Introduction

The epidermis is the first line of defence protecting the human body from its external environment and the keratinocytes that make up this important barrier tissue express specialized structural molecules in order to resist mechanical insult (Lane and McLean, 2004; Omary et al., 2004). The intermediate filament cytoskeleton extending throughout the cytoplasm of keratinocytes is the main stress-bearing structure in epithelial tissues and is composed of keratins (Fuchs and Cleveland, 1998). These rod-like fibrous proteins form heteropolymers consisting of at least one type I and one type II keratin (Quinlan et al., 1994; Coulombe and Omary, 2002). Keratins share a conserved protein domain organization with other intermediate filament proteins, consisting of a central -helical rod domain flanked by globular head and tail domains. Two short amino-acid sequences at the start and end of the rod domain, termed the helix boundary motifs, are hotspots for pathogenic mutations associated with keratinizing disorders (Irvine and McLean, 1999). Keratins exhibit tissue- and differentiation-specific expression patterns, for example, K1 and K10 are expressed in suprabasal cells throughout the epidermis (Lane, 1993), whereas keratin 9 (K9) is a suprabasal keratin found only in palmoplantar skin (Langbein et al., 1993).

The palmoplantar keratodermas are a highly heterogeneous group of genodermatoses where the primary clinical phenotype is hyperkeratosis of thick skin (Stevens et al., 1996). Classification of these disorders on the basis of phenotypic and morphological criteria has been difficult; however, a revised classification system based on the knowledge of the underlying genetic defects is now beginning to emerge (McLean, 2003). Epidermolytic palmoplantar keratoderma (EPPK; OMIM number 144200) is an autosomal dominant disorder characterized by cytolysis, blistering, and a diffuse pattern of keratoderma affecting both palms and soles (Vörner, 1901). The keratoderma is strictly limited to the thick skin and typically has a well-circumscribed erythematous border. The gene for this disease was mapped to the type I keratin gene cluster where the K9 gene (gene symbol KRT9) was localized (Reis et al., 1992). Subsequently, K9 mutations were identified in EPPK families (Reis et al., 1994) and K9 defects have been reported in many populations (McLean, 2003). Most of the reported EPPK mutations have been missense or small in-frame insertion–deletion mutations in the helix boundary motifs of the K9 polypeptide (Human Intermediate Filament Mutation Database, http://www.interfil.org). More recently, a few mutations have been reported in the K1 gene causing a phenotype that can be clinically and histologically difficult to distinguish from EPPK (Hatsell et al., 2001; Terron-Kwiatkowski et al., 2002, 2004; Whittock et al., 2002). The K1 mutations reported in families with a presentation predominantly consisting of palmoplantar keratoderma are larger deletions in the K1 rod domain (Hatsell et al., 2001; Terron-Kwiatkowski et al., 2002) or affect domains with a more subtle role in keratin filament assembly, such as the ISIS motif within the V1 domain (Kimonis et al., 1994), or affect the V2 domain (Whittock et al., 2002). More recently, we have shown that certain missense mutations in the helix termination motif of K1 can lead to a phenotype easily confused with Vörner-type EPPK (Terron-Kwiatkowski et al., 2004).

Here, we studied two apparently unrelated families with EPPK where the keratin cytoskeleton of palmoplantar suprabasal cells had a remarkable tubular morphology and where the underlying genetic defect was a missense mutation in the 1B domain of K1.

Results

Clinical description

Two apparently unrelated Dutch kindreds (designated Families L and J) presented with a phenotype entirely typical of EPPK of the Vörner type (Vörner, 1901). Specifically, there was a diffuse pattern of keratoderma affecting both the palmar and plantar surfaces, with a well-circumscribed erythematous margin (Figure 1). A plantar biopsy of an affected member of Family L revealed epidermolytic changes consistent with a diagnosis of EPPK (Figure 2a) and essentially identical histology was observed in biopsy material from the proband in Family J (not shown). In each of the families, several generations were affected in a clear autosomal pattern of inheritance, with many examples of male-to-male transmission seen in both kindreds (pedigrees not shown).

Figure 1.

Clinical presentation of epidermolytic palmoplantar keratoderma (EPPK) in the proband from Family J. Affected persons in both EPPK Families J and L studied here exhibit diffuse palmoplantar keratoderma with a well-demarcated erythematous margin, shown here on the palm. Significantly, there was no spread of the keratoderma onto the wrists as observed in other "K1 keratoderma" kindreds and there was no occurrence of hyperkeratosis affecting other body sites, notably the flexures. Together with the epidermolytic histology (see Figure 2), these features are quite typical of the Vörner form of epidermolytic palmoplantar keratoderma, commonly caused by mutations in keratin 9.

