Full-length CYLD is predicted to be a protein of 956 amino acids, with the exon −7 splice variant encoding 953 amino acids (Fig. 46.1-4). Likely orthologues, identified by regions of extensive amino acid identity, are present in Drosophila melanogaster and Caenorhabditis elegans. Although the predicted protein does not show extensive sequence similarity to other known human proteins, a number of previously defined motifs are present (Fig. 46.1-4).
Representation of the CYLD gene. The complete coding sequence (CDS) from codon 1 to 956 is represented with the exons numbered. The positions of germ line and somatic mutations are shown as dots above the CDS (nonsense mutations in red, frameshift mutations in yellow, and splice site mutations in blue). The relative positions of the known motifs are shown as rectangles below the CDS (CAP-GLY domains in yellow, putative SH3 binding domain in red and the UCH domains in blue).
There are three CAP-GLY domains.10 Two of these domains (approximately from amino acids 127 to 203 and amino acids 472 to 540) show strong similarity to previously described CAP-GLY domains, while the third (from amino acids 232 to 285) exhibits weaker similarity. The third CAP-GLY domain is present in the D. melanogaster CYLD orthologue.
Comparison of the CYLD sequence against itself reveals a short repeated segment of approximately 25 amino acids (between amino acids 388 to 413 and 446 to 471) that is rich in proline residues. This proline-rich region may constitute an SH3 binding domain that mediates protein-protein interactions in signal transduction or vesicle transport pathways.11
There are two short regions that exhibit homology to the two halves of a split ubiquitin C-terminal hydrolase (UCH) catalytic domain (reference12 and www.expasy.ch/cgi-bin/get-prodoc-entry?PDOC00127; www.expasy.ch/cgi-bin/get-prodoc-entry?PDOC00750) amino acids 871 to 889 showing homology to a UCH2-2 domain and amino acids 593 to 610 showing weaker homology to a UCH2-1 domain. These domains are highly conserved in both D. melanogaster and C. elegans orthologues.
There are four Cys-X-X-Cys pairs (between amino acids 788 and 856) that may represent finger-like metal-binding domains. Three of the four Cys-X-X-Cys pairs are conserved in D. melanogaster and C. elegans, despite relative lack of conservation of the surrounding amino acids.
CAP-GLY domains were originally identified in CLIP170/Restin, a protein that acts as a linker between endocytic vesicles and microtubules.13,14 Subsequently, CAP-GLY domains have been identified in a number of proteins that are believed to coordinate the attachment of cellular organelles, such as vesicles or chromosomes, to microtubules and thus assist their movements within the cell.11 The CAP-GLY domains themselves are responsible for microtubule binding.15 Included among this group of proteins are p150Glued which forms part of the dynactin complex16 (necessary for the docking of cargoes to the minus end directed microtubule associated motor, dynein) and BIK1 (mutations of which cause disorders of mitotic segregation in yeast).17,18 CLIP170 and a number of other proteins have two CAP-GLY domains, while the majority of CAP-GLY proteins contain a single copy. CYLD is the first protein, to our knowledge, that has three CAP-GLY motifs (although one of these is quite divergent from the consensus sequence). All truncating CYLD mutations are predicted to leave intact the N-terminal two CAP-GLY domains. Similar to CYLD, several other proteins with CAP-GLY domains have C-terminal metal-binding finger-like domains.11
The mechanism by which inactivation of a protein containing CAP-GLY domains might contribute to neoplastic transformation is not clear. Several CAP-GLY proteins are implicated in the attachment of the mitotic spindle to chromosomes.19–21 It is, therefore, plausible that CYLD is required for appropriate segregation of chromosomes during mitosis and that abnormalities of CYLD result in gains or losses of chromosomes. However, examination of 25 familial cylindromas for LOH on almost all chromosomal arms failed to reveal any evidence of allele loss other than on chromosome 16q.8 Similarly, comparative genomic hybridization (CGH) performed on 10 familial cylindromas showed very few copy number changes (Shipley and Stratton, unpublished data). These results therefore provide no evidence in favor of a phenotype of chromosomal instability in familial cylindromas carrying CYLD mutations.
Another major clue to the function of CYLD is the presence of two regions that could form a putative UCH Type 2 catalytic domain toward the C-terminus of the protein (www.expasy.ch/cgi-bin/get-prodoc-entry?PDOC00127; www.expasy.ch/cgi-bin/get-prodoc-entry?PDOC00750). UCH catalyzes the hydrolysis of ubiquitin resulting in deubiquitination, reducing degradation of target proteins by the proteasome.12 Therefore, inactivation of CYLD could conceivably contribute to oncogenesis by enhancing the degradation of proteins that, for example, suppress cell proliferation or promote apoptosis.22
Germ line mutations in the CYLD gene are remarkable for their tumor-type specificity. With the exceptions of trichoepitheliomas and eccrine spiradenomas (both of which also show features of skin adnexal structures), there is no evidence for predisposition to other types of neoplasm, either in the skin or other organs. Similarly, somatic CYLD mutations appear restricted to cylindromas and are not found in other tumors. Nevertheless, CYLD is highly expressed in fetal brain, testis, and skeletal muscle, and at a lower level in adult brain, leukocytes, liver, heart, kidney, spleen, ovary, and lung. It therefore seems unlikely that the propensity of CYLD mutations to cause neoplastic transformation in skin appendageal tissues is attributable simply to the tissue or cell-type specific expression pattern of the gene.
The biological activities of CYLD predicted on the basis of amino acid sequence require direct experimental confirmation. However, the presence of functional motifs that have not previously been found in proteins implicated in oncogenesis promises new insights into mechanisms of neoplastic transformation.