can be argued that the minor product
is due to the presence of capped K13 mRNA. However,
Fig. 4. Human genomic Southern hybridisation. Human leukocyte
DNA (10 mg) was digested separately with HindIII, BamHI, ApaI,
HincII, SacI and KpnI at 37°C for 36 h. The digest was analysed on a
0.8% agarose gel and after transfer the membrane was probed with
33P-labelled cRNA prepared by in vitro transcription of the linearised
K13-3∞ construct. The membrane was washed and exposed to a film
for 24 h at −70°C. Lane 1, HindIII; lane 2, BamHI; lane 3, ApaI; lane
4, HincII; lane 5, SacI and lane 6, KpnI. Each digest contained only
one K13-positive band. The band in lane 3 was too faint and is marked
by an asterisk. The molecular sizes of l HindIII markers are given on
Fig. 5. 5∞ Upstream region of the human K13 gene. The 5∞ upstream
sequence of the K13 gene is shown beginning at theATG and extending
511 nucleotides upstream from this site. Consensus motifs for possible
binding of transcription factors are underlined. The four half site-
motifs specific for retinoids are shown as TRE. The transcription-start
site is indicated by an arrow. The nucleotide positions relative to the
transcription-start site are given on the right.
A. Waseem et al. / Gene 215 (1998) 269–279
role in the activation of keratin genes. At 160 nucleotides
upstream of the TATA box, we identified TGAATCA
which was very close to the TGAGTCA sequence recog-
nised by the proto-oncoprotein AP-1.
Treatment of epidermal keratinocytes, which normally
do not express K13, with retinoids can induce K13
transcription in vivo and in vitro (Rosenthal et al.,
1992). The action of retinoids is normally mediated by
nuclear receptors, RAR and RXR, which in the presence
of ligand bind to a consensus DNA sequence on the
target gene to induce transcription. The K13 gene does
not contain an authentic
(GGTCATGACC); however, there are four motifs at
−18 (GGTGTCC), −250 (GGTGA), −384 (GGTGA-
CC) and at −399 (GGTCG) which can act as half sites
for retinoid action. It is therefore possible that these
sites in co-operation with other transcription factors,
perhaps AP-1 and AP-2 or others located outside the
gene, play a role in the tissue-specific expression of K13.
Fig. 6. Primer extension analysis of K13 mRNA. Highly purified cyto-
plasmic RNA was isolated from A431 cells, annealed with 33P-labelled
primer and extended in the presence (+) or absence (−) of reverse
transcriptase as described in Section 2.5. The labelled primer was also
used to sequence pUNC-K13F containing the K13 gene. The reverse
transcription products and the sequencing reaction were separated on
a sequencing gel and exposed to phospho-image screen for 14 days
before analysing on a Storm 840 image analyser. The arrows indicate
the presence of specific cDNA products in the RT reaction mixture.
The nucleotide sequence in the region and its direction is shown on
the left. The transcription-start side corresponding to the major cDNA
band is shown by ‘+1’.
3.7. Expression of K13 in cultured keratinocytes
To study K13 expression in keratinocytes we analysed
RNA from A431, Detroit 562, HeLa, KB, MCF-7,
SCC4, SCC27 and SVK14 cell lines by northern blotting
using a 32P-labelled K13-3∞ cRNA probe. As shown in
Fig. 7A, we found differing expression of K13 mRNA
in various keratinocytes grown in culture medium with
10% FCS. Specific K13 signal could be detected only in
A431, Detroit 562, KB and SVK14 cells, and the rest
did not give a signal. In KB cells, we detected K13
mRNA but could not detect protein with MAb IC7. It
is therefore possible that K13 expression is not only
regulated at transcriptional but also at post-transcrip-
tional levels. The overall level of K13 mRNA in A431
cells was consistent with the suprabasal expression in
mucosal tissues studied by in situ hybridisation (data
not shown). Studies on mouse K13 have suggested that
it is expressed in epidermal keratinocytes derived from
tumours induced by DMBA/TPA (Nischt et al., 1988).