Full figure and legend (137K)

Figure 2.

Histological and ultrastructural features of "tonotubular" keratin aggregation. (a) A semithin section of plantar epidermis from the proband in Family L, stained with Alcian blue, shows extensive cytolysis of suprabasal cells diagnostic of epidermolytic palmoplantar keratoderma. Some suprabasal cells appear to have large, densely staining inclusions (arrows). Basal cells (B) are normal. The histology of biopsy material from affected persons in Family J was essentially identical (not shown). Scale bar=20 m. (b) Low-power electron micrograph of a suprabasal keratinocyte from the proband in Family J showing extensive vacuolation and destruction of the cytoplasm (^*). There are some very electron-dense aggregates in the cytoplasm (arrows) and a larger number of "whorls" of aggregated keratin (W). N=nucleus. Scale bar=2 m. (c) High-power electron micrograph from the proband in Family J, showing details of a cytoplasmic keratin "whorl", as seen in (b), above. At high power, the keratin is seen to form tubules that can be observed in transverse section (arrows) or in longitudinal section (arrowheads). Scale bar=100 nm. (d) High-power electron micrograph from the proband in Family L, again showing details of cytoplasmic keratin "whorls". Keratin tubules can be seen in transverse section (arrows) or in longitudinal section (arrowheads). The bizarre morphology of the keratin in this unrelated patient is strikingly similar to that seen in Family J, (c) above. Bar=100 nm.

Full figure and legend (213K)

Electron microscopy reveals tonotubular keratin

Very unusual ultrastructural changes in keratin morphology were observed in suprabasal keratinocytes of plantar biopsy material derived from affected members of both Families L and J (Figure 2b–d). Rather than forming bundles of tonofilaments, the suprabasal keratin instead formed tubular structures, seen in both longitudinal and transverse sections (Figure 2c and d). In some places, there was aggregation of these structures, but within these conglomerates, the tubular morphology could still be seen. There were electron-dense aggregates more typical of EPPK in some places; however, the "whorls" of keratin exhibiting the tubular phenotype predominated (Figure 2b). Examination of the literature revealed that virtually identical keratin morphology had been described previously in a German kindred who also presented with EPPK (Wevers et al., 1991). These authors had coined the term "tonotubular" keratin for these structures.

A novel 1B domain mutation in K1

As EPPK of the Vörner type is known to be caused by mutations in K9 (Reis et al., 1994), we initially screened this gene in its entirety using DNA from the proband in Family L. Although we detected a few previously reported polymorphisms (data not shown), we did not identify any potentially pathogenic sequence changes in the coding sequences or intronic splice sites of this gene.

As a subset of mutations in K1 have been associated with different forms of palmoplantar keratoderma (Kimonis et al., 1994; Whittock et al., 2002; Hatsell et al., 2001; Sprecher et al., 2001; Terron-Kwiatkowski et al., 2002, 2004), we screened the entire K1 gene in Family L and identified a heterozygous transition mutation 698C>T in exon 2 (Figure 3a and b). This sequence change was present in the affected individuals, but was absent from unaffected family members and leads to a leucine to serine substitution at position 233 of the K1 polypeptide, designated S233L. This mutation lies at the beginning of the predicted helix 1B domain of the central -helical coiled-coil rod domain (Smith et al., 2002). Only one other homozygous change 1072A>T was found in exon 5 of the proband of Family L, his unaffected son and one unaffected control (data not shown). As it did not co-segregate with the disease phenotype in the family and was detected in unaffected controls, this could be excluded as a pathogenic mutation, even though it predicts the amino-acid change N358Y in the 2A domain of the K1 protein. The same S233L mutation in K1 was also identified in all affected members of Family J by direct sequencing (data not shown).

Figure 3.

Molecular genetic analysis reveals a new mutation in the K1 helix 1B domain. (a) Normal K1 exon 2 sequence, showing forward strand sequence covering codons 231–235, inclusive. (b) The same region of K1 exon 2 as shown in (a), above, derived from the proband in Family J. A heterozygous 698C>T transition (arrow) was detected, predicting the S233L missense mutation in the 1B domain of the K1 polypeptide. The same mutation was found to co-segregate with the EPPK phenotype in both families. (c) Confirmation of the S233L mutation in exon 2 of the K1 gene by denaturing high-performance liquid chromatography in affected members of Families J and L. Double peaks corresponding to the heteroduplexes formed between the wild-type and mutant alleles are observed in samples carrying the mutation.

