-3 Associated with Autosomal Dominant Erythrokeratoderma VariabilisThe erythrokeratodermas represent a group of disorders in which the common denominator is the presence of fixed or slowly moving erythematous hyperkeratotic plaques (Rook et al. 1998). Distinction between the entities is on clinical grounds and not always straight forward. Erythrokeratoderma variabilis (EKV; OMIM 133200) is a rare autosomal dominantly inherited skin disease that usually presents at birth or soon after. It is characterized by fixed brownish red hyperkeratotic plaques mainly on the extensor surfaces of the limbs and transient erythematous macules that may be of unusual shape, and some cases can have palmoplantar keratoderma. There is considerable intra- and interfamilial variability in severity of the skin disease; for example, some red patches may only last a few hours whereas other persist for a few days. The patches may migrate followed by a fine white scale. This variability in disease expressivity may be due to trauma, including mechanical irritation and excessive UV exposure, emotional stress, and, possibly, genetic background (Rook et al. 1998). Histopathologic features are largely nonspecific but include acanthosis and papillomatosis. Due to its variable penetrance, the classification of EKV can often be difficult particularly as it shares clinical features with other erythrokeratodermas. The disease in some families with autosomal dominantly inherited EKV has been linked to the chromosomal region 1p34–1p35, where there is a cluster of connexin genes (van der Schroeff et al. 1984;Richard et al. 1997). The gap junction protein 
-3 (GJB3: encoding Connexin 31), which maps in the minimal region on chromosome 1, has been demonstrated to be expressed in human keratinocytes (Wenzel et al. 1998). Recently, three distinct heterozygous missense mutations GJB3 have been shown to underlie EKV in four families (Richard et al. 1998). In this paper, we describe a novel substitution R42P in GJB3 in a family with erythrokeratoderma variabilis, adding further support for germline GJB3 mutations underlying the pathogenesis of this skin disease and the importance of gap junctions in epidermal communication and differentiation.
MATERIALS AND METHODS
Patients and biologic material
A single pedigree containing four individuals affected with EKV was studied. Genomic DNA was extracted from the venous blood of the three surviving affected and two unaffected family members using standard procedures (Kelsell et al. 1995). The clinical status of each individual was established after examination by two dermatologists (CTK and IML). Skin biopsy material was obtained and processed from one affected family member for histologic analyses. Genomic DNA was extracted from a control population of 125 unrelated individuals.
DNA amplification and sequencing of the Cx31 gene
We amplified a 1107 bp fragment from genomic DNA in which the entire coding sequence of human connexin 31 (GJB3) is contained using the primer pair 5' GTCAGAACTCAGAACACTGCC 3' corresponding to bases 681–701 and 5' CCTATACCCGGCTAGACAGC 3' corresponding to bases 1787–1768 (Genbank AJ004856). The DNA fragment was amplified under the following PCR conditions: 35 cycles of 30 s at 94°C, 30 s at 62°C, and 30 s at 72°C. The PCR product was purified using a Qiaquick PCR column (Qiagen, Crawley, West Sussex, U.K.) and directly sequenced by polymerase chain reaction (PCR)-cycle sequencing (Big-Dye terminator system, PE Applied Biosytems, Warrington, Cheshire, U.K.) and analyzed on a ABI 377 automated sequencer (PE Applied Biosytems). Sequence traces were analyzed by eye.
Restriction enzyme analysis of the R42P mutation
The nucleotide change in codon 42 of GJB3 abolishes a HinP1I cleavage site. A 498 bp DNA fragment (amplified by PCR using the primer pair 5' GCCTGGTACATAGTAAATGC 3' corresponding to bases 698–718 and 5' CTGAACAGGTAGGTCCACCA 3' corresponding to bases 1197–1178:Genbank AJ004856) was digested according to the manufacturer's recommended conditions (New England Biolabs, Hitchin, Herts, U.K.). Products were separated on a 2% agarose gel containing 0.5 mg ethidium bromide per ml and visualized using a UVP transilluminator. This assay was also used to screen the control population for the presence of R42P.
Heteroduplex formation and denaturing high performance liquid chromatography
All PCRs were heated to 95°C for 5 min followed by cooling at -1.5°C per min to a final temperature of 25°C. The samples were then transferred to the autosampler block, maintained at 7°C, for analysis. Denaturing high performance liquid chromatography (DHPLC) was performed using a WAVE DNA Fragment analysis system (Transgenomic, Crewe, Cheshire, U.K.). Five microliters of the 498 bp PCR product was eluted from the DNASep cartridge (Transgenomic) using a gradient created by a mixture of Buffer A (0.1 M Triethylammonium acetate pH 7) and Buffer B (25% Acetonitrile and 0.1 M Triethylammonium acetate) at a constant flow rate of 0.9 ml per min. For analysis of this DNA fragment the column temperature was set to 64°C and the concentration of eluent B was increased from 53% to 58% over 6 s and from 58% to 67% over 4.5 min. The eluted DNA fragments were detected using the WAVE systems UV detector in conjunction with the HPLC System Manager software (Hitachi, Tokyo, Japan).
