Clustered cadherin genes: a sequence-ready contig for the desmosomal cadherin locus on human chromosome 18.

We describe the assembly of a cosmid and PAC contig of approximately 700 kb on human chromosome 18q12 spanning the DSC and DSG genes coding for the desmocollins and desmogleins. These are members of the cadherin superfamily of calcium-dependent cell adhesion proteins present in the desmosome type of cell junction found especially in epithelial cells. They provide the strong cell-cell adhesion generated by this type of cell junction for which expression of both a desmocollin and a desmoglein is required. In the autoimmune skin diseases pemphigus foliaceous and pemphigus vulgaris (PV), where the autoantigens are, respectively, encoded by the DSG1 and DSG3 genes, severe areas of acantholysis (cell separation), potentially life-threatening in the case of PV, are evident. Dominant mutations in the DSG1 gene causing striate palmoplantar keratoderma result in hyperkeratosis of the skin on the parts of the body where pressure and abrasion are greatest, viz., on the palms and soles. These genes are also candidate tumor suppressor genes in squamous cell carcinomas and other epithelial cancers. We have screened two chromosome 18-specific cosmid libraries by hybridization with previously isolated YAC clones and DSC and DSG cDNAs, and a whole genome PAC library, both by hybridization with the YACs and by screening by PCR using cDNA sequences and YAC end sequence. The contigs were extended by further PCR screens using STSs generated by vectorette walking from the ends of the cosmids and PACs, together with sequence from PAC ends. Despite screening of two libraries, the cosmid contig still had four gaps. The PAC contig filled these gaps and in fact covered the whole locus. The positions of 45 STSs covering the whole of this region are presented. The desmocollin and desmoglein genes, which are about 30-35 kb in size, are quite well separated at approximately 20-30 kb apart and are arranged in two clusters, one DSC cluster and one DSG cluster, which are transcribed outward from the interlocus region. The order of the genes is correlated with the spatial order of gene expression in the developing mouse embryo, and this, and previous transgenic experiments, suggests that long-range genetic elements that coordinate expression of these genes may be present. The complete bacterial clone contig described in this paper is thus a resource not only for future sequencing but also for investigations into the control of expression of these clustered genes. Copyright 1999 Academic Press.