Molecular cloning, characterization and alternative splicing of the human cytoplasmic serine hydroxymethyltransferase gene.

The human cytoplasmic serine hydroxymethyltransferase (CSHMT) gene was isolated, sequenced and its expression characterized in human MCF-7 mammary carcinoma and SH_5Y5Y neuroblastoma cells. The 23-kb gene contains 12 introns and 13 exons; all splice junctions conform to the gt/ag rule. The open reading frame is interrupted by 10 introns, two of which are positionally conserved within the human mitochondrial SHMT gene. The gene is expressed with 330 nucleotides of 5' untranslated message within three exons. The 5' promoter region does not contain a consensus TATA, and primer extension and 5'-RACE studies suggest that transcription initiation occurs at multiple sites. Consensus motifs for several regulatory proteins, including SP1, mammary and neuronal-specific elements, NF1, a Y-box, and two steroid hormone response elements, are present within the first 408 nucleotides of the 5' promoter region. The human gene is expressed as multiple splice variants in both the 5' untranslated region and within the open reading frame, all due to exon excision. The splicing pattern is cell-specific. At least six CSHMT mRNA splice forms are present in MCF-7 cells; the gene is expressed as a full-length message as well as splice forms that lack exon(s) 2, 9 and 10. In 5Y cells, the predominant form of the message lacks exon 2, which encodes part of the 5' untranslated region, but does not contain deletions within the open reading frame. Western analysis suggests that the CSHMT gene is expressed as a single full-length protein in 5Y cells, but as multiple forms in MCF-7 cells. Multiple tissue Northern blots suggest that the CSHMT message levels and alternative splicing patterns display tissue-specific variations.