Identification of Novel and Known Mutations in the Genes forKeratin 5 and 14 in Danish Patients with EpidermolysisBullosa Simplex: Correlation Between Genotype andPhenotype

Charlotte B. Sørensen,*† Anne-Sofie Ladekjær-Mikkelsen,*† Brage S. Andresen,*† Flemming Brandrup,‡Niels K. Veien,§ Sanne K. Buus,§ Ingrun Anton-Lamprecht,1 Torben A. Kruse,† Peter K.A. Jensen,**Hans Eiberg,†† Lars Bolund,† and Niels Gregersen**Research Unit for Molecular Medicine, Aarhus University Hospital and Faculty of Health Sciences, Skejby Sygehus, Aarhus, Denmark; †Institute ofHuman Genetics, University of Aarhus, Aarhus, Denmark; ‡Department of Dermatology, Odense University Hospital, Odense, Denmark;§Dermatologic Clinic, Aalborg, Denmark; 1Institute for Ultrastructure Research of the Skin, Department of Dermatology, Ruprecht-Karls University,Heidelberg, Germany; ** Department of Clinical Genetics, Aarhus University Hospital, Aarhus Kommunehospital, Aarhus, Denmark; ††GenomeGroup/RC Link, Panum Institute, Copenhagen, Denmark

Epidermolysis bullosa simplex (EBS) is a group ofautosomal dominant inherited skin diseases caused bymutations in either the keratin 5 (K5) or the keratin14 (K14) genes and characterized by development ofintraepidermal skin blisters. The three major subtypesof EBS are Weber-Cockayne, Koebner, and Dowling-Meara, of which the Dowling-Meara form is the mostsevere. We have investigated five large Danish familieswith EBS and two sporadic patients with the Dowling-Meara form of EBS. In the sporadic Dowling-MearaEBS patients, a novel K14 mutation (N123S) anda previously published K5 mutation (N176S) wereidentified, respectively. A novel K14 mutation (K116N)was found in three seemingly unrelated families,whereas another family harbored a different novel K14mutation (L143P). The last family harbored a novelK5 mutation (L325P). The identified mutations were

Epidermolysis bullosa simplex (EBS) is a group of raregenetic skin disorders with intraepidermal cytolyticblister formation on mechanical trauma (Pearson andSpargo, 1961). Three major subtypes and various raretypes all with autosomal dominant inheritance and

additional rarer subtypes have been identified (Gedde-Dahl andAnton-Lamprecht, 1996). The Weber-Cockayne type of EBS(EBS-WC), with blistering restricted to hands and feet, the Koebnertyoe (EBS-K) with generalized blister formation, and the Dowling-Meara type (EBS-DM) characterized by severe involvement andclumping of the basal keratin network (Anton-Lamprecht and

Manuscript received March 27, 1998; revised October 20, 1998; acceptedfor publication October 21, 1998.

Reprint requests to: Dr. Charlotte B. Sørensen, Research Unit forMolecular Medicine, Skejby Sygehus, 8200 Aarhus N, Denmark.

Abbreviations: EBS-DM, Dowling-Meara type of epidermolysis bullosasimplex; EBS-K, Koebner type of epidermolysis bullosa simplex; EBS-WC, Weber-Cockayne type of epidermolysis bullosa simplex.

0022-202X/99/$10.50 · Copyright © 1999 by The Society for Investigative Dermatology, Inc.

184

not present in more than 100 normal chromosomes.Six polymorphisms were identified in the K14 gene andtheir frequencies were determined in normal controls.These polymorphisms were used to show that the K14K116N mutation was located in chromosomes withthe same haplotype in all three families, suggesting acommon ancestor. We observed a strict genotype-phenotype correlation in the investigated patients asthe same mutation always resulted in a similar pheno-type in all individuals with the mutation, but ourresults also show that it is not possible to predict theEBS phenotype merely by the location (i.e., head, rod,or linker domains) of a mutation. The nature ofthe amino acid substitution must also be taken intoaccount. Key words: avoidance of pseudogene co-amplifica-tion/founder effect/skin disease. J Invest Dermatol 112:184–190, 1999

Schnyder, 1982; Anton-Lamprecht, 1994), are the most prevalenttypes of EBS. All EBS subtypes are caused by mutations in eitherK5 or K14, the major keratins expressed in the basal layer of theepidermis (Albers and Fuchs, 1987; Kitajima et al, 1989; Bonifaset al, 1991; Vassar et al, 1991; Lane et al, 1992).

The keratins have been divided into two types according tomolecular weight and isoelectric points (Moll et al, 1982). Thetype I and type II keratin genes are located in clusters onchromosome 17 and chromosome 12, respectively (Bader et al,1988; Lessin et al, 1988; Romano et al, 1988; Rosenberg et al,1988, 1991; Popescu et al, 1989; Waseem et al, 1990). The keratinshave a similar protein structure consisting of a highly conservedalpha-helical rod domain flanked by nonhelical ‘‘head’’ and ‘‘tail’’regions in the N-terminal and C-terminal part of the protein,respectively. Keratin molecules of type I and II readily associate inan obligate fashion into heteromeric molecules that make up theintermediate filaments also known as tonofilaments (reviewed bySmack et al, 1994; Fuchs, 1996).

More than 20 missense mutations associated with EBS disordershave been identified in K5 and K14 (Corden and McLean, 1996;

VOL. 112, NO. 2 FEBRUARY 1999 NOVEL AND KNOWN MUTATIONS IN THE GENES FOR KERATIN 5 AND 14 185

Figure 1. Pedigrees of the five families investigated in this study. A black circle or square illustrates individuals with the disease. Unaffectedindividuals are represented by open symbols. Asterisks denote members of the families that have been investigated for the presence of mutations. Anarrow indicates the index patients.

