Eukaryotic core promoters and the functional basis of transcription initiation

Abstract

RNA polymerase II (Pol II) core promoters are specialized DNA sequences at transcription start sites of protein-coding and non-coding genes that support the assembly of the transcription machinery and transcription initiation. They enable the highly regulated transcription of genes by selectively integrating regulatory cues from distal enhancers and their associated regulatory proteins. In this Review, we discuss the defining properties of gene core promoters, including their sequence features, chromatin architecture and transcription initiation patterns. We provide an overview of molecular mechanisms underlying the function and regulation of core promoters and their emerging functional diversity, which defines distinct transcription programmes. On the basis of the established properties of gene core promoters, we discuss transcription start sites within enhancers and integrate recent results obtained from dedicated functional assays to propose a functional model of transcription initiation. This model can explain the nature and function of transcription initiation at gene starts and at enhancers and can explain the different roles of core promoters, of Pol II and its associated factors and of the activating cues provided by enhancers and the transcription factors and cofactors they recruit.

Box 1 figure

Box 2 figure

Figure 1. Properties and function of core promoters and enhancers.

a) The traditional view of transcription initiation postulates that transcription initiates at gene core promoters, which recruit the transcription machinery consisting of RNA polymerase II (Pol II) and general transcription factors (GTFs), thereby leading to the formation of the pre-initiation-complex (PIC) and transcription initiation. Transcription from core promoters is activated by enhancers, which can be located distally and bind sequence-specific transcription factors (TF), which recruit cofactors (COF) that convey the activating cues to the PIC at the core promoter. (b) Active enhancers exhibit divergent transcription of short, unstable enhancer RNAs (eRNAs) from two separate transcription start sites (TSSs) located at the edges of the nucleosome-depleted region where the enhancer resides. (c) Promoters produce long, stable mRNAs from a gene core promoter in the sense direction (orientation of the gene) and short, unstable upstream antisense RNAs (uaRNAs) from the upstream edge of a nucleosome depleted region that contains the transcription factor-bound proximal promoter. Separate pre-initiation complexes drive unidirectional transcription from each of the two TSSs.

Figure 2. Regulation of different steps of transcription from core promoters.

a) Pre-initiation complex (PIC) assembly and RNA polymerase II (Pol II) recruitment. The first step of transcription initiation is the assembly of the PIC consisting of Pol II and six general transcription factors (GTFs): transcription factor IIA (TFIIA), TFIIB, TFIID, TFIIE, TFIIF and TFIIH (left). Enhancers can promote PIC assembly by recruiting transcription factors (TFs) and cofactors (COFs) that directly interact with GTFs or Pol II (right). b) Initiation by Pol II and ’promoter escape’. After PIC assembly, the DNA duplex at core promoters melts (not shown) and allows Pol II to initiate transcription at the transcription start site (TSS). To continue transcribing, Pol II has to dissociate (escape) from the TSS-binding GTFs, which is mediated by phosphorylation of Ser 5 and Ser 7 of the Pol II carboxy-terminal domain (CTD) by TFIIH. Enhancers can aid this process by recruiting cofactors such as the Mediator complex (MED) or the acetyltransferase CBP/P300 (see main text for these and other cofactors’ functions). c) Pol II promoter-proximal pausing. After escaping from the TSS, Pol II synthetizes a short stretch of nascent RNA (30-50 nucleotides) and then pauses downstream of the TSS. DRB sensitivity inducing factor (DSIF) and negative elongation factor (NELF) bind to Pol II and the nascent RNA and promote Pol II pausing. Pause-release is mediated by cyclin-dependent kinase 9 (CDK9), which is a subunit of the positive transcription elongation factor b (P-TEFb) that phosphorylates DSIF, NELF and Ser 2 of the Pol II CTD. This leads to dissociation of NELF and entry of Pol II into productive elongation. Enhancers promote this process by recruiting cofactors that either recruit and stimulate CDK9 or directly affect pause-release, such as Brd4 and p300. d) Regulation of transcription bursting. Transcription occurs in short ‘bursts’, which comprise groups of initiation events separated by periods of inactivity. The core promoter sequence determines burst size, that is the number of transcribing Pol II molecules per burst (left), while enhancers increase bursting frequency from their target core promoter (right). ‘+’ denotes target activation and ‘-‘ denotes target inhibition.

Figure 3. Sequence-Bencoded specificity of core promoters towards enhancers and activation by specific transcription (co)factors.

Different types of core promoters respond differentially to distal enhancers, that is an enhancer can activate them (solid arrows) or not (dashed arrows). This selectivity or specificity is mediated by different transcription factors (TF) and cofactors (COF), which display core promoter preferences likely based on biochemical compatibilities between the cofactors and core promoter-bound general transcription factors (GTFs). Mapping and understanding preferences and compatibilities between cofactors and core promoters is an important goal for future research. Pol II, RNA polymerase II; TBP, TATA-box binding protein; TRF2, TBP-related factor 2.

Figure 4. Functional model of transcription initiation at genomic promoters and enhancers.

a) Model of transcription initiation at enhancers (left) and promoters (right) arising from their distinct sequence-encoded activities. Enhancers bind transcription factors (TF) and recruit cofactors (COF), thereby creating a high local concentration of transcription activators. This should lead to fortuitous transcription initiation at proximal sites that resemble bona fide core promoters (“best-of-random sites”), resulting in divergent transcription of short unstable enhancer RNAs (eRNAs). Promoters transcribe stable mRNAs from a dedicated gene core promoter and – due to high activator concentration – will also show fortuitous transcription initiation in the antisense direction. b) Model of evolution of a functional core promoter or an enhancer. Newly emerging transcription-factor binding sites (blue) create enhancer-like activity and exhibit low levels of bidirectional transcription at best-of-random sites. If such transcription is harmful, it might be actively suppressed by DNA methylation (pins), repressive factors or repressive chromatin and the transcription factor binding sites will degenerate over time. If by contrast the transcription in one or both directions is beneficial, the respective transcription start site will be positively selected and evolve to a fully functional core promoter (red) with strong core promoter motifs, able to support high levels of regulated and productive transcription. Transcription in the non-beneficial direction will remain low and yield non-stable upstream-antisense RNAs (uaRNAs). The activator binding sites near core promoters are often referred to as ‘proximal promoter’. Finally, if the transcription from a putative regulatory sequence is neutral and its enhancer activity is beneficial, the enhancer function should be strengthened and the enhancer will transcribe low levels of bidirectional eRNAs from best-of-random sites.