Genomic structure of the human caldesmon gene.

The high molecular weight caldesmon (h-CaD) is predominantly expressed in smooth muscles, whereas the low molecular weight caldesmon (l-CaD) is widely distributed in nonmuscle tissues and cells. The changes in CaD isoform expression are closely correlated with the phenotypic modulation of smooth muscle cells. During a search for isoform diversity of human CaDs, l-CaD cDNAs were cloned from HeLa S3 cells. HeLa l-CaD I is composed of 558 amino acids, whereas 26 amino acids (residues 202-227 for HeLa l-CaD I) are deleted in HeLa l-CaD II. The short amino-terminal sequence of HeLa l-CaDs is different from that of fibroblast (WI-38) l-CaD II and human aorta h-CaD. We have also identified WI-38 l-CaD I, which contains a 26-amino acid insertion relative to WI-38 l-CaD II. To reveal the molecular events of the expressional regulation of the CaD isoforms, the genomic structure of the human CaD gene was determined. The human CaD gene is composed of 14 exons and was mapped to a single locus, 7q33-q34. The 26-amino acid insertion is encoded in exon 4 and is specifically spliced in the mRNAs for both h-CaD and l-CaDs I. Exon 3 is the exon that encodes the central repeating domain specific to h-CaD (residues 208-436) together with the common domain in all CaD (residues 73-207 for h-CaD and WI-38 l-CaDs, and residues 68-201 for HeLa l-CaDs). The regulation of h- and l-CaD expression is thought to depend on selection of the two 5' splice sites within exon 3. Thus, the change in expression between l-CaD and h-CaD might be caused by this splicing pathway.