Fuchs, 1996; Nomura et al, 1996; Stephens et al, 1997; OnlineMendelian Inheritance in Man). In general, these mutations concen-trate in different clusters, namely, the ends of the rod domain (1Aand 2B) and the L12 linker region. In addition, mutations havebeen found in the H1 segment in the head region of K5. It haspreviously been described that there is a correlation between therespective EBS subtype and the location of the K5 or K14 mutations(Letai et al, 1993).

In this study, we have investigated members from five DanishEBS families (Fig 1), three of the WC type, two of the K type,and two sporadic patients with the DM type of EBS. We identifiedone new and one previously identified disease-associated mutationin the K5 gene and three new disease-associated mutations inK14. Furthermore, we investigated if a strict genotype-phenotypecorrelation could be established in our patient material. In addition,a number of different polymorphisms in K14 were identified andtheir frequencies were determined in the normal population.

MATERIALS AND METHODS

Patients Five seemingly unrelated Danish families with cases of EBS-Kor EBS-WC and two sporadic Danish EBS patients were studied. In thetwo sporadic EBS patients, clumping of basal keratins, the ultrastructuralcharacteristic of EBS-DM, was demonstrated by electron microscopy. Inthe EBS-K families, development of bullae was observed at birth or in theneonatal period, located mainly on the hands and feet including the dorsalsurfaces. In adulthood, patients had experienced blistering on other partsof the body, e.g., the thighs or buttocks, provoked by friction. In theEBS-WC families, development of bullae was restricted to hands and feet.The investigated family members are indicated in the pedigrees (Fig 1). Asummary of the clinical and genetic data of the seven index patients fromthis study is listed in Table I.

PCR amplification Genomic DNA was isolated from blood samplesaccording to standard methods (Gustafson et al, 1987) and kept at 4°C.Using genomic DNA as template, PCR amplifications of all exons,including part of the flanking intron sequences, in the genes for K5 andK14 were carried out in an automated thermal cycler (Perkin Elmer, FosterCity, CA). Initially, primer pairs were used for the PCR, which turnedout to result in coamplification of the K14 pseudogene. The presence ofthe pseudogene was evident by sequence analyses of these PCR products,because both the pseudogene and the functional K14 gene sequences werefound in nonidentical regions of the two genes. To avoid this coamplificationof the K14 pseudogene during PCR amplification of the K14 functionalgene, all primer pairs subsequently used for PCR comprised mismatchescompared with the K14 pseudogene sequence. The resulting sequenceswere unambiguously identified as the functional K14 gene sequence.Likewise, primers used for amplification of K5 were designed to preventcoamplification of the homologous gene K6 (Stephens et al, 1997). AllPCR amplifications were carried out under standard conditions in a volumeof 50 µl, except for amplification of K14, for which the PCR reactionvolume contained 4% dimethyl sulfoxide.

The resulting PCR products were subjected to bi-directional sequencingusing M13(–21) forward and M13 reverse primers as well as gene specificprimers. Sequence analyses were performed on an automated ABI 373Asequencer (PE Applied Biosystems, Foster City, CA). The nucleotidesequences of the primers used for PCR amplification of K5 and K14 arelisted below. All references to nucleotides or amino acids in the text arebased upon the cDNA sequence of K5 (Eckert and Rorke, 1988) and thesequence of the K14 gene (Marchuk et al, 1984). The initiating ATGcodon is numbered as bp 1–3, and the initiating methionine is numberedas amino acid 1.

Primers for amplification of K5 containing M13(–21) forward or M13reverse tails (tails are underlined):

Exon 1(1): forward 59-TGTAAAACGACGGCCAGTGAGCTCTG-TTCTCTCCAGCA-39

186 SØRENSEN ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Tab

leI.

Clinic

alan

dgen

etic

dat

aofEB

Sfa

milie

san

din

dex

pat

ients

Fam

ilym

embe

rsFa

mily

mem

bers

Dia

g.In

dex

Pat.

exam

ined

clin

ical

lyte

sted

for

mut

atio

nsM

utat

ion

Res

tr.

Blis

ters

,Fa

mno

.N

o.(t

otal

/affe

cted

)(a

ffect

ed/u

naffe

cted

)G

ene

DN

A/a

min

oac

idE

nz.

EB

Sty

peb

Blis

ter

siteb

age

ofon

setb

Com

men

ts

1(I

I-6)

13/7

7/6

K14

348G

.C

,K

116N

Mae

IIW

CFe

etIn

fanc

yLæ

sø2

(I-2

)20

/11

11/9

K14

348G

.C

,K

116N

Mae

IIW

CH

ands

and

feet

Infa

ncy

Læsø

3(I

II-6

)15

/88/

4K

1442

8T.

C,

L143

PH

paII

KH

ands

,fe

et,

nate

s,th

igh

Neo

nata

l4

(III

-4)

9/3

3/4

K5

974T

.C

,L3

25P

Apa

IaK

Han

ds,

feet

and

else

whe

reN

eona

tal

5(I

II-4

)15

/64/

3K

1434

8G.

C,

K11

6NM

aeII

WC

Han

dsan

dfe

etE

arly

infa

ncy

Læsø

Spor

adic

1K

552

7A.

