Transcriptional Regulation of Gene Expression

Eunice Kennedy Shriver National Institute Of Child Health & Human Development

Investigators

Linked publications, trials & patents

Abstract

We study the regulation of amino acid biosynthetic genes in budding yeast as a means of dissecting mechanisms of transcriptional control of gene expression, whose transcription is coordinately induced by activator Gcn4 during amino acid limitation. We and others showed previously that transcriptional activation by Gcn4 is enhanced by its recruitment of multiple cofactor complexes including SAGA, which contains the histone acetyltransferase (HAT) Gcn5, the chromatin remodeling (CR) complexes SWI/SNF and RSC, and the RNA Polymerase II (Pol II) Mediator, which in turn stimulate recruitment of general transcription factors (GTFs) and Pol II to the promoter to stimulate preinitiation complex (PIC) assembly. Seeking to identify all cofactors that mediate eviction of the -1 and +1 nucleosomes that occlude promoter DNA and impede binding of the GTF TATA-binding protein (TBP) and subsequent PIC assembly, we uncovered extensive cooperation among the chromatin remodeling complexes SWI/SNF, RSC, and Ino80C in evicting and repositioning promoter nucleosomes at genes activated by Gcn4 and other highly expressed genes in starved cells. We also showed that Ino80C acts in nucleosome eviction independently of its proposed function in replacing the specialized histone H2A.Z with conventional H2A in promoter nucleosomes. We further discovered that most of the occupied Gcn4 binding sites in the genome reside inside coding sequences (CDSs) rather than upstream of promoters and that such non-canonical binding events frequently induce bidirectional transcription from within the coding sequences in addition to the conventional full-length mRNA transcripts of the gene. Recently, we established that the NuA4 HAT complex acts on par with Gcn5, and functions additively, in evicting and repositioning promoter nucleosomes, supporting TBP recruitment, and stimulating transcription at both starvation-induced genes and the highly expressed ribosomal protein genes, whereas NuA4 is more important than Gcn5 at most other genes expressed constitutively that are principally dependent on TFIID versus SAGA for PIC assembly. SAGA subunits Spt3 and Spt8 act directly and non-redundantly with TFIID in TBP recruitment in the Gcn4 transcriptome. TFIID and SAGA are the principal coactivator complexes involved in recruitment of TBP to promoters. Published findings indicate that transcription of ~90% of yeast genes depends exclusively on TFIID, while the remaining ~10% require cumulative contributions from SAGA and TFIID (the coactivator-redundant or “CR” genes), presumably at the step of TBP recruitment. Both groups of genes additionally require the chromatin-modifying activities of SAGA, including Gcn5 HAT function. Other work suggests that genes induced by Gcn4 depend on the HAT but not TBP-recruitment function of SAGA, with TFIID executing TBP recruitment. To test this model, we examined the effects of deleting the genes encoding SAGA subunits directly involved in TBP recruitment (SPT3 and SPT8) or HAT activity (GCN5) on TBP recruitment and transcription at Gcn4 target genes. We found that deleting SPT3 and SPT8, but not GCN5, reduced induction of TBP binding whereas deleting either SPT3/SPT8 or GCN5 reduced Pol II occupancies at many SM-induced genes. In contrast, eliminating Gcn5, but not Spt3/Spt8, impaired promoter nucleosome eviction at many induced genes. Importantly, conditional nuclear depletion of the essential TFIID subunit Taf1 in cells containing or lacking Spt3/Spt8 revealed that TFIID contributes strongly to TBP recruitment only in the absence of Spt3/Spt8 at SM-induced genes and most highly-expressed “CR” genes. Thus, SAGA plays a critical, non-redundant role in TBP recruitment in the Gcn4 transcriptome via Spt3/Spt8 beyond its role in histone acetylation by Gcn5, which promotes transcription primarily via histone eviction. Mot1 regulation of promoter binding by TBP varies with stress and gene expression levels independently of coactivator-dependence It is thought that SAGA recruits TBP to promoters of stress-responsive genes with consensus TATA sequences while TFIID is more involved with housekeeping genes containing non-consensus TATA-like elements. Mot1 is a DNA-dependent ATPase that dissociates TBP from incomplete PICs in vitro. Previous studies indicated that in cells Mot1 favors PIC assembly at TATA-like TFIID-dependent promoters by preferentially dissociating TBP from consensus-TATA SAGA-dependent promoters. It was proposed that TBP bound to DNA in association with TFIID would be less accessible to Mot1 compared to TBP delivered by SAGA. However, more recent structures of TBP-TFIID-TFIIA-DNA complexes revealed only indirect interaction between TBP and one TFIID subunit (Taf4) that is bridged by TFIIA, making it unclear how Mot1 differentiates between SAGA- and TFIID-dependent promoters. We determined the effect of depleting Mot1 from nuclei on TBP binding at genes activated by Gcn4 during amino acid starvation, finding that that Mot1 is needed for strong TBP binding at the most highly activated Gcn4 target genes, which are generally SAGA-dependent. Mot1 depletion also reduced TBP binding at other highly expressed genes transcribed constitutively, while increasing TBP occupancies at weