Plastins are a family of at least three cytoplasmic protein isoforms that are expressed differentially between cells of the hematopoietic lineages and cells of solid tissues. Expression of the L-plastin isoform appears to be restricted to replicating blood cells, and the two T-plastin isoforms appear to be restricted to replicating cells of solid tissues. However, L-plastin is induced in many human solid tumor-derived cells. We used the anchored polymerase chain reaction technique to amplify and clone the missing 5' ends of plastin mRNAs. We found that both plastin isoforms contain a potential calcium binding site near the N terminus.Plastin was first noted as a polypeptide that appeared to be induced in human fibrosarcomas and other human tumorderived cell lines (2,(6)(7)(8). Initially, expression of a single plastin species was thought to be restricted in normal cells to hematopoietic cell lineages and activated in other cell types as a consequence of neoplastic transformation (2, 8). When plastin cDNAs were isolated from the human fibrosarcoma cell line HuT-14, we discovered that at least two distinct isoforms were coexpressed in this neoplastic cell line: Lplastin, which was normally expressed in blood leukocytes, and T-plastin, which was found in normal cells derived from solid tissues such as fibroblasts, endothelial and epithelial cells, keratinocytes, and melanocytes (8; J. Leavitt and C.-S. Lin, unpublished results). The high cytoplasmic abundance of L-plastin and its normal restriction to expression in circulatory cells led to speculation that its induction in tumor cells may govern some phenotypic characteristics of neoplastic cells, such as reduced cytoplasmic spreading on substrates and reduced anchorage dependence, both of which are normal properties of blood cells (2, 4). Initially, no clues as to the cellular function of plastin isoforms were apparent when we examined the amino acid sequences of the two isoforms. Other than the appearance of L-plastin in human cancer cells and its cytoplasmic location, our knowledge of plastins was limited to the observation that L-plastin was a substrate for phosphorylation at serine residues (1,2,5,12).A number of subsequent observations led us to question whether we had correctly identified the N-terminal amino acid sequence and the 5' ends of these cDNAs for both plastin isoforms. Initially, we identified the N-terminal sequence (8) on the basis of the finding that the open reading frames for both isoform cDNAs ended at a methionine residue (amino acid 58 in Fig. 1). We were unable to identify the N terminus of L-plastin by microsequencing because of N blocking (8). Second, the L-plastin clone contained an insert cDNA of 3.7 kilobases, the size of the L-plastin mRNA. Because the L-plastin cDNA was the size of the mRNA, we assumed that at least part of the upstream noncoding sequence of this cDNA represented the 5' untranslated region of the L-plastin mRNA. We then attempted * Corresponding author. to use this putative 5' sequence to map the transcription star...