The unexpected traits associated with core promoter elements

Abstract

The core promoter of eukaryotic coding and non-coding genes that are transcribed by RNA polymerase II (RNAP II) is composed of DNA elements surrounding the transcription start site. These elements serve as the docking site of the basal transcription machinery and have an important role in determining the position and directing the rate of transcription initiation. This review summarizes the current knowledge about core promoter elements and focuses on several unexpected links between core promoter structure and certain gene features. These include the association between the presence or absence of a TATA-box and gene length, gene structure, gene function, evolution rate and transcription elongation.

Introduction

Diversity in rates of gene expression is essential for basic cell functions and is controlled through several intricate mechanisms. Major contributors to gene expression rates are DNA cis-regulatory elements that vary between promoters of individual genes. Two types of DNA regulatory sequences control transcription of protein-encoding and non-coding genes in eukaryotes. The first type is gene specific enhancer elements that serve as the binding sites of transcription regulatory factors and can be divided into two classes: those that function independently of their position relative to the transcription start site (TSS) and those that can activate transcription only when located proximal to the TSS. The second type is the core promoter, which consists of sequence elements that surround the TSS. These elements serve as a docking site for general transcription factors (GTFs) and RNA polymerase II that assemble into a pre-initiation complex (PIC).^1,2 The core promoter has a crucial role in transcription as it serves as the acceptor site for the effects exerted by enhancer-bound transcription factors; it contributes to the overall transcription level and it determines the site of transcription initiation. Thus, the information encoded in the core promoter ensures proper regulation of gene expression. This review summarizes the current knowledge about core promoter types and their mechanism of action and focuses on several traits associated with the core promoter, some of which run beyond transcription initiation.

Core Promoter Elements

The first core promoter element to be described in eukaryotes was the TATA-box.³ The TATA-box is a highly conserved element, strictly located between −35 to −25 relative to the TSS (designated +1) in most eukaryotes. The TATA-box, which was once thought to be a universal element, is present in a smaller fraction of RNAP II genes than initially estimated (Table 1): between 20–46% in yeast (depending on the definition of the TATA-box sequence),^4,5 ∼30% of Drosophila genes⁶ and up to 35% of human genes.^7–10 The major core promoter-binding factor is TFIID, a large complex consisting of the TATA-binding protein (TBP) and 13 associated factors called TAFs. The TATA-box is directly recognized and bound by TBP, while the TAFs interact with sequences upstream and downstream to the TATA-box.^11–19 In certain promoters, the TATA-box cooperates with one or more elements to direct efficient transcription initiation (Fig. 1). For example, two TFIIB recognition elements (BRE), which are located either upstream (BREu) or downstream (BREd) of the TATA-box^20,21 function only together with the TATA-box. Similarly, the contribution to promoter strength of the TAF1 recognition element DCE that is located downstream relative to the TSS is also dependent on the presence of a TATA-box.²² The presence or absence of a TATA-box in core promoters has been linked in yeast^4,5,23–25 and humans²⁶ to two pathways of pre-initiation complex assembly, one being TFIID dependent (weak TATA or TATA-less) and the other TFIID independent and SAGA dependent.

Table 1.

Prevalence of the TATA box in various species

Species	Sequence	Fraction (%)	Ref.
Saccharomyces cerevisiae	TATAWAWR	∼20	3
Saccharomyces cerevisiae	TATAWA	45.8	4
Drosophila	TATAAA (up to one mismatch)	29.3	5
Homo sapiens	TATAWA (up to one mismatch)	8.3	8
Homo sapiens	TATAWA (two mismatches)	27	8

Figure 1.

A scheme of several core promoters with a major TSS(s) which are governed by specific functional combinations of core promoter motifs.

The initiator (INR) is a metazoan conserved element, strictly located around the TSS, with a consensus of YYANWYY. The INR can be weakly bound by RNAP II itself²⁷ and more strongly by a complex consisting of TFIIB, TFIID, TFIIF and RNAP II.^27,28 Within TFIID the subunits that form direct and specific contacts with the INR are TAF1 and TAF2.^12,29,30 The INR can function alone, together with the TATA-box or in conjunction with two specific core promoter elements, the DPE and the MTE (Fig. 1). The DPE and the MTE are mostly found in Drosophila promoters and both have a strict downstream location at +28 and +18, respectively, relative to the TSS, and both are recognized by TFIID through the TAF6 and TAF9 subunits.^31–34 Following computational analysis of a large number of mammalian TSSs, the INR consensus in mammals was recently suggested to be composed of only YR, where R corresponds to the TSS.^35,36 On the other hand, another study has identified, also by computational analysis a version of the INR, called “strict Initiator” (sINR), that is much less divergent than the INR as, unlike the INR, its core sequence is very strict and is flanked by additional conserved sequences, not shared by the INR.³⁷ sINR is specifically enriched in TATA-less promoters and functions in cooperation with a nearby Sp1 site (Fig. 1). Interestingly, while sINR can substitute for a canonical INR, it cannot be replaced by an INR, indicating that the small sequence variations are functionally very important.³⁷ Another element with an INR positional bias is the pyrimidine-rich TOP element present in many protein biogenesis genes.^38,39 The TOP element was recently reported to be active in Drosophila as well.⁴⁰