Full figure and legend (117K)

The presence of the K1 mutation S233L was confirmed by denaturing high-performance liquid chromatography in all affected members of Families L and J. In this assay, the mutation produced clearly identifiable peaks corresponding to heteroduplexes formed between the wild-type and mutant alleles (Figure 3c). Subsequently, denaturing high-performance liquid chromatography was used to rapidly exclude the mutation from 94 Dutch and 74 British unaffected controls. Affected and unaffected samples of Family L were used as positive and negative controls, respectively. From the populations screened, a total of 13 control samples presented elution profiles different from that of a sequence-verified unaffected control and were therefore sequenced. All samples sequenced, including all samples from the two EPPK families, had a silent polymorphism 720A>G (R240R in the 1B domain) that differed from the published sequence (GenBank Accession number AF237621). In addition, a heterozygous 1762G>A polymorphism (G488R in the variable V2 domain) was found to be common in both Dutch and British populations (5/13 samples sequenced). Remarkably, an additional heterozygous 800G>A change, predicting the non-conservative amino-acid substitution R267Y in the helix 1B domain, was found in one British control who had no history of skin disease. This must therefore be a rare non-pathogenic substitution polymorphism in K1.

Functional analysis reveals a subtle filament aggregation phenotype

As the mutation occurred in a part of the K1 protein not previously associated with pathogenic mutations, we performed transient transfection experiments using normal and mutant keratin constructs in a keratinocyte cell line in order to confirm that the S233L mutation in the 1B domain of K1 is pathogenic and perhaps recapitulate the tubular keratin morphology in a cell culture system. We used the human keratinocyte cell line HaCaT (Boukamp et al., 1988), which expresses the basal cell keratins K5 and K14 (Ryle et al., 1989), reported to be a pre-requisite for the assembly of keratins K1 and K10 (Kartasova et al., 1993; Paramio and Jocarno, 1994). K1 and K10 are localized in the suprabasal layer of the skin, and expression of these differentiation-specific keratins in HaCaT cells appears to be dependent on cell density (Ryle et al., 1989). Here, we found that using normal keratinocyte growth medium with a standard concentration of calcium (which affects differentiation), there was reasonable expression of K1 and K10 (data not shown). The transfected K1 could pair another type I keratin, such as the more widely expressed K14, and sub-optimal keratin pairing might mask the effect of the mutation in filament formation. To exclude this possibility, we reproduced the equivalent of the K1 mutation S233L in type II keratin K5 (Q221L), the latter protein being constitutively expressed in HaCaT cells (Ryle et al., 1989). To enable visualization of the transfected keratins, all keratin constructs were tagged at the N-terminus with enhanced green fluorescent protein (EGFP).

HaCaT cells were transiently transfected with wild-type and mutant versions of the K1 and K5 constructs. At 72 hours after transfection, morphologically normal keratin filaments were present in most of the cells transfected with either wild-type or the 1B mutant keratins (Figure 4a). In some cells, and for all constructs regardless of mutation status, thicker filaments could be observed in a perinuclear distribution (not shown). These may correspond to the newly assembled keratin filaments as described previously (Albers and Fuchs, 1987). In addition, with all constructs, some cells presented small, irregular, densely staining punctate aggregates (Figure 4b). Electron microscopy was performed on transfected cells to see if the mutant keratins recapitulated the tubular morphology in this cell culture model; however, no tubular structures were observed (data not shown). Nevertheless, a lower proportion of cells with aggregates were observed among those transfected with K1 wild-type (2.90.7%) compared to the K1-S233L mutant constructs (3.91.3%). Similar results were obtained for the equivalent K5 constructs. Specifically, K5 wild-type produced 11.32.2% aggregates, versus K5 Q221L, which gave 15.32.1% aggregates. Although these differences were not statistically significant (P-value >0.10 combining data for K1 and K5; K1 ²=0.243; or K5 ²=1.410), these functional data, together with exclusion of S233L from 336 normal chromosomes, is indicative that this is a pathogenic mutation in K1.

Figure 4.

The 1B domain mutation is detrimental to filament formation in cultured cells. (a) HaCaT cells were transfected with K1 or K5 wild-type and 1B mutant GFP constructs. At 72 hours after transfection, keratin filaments of essentially normal morphology were observed in cells transfected with all constructs, as shown here with wild-type K1. (b) At shorter times following transfection, cells with prominent, punctuate aggregates were also present in all transfection experiments, exemplified here by K1 mutant S233L, at 48 hours post-transfection. Higher percentages of aggregates were observed with both K1 and K5 mutants compared to their wild-type controls, indicative of the pathogenicity of this mutation in the 1B domain.

