Structure and expression of human cell adhesion genes : the desmosomal cadherins

The desmocollins and desmogleins are transmembrane glycoproteins of the cadherin superfamily of adhesion proteins. They are present in desmosomes, the specialised intercellular junctions characteristically found in epithelial tissues, where they provide strong and stable cellular interactions. Both types of desmosomal cadherin, together with the desmosomal plaque protein plakoglobin, are required for these strong intercellular interactions. Three types of desmocollin (DSCl, 2, and 3) and three types of desmoglein (DSGl, 2, and 3) have been identified and their genes are clustered in a region of not more than 700kb on human chromosome 18q12.1, the DSC and DSG genes being arranged separately in two tandem arrays. Desmocollins most strongly resemble the classical cadherins in their structure, although they are unique amongst the cadherins in that there are two protein forms, a long 'a' form and a truncated 'b' form, generated by alternative splicing. Desmogleins differ in that they have a number of unique repeats in the cytoplasmic domain, varying between the different types, which is predicted to form a β-pleated sheet structure. Given the clustered position of the genes, it seems likely that they evolved by gene duplication and mutation from an ancestral gene on this chromosome. I have determined the exon-intron boundaries of the human DSC2 gene and shown it to consist of more than 33kb of DNA arranged into 17 exons ranging in size from 46 to 258bp; exon 16 is alternatively spliced giving rise to the 'a' and 'b' forms of the protein. This has revealed a most remarkable degree of conservation of intron position with other cadherins, which strongly suggests that they all evolved from a common ancestral gene. The desmocollin exon-intron organisation is more similar to the classical cadherins than to the desmogleins,especially in the cytoplasmic domain. Intron 1 is the largest, as it is in the desmogleins, in contrast to the classical cadherins, where intron 2 is extremely large; this latter intron is missing from the desmogleins. The intron positions occur at different positions in consecutive repeats in the extracellular domain, which may have evolved to constrain intragenic crossing-over events that could otherwise take place and lead to proteins with differing numbers of extracellular repeats.The desmosomal cadherins have varying spatial and temporal distributions. The type 2 isoforms have the most widespread tissue distribution in epithelia and also in desmosome-bearing non-epithelial tissues, being found for example as the only isoforms in the heart. Dsc2 is the first type expressed in the developing mouse embryo from the 16/32-cell stage. A luciferase reporter gene system has been used to analyse regions of the gene to determine if they contain elements required for tissue specific expression. It is known that a 3.5kb upstream fragment mediates strong expression of the reporter in epithelial cells with reduced expression in non-epithelial cells, and deletion derivatives have defined a minimal promoter of 525bp, Using a primer extension experiment I have determined the position of transcription initiation in DSC2 to be 201 bp upstream of the translation start site. As distinct from E-cadherin, which has enhancer activity in intron 1, I have found that intron 1 of DSC2 has a possible silencer element which may be responsible for restricting expression in nonexpressing cell types. No enhancer activity was found in any regions examined, which included the 5' and 3' untranslated regions, intron 1 and intron 2, and the coding region.The DSC3 gene differs from the DSC2 gene in being expressed only in stratified epitheha, mainly in the lower layers. I have cloned and sequenced the promoter region of human DSC3. No useful homologies were found with the DSC2 promoter, but comparison with the mouse DSC3 promoter has revealed several highly conserved regions which may indicate the sites of important regulatory functions. In common with DSC2, the DSC3 gene had no obvious TATA box, but contained a number of possible sites for transcription factors such as Ap-1 and Ap-2, implicated in epithelial specificity of gene expression. Deletion derivatives were cloned into a luciferase reporter vector and reporter enzyme assays were used to define a minimal promoter of 438bpIt is speculated that the failure to produce absolute in vivo tissue specificity using short-range genetic elements such as promoters may be because of the absence of more upstream genetic elements. It is possible that the specificity of expression of the desmosomal cadherins may be controlled within a multi-layered transcriptional unit in which each gene has proximal control elements which are overlayed by a global regulation system for the whole desmosomal cadherin locus. It is also therefore possible that gene position within the cluster plays a part in the regulation of gene expression.