G,

N17

6S–

DM

Gen

eral

ized

atbi

rth,

Con

geni

tal

incl

udin

gor

alm

ucos

a,th

icke

ning

ofna

ils,

palm

o-pl

anta

rke

rato

derm

aSp

orad

ic2

K14

368A

.G

,N

123S

–D

MG

ener

aliz

edat

birt

h,C

onge

nita

lFo

ster

child

dyst

roph

icna

ils,

palm

o-pl

anta

rke

rato

derm

a

a Art

ifici

ally

intr

oduc

edre

stri

ctio

nsit

e.b T

hecl

inic

alsig

nsde

scri

bed

inth

ein

dex

patie

nts

from

each

fam

ilyw

ere

also

typi

cal

inal

lot

her

affe

cted

mem

bers

ofth

atfa

mily

.

VOL. 112, NO. 2 FEBRUARY 1999 NOVEL AND KNOWN MUTATIONS IN THE GENES FOR KERATIN 5 AND 14 187

reverse 59-CAGGAAACAGCTATGACCCTCCACCGCCGAAACC-AAAT-39

Exon 1(2): forward 59-TGTAAAACGACGGCCAGTGCTATGGCT-TTGGAGGTGGT-39

reverse 59-CAGGAAACAGCTATGACCCCTTCTTTCTCTCTCT-TTGGC-39

Exon 2: forward 59-TGTAAAACGACGGCCAGTCTCTATCTT-CAAACCCTGCT-39

reverse 59-CAGGAAACAGCTATGACCCCATCTGGTACCAAGA-AGAC-39

Exon 3: forward 59-TGTAAAACGACGGCCAGTTGGCCAGAGG-TTCATGCTAC-39

reverse 59-CAGGAAACAGCTATGACCTCAACCTTGGCCTCCA-GCTCC-39

Exon 4: forward 59-TGTAAAACGACGGCCAGTGAGAACCAGC-AGCCTGCAG-39

reverse 59-CAGGAAACAGCTATGACCTGAGGTGTCAGAGACA-TGC-39

Exon 5: forward 59-TGTAAAACGACGGCCAGTATGAGATTAA-CTTCATGAAGATG-39

reverse 59-CAGGAAACAGCTATGACCCCATTCTTAGTGTCGT-CATG-39

Exon 6: forward 59-TGTAAAACGACGGCCAGTAATTTCCATCT-AAACCCAAG-39

reverse 59-CAGGAAACAGCTATGACCTTTAGAACTCAGGCCC-CTTC-39

Exon 7: forward 59-TGTAAAACGACGGCCAGTGACCCAGAAA-CTCAGAAGGA-39

reverse 59-CAGGAAACAGCTATGACCTAGAGCAGCCTCGCTT-TATC-39

Exon 8: forward 59-TGTAAAACGACGGCCAGTTCGAATCATGA-GGATGGGAG-39

reverse 59-CAGGAAACAGCTATGACCTGAGACATAAGCCACAT-TGC-39

Exon 9(1): forward 59-TGTAAAACGACGGCCAGTAAGGGGGT-CCAGTAGAGTGC-39

reverse 59-CAGGAAACAGCTATGACCATTTGACGCTGGAGCT-GCTA-39

Exon 9(2): forward 59-TGTAAAACGACGGCCAGTCCTAGGTGG-TGGGCTCAGTG-39

reverse 59-CAGGAAACAGCTATGACCTTGGGTTCTCGTGTCA-GCAG-39

Primers for amplification of K14 containing M13(–21) forward or M13reverse tails (tails are underlined):

Exon 1: forward 59-TGTAAAACGACGGCCAGTGCAATTTACCC-GAGCACCTTCTCTTCACTCA-39

reverse 59-CAGGAAACAGCTATGACCATCTTAAGGTCTCAGC-GGCCTGGGGCAT-39

Exon 2: forward 59-TGTAAAACGACGGCCAGTAGGCTACAG-TGAAGTCCAGCTTGTGAAGTCCA-39

reverse 59-CAGGAAACAGCTATGACCGGAAACACTGCTCCAA-AAATGCCCTACTCTG-39

Exon 3: forward 59-TGTAAAACGACGGCCAGTGCACTGTGTT-CAACCACGCCATTTTTCAA-39

reverse 59-CAGGAAACAGCTATGACCTCCTGTCTCAGCCTC-CCAAGTAGCTGGG-39

Exon 4: forward 59-TGTAAAACGACGGCCAGTACCAATCCG-CTGCCATGGTGGAACTCCTG-39

reverse 59-CAGGAAACAGCTATGACCGAATGCCATTCACACC-AGAAGGCCCCAGA-39

Exon 5: forward 59-TGTAAAACGACGGCCAGTCTGCCTTCT-GGGGCCTTCTGGTGTGAATG-39

reverse 59-CAGGAAACAGCTATGACCAGTGTGGCCGTTCTCT-CCCTGCCAGTCCT-39

Exon 6: forward 59-TGTAAAACGACGGCCAGTTGCACCCAG-GACTGGCAGGGAGAGAACGG-39

reverse 59-CAGGAAACAGCTATGACCTGAGAGTGCCATGGG-GGGGGCGGACTAAG-39

Exon 7: forward 59-TGTAAAACGACGGCCAGTGTGAAGACGC-GGCTGGAGCAGGAGAT-39

reverse 59-CAGGAAACAGCTATGACCGCCTAGACCTGCTTGG-GGTACAGAGGGTG-39

Exon 8: forward 59-TGTAAAACGACGGCCAGTTTTCCTCAC-CTTCTTGGCCTCCTTACTCCTG-39

reverse 59-CAGGAAACAGCTATGACCAGAGGGGATCTTCCAG-TGGGATCTGTGTC-39

Primers for amplification of K14 cDNA [M13(–21) forward tailunderlined]:

cDNA: forward 59-TGTAAAACGACGGCCAGTGCAATTTACCC-GAGCACCTTCTCTTCACTCA-39

reverse 59-AGGCCTGAGCGGGGCTGGGCAG

Gene-specific primers for sequencing of K14-specific PCR products:Exon 1(1): forward: M13(–21) forward primerreverse: 59-GCCCACCAGAAGCCCATC-39Exon 1(2): forward: 59-GGCTATGGCGGTGGCTTC-39reverse: M13 reverse primer