Full figure and legend (54K)

Discussion

Here, we have studied two large unrelated families who presented with all the usual clinical hallmarks of Vörner-type EPPK: keratoderma in a diffuse pattern on palms and soles with a well-defined erythematous margin; histology showing epidermolytic changes; lack of involvement of non-glabrous skin; and canonical signs of autosomal dominant inheritance (Figures 1 and 2). Following exclusion of a mutation in the K9 gene by full genomic sequence analysis and genetic linkage analysis, we identified a novel missense mutation S233L in the 1B domain of the K1 polypeptide, which co-segregated with disease phenotype in both families and was excluded from 168 normal controls (Figure 3). Protein expression analysis in cultured cells provided further evidence for pathogenicity of this unique mutation. The Vörner form EPPK is predominantly caused by missense or small in-frame insertion–deletion mutations in the helix boundary motifs of K9 (Reis et al., 1994; Coleman et al., 1999). There are currently 48 independently occurring mutations in K9 logged in the Human Intermediate Filament Mutation Database (http://www.interfil.org). Recently, a small number of families have come to light exhibiting phenotypes that are very similar to or are clinically indistinguishable from EPPK, but who carry a mutation in K1 instead of the K9 gene (Hatsell et al., 2001; Terron-Kwiatkowski et al., 2002, 2004). Some of these families have, upon close examination, small areas of epidermolytic hyperkeratosis on sites other than palm and sole, typically this is seen as hyperkeratosis spreading onto the wrist, or in a few cases, the flexures and other sites (Terron-Kwiatkowski et al., 2004). Families J and L studied here did not have any hyperkeratosis outside of the palm and sole, and so were classified as Vörner-type EPPK. Expression of K1 and its type I keratin polymerization partner K10 is not limited to thick skin, but instead, these keratins are expressed in all regions of the epidermis (Lane, 1993). Dominant-negative mutations in either of these keratins have largely been associated with a more severe and widespread phenotype, bullous congenital ichthyosiform erythroderma, also known as epidermolytic hyperkeratosis (Cheng et al., 1992; Chipev et al., 1992; Rothnagel et al., 1992). It was recognized soon after the discovery that K1 and K10 are the causative genes for bullous congenital ichthyosiform erythroderma/epidermolytic hyperkeratosis, and that this disorder could be sub-divided into mainly palm/sole-affected (PS) and non-palm/sole-affected (NPS) forms (DiGiovanna and Bale, 1994). Furthermore, the PS subtype could be further divided into PS-1 (predominantly palm/sole) or PS-2 (palm/sole with more widespread involvement and neonatal blistering). The PS-1 and PS-2 presentations were caused by mutations in K1, whereas the NPS presentation was caused by K10 mutations (DiGiovanna and Bale, 1994). Using this very informative DiGiovanni–Bale classification, the families studied here fit clearly into the PS-1 category, and this is consistent with a K1 defect. It is difficult to explain why some mutations in K1 lead to hyperkeratosis restricted to palm and sole epidermis. One obvious explanation is that these sites undergo substantially more physical trauma during everyday life. However, given that the pattern of keratoderma in the patients studied here is diffuse with well-demarcated boundaries rather than focal in nature (Figure 1) is suggestive of a phenomenon related to regionalized gene expression in the palm and sole, for example, certain mutations may somehow perturb the interaction of K1 with K9 more than its interaction with K10.