Even though a substantial fraction of promoters lack both TATA-box and INR there is only a limited number of characterized core elements that function independently of a TATA-box or INR. Two such elements called, XCPE1 and XCPE2, were identified in the hepatitis B virus X gene, each one directing a distinct TSS.^41,42 XCPE1 has a consensus sequence of DSGYGGRASM and is located from −8 to +2 relative to the TSS. It is present in ∼1% of human core promoters and acts only in conjunction with other sequence-specific activators such as the NRF1, NF-1 and Sp1.⁴² XCPE2 has a VCYCRTTRCMY consensus and drives transcription that is independent of TAFs but dependent on TBP and the mediator.⁴¹ A third element is MED-1 found in TATA-less promoter with unclustered, multiple start sites.⁴³

Several bioinformatics studies found that most of the mammalian TATA-less promoters are associated with CpG islands.^44–47 Another poorly investigated phenomenon characteristic to a large number of TATA-less and INR-less promoters is transcription initiation from multiple sites, as opposed to the single major site that is characteristic of promoters driven by a TATA-box or INR.³⁶ It is very likely that in such promoters the mechanism of transcription initiation is quite different. With a strict TSS site the PIC associates with the promoter through a specific site (TATA, INR or other), while with dispersed TSSs the PIC may not be associated with one specific element. One possibility is that the general machinery is recruited by a transcription factor bound to a proximal promoter element. In the absence of a direct docking site, RNAP II is more flexible and can initiate transcription at favorable nucleotides in the vicinity of the element. Another possibility is the presence of a number of docking sites on the same promoter to which the general machinery can weakly bind and direct transcription initiation. The two possibilities are not necessarily mutually exclusive.

Proximal Promoter Elements and their Impact on mRNA Translation

Another group of promoter elements that might contribute to the basal activity of promoters are proximal promoter elements that are specifically located near the TSS. Among these, Sp1 is likely to have a general role as our computational analysis revealed that there are one or more Sp1 binding sites in the proximal promoter region of ∼60% of human genes, making it the most prevalent motif in human promoters (Elfakess R and Dikstein R, unpublished data). Bioinformatics studies have identified numerous over-represented motifs that are located very close to the transcription start site. Some of these motifs are the binding sites of well-known transcription factors such as Ets-1, NRF1, NRF2 and CREB, while others have not yet been characterized.^7,10

A recently described interesting proximal element is TISU, an element that is located downstream relative to the TSS between +5 to +30, specifically in TATA-less promoters.⁴⁸ Remarkably, in addition to its contribution to promoter strength it also serves as a strong translation initiator that is optimized to function in mRNAs with very short 5′UTR.⁴⁹ The pyrimidine-rich TOP element described above, which encompasses the mRNA 5′end, was also characterized as a translation regulatory element that plays an important role under various stress conditions.^50,51 Thus, it is possible that regulatory elements common to the promoter and the mRNA are involved in coordination between these two major steps of gene expression.

Association of Core Promoter Elements with Specific Gene Functions

As the number of distinct core promoter elements being revealed is gradually increasing, it is clear that the core promoter is not the ubiquitous element once considered to be common to most genes. In fact, much like gene specific enhancer elements, certain core promoter elements were found to be associated with specific biological activities. For example the TATA-box (TATAWA, with up to one mismatch) is specifically enriched among tissue specific genes and genes involved in development, response to wounding, response to external stimulus, inflammatory response and chromatin assembly.^9,10,47 In yeast, it is overrepresented among inducible and stress response genes.⁴ In Drosophila, the DPE is specifically enriched in homeotic (Hox) gene promoters and has been demonstrated to be functionally important for the transcription activation by Caudal, a key regulator of the Hox genes that displays a preference for the DPE.⁵² The sINR-containing TATA-less genes are specifically enriched with biological activities related to RNA metabolism, such as RNA processing and synthesis, nucleic acid metabolism and chromosome organization and biogenesis.³⁷ These findings indicate that core promoter elements are an integral part of the combinatorial strategy of transcription regulatory programs.

The Core Promoter and the Transcription Elongation Phase

While the central role of the core promoter is related to the initiation phase of transcription, it can also impact on transcription elongation. Early studies suggested that the core promoter has a role in RNA RNAP II processivity. It was found that a paused but not processive transcription is dependent on an intact TATA-box for HIV-LTR and c-myc transcription.^53,54 More recently, the importance of the arrangement of an NFκB enhancer with a core promoter for elongation control was examined. The transcriptional activity of NFκB affects both the re-initiation rate and transcription elongation.^55–59 At the elongation level, NFκB target genes are divided into two groups: those regulated by the positive elongation factor P-TEFb and those regulated by the negative elongation factor DSIF.²⁶ Whether P-TEFb or DSIF will regulate elongation depends on NFκB itself, the core promoter type and the initiation complex: P-TEFb controls NFκB target genes driven by TATA-containing promoters, which are also TFIID independent, whereas DSIF controls weak TATA and TATA-less genes that are dependent on TFIID. Changing a TATA-less to a TATA core promoter was sufficient to switch the regulation from DSIF to P-TEFb, and from TFIID dependency to TFIID independency.²⁶ This mode of regulation may not be restricted to the NFκB pathway as a another study reported that regulation of the c-myc gene by P-TEFb is also dependent on the TATA-box.⁶⁰ Considering that RNAP II is the major component common to the initiation and elongation phases, it is likely that the specific type of pre-initiation complex assembled on the distinct core promoters influences the type of elongation factor that is recruited.