Mutation specific restriction enzyme cleavage assays A total of 55control individuals (110 alleles) were tested for the K5 L325P mutation.Primers introducing an ApaI-site surrounding the mutated base in exon 5were designed (forward primer 59-ATGTCTCTGACACCTCAGTGG-GCC-39 and reverse primer 59-GTCATCAGAGGGCCCACCTT-GG-39) and PCR was performed, thereby amplifying both the normal andthe mutated allele if present. Restriction enzyme digestion of the resultingPCR products with ApaI confirmed the absence of the K5 L325P mutationin the control individuals.

The K14 mutations K116N and L143P create new restriction sites forthe enzymes MaeII and HpaII, respectively. Affected and unaffectedmembers of family 1 and 2 were analyzed for the K14 K116N mutationby MaeII-digestion of K14 exon 1 amplified products confirming that onlyaffected individuals harbor the mutation. Likewise, affected and unaffectedmembers of family 3 were tested for the K14 L143P mutation by HpaII-digestion of K14 exon 1 amplified products confirming the presence ofthe mutation exclusively in affected relatives.

RESULTS

Linkage analysis Following clinical diagnosis of the EBS patients,four of the five Danish families were tested in a linkage analysisstudy in order to determine if a possible mutation was located inthe gene for K5 or in the K14 gene. In families 1–3, the diseasecosegregated with the human keratin 10 (K10) intragenic marker(Korge et al, 1992; Mischke, 1993) located on chromosome 17q21and did not cosegregate with chromosome 12 microsatellite markersD12S87, D12S85, D12S90, D12S80, and D12S81 covering the K5region, indicating a defect in K14. In family 4, on the other hand,the situation was reversed, suggesting a mutation in the K5 gene.Linkage analysis was not performed in family 5, because this family,as with families 1 and 2, was found to be connected to the smallisland of Læsø in Denmark. Because linkage analysis could not becarried out in the two sporadic EBS-DM patients, mutation-screening analyses were performed for both K5 and K14.

Mutations in the K5 gene K5 mutations were associated withthe disease in family 4 (see pedigree in Fig 1) and in case 1 of thesporadic EBS-DM patients. Patients from family 4, all of whomexhibit clinical signs consistent with the EBS-K type, were foundto be heterozygous for a novel mutation in the K5 gene. Themutation, a T to C transition in exon 5 at nucleotide position 974,results in a change from leucine to proline at codon 325 (L325P)(Table I). Screenings of the unaffected family members by sequen-cing showed that the mutation was absent in these individuals.Fifty-five control individuals were tested for the L325P mutationby a mutation-specific restriction enzyme cleavage assay using theenzyme ApaI. None of the control individuals carried the mutation.

A previously published mutation (Stephens et al, 1997) wasidentified in exon 1 of the K5 gene in case 1 of the sporadic EBS-DM patients. This patient was found to be heterozygous for the Ato G transition at nucleotide position 527, changing an asparagineto serine at codon 176 (N176S) (Table I). As expected, the parentsof this patient did not carry the mutation. The K14 gene sequenceof this patient was normal.

Mutations in the K14 gene In families 1–3 and 5, two novelmutations in exon 1 of the K14 gene were found to be associatedwith the disease. In the second sporadic EBS-DM patient (case 2),a third novel K14 mutation was identified. The index patients infamilies 1 (II-6), 2 (I-2), 3 (III-6), and the sporadic patient (case

188 SØRENSEN ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Table II. KRT14 polymorphisms and frequenciesa

Base change Amino acid change Frequency

6T.C T2T (silent) 0.53189T.C Y63Y (silent)b 0.48193T.C L65 L (silent)b 0.48231C.T S77S (silent)b 0.37280G.A A94Tb 0.48369T.C N123N (silent)b 0.52

aThe frequencies are based upon 60 unrelated control individuals (120 alleles),except for the T2T polymorphism that was tested in 43 individuals (86 alleles).

bThese mutations have been described previously by Marchuk et al, 1984;Coulombe et al, 1991, and Chen et al, 1995.

2) all had a normal K5 gene sequence. The index patient III-4 infamily 5 was not screened for K5 abnormalities.

All 22 patients tested in families 1, 2, and 5 harbored the sameheterozygous base substitution in exon 1 of the K14 gene. Themutation was a G to C transversion at nucleotide position 348,changing lysine 116 to asparagine (K116N) (Table I). This G toC mutation generates a new MaeII site.

All patients from the three families exhibit clinical signs consistentwith EBS-WC. PCR products from the unaffected relatives infamilies 1 and 2 were tested by specific restriction enzyme cleavageassay with MaeII in order to see if any of these individuals carriedthe mutation. The unaffected relatives in family 5 were screenedfor the mutation by sequencing. The K116N mutation was notpresent in unaffected individuals from any of the three families.These three families (1, 2, and 5) all turned out to originate fromthe small island of Læsø in Denmark. Although the families arenot related by the pedigrees available, the identified mutationsuggests that the families are related further back than the pedi-grees show.

In patients from family 3 (see pedigree in Fig 1) a heterozygousT to C transition at nucleotide position 428 was identified, changingleucine 143 to proline (L143P) (Table I). All eight examinedpatients from this family exhibited clinical signs of EBS-K. Themutation generates a new HpaII-site, which was utilized for testingof the unaffected family members. The L143P mutation was notpresent in any of the unaffected relatives in the family.

