UC San Diego

The manifold species we marvel today reflect the diverse regulatory mechanisms facilitating differential gene expression and therefore cell differentiation. Transcription is the initial step in gene expression and three RNA polymerases (RNAPs) with reportedly non- overlapping functions reside in the human nucleus. All three RNAPs share one common ancestor, which may be reflected in the finding that RNAP III accurately initiates transcription from RNAP II promoters. These results propose a functional conservation among RNAP II & III and further imply polymerase specificity as not constant, but a variable that depends on promoter properties and transcription conditions. The regulation of gene expression was previously assumed to be primarily mediated by sequence-specific transcription factors. It has now become evident that also the basal transcription machinery greatly diversified during evolution. Independent duplications of the TATA-binding protein (TBP) yielded paralogs, termed TBP-related factors (TRFs) which support different transcription systems. For example, TRF2 but not TBP is required for TCT core promoter motif- dependent transcription in Drosophila melanogaster while TBP but not TRF2 is essential for the initiation of transcription from TATA box-dependent promoters. The observation that TBP family proteins each support distinct sets of transcriptional programs led to the proposal to refer to them as "system factors". Notably, subdivision of the transcription systems likely reduced constraints, thereby facilitating evolvability and ultimately specialization. One example for specialization was encountered by the functional characterization of the TCT motif that was found to be highly enriched at the promoters of genes involved in translation. These findings propose TCT-dependent transcription to constitute a conserved and specialized RNAP II system complementing RNAP I and III to ensure translation, thereby proving an example how distinct core promoter motifs and TRFs facilitates regulatory specialization. In addition, the directionality of the human transcription apparatus was analyzed which revealed core promoter and the basal transcription apparatus to be highly directional. Using DNase I accessibility defined promoter borders revealed that about half of the promoters are unidirectional. Reverse-oriented transcripts of bidirectional or divergent promoters originated from their own cognate reverse- directed core promoter. Unidirectional promoters are depleted of reverse core promoter sequences or associated chromatin features at their upstream edges