In these keratinocytes expression of K13 has been shown
to be regulated by hypermethylation of the CpG island
(Winter et al., 1990). In order to see if human K13 may
be regulated via a similar mechanism, we measured the
levels of methylation in A431 cells (which express large
quantities of K13) and MCF-7 cells (which do not
express K13) by digesting genomic DNA from the
respective cell line with a methylation-sensitive enzyme
HpaII. We did not find a detectable difference in DNA
methylation in the two cell types. Furthermore, treat-
ment of MCF-7 cells with 5-azacytidine, which demethy-
lates DNA, did not induce K13 expression (not shown),
suggesting that the regulation of mouse K13 by methyla-
tion could be a species-specific feature.
Since K13 expression is known to change when differ-
entiation is induced by calcium (Breitkreutz et al., 1993),
this is unlikely because capped and uncapped RNA
should differ by only one nucleotide and the reverse
transcriptase enzyme does not extend the guanidine
nucleotide which caps the 5∞ end of the mRNA. Taking
these facts together we conclude that transcription of
the K13 gene starts from two sites, but the site located
at 61 nucleotide upstream of the translation initiation
site produces the majority of the K13 mRNAs (shown
by +1 in Fig. 6).
The 5∞ upstream sequence was screened using the
Transcription Element Search Software (TESS) avail-
able on-line from the HGMP Resource Centre, Hinxton,
UK. At 26 nucleotides upstream of the major transcrip-
tion-start site, a TATAA box was identified. Although
this sequence is not universally present in eukaryotic
genes, it has been identified in other keratin genes (Bader
et al., 1988). There were two AP-2 sites (TGGGGA) in
reverse orientations located on either side of the TATA
box. AP-2 activation has been shown to play an impor-
tant role in epithelial differentiation and activation of
several keratin genes (Wanner et al., 1996; Chen et al.,
1997). We have also identified three motifs located at
11 (ACCAGCCCC), 188 (AGGGCAGTCTC) and 256
(TGGGTGGGC) nucleotides upstream of the TATA
box with strong homology with the SP-1 binding core
element (CCGCCC or GGGCGG). SP-1 binds DNA
with high affinity and has been shown to play a vital
A. Waseem et al. / Gene 215 (1998) 269–279
Fig. 7. Expression of K13 mRNA in cultured keratinocyte cell lines by northern hybridisation. (A) Purified total RNA from different cell lines
was separated on 0.8% (w/v) agarose gel in the presence of formaldehyde and transferred onto a nylon membrane and probed with a cRNA probe
specific for K13 and 18S RNA. (B) A431 cells were grown in 0.5 mM (L) or 1.5 mM (H) calcium for 3 days before RNA was extracted and
processed for northern hybridisation.
we determined whether K13 mRNA expression in A431
cells would also be sensitive to changes in calcium in
the medium. As shown in Fig. 7B, the K13 mRNA
levels in A431 cells grown at low (0.5 mM) and at high
(1.5 mM) calcium concentrations were virtually iden-
tical, despite the fact that the cells clearly changed
intercellular contacts as well as cell shape, which are
characteristic responses of keratinocytes to extracellular
calcium. A431 cells growing in low calcium also
expressed K13 protein as detected by MAb IC7 (not
shown), but under similar conditions other keratinocytes
express little or no K13 (Jetten et al., 1989). These
observations suggest that the effect of extracellular
calcium is cell type-specific and may not be evident in
A431 cells under our experimental conditions.
In conclusion, we have isolated and sequenced the
gene encoding human K13. The gene shows all the
characteristics of a typical type I keratin gene. We have
demonstrated that the gene is functional and in the
human genome there is only a single copy of this gene.
Isolation of the human K13 gene will enable an analysis
of the regulatory elements and the cellular factors which
modulate expression of this gene.
sequence analyses. We also thank Dr R. Evans, Division
of Biochemistry and Molecular Biology, UMDS, for
helpful suggestions and comments. Finally, we would
like to thank the Wellcome Trust (grant no. 037824/1.5
to A.W.) and the Guy’s Special Trustees for financial
support during the course of this study.
Asselineau, D., Darmon, M., 1995. Retinoic acid provokes metaplasia
of epithelium formed in vitro by adult human epidermal keratino-
cytes. Differentiation 58, 297–306.
Bader, B.L., Jahn, L., Franke, W.W., Low level expression of cytokera-
tins 8, 18 and 19 in vascular smooth muscle cells of human umbilical
cord and in cultured cells derived therefrom, with an analysis of the
chromosomal locus containing the cytokeratin 19 gene. 1988. Eur.
J. Cell Biol. 47, 300–319.
Ben-Arie, N., Lancet, D., Taylor, C., Khen, M., Walker, N., Ledbetter,
D.H., Carrozzo, R., Patel, K., Sheer, D., Lehrach, H., North, M.A.,
1994. Olfactory receptor gene cluster on human chromosome 17:
possible duplication of an ancestral receptor repertoire. Hum. Mol.
Genet. 3, 229–235.
Breitkreutz, D., Stark, H.J., Plein, P., Baur, M., Fusenig, N.E., 1993.
Differential modulation of epidermal keratinization in immortalized
(HaCaT) and tumorigenic human skin keratinocytes (HaCaT-ras)
by retinoic acid and extracellular Ca2+. Differentiation 54, 201–217.
Chen, T.T., Wu, R.L., Castro-Munozledo, F., Sun, T.T., 1997. Regula-
tion of K3 keratin gene transcription by Sp1 and AP-2 in differenti-
ating rabbit corneal epithelial cells. Mol. Cell. Biol. 17, 3056–3064.
Coulombe, P.A., Fuchs, E., 1990. Elucidating the early stages of kera-
tin filament assembly. J. Cell Biol. 111, 153–169.
Davies, C.J., Hutchison III, C.A., 1991. A directed DNA sequencing
strategy based upon Tn3 transposon mutagenesis: application to the
ADE1 locus on Saccharomyces cerevisiae chromosome I. Nucleic
Acids Res. 19, 5731–5738.
Davies, C.J., Hutchison III, C.A., 1995. Insertion site specificity of the
transposon Tn3. Nucleic Acids Res. 23, 507–514.
Eckert, R.L., Green, H., 1984. Cloning of cDNAs specifying vitamin
A-responsive human keratins. Proc. Natl. Acad. Sci. USA 81,
Fuchs, E., Weber, K., 1994. Intermediate filaments: structure,
dynamics, function, and disease. Annu. Rev. Biochem. 64, 345–382.
Jetten, A.M., George, M.A., Smits, H.L., Vollberg, T.M., 1989. Kera-
We are grateful to Dr G. Zehetner (ICRF, London,
UK) for the chromosome 17 library, Professor F.C.S.
Raemaekers (Maastricht, The Netherlands) for MAb
IC7, Professor W.W. Franke (Heidelberg, Germany) for
K13 cDNA, Dr C. Davies (University of North
Carolina, Chapel Hill, USA) for Tn3 mutagenesis, Dr
Ian Goldsmith (Oligo Laboratory, Clare Hall, ICRF)
for supplying synthetic oligonucleotides, and to Mrs P.
Purkis (Royal London Hospital, London) and Mr
Anand Lalli (UMDS, London) for technical assistance.
We wish to thank Dr Peter Green, Division of Molecular
and Medical Genetics, UMDS, for his help with
A. Waseem et al. / Gene 215 (1998) 269–279
tin 13 expression is linked to squamous differentiation in rabbit
tracheal epithelial cells and down-regulated by retinoic acid. Exp.
Cell Res. 182, 622–634.
Kallioinen, M., Koivukangas, V., Jarvinen, M., Oikarinen, A., 1995.
Expression of cytokeratins in regenerating human epidermis. Br.
J. Dermatol. 133, 830–835.
King, I.A., Hounsell, E.F., 1989. Cytokeratin 13 contains O-glycosidi-
cally linked N-acetylglucosamine residues. J. Biol. Chem. 264,
Kulesh, D.A., Oshima, R.G., 1988. Cloning of the human keratin 18
gene and its expression in non-epithelial mouse cells. Mol. Cell. Biol.
Kuruc, N., Leube, R.E., Moll, I., Bader, B.L., Franke, W.W., 1989.
Synthesis of cytokeratin 13, a component characteristic of internal
stratified epithelia, is not induced in human epidermal tumors.
Differentiation 42, 111–123.
Leigh, I.M., Purkis, P.E., Whitehead, P., Lane, E.B., 1993. Monospec-
ific monoclonal antibodies to keratin 1 carboxy terminal (synthetic
peptide) and to keratin 10 as marker of epidermal differentiation.
Br. J. Dermatol. 129, 110–119.
Levy, R., Czernobilsky, B., Geiger, B., 1988. Subtyping of epithelial
cells of normal and metaplastic human uterine cervix, using polypep-
tide-specific cytokeratin antibodies. Differentiation 39, 185–196.
Machesney, M., Tidman, N., Waseem, A., Kirby, L., Leigh, I., 1998.
Activated keratinocytes in the epidermis of hypertrophic scars. Am.
J. Pathol. 152, 1133–1141.
Malecha, M.J., Miettinen, M., 1991. Expression of keratin 13 in human
epithelial neoplasms. Virchows Arch. A Pathol. Anat. Histopathol.
Mendelsohn, M.G., Dilorenzo, T.P., Abramson, A.L., Steinberg,
B.M., 1991. Retinoic acid regulates, in vitro, the two normal path-
ways of differentiation of human laryngeal keratinocytes. In vitro
Cell Dev. Biol. 27A, 137–141.
Mischke, D., Wachter, E., Hochstrasser, K., Wild, A.G., Schulz, P.,
1989. The N-, but not the C-terminal domains of human keratins
13 and 15 are closely related. Nucleic Acids Res. 17, 7984
Moll, R., Achtstatter, T., Balcarova-Stander, J., Ittensohn, M.,
Franke, W.W., 1988. Cytokeratin in normal and malignant epithe-
lium. Maintenance of expression of urothelial differentiation features
in transitional cell carcinomas and bladder carcinoma cell culture
lines. Am. J. Pathol. 132, 123–144.
Nelson, W.G., Sun, T.T., 1983. The 50- and 58-kdalton keratin classes
as molecular markers for statified squamous epithelia: cell culture
studies. J. Cell Biol. 97, 244–251.
Nischt, R., Roop, D.R., Mehrel, T., Yuspa, S.H., Rentrop, M., Winter,
H., Schweizer, J., 1988. Abberrant expression during two-stage
mouse skin carcinogenesis of type I 47-kDa keratin, K13, normally
associated with terminal differentiation of internal stratified epithe-
lia. Mol. Carcinog. 1, 96–108.
Quinlan, R.A., Schiller, D.L., Hatzfeld, M., Achtstatter, T., Moll, R.,
Jorcano, J.L., Magin, T.M., Franke, W.W., 1985. Patterns of expres-
sion and organisation of cytokeratin intermediate filaments. Ann.
NY Acad. Sci. 455, 282–306.
Rieger, M., Franke, W.W., 1988. Identification of an orthologous
mammalian cytokeratin gene: high degree of intron sequence conser-
vation during evolution of human cytokeratin 10. J. Mol. Biol.
Romano, V., Raimondi, E., Bosco, P., Feo, S., Di Pietro, C., Leube,
R.E., Troyanovsky, S.M., Ceratto, N., 1992. Chromosomal mapping
of human cytokeratin 13 gene (KRT13). Genomics 14, 495–497.