The second sporadic EBS-DM patient (case 2) carried a hetero-zygous A to G transition at nucleotide position 368, changingasparagine to serine at codon 123 (N123S) (Table I). Because thispatient is a foster child, we were not able to test the biologicparents for the identified mutation.

Screening for the three mutations found in the K14 gene(K116N, L143P, and N123S) was carried out on 60 controlindividuals (120 alleles) by direct sequencing of K14 exon 1-specificPCR products. None of the control individuals harbored themutations.

Polymorphisms in the KRT14 gene During the course ofsequencing the K14 gene in the families, a number of basesubstitutions were identified in affected as well as unaffectedindividuals in exon 1 of K14. A total of six variations was identified(Table II). Five of these variations have previously been identified,namely the T to C transitions at nucleotide positions 189, 193,231, and 369 and the G to A transition at position 280 in the K14gene sequence corresponding to Tyr63, Leu65, Ser77, Asn123, andAla94, respectively (Marchuk et al, 1984; Coulombe et al, 1991;Chen et al, 1995). All these transitions cause silent mutations,except for codon 94, in which the transition changes alanine to athreonine residue. Alanine 94 is located in the nonhelical headregion of K14, which has been found to be dispensable for filamentassembly in in vitro studies (Coulombe et al, 1990). Thus, the A94Tmutation is not expected to have any functional effect on the K14protein. The remaining new variation is found at position 6 andcauses a silent mutation of Thr2 due to another T to C transition.

In order to determine the frequency of these variations, 60unrelated control individuals (120 alleles) were tested by direct

sequencing of K14 exon 1-specific PCR products. The frequenciesof the six variations are listed in Table II, showing that thesevariations are common polymorphisms in the Danish population.

In addition to the above-mentioned polymorphisms, we founda few discrepancies between the published K14 gene sequence(GenBank acc.no. J00124) and our K14 gene sequences in exon1. In all the Danish patients and their healthy relatives examinedin this study, we found the sequence AA instead of TT in thenontranslated region at nucleotide position –33 and –34 upstreamfor the ATG. Likewise, the sequence GC instead of CG was foundat positions 77 and 78, resulting in a change from alanine to glycinein codon 26. This could be due to compression during the originalsequencing of the K14 gene. Alignment of the type I and II keratinsreveals that this position often is occupied by a glycine residue.Finally, a G instead of A was always found at nucleotide position131, changing Asn44 to a serine. Both Gly26 and Ser44 werefound in all the 60 controls.

To confirm that the identified mutations and polymorphisms areexpressed and do not originate from coamplification of the K14pseudogene (Savtchenko et al, 1988; GenBank acc.nos. M22927and M22928), we isolated RNA from keratinocytes and performedoligo-dT cDNA synthesis prior to PCR amplification of the K14cDNA with K14 specific primers. Sequencing of the resulting K14cDNA from both a normal individual and patients representativefor the K14 K116N and L143P mutations, as well as the cDNAfrom the sporadic case 2 (N123S), were carried out, confirmingthe presence of the identified mutations and polymorphisms, aswell as the described K14 sequence discrepancies. Furthermore,we subjected K14 exon 1 PCR products amplified from genomicDNA to digestion with the restriction enzyme TaqI. This enzymecleaves exon 1 in the K14 pseudogene, if present, but leaves exon1 in the functional K14 gene intact. Subsequent PAGE of thedigested PCR products showed no bands corresponding to theexpected length of the pseudogene digestion products, supportingthat the K14 pseudogene was not coamplified during the PCRreaction. In accordance with this, we see no contaminatingpseudogene sequence surrounding the two positions in exon 1,where the pseudogene has 3-nucleotide deletions compared withthe K14 gene sequence.

Variations also appeared in the coding sequence of the K5 genein both affected and unaffected individuals. These variations have,however, not been investigated further.

Determination of haplotypes in EBS families with K14mutations Characterization of the six K14 polymorphismsdescribed above facilitated determination of the haplotypes inaffected individuals carrying the K14 K116N mutation in families1, 2, and 5 from the island of Læsø. In all three families, an identicalhaplotype of the mutant chromosome was found in affectedindividuals indicating a common ancestor with the mutation. Adifferent haplotype of the mutant chromosome was found in family3, harboring the L143P mutation.

DISCUSSION

At present 15 (K5) and 16 (K14) different disease-causing missensemutations have been identified in the genes for K5 or K14 inpatients with EBS, and a correlation between the location of themutation in the keratin gene and the severity of the disease hasbeen suggested (Letai et al, 1993; Humphries et al, 1996; Uttamet al, 1996; Fuchs, 1997; Irvine et al, 1997; Stephens et al, 1997;Online Mendelian Inheritance in Man). Our purpose with thisstudy was to investigate the mutational spectrum in Danish EBSpatients, to evaluate the genotype-phenotype relationship, and inparticular to analyze whether patients from large kindreds withidentical disease-causing mutations always exhibit the same EBStype.

All previously characterized EBS-causing mutations are clusteredin four regions in the keratin 5 and 14 molecules: the H1 domainof the head region (only for type II keratins), the two segments(1A and 2B) of the rod domain, and the linker region L12. Typically,

VOL. 112, NO. 2 FEBRUARY 1999 NOVEL AND KNOWN MUTATIONS IN THE GENES FOR KERATIN 5 AND 14 189

EBS-DM-associated mutations reside in the highly conserved endsof the 1A or 2B segments, whereas EBS-K-associated mutationsare located more centrally in the rod domain. EBS-WC mutationsare most frequently found in the linker region L12, in the H1domain of the molecule (K5), or in the 2B segment (K14).

Consistent with these earlier findings, the mutations identifiedin our two sporadic EBS-DM patients, N176S in K5 and N123Sin K14, are located at homologous positions in the beginning ofthe 1A domain in the two genes. This asparagine residue is highlyconserved in the keratins and other intermediate filament proteinssuch as vimentin, lamin C, and snail intermediate filament protein.

The K14 N123S mutation is situated in a cluster of previouslydescribed K14 mutations (Q120R, R125C/H/S, and Y129D) thatall cause EBS-DM (Coulombe et al, 1991; Chen et al, 1995; Chanet al, 1996). Because both N176S and N123S mutations cause anidentical amino acid substitution, one would expect the functionalsignificance of the two mutations to be similar, resulting in anidentical disease phenotype, as observed in the two sporadic EBS-DM patients in this study. The K5 N176S mutation, however, hasrecently been described in a sporadic patient with clinical findingsconsistent with EBS-DM, except that no clumping of keratinfilaments was seen by electron microscopy (Stephens et al, 1997).Our patient carrying this mutation, as well as our patient with theK14 N123S mutation, both showed the typical clinical features ofthe EBS-DM type and clumping of the basal keratins.

Both our families (3 and 4) with EBS-K have mutations that arelocated more centrally in the rod domain, where mutations associ-ated with EBS-WC have previously been identified. TheK14 L143P mutation observed in family 3 is located at the ‘‘d’’position in the heptad repeat N-X-X-L-E-X-K (a-b-c-d-e-f-g) ofthe 1A segment. This leucine residue is highly conserved amongthe keratins and is thought to have hydrophobic interaction withthe residue at the ‘‘a’’ position in K5, meaning that the proline atposition 143 should interact with asparagine 193 in K5 (Hovnanianet al, 1993). Interestingly, a N193K mutation involving the proposedpartner residue of the substituted leucine 143 has been identifiedin the K5 gene of EBS patients with a milder phenotype (Humphrieset al, 1996).1 In addition, an E144A mutation associated withrecessive EBS-WC has been identified in this region of K14(Hovnanian et al, 1993).

The K5 L325P mutation that we observed in family 4 with theEBS-K type, is also located in a region harboring a cluster ofmutations that all have been observed in patients with EBS-WC.In fact, the L325P mutation is the fifth mutation identified in acluster of K5 EBS-WC mutations (M327T, D328 V, N329K,R331C) situated in the L12 linker region (Rugg et al, 1993; Chanet al, 1994; Matsuki et al, 1995). Leu325 is highly conserved amongthe type II keratins and is homologous to Val270 situated in L12in K14, which has also been found to be mutated in EBS-WC(V270M) (Rugg et al, 1993). The L325P mutation in the L12region is likely to affect the flexibility and structure of this nonhelicalregion, and thereby disrupt the lateral associations between keratinheterodimers. Interestingly, a very recent publication describesanother EBS-K mutation in this region of K5, changing Val323 toan alanine (Galligan et al, 1998).

At first glance, it may seem strange that both our families withthe EBS-K have mutations that on the basis of their location wouldbe suspected to result in the milder localized EBS-WC. The factthat all 12 investigated members of families 3 and 4 have the sameclinical phenotype, namely EBS-K, excludes the possibility ofa misclassification. This observation supports a strict genotype-phenotype correlation in EBS, namely in the sense that the samemutation will always result in a single phenotype in all individualswith the mutation. Furthermore, our findings illustrate that it isnot possible to predict the resulting phenotype merely by the

1Smith FJD, Morley SM, Rugg EL, et al: Clustering of epidermolysisbullosa simplex mutations in relation to disease phenotype: Data fromWeber-Cockayne EBS. J Invest Dermatol 101:481A, 1993 (abstr.)

position of a mutation in one of the traditional domains. The natureof the amino acid substitution is very important in determining themolecular consequences of the mutation, and thus the EBS type.The K14 L143P mutation is expected to severely impair or preventstabilization of the coiled-coil structure of the K14/K5 heterodimerdue to kinks in the alpha-helix, and thus have drastic functionalconsequences on the keratin molecule, whereas the K14 E144Aand K5 N193K merely affect the hydrophobic interactions betweenthe heptad repeats of the alpha-helical segments, and thus have aless drastic functional effect. Hence, it is not surprising that a moresevere phenotype is seen in patients with the K14 L143P mutation.In a similar way, it can be argued that the K5 L325P mutationmust have a more severe effect than the EBS-WC mutationspreviously observed in the L12 region.

The K14 K116N mutation that we identified in the three familieswith EBS-WC is located in the beginning of the 1A segment and,consequently, it cannot be defined as a typical K14 EBS-WCmutation. Lys116 is highly conserved among the type I keratins,whereas the homologous positions in type II keratins are occupiedby an arginine, suggesting that a basic amino acid in this positionmay be important. Another EBS-WC mutation (M119I), however,has been found in this region of K14 (Chen et al, 1995; Hu et al,1997). Patients heterozygous for this M119I mutation have theEBS-WC phenotype, whereas patients homozygous for M119Ihave a more severe phenotype resembling EBS-DM. Interestingly,as for Lys116 in K14, Met119 is conserved among type I keratins,whereas an isoleucine is conserved at the homologous positionamong type II keratins. Thus, methionine and isoleucine are notinterchangeable at these homologous positions in the two moleculeswithout disturbing the keratin filament assembly.