The most striking feature of Families J and L was the bizarre tubular morphology of the tonofilaments observed by electron microscopy (Figure 2b–d). This was essentially identical in appearance to a previous report describing "tonotubular" keratin in a patient, which like those studied here was diagnosed as having EPPK (Wevers et al., 1991). The previously reported family originated in Germany and therefore it is possible that they may represent distant relatives of our two Dutch kindreds. We were unable to obtain DNA for analysis from the German kindred, but in view of the ultrastructural similarities, it seems very likely that they will harbor the S233L mutation or a very similar defect in K1. It is remarkable that a single amino-acid substitution in the 1B domain of the K1 rod domain is responsible for generating the tonotubular phenotype. Keratin rod domains adopt a coiled-coil secondary structure owing to a heptad repeat sequences (Crewther et al., 1983; Smith et al., 2002). Within a heptad, the a and d positions are usually occupied by hydrophobic residues that mediate leucine zipper interactions in the formation of a keratin type I–type II dimer. The serine mutated here in K1 (S233) occupies the f position within the heptad repeat thus: ...ESF INNLRRR VDQLKSD..., where S233 is in bold, and hydrophobic a and d residues underlined. In the coiled-coil structure, amino acids in the f position are on the outer face of the dimer. Thus, the mutant leucine residue is in a position where it could cause an aberrant hydrophobic interaction between adjacent keratin dimers. We therefore postulate that this gain-of-function mutation leads to tubular aggregation because of an abnormal hydrophobic dimer–dimer and/or filament–filament interaction. The overall result of the tubular aggregation and, to a lesser extent, dense aggregation of keratins (Figure 2b) are mechanical fragility of the suprabasal cells (Figure 2a), leading to cytolysis, and ultimately epidermolytic hyperkeratosis of palms and soles in the patients (Figure 1). It was disappointing that expression of the K1 mutant and the equivalent K5 mutation in HaCaT cells did not appear to recapitulate the tubular morphology upon electron microscopic analysis (data not shown). We postulate that this is because of the obvious differences in morphology and gene expression between a monolayer culture of a transformed cell line compared to the living multilayered epidermis in vivo. Perhaps, in stratifying cultures of stably transfected keratinocytes, the tubular keratin phenotype could be modelled, allowing further study of this unusual structural phenomenon.

In summary, molecular genetic analysis has identified a unique K1 mutation as the underlying cause of EPPK in a subset of patients with "tonotubular" keratin (Wevers et al., 1991). This sheds further light on the spectrum of phenotypes that can arise from K1 mutations and helps in future molecular classification of the keratodermas, which owing to their extensive heterogeneity are a difficult group of genodermatoses to define and diagnose on the basis of clinical features alone.

Materials and Methods

Clinical material

Patient materials were collected with informed consent using protocols approved by an Institutional Review Board that complies with all principles of the Declaration of Helsinki Principles Accord. A punch biopsy was taken from palmoplantar skin of the probands in Families L and J. Part of the biopsy was processed for standard hematoxylin and eosin staining. The other part was processed for electron microscopy as described previously. Genomic DNA was obtained with informed consent from peripheral blood lymphocytes using the Nucleon DNA Extraction Kit (Amersham Pharmacia Biotech, Bucks, UK).

K9 mutation detection

Amplification of the entire 4.8 kb K9 gene (gene symbol KRT9) was performed using 200 ng of each genomic DNA, with 1 M each of primers K9.C1 and K9.C2 (Table 1), 0.25 mM of each dNTP in 1 high-fidelity polymerase buffer with 1.5 mM MgCl₂ and 4% DMSO. High-fidelity Taq polymerase (1 U) (Roche Diagnostics, Mannheim, Germany) was added to each sample after DNA denaturation at 94°C for 6 minutes. Long-range PCR was carried out initially for 10 cycles of 94°C, 30 seconds; 55°C, 30 seconds; and 68°C, 4 minutes. This was followed by another 25 cycles with increasing extension times of 20 seconds/cycle and a final 10 minutes extension at 68°C. PCR products were purified using QIA quick PCR purification method (Qiagen, Crawley, UK) before DNA sequencing on a ABI DNA sequencer (Applied Biosystems, Foster City, CA) with internal primers: K9.p6, K9e7.L, and K9.C2 (Table 1).

Table 1 - Primers used for PCR and DNA sequencing.

Full table

K1 mutation detection

Exons 1 and 7 of the K1 gene (KRT1) were amplified using 200 ng of each genomic DNA, 1 M each of primer pairs K1e1.F&R or K1e7.F&2R (Table 1), 0.25 mM of each dNTP, and 1 U Taq polymerase (Promega, Southampton, UK) in 1 polymerase buffer with 1.5 mM MgCl₂ and 4% DMSO. Following DNA denaturation at 94°C for 5 minutes, PCR was carried out for 35 cycles of 94°C, 30 seconds; 55°C, 45 seconds; and 72°C, 90 seconds; with a final 5 minutes extension. Amplification of exons 2–6 and 8–9 was performed similarly with primers described elsewhere (Whittock et al., 2000) at an annealing temperature of 56°C for 30 cycles. PCR products were purified and sequenced as indicated above with each forward and/or reverse primers.

K1 mutation confirmation and population screening

For analysis by denaturing high-performance liquid chromatography, exon 2 of the K1 gene was performed using HotStar Taq system (Qiagen), but otherwise the conditions were as stated above. PCR products were subjected to denaturation at 95°C for 5 minutes, and thereafter renatured in 70 cycles of 22 seconds starting at 95°C and cooling 1°C/cycle to allow heteroduplex formation. Denaturing high-performance liquid chromatography was carried out using the Wave System (Transgenomics, Crewe, UK) with 5–10 l of each PCR product at a melting temperature of 59°C, as predicted for the 410 bp exon 2 fragment by the WaveMaker program (Transgenomics). Any sample that showed a different elution profile compared to a normal control was directly sequenced.

EGFP constructs

A wild-type human K1 construct tagged at the N-terminus with EGFP, consisting of the entire K1 open-reading frame between EcoRI and PstI sites in the pEGFP-C3 vector (Clontech, Palo Alto, CA), was kindly donated by Dr Sarah Hatsell and Professor David Kelsell (Centre for Cutaneous Research, St Bartholomew's and The Royal London School of Medicine and Dentistry, Queen Mary and Westfield College, London, UK). A similar EGFP-K1 construct carrying the S233L mutation construct was made by site-directed mutagenesis using the QuickChange Kit (Stratagene, La Jolla, CA) using K1-S233L/m-F and -R primers (Table 1), following the manufacturer's instructions.

A human K5 wild-type EGFP construct, consisting of the full-length K5 cDNA between HindIII and EcoRI sites in the pEGFP-C2 vector (Clontech), was a gift from Dr Dan Gibbs (CRUK Laboratories, School of Life Sciences, University of Dundee, Dundee, UK). A similar EGFP-K5 construct carrying the missense mutation Q221L (equivalent to the S233L mutation in K1) was made by site-directed mutagenesis using K5-Q221L/m-F and -R primers (described in Table 1). Wild-type and mutant plasmids were isolated using a Plasmid Maxi kit (Qiagen) and further purified for transfection by the standard phenol:chloroform method.

Keratinocyte transfection

HaCaT human keratinocyte cell line (Boukamp et al., 1988) was grown at 37°C and 5% CO₂ in keratinocyte serum-free medium (GIBCO-Invitrogen, Paisley, UK) supplemented with epidermal growth factor (at 5 ng/ml), bovine pituitary extract (at 25 mg/ml), penicillin (100 U/ml), and streptomycin (100 mg/ml). Transient transfections were performed at 50% confluency with the lipofectin–integrin–DNA method described previously (Hart et al., 1998). The integrin receptor-specific peptide 6 described previously (Hart et al., 1998) was obtained from Zinsser Analytic (Berks, UK) and the Lipofectin from GIBCO-Invitrogen. Lipofectin–integrin–DNA complex ratios (l lipofectin:g integrin:g DNA) of 4:12:2, 3:8:2, 2:8:2, and 2.25:12:3 were used to transfect the K1 wild-type and S233L mutant EGFP constructs in four independent experiments. For transfection, HaCaT cells were grown on 13 mm diameter glass coverslips in 35 mm culture plates. Other transfection experiments with K5 wild-type and Q221L mutant EGFP constructs were performed in four-chamber slides at lipofectin–integrin–DNA complex ratios of 1:2:0.5, 2:2:0.5, and 2:4:1. All lipofectin–integrin–DNA mixtures were allowed to form complexes for 2 hours at room temperature before transfection and the cells were washed 5 hours after transfection. Thereafter, cells were grown in fresh keratinocyte serum-free medium up to 72 hours post-transfection. For staining, cells were fixed with 3% paraformaldehyde for 10 minutes, and permeabilized with 0.2% Triton X-100 for 5 minutes. Nuclei were stained with 1 mg/ml DAPI (4,6-diamidino-2-phenylindole; Sigma Chemicals, Poole, UK) in phosphate-buffered saline. Hydromount containing 2.5% DABCO (diazabicyclo[2,2,2]octane; Sigma) was used to mount coverslips and slides. Cells were visualized using an Olympus BX60 epifluorescent microscope and images captured using the SmartCapture 2.1 (Digital Scientific, Cambridge, UK).

Statistical analysis of transfection experiments

Statistical analysis was carried out using the Microsoft Excel program. The proportion of cells with keratin aggregates among the transfected cell population for each K1 and K5 construct was expressed as the percentagethe standard error. Comparison of cells transfected with either wild-type or 1B mutant versions of K1 or K5 was performed using the ² test to calculate the P-value.

Conflict of Interest

The authors state no conflict of interest.

References

1. Albers K, Fuchs E (1987) The expression of mutant epidermal keratin cDNAs transfected in simple epithelial and squamous cell carcinoma lines. J Cell Biol 105:791–806 | Article | PubMed | ISI | ChemPort |
2. Boukamp P, Petrussevska RT, Breitkreutz D, Hornung J, Markham A, Fusenig NE (1988) Normal keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line. J Cell Biol 106:761–771 | Article | PubMed | ISI | ChemPort |
3. Cheng J, Syder AJ, Yu Q-C, Letai A, Paller A, Fuchs E (1992) The genetic basis of epidermolytic hyperkeratosis: a disorder of differentiation-specific epidermal keratin genes. Cell 70:811–819 | Article | PubMed | ISI | ChemPort |
4. Chipev CC, Korge BP, Markova N, Bale SJ, DiGiovanna JJ, Compton JC et al. (1992) A leucine–proline mutation in the H1 subdomain of keratin 1 causes epidermolytic hyperkeratosis. Cell 70:821–828 | Article | PubMed | ISI | ChemPort |
5. Coleman CM, Munro CS, Smith FJD, Uitto J, McLean WHI (1999) Epidermolytic palmoplantar keratoderma due to a novel type of keratin mutation, a 3 bp insertion in the keratin 9 helix termination motif. Br J Dermatol 140:486–490 | Article | PubMed | ISI | ChemPort |
6. Coulombe PA, Omary MB (2002) "Hard" and "soft" principles defining the structure, function and regulation of keratin intermediate filaments. Curr Opin Cell Biol 14:110–122 | Article | PubMed | ISI | ChemPort |
7. Crewther WG, Dowling LM, Steinert PM, Parry DAD (1983) Structure of intermediate filaments. Int J Biol Macromol 5:267–274 | Article | ISI | ChemPort |
8. DiGiovanna JJ, Bale SJ (1994) Clinical heterogeneity in epidermolytic hyperkeratosis. Arch Dermatol 130:1026–1035 | Article | PubMed | ISI | ChemPort |
9. Fuchs E, Cleveland DW (1998) A structural scaffolding of intermediate filaments in health and disease. Science 279:514–519 | Article | PubMed | ISI | ChemPort |
10. Hart SL, Arancibia-Carcamo CV, Wolfert MA, Mailhos C, O'Reilly NJ, Ali RR et al. (1998) Lipid-mediated enhancement of transfection by a nonviral integrin-targeting vector. Hum Gene Ther 9:575–585 | PubMed | ISI | ChemPort |
11. Hatsell SJ, Eady RA, Wennerstrand L, Dopping-Hepenstal P, Leigh IM, Munro C et al. (2001) Novel splice site mutation in keratin 1 underlies mild epidermolytic palmoplantar keratoderma in three kindreds. J Invest Dermatol 116:606–609 | Article | PubMed | ISI | ChemPort |
12. Irvine AD, McLean WHI (1999) Human keratin diseases: the increasing spectrum of disease and subtlety of the phenotype–genotype correlation. Br J Dermatol 140:815–828 | Article | PubMed | ISI | ChemPort |
13. Kartasova T, Roop DR, Holbrook KA, Yuspa SH (1993) Mouse differentiation-specific keratins 1 and 10 require a preexisting keratin scaffold to form a filament network. J Cell Biol 120:1251–1261 | Article | PubMed | ISI | ChemPort |
14. Kimonis V, DiGiovanna JJ, Yang J-M, Doyle SZ, Bale SJ, Compton JG (1994) A mutation in the V1 end domain of keratin 1 in non-epidermolytic palmar–plantar keratoderma. J Invest Dermatol 103:764–769 | Article | PubMed | ISI | ChemPort |
15. Lane EB (1993) Keratins. In: Connective Tissue and its Heritable Disorders Molecular, Genetic and Medical Aspects. (Royce PM, Steinmann B, eds), Wiley-Liss Inc., New York pp 237–247
16. Lane EB, McLean WHI (2004) Keratins and skin disorders. J Pathol 204:355–366 | Article | PubMed | ISI | ChemPort |
17. Langbein L, Heid HW, Moll I, Franke WW (1993) Molecular characterization of the body site-specific human epidermal cytokeratin 9: cDNA cloning, amino acid sequence, and tissue specificity of gene expression. Differentiation 55:57–71 | Article | PubMed | ISI | ChemPort |
18. McLean WHI (2003) Genetic disorders of palm skin and nail. J Anat 202:133–141 | Article | PubMed | ISI |
19. Omary MB, Coulombe PA, McLean WHI (2004) Intermediate filament proteins and their associated diseases. N Engl J Med 351:2087–2100 | Article | PubMed | ISI | ChemPort |