Light and electron microscopy
A skin biopsy from the palm of an affected family member was processed using standard procedures (Eady 1985). Part of the sample was fixed in buffered half-strength Karnovsky fixative, rinsed thoroughly in buffer, further fixed in 1.3% osmium tetroxide in distilled water, dehydrated, and embedded in epoxy resin (Taab Laboratories, Aldermaston, Berks, U.K.). For light microscopy, semi-thin sections were stained with Richardson's stain. For electron microscopy, ultra-thin sections were stained with uranyl acetate and lead citrate, and viewed in a JEOL 100 CX transmission electron microscope operating at 80 kV.
Protein modeling experiments
Using the transmembrane protein pre-diction tools PHD (http://www.embl-heidelberg.de/predictprotein/), SOSUI (http://www.tuat.ac.jp/~mitaku/adv_sosui/), DAS (http://www.biokemi.su.se/~server/DAS/), SPLIT (drava.etfos.hr/
zucic/split.html) and TopPredII (http://www.biokemi.su.se/~server/toppred2/), we compared the transmembrane configuration of the normal Cx31 protein versus mutant R42P protein. Helical wheel representation of M1 of Cx31 was produced from the SOSUI output.
RESULTS AND DISCUSSION
Clinical and histologic description
The three surviving members of this pedigree have very similar clinical features. The pedigree structure is shown in (Figure 2a). Both fixed erythematous keratotic plaques and transient red patches began in early infancy. The plaques have involved knees, elbows, and progressively the calves, thighs, buttocks and lower back, arms, and neck. The transient red patches occur on any body site, occur more frequently soon after exposure to cold and wet climate, and with physical pressure on the skin. In II and III(2) they become finely scaly after resolving. The palms and soles are affected by diffuse keratoderma (Figure 1). Case III(1) has improved spontaneously from age 10 and has not sought treatment; II and III(2) have taken etretinate from the early 1980s and more recently Acitretin, with near clearance of all manifestations of EKV. No member of the pedigree suffers from deafness.
Figure 2.
Identification of the R42P mutation in affected members of the EKV pedigree. (a) Pedigree of the family; (b) DNA heteroduplex formation of the PCR-amplified GJB3 DNA from an affected family member compared with that of a homoduplex from an unaffected family member; (c) DNA sequence traces showing the G to C change, which introduces an R42P substitution in the M1/E2 boundary of the gap junction protein; (d) Hinp1I restriction enzyme digestion of PCR-amplified DNA of the family members, the mutation abolishes a Hinp1I cleavage site.
Full figure and legend (73K)Figure 1.
Clinical photograph of palm from affected family member. (a) Pattern of diffuse keratoderma on the palm and (b) well-defined plaques of hyperkeratosis over dorsal metacarpophalangeal and interphalangeal joints.
Full figure and legend (48K)No major changes were seen in the epidermis by light microscopy. The epidermal keratinocytes were normal in both shape and size and were normally compacted throughout the different epidermal layers. Electron microscopy showed a normal distribution and population density of tonofilaments (keratin intermediate filaments) and desmosomes throughout the epidermis, and normal connections between the tonofilaments and both desmosomes and hemidesmosomes. Gap junctions were also seen in different epidermal layers, but mainly in the upper spinous layer where they were in close proximity to desmosomes. The ultrastructure of the desmosomes appeared normal. Of the 20 gap junctions analyzed, no gross abnormalities were apparent by this detection system. Lamellar granules, also with a normal ultrastructure, were abundant in the upper spinous and granular layers, and there was evidence of normal secretion or extrusion of the contents of the granules into the extracellular space between upper granular and lower cornified cells.
Identification of a novel arginine to proline substitution on the border of the first transmembrane domain with the first extracellular domain in GJB3
Sequence analysis of the whole coding region of GJB3 revealed only one nucleotide substitution (CGC to CCC), which was present in codon 42 and was only present in the affected members of the family (Figure 2c). This nucleotide substitution abolishes a HinP1I restriction enzyme site and was used to confirm carrier status in affected members of the family (Figure 2d). In addition, the restriction enzyme assay did not reveal the sequence variant in 250 alleles from the control population. To further establish that there were no other GJB3 sequence variants segregating in the family, heteroduplex analysis using DHPLC was performed and revealed only the nucleotide substitution described above (Figure 2b). As the unaffected family members and the control population were not carriers of this sequence variant, this substitution is likely to represent the molecular basis of EKV in this family.
R42P may disrupt the normal function of the connexin 31-associated channels. The mutant allele encodes a proline residue instead of an arginine at position 42 and is the amino acid in the first extracellular domain of the protein adjacent to the first transmembrane domain. The positively charged arginine is conserved in all Cx31 and Cx31.1 species; however, other connexin family members either have a negatively charged glutamic acid, e.g., Cx26, or a serine residue, e.g., Cx43. Experimental data suggest that the residues at the M1/E2 boundary may have a key role in the voltage gating polarity of the connexon (Rubin et al. 1992;Verselis et al. 1994;White et al. 1995). A helical wheel representation of Cx31 M1 models E41, between L34 and V33 (Figure 3). Therefore, it is possible that the R42P mutation in GJB3 alters the gating polarity of the mutant connexin molecules leading to defective homo- and heterotypic channel formation and regulation. In addition to charge-related effects, substitution of arginine 42 by proline may also have a wider structural destabilization effect in the connexon; however, further studies need to be performed to elucidate the disruptive nature of the R42P mutation and those previously described (Richard et al. 1998).
Figure 3.
Helical wheel model of the first transmembrane domain of Cx31. Residues 21–41 of Cx31 are portrayed in an axial projection of an 
-helical backbone. Hydrophobic residues are represented in black, polar residues in green, positively charged residues in blue, negatively charged residues in red. The dotted line and asterisks mark the putative area of residues exposed within the connexon hemichannel pore, based on data from Cx43 (Unger et al. 1997).
This is a novel mutation as the only other described EKV associated GJB3 mutations were found either in the cytoplasmic amino terminus or in the second transmembrane domain (Richard et al. 1998). These are regions of the protein thought to be important for polarity of voltage gating and voltage gating activity, respectively. The Cx31 mutations associated with high frequency hearing loss (Xia et al. 1998) were present in the second extracellular domain, which is thought to be important for the specificity of connexon–connexon interactions. This suggests that the site of mutation in GJB3 may determine whether the individual develops EKV or high frequency deafness. The identification of additional GJB3 mutations associated with hearing loss, however, has yet to be identified (Coucke et al. 1999;Van Hauwe et al. 1999). Further studies are in progress to investigate the effect of the R42P mutation on keratinocyte gap junction communication, in particular investigating its effect on the other known epidermally expressed gap junction proteins, Cx26 and Cx43.
References
- Coucke, PJ, Hauwe, PV, Kelley, PM et al. Mutations in the KCNQ4 gene are responsible for autosomal dominant deafness in four DFNA2 families. Hum Mol Genet, 1999 8, 1321–1328, | Article | PubMed | ISI | ChemPort |
- Eady, RAJ Transmission electron microscopy In: Methods in Skin Research, 1985 eds D Skerrow, C Skerrow, Chichester: Wiley, pp. 1–36,
- Kelsell, DP, Stevens, HP, Ratnavel, R, Bryant, SP, Bishop, DT, Leigh, IM, Spurr, NK Genetic linkage studies in non-epidermolytic palmoplantar keratoderma: evidence for heterogeneity. Hum Mol Genet, 1995 4, 1021–1025, | PubMed | ISI | ChemPort |
- Richard, G, Lin, JPJ, Smith, L et al. Linkage studies in erythrokeratodermias; fine mapping, genetic heterogeneity and analysis of candidate genes. J Invest Dermatol, 1997 109, 666–671, | Article | PubMed | ISI | ChemPort |
- Richard, G, Smith, LE, Bailey, RA et al. Mutations in the human connexin gene GJB3 cause erythrokeratodermia variabilis. Nat Genet, 1998 20, 366–369, | Article | PubMed | ISI | ChemPort |
- Rook, A, Wilkinson, D, Ebling, F, Champion, R, Burton, J Textbook of Dermatology 1998 London. Blackwell Scientific, pp. 1533–1536,
- Rubin, JB, Verselis, VK, Bennett, MV, Bargiello, TA A domain substitution procedure and its use to analyze voltage dependence of homotypic gap junctions formed by connexins 26 and 32. Proc Natl Acad Sci USA, 1992 89, 3820–3824, | PubMed | ChemPort |
- van der Schroeff, JG, Nijenhuis, LE, Meera Khan, P et al. Genetic linkage between erythrokeratrodermia variabilis and Rh locus. Hum Genet, 1984 68, 165–168, | PubMed | ChemPort |
- Unger, VM, Kumar, NM, Gilula, NB, Yeager, M Projection structure of a gap junction membrane channel at 7 A resolution. Nat Struct Biol, 1997 4, 39–43, | Article | PubMed | ISI | ChemPort |
- Van Hauwe, P, Coucke, PJ, Declau, F et al. Deafness linked to DFNA2: one locus but how many genes?. Nat Genet, 1999 21, 263,
- Verselis, VK, Ginter, CS, Bargiello, TA Opposite voltage gating polarities of two closely related connexins. Nature, 1994 368, 348–351, | Article | PubMed | ISI | ChemPort |
- Wenzel, K, Manthey, D, Willecke, K, Grzeschik, KH, Traub, O Human gap junction protein connexin31: molecular cloning and expression analysis. Biochem Biophys Res Commun, 1998 248, 910–915, | Article | PubMed | ISI | ChemPort |
- White, TW, Paul, DL, Goodenough, DA, Bruzzone, R Functional analysis of selective interactions among rodent connexins. Mol Biol Cell, 1995 6, 459–470, | PubMed | ISI | ChemPort |
- Xia, JH, Liu, CY, Tang, BS et al. Mutations in the gene encoding gap junction protein beta-3 associated with autosomal dominant hearing impairment. Nat Genet, 1998 20, 370–373, | Article | PubMed | ISI | ChemPort |