Just as observed in families 3 and 4 with the EBS-K form, all24 investigated patients from the three families (1, 2, and 5) exhibitsimilar clinical symptoms (EBS-WC). The fact that the K14 K116Nmutation was found in three of the five Danish families couldmean that this mutation accounts for a significant proportion ofthe EBS mutations in Denmark. All three families, however, turnedout to be connected to the small island of Læsø, and share thesame haplotype of the mutant chromosome, suggesting that this ismerely the result of a local founder effect. In fact, clinical EBSstudies have been conducted in Denmark focusing on the highincidence of EBS on the island of Læsø, where the disease forgenerations has been referred to as the ‘‘Læsø-disease’’ (Bartels,1939; Bulow and Nørholm-Pedersen, 1953; Nørholm-Pedersenand Nielsen, 1953). It was suggested that these EBS families witha connection to Læsø had a common ancestor with the disease,although the available pedigrees of the families could not beconnected. The EBS-WC families in this study are most probablydescendants of the investigated Læsø families in this old study, andour results provide a molecular support for the relatedness betweenthe families and illustrates the usefulness of the six characterizedpolymorphisms in exon 1 of the K14 gene in this study.

In conclusion, this study shows that there is a strict genotype-phenotype correlation in EBS, but that a more detailed evaluationof the nature of the mutations than merely assignment to thedomains of the keratin protein is needed to make genotype-phenotype predictions possible.

We would like to thank Kirsten Rasmussen, Department of Clinical Genetics,University Hospital of Odense, Denmark for help with the survey of family 3.This work was supported by grants from the Karen Elise Jensen Foundation; theInstitute of Experimental Clinical Research, Aarhus University, and AarhusUniversity Hospital Research Initiative.

REFERENCESAlbers K, Fuchs E: The expression of mutant epidermal keratin cDNAs transfected

in simple epithelial and squamous cell carcinoma lines. J Cell Biol 105:791–806, 1987

190 SØRENSEN ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Anton-Lamprecht I: Ultrastructural identification of basic abnormalities as clues togenetic disorders of the epidermis. J Invest Dermatol 103:6S–12S, 1994

Anton-Lamprecht I, Schnyder UW: Epidermolysis bullosa herpetiformis Dowling-Meara: Report of a case and pathogenesis. Dermatologica 164:221–235, 1982

Bader BL, Jahn L, Franke WW: Low level expression of cytokeratins 8, 18, and 19in vascular smooth muscle cells of human umbilical cord and in cultured cellsderived therefrom, with an analysis of the chromosomal locus containing thecytokeratin 19 gene. Eur J Cell Biol 47:300–319, 1988

Bartels ED: En familie med epidermolysis bullosa hereditaria. Ugeskr Laeger 101:141–144, 1939

Bonifas JM, Rothman AL, Epstein EH: Epidermolysis bullosa simplex: Evidence intwo families for keratin gene abnormalities. Science 254:1202–1205, 1991

Bulow K, Nørholm-Pedersen A: Epidermolysis bullosa hereditaria.Arvelighedsforhold, prognose og forekomst i Danmark. Ugeskr Laeger 115:479–487, 1953

Chan Y-M, Yu Q-C, LeBlanc-Straceski J, et al: Mutations in the non-helical linkersegment L1–2 of keratin 5 in patients with Weber-Cockayne epidermolysisbullosa simplex. J Cell Sci 107:765–774, 1994

Chan Y-M, Cheng J, Gedde-Dahl T Jr, Niemi K-M, Fuchs E: Genetic analysis ofa severe case of Dowling-Meara epidermolysis bullosa simplex. J Invest Dermatol106:327–334, 1996

Chen H, Bonifas JM, Matsumura K, Ikeda S, Leyden WA, Epstein EH Jr: Keratin14 gene mutations in patients with epidermolysis bullosa simplex. J InvestDermatol 105:629–632, 1995

Corden LD, McLean WHI: Human keratin diseases: Hereditary fragility of specificepithelial tissues. Exp Dermatol 5:297–307, 1996

Coulombe PA, Chan Y-M, Albers K, Fuchs E: Deletions in epidermal keratinsleading to alterations in filament organization in vivo and in vitro. J Cell Biol111:3049–3064, 1990

Coulombe PA, Hutton ME, Letai A, Hebert A, Paller AS, Fuchs E: Point mutationsin human keratin 14 genes of epidermolysis bullosa simplex patients: geneticand functional analyses. Cell 66:1301–1311, 1991

Eckert RL, Rorke E: The sequence of the human epidermal 58-kD (#5) type IIkeratin reveals an absence of 59 upstream sequence conservation betweencoexpressed epidermal keratins. DNA 7:337–345, 1988

Fuchs E: The cytoskeleton and disease: Genetic disorders of intermediate filaments.Annu Rev Genet 30:197–231, 1996

Fuchs E: Of mice and men: Genetic disorders of the cytoskeleton. Mol Biol Cell8:189–203, 1997

Galligan P, Listwan P, Siller GM, Rothnagel JA: A novel mutation in the L12domain of keratin 5 in the Kobner variant of epidermolysis bullosa simplex.J Invest Dermatol 111:524–527, 1998

Gedde-Dahl T Jr, Anton-Lamprecht I: Epidermolysis bullosa. In: Rimoin DL,Connor JM, Pyeritz RE (eds). Emery and Rimoin’s Principles and Practiceof Medical Genetics, 3rd edn. Edinburgh: Churchill Livingstone, 1996, pp.1225–1278

Gustafson S, Proper JA, Bowie EJW, Sommer SS: Parameters affecting the yield ofDNA from human blood. Anal Biochem 165:294–299, 1987

Hovnanian A, Pollack E, Hilal L, Rochat A, Prost C, Barrandon Y, Goossens M: Amissense mutation in the rod domain of keratin 14 associated with recessiveepidermolysis bullosa simplex. Nature Genet 3:327–332, 1993