20. Paramio J, Jocarno J (1994) Assembly dynamics of epidermal keratins K1 and K10 in transfected cells. Exp Cell Res 215:319–331 | Article | PubMed | ISI | ChemPort |
21. Quinlan RA, Hutchison CJ, Lane EB (1994) Intermediate Filaments. Academic Press: London
22. Reis A, Hennies H-C, Langbein L, Digweed M, Mischke D, Drechsler M et al. (1994) Keratin 9 gene mutations in epidermolytic palmoplantar keratoderma (EPPK). Nat Genet 6:174–179 | Article | PubMed | ISI | ChemPort |
23. Reis A, Kuster W, Eckardt R, Sperling K (1992) Mapping of a gene for epidermolytic palmoplantar keratoderma to the region of acidic keratin gene cluster at 17q12–q21. Hum Genet 90:113–116 | PubMed | ISI | ChemPort |
24. Rothnagel JA, Dominey AM, Dempsey LD, Longley MA, Greenhalg DA, Gagne TA et al. (1992) Mutations in the rod domains of keratins 1 and 10 in epidermolytic hyperkeratosis. Science 257:1128–1130 | PubMed | ISI | ChemPort |
25. Ryle CM, Breitkreutz D, Stark HJ, Leigh IM, Steinert PM, Roop D et al. (1989) Density-dependent modulation of synthesis of keratins 1 and 10 in the human keratinocyte line HACAT and in ras-transfected tumorigenic clones. Differentiation 40:42–54 | PubMed | ISI | ChemPort |
26. Smith TA, Strelkov SV, Burkhard P, Aebi U, Parry DA (2002) Sequence comparisons of intermediate filament chains: evidence of a unique functional/structural role for coiled-coil segment 1A and linker L1. J Struct Biol 137:128–145 | Article | PubMed | ISI | ChemPort |
27. Sprecher E, Ishida-Yamamoto A, Becker OM, Marekov L, Miller CJ, Steinert PM et al. (2001) Evidence for novel functions of the keratin tail emerging from a mutation causing ichthyosis hystrix. J Invest Dermatol 116:511–519 | Article | PubMed | ISI | ChemPort |
28. Stevens HP, Kelsell DP, Bryant SP, Bishop DT, Spurr NK, Weissensbach J et al. (1996) Linkage of an American pedigree with palmoplantar keratoderma and malignancy (palmoplantar ectodermal dysplasia type III) to 17q24. Literature survey and proposed updated classification of the keratodermas. Arch Dermatol 132:640–651 | Article | PubMed | ISI | ChemPort |
29. Terron-Kwiatkowski A, Paller AS, Compton J, Atherton DJ, McLean WHI, Irvine AD (2002) Two cases of primarily palmoplantar keratoderma associated with novel mutations in keratin 1. J Invest Dermatol 119:966–971 | Article | PubMed | ISI | ChemPort |
30. Terron-Kwiatkowski A, Terrinoni A, Didona B, Melino G, Atherton DJ, Irvine AD et al. (2004) Atypical epidermolytic palmoplantar keratoderma presentation associated with a mutation in the keratin 1 gene. Br J Dermatol 150:1096–1103 | Article | PubMed | ISI | ChemPort |
31. Vörner H (1901) Zur Kenntniss des Keratoma hereditarium palmare et plantare. Arch Dermatol Syphilol 56:3–31 | Article |
32. Wevers A, Kuhn A, Mahrle G (1991) Palmoplantar keratoderma with tonotubular keratin. J Am Acad Dermatol 24:638–642 | PubMed | ISI | ChemPort |
33. Whittock NV, Eady RA, McGrath JA (2000) Genomic organization and amplification of the human epidermal type II keratin genes K1 and K5. Biochem Biophys Res Commun 274:149–152 | Article | PubMed | ISI | ChemPort |
34. Whittock NV, Smith FJ, Wan H, Mallipeddi R, Griffiths WA, Dopping-Hepenstal P et al. (2002) Frameshift mutation in the V2 domain of human keratin 1 results in striate palmoplantar keratoderma. J Invest Dermatol 118:838–844 | Article | PubMed | ISI | ChemPort |

Acknowledgments

We thank the patients and their families for their participation in this study. We also thank the Molecular Genetics Laboratory, Ninewells University Hospitals NHS Trust for genomic DNA extraction. This work was funded by a Wellcome Trust Senior Research Fellowship, and grants from the Dystrophic Epidermolysis Bullosa Research Association (DEBRA) UK and The Pachyonychia Congenita Project (to W.H.I.M.). M.A.M.v.S. is supported by The Pachyonychia Congenita Project, Barrier Therapeutics NV, and the GROW Research Institute.