The Noisy TATA-Box

Genetically identical populations of cells in bacteria, yeast and mammals exhibit cell-to-cell variations in the amount of proteins, a phenomenon termed gene expression noise.^61–65 This inherent variation is thought to contribute to their phenotypic diversity and to increase fitness.^66,67 It has been found that different genes exhibit different levels of noise indicating that there are mechanisms that enhance or suppress the intrinsic noise in gene expression. Cell-to-cell variations can be derived from two sources, extrinsic or intrinsic: extrinsic sources are global factors such as transcription factors, cell size, cell cycle, the cell local environment and others; intrinsic sources derive from molecular events involving small numbers of molecules such as mRNA transcription or protein translation.^62,68 Each step of gene expression can contribute variations,^64,69–71 and in yeast it has been shown that the presence of a strong TATA-box increases the noise in gene expression.^69,72,73 The reported intrinsic variability is suggested to result from random transitions between an active state, in which mRNA is transcribed at a high rate (burst), and an inactive state, in which mRNA is transcribed at a much lower rate. The contribution of other core promoter elements to noise is presently unknown.

Association of Core Promoter with Gene Length and the Relationship with Expression

The core promoter has been recently reported to be linked to structural features of genes. A statistical analysis of more than 14,000 human genes that were classified into groups according to their core promoter type revealed a remarkable observation. Genes with a TATA-box are, on average, 3-fold shorter than TATA-less genes. Furthermore, within the TATA containing genes, length is inversely correlated to the strength of the TATA-box, with one mismatch from the TATA-box consensus being associated with a more than 2-fold increase in gene length over canonical TATA genes. Differences in gene length are primarily due to the size and number of introns.⁹

Analysis of gene expression data of genes in the different core-promoter groups revealed the expected correlation between the strength of the TATA-box and expression levels,⁹ confirming previous gene-specific studies.^69,74,75 In general an increase in gene length also correlates with reduced levels of expression, but the impact of gene length varies according to the core promoter. The inverse correlation of gene length with expression was found to be the highest for genes with a TATA-box and lowest for TATA-less genes. Having a TATA-box in the core promoter seems beneficial for expression of short genes, while its advantage diminishes with longer genes. We have therefore proposed that substantial variation in gene expression levels can be achieved through different combinations of TATA promoters with varying intron length. On the other hand, a TATA-less promoter ensures similar levels of expression regardless of gene length.

The sensitivity of TATA-box genes to increased gene length may be related to the bursty nature of transcription initiation directed by the TATA-box element.^69,76 Theoretical calculations predicted that bursty transcription initiation would be highly sensitive to elongation interruptions⁷⁷ because transcription initiation bursts create localized pools of RNAP II molecules over the gene. When elongation is interrupted, the leading RNAP II pauses or stalls, thereby increasing the risk that the following RNAP II molecules will collide and destroy the burst. Clearly, the chances of RNAP II to encounter an obstacle during elongation increase in proportion to gene length, which may affect initiation bursts more frequently. In contrast, in non-bursty transcription, the distance between each RNAP II molecule allows sufficient time for RNAP II to clear the block so risk of RNAP II collision is lower.

The TATA-Box and Evolution

The presence or absence of a TATA-box has been also linked to the rate at which genes are evolved. Examination of the transcriptional responses of four closely related yeast species to a variety of environmental stresses revealed that genes containing a TATA-box show an increase in interspecies variability in expression. This enhanced expression divergence of TATA-containing genes was confirmed in all eukaryotes.⁷⁸ Another study that examined the effects of naturally occurring mutations on gene expression in yeast also found that the sensitivity of gene expression to mutations is higher for genes with a TATA-box.⁷⁹ Thus, the transcription initiation mechanism associated with the TATA-box may facilitate evolution in gene expression.

Conclusions and Perspectives

Although the core promoter is a key element in gene transcription, we still know very little about its structure and function in most of the TATA-less genes. It is likely that there are several mechanisms, yet to be discovered, by which distinct core promoter elements cooperate with distal enhancer elements to increase the rate of transcription. Additional unresolved issues that are of interest include: the identity of the trans-acting factors that bind to the different core promoter elements; the chromatin features associated with the different core promoter elements; the interplay between new core promoter elements and the succeeding stages of gene expression; the association of core promoter elements, other than the TATA-box, and gene features directly and indirectly related to transcription. Undoubtedly, a more systematic analysis of this basic component of transcriptional control is required in order to increase our ability to read the regulatory information encoded by the genome.