Structural and functional analysis of the human TAF1/DYT3 multiple transcript system.

We analyzed TAF1/DYT3, a complex transcript system that is composed of at least 43 exons. Thirty-eight exons code for TATA box binding protein associated factor I (TAF1). Five downstream exons (d1-d5) of yet unknown function can either form transcripts with TAF1 exons or be transcribed independently. Splice variants can include d (notably d3 and d4) plus at least 12 TAF1 exons (exons 26-37 but not exon 38). These splice variants are highly polymorphic and include alternative exons (e.g., exons 30b, 31b, 32', 34', 35'). The frequency of these splice variants differs greatly in human fetal brain. Data were obtained by both RT-PCR and construction of a plasmid cDNA library. Promoter assays performed in NT2/D1 and in U87 cells demonstrate that TAF1-independent transcription of exons d2-d4 is driven by a TATA box-less promoter that is regulated by transcription factor Ikaros. Antisense transcription of exon d4 is under the control of a LTR promoter. While the 38 exons encoding TAF1 have been highly conserved in eukaryotes, the downstream exons d1-d5 were added to the transcript system much later during evolution and first appear in primates. The study demonstrates the structural and functional evolution of a complex transcript system.