Rosenthal, D.S., Griffiths, C.E., Yuspa, S.H., Roop, D.R., Voorhees,
J.J., Acute or chronic topical retinoic acid treatment of human skin
in vivo alters the expression of epidermal transglutaminase, loricrin,
involucrin, filaggrin, and keratins 6 and 13 but not keratins 1, 10,
and 14. 1992. J. Invest. Dermatol. 98, 343–350.
Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning—
A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold
Spring Harbor, NY.
Sawaf, M.H., Ouhayoun, J.P., Forest, N., 1991. Cytokeratin profiles
in oral epithelial: a review and a new classification. J. Biol. Buccale
Stewart, M., 1993. Intermediate filament structure and assembly. Curr.
Opin. Cell Biol. 5, 3–11.
Stoler, A., Kopan, R., Duvic, M., Fuchs, E., 1988. Use of monospecific
antisera and cRNA probes to localise the major changes in keratin
expression during normal and abnormal epidermal differentiation.
J. Cell Biol. 107, 427–446.
Sutter, C., Nischt, R., Winter, H., Schweizer, J., 1991a. Aberrant in
vitro expression of keratin K13 induced by Ca2+ and vitamin A acid
in mouse epidermal cell lines. Exp. Cell Res. 195, 183–193.
Sutter, C., Strickland, J.E., Welty, D.J., Yuspa, S.H., Winter, H.,
Schweizer, J., 1991b. v-Ha-ras-induced mouse skin papillomas
tetradecanoylphorbol-13-acetate-induced analogues. Mol. Carcinog.
Takahashi, K., Paladini, R.D., Coulombe, P.A., 1995. Cloning and
characterization of multiple human genes and cDNAs encoding
highly related type II keratin 6 isoforms. J. Biol. Chem. 270,
Tomie-Canie, M., Day, D., Samuels, H.H., Freedberg, I.M., Blumenb-
erg, M., 1996. Novel regulation of keratin gene expression by thyroid
hormone and retinoid receptors. J. Biol. Chem. 271, 1416–1423.
Van Muijen, G.N.P., Ruiter, D.J., Franke, W.W., Achtstatter, T.,
Haasnoot, W.H.B., Ponec, M., Warnaar, S.O., 1986. Cell type
heterogeneity of cytokeratin expression in complex epithelia and
carcinoma as demonstrated by monoclonal antibodies specific for
cytokeratins nos 4 and 13. Exp. Cell Res. 162, 97–113.
Van Muijen, G.N.P., Warnaar, S.O., Ponec, M., 1987. Differentiation-
related changes of cytokeratin expression in cultured keratinocytes
and fetal, newborn, and adult epidermis. Exp. Cell Res. 171,
Wanner, R., Zhang, J., Henz, B.M., Rosenbach, T., 1996. AP-2 gene
expression and modulation by retinoic acid during keratinocyte
differentiation. Biochem. Biophys. Res. Commun. 223, 666–669.
Waseem, A., Alexander, C.M., Steel, J.B., Lane,E.B., 1990. Embryonic
simple epithelial keratins 8 and 18: chromosomal location empha-
sizes difference from other keratin pairs. New Biol. 2, 464–478.
Waseem, N.H., Labib, K., Nurse, P., Lane, D.P., 1992. Isolation and
analysis of the fission yeast gene encoding polymerase delta acces-
sory protein PCNA. EMBO J. 11, 5111–5120.
Winter, H., Rentrop, M., Nischt, R., Schweizer, J., Tissue-specific
expression of murine keratin K13 in internal stratified squamous
epithelia and its aberrant expression during two-stage mouse skin
carcinogenesis is associated with the methylation state of a distinct
CpG site in the remote 5∞-flanking region of the gene. 1990. Differen-
tiation 43, 105–114.
of keratin K13asdo their