Hu Z, Smith L, Martins S, Bonifas JM, Chen H, Epstein EH Jr: Partial dominanceof a keratin 14 mutation in epidermolysis bullosa simplex – increased severityof disease in a homozygote. J Invest Dermatol 109:360–364, 1997

Humphries MM, Mansergh FC, Kiang A-S, et al: Three keratin gene mutationsaccount for the majority of dominant simplex epidermolysis bullosa caseswithin the population of Ireland. Hum Mutat 8:57–63, 1996

Irvine AD, McKenna KE, Jenkinson H, Hughes AE: A mutation in the V1 domainof keratin 5 causes epidermolysis bullosa simplex with mottled pigmentation.J Invest Dermatol 108:809–810, 1997

Kitajima Y, Inoue S, Yaoita H: Abnormal organization of keratin intermediatefilaments in cultured keratinocytes of epidermolysis bullosa simplex. ArchDermatol Res 281:5–10, 1989

Korge BP, Gan SQ, McBride OW, Mischke D, Steinert PM: Extensive sizepolymorphism of the human keratin 10 chain resides in the C-terminal V2subdomain due to variable numbers and sizes of glycine loops. Proc Natl AcadSci USA 89:910–914, 1992

Lane EB, Rugg EL, Navsaria H, Leigh IM, Heagerty AHM, Ishida-Yamamoto A,Eady RAJ: A mutation in the conserved helix termination peptide of keratin5 in hereditary skin blistering. Nature 356:244–246, 1992

Lessin SR, Huebner K, Isobe M, Croce CM, Steinert PM: Chromosomal mappingof human keratin genes. Evidence of non-linkage. J Invest Dermatol 91:572–578, 1988

Letai A, Coulombe PA, McCormick MB, Yu Q-C, Hutton E, Fuchs E: Diseaseseverity correlates with position of keratin point mutations in patients withepidermolysis bullosa simplex. Proc Natl Acad Sci USA 90:3197–3201, 1993

Marchuk D, McCrohon S, Fuchs E: Remarkable conservation of structure amongintermediate filament genes. Cell 39:491–498, 1984

Matsuki M, Hashimoto K, Yoshikawa K, Yasuno H, Yamanishi K: Epidermolysisbullosa simplex (Weber-Cockayne) associated with a novel missense mutationof Asp328 to Val in linker 12 domain of keratin 5. Hum Mol Genet 4:1999–2000, 1995

Mischke D: Frequencies of human keratin 10 alleles. Hum Mol Genet 2:618, 1993Moll R, Franke WW, Schiller DL, Geiger B, Krepler R: The catalog of human

cytokeratins: Patterns of expression in normal epithelia, tumors, and culturedcells. Cell 31:11–24, 1982

Nomura K, Shimizu H, Meng X, et al: A novel keratin 5 gene mutation in Dowling-Meara epidermolysis bullosa simplex. J Invest Dermatol 107:253–254, 1996

Nørholm-Pedersen A, Nielsen NB: ‘‘Laesø disease’’ – epidermolysis bullosa simplex.Acta Genet Stat Med 4:417–423, 1953

Online Mendelian Inheritance in Man, OMIM (TM): Johns Hopkins University,Baltimore, MD. MIM Numbers 148040 (K5) and 148066 (K14). World WideWeb URL: http: //www.ncbi.nlm.nih.gov/omim/

Pearson RW, Spargo B: Electron microscope studies of dermal-epidermal separationin human skin. J Invest Dermatol 36:213–224, 1961

Popescu NC, Bowden PE, DiPaolo JA: Two type II keratin genes are localized onhuman chromosome 12. Hum Genet 82:109–112, 1989

Romano V, Bosco P, Rocchi M, Costa G, Leube RE, Franke WW, Romeo G:Chromosomal assignments of human type I and type II cytokeratin genes todifferent chromosomes. Cytogenet Cell Genet 48:148–151, 1988

Rosenberg M, RayChaudhury A, Shows TB, Le Beau MM, Fuchs E: A group of typeI keratin genes on human chromosome 17: Characterization and expression. MolCell Biol 8:722–736, 1988

Rosenberg M, Fuchs E, Le Beau MM, Eddy RL, Shows TB: Three epidermal andone simple epithelial type II keratin genes map to human chromosome 12.Cytogenet Cell Genet 57:33–38, 1991

Rugg EL, Morley SM, Smith FJD, et al: Missing links: Weber-Cockayne keratinmutations implicate the L12 linker domain in effective cytoskeleton function.Nature Genet 5:294–300, 1993

Savtchenko ES, Freedberg IM, Choi I-Y, Blumenberg M: Inactivation of humankeratin genes: The spectrum of mutations in the sequence of an acidic keratinpseudogene. Mol Biol Evol 5:97–108, 1988

Smack DP, Korge BP, James WD: Keratin and keratinization. J Am Acad Dermatol30:85–102, 1994

Stephens K, Ehrlich P, Weaver M, Le R, Spencer A, Sybert VP: Primers for exon-specific amplification of the KRT5 gene: Identification of novel and recurrentmutations in epidermolysis bullosa simplex patients. J Invest Dermatol 108:349–353, 1997

Uttam J, Hutton E, Coulombe PA, et al: The genetic basis of epidermolysis bullosasimplex with mottled pigmentation. Proc Natl Acad Sci (USA) 93:9079–9084, 1996

Vassar R, Coulombe PA, Degenstein L, Albers K, Fuchs E: Mutant keratin expressionin transgenic mice causes marked abnormalities resembling a human geneticskin disease. Cell 64:365–380, 1991

Waseem A, Gough AC, Spurr NK, Lane B: Localization of the gene for humansimple epithelial keratin 18 to chromosome 12 using polymerase chain reaction.Genomics 7:188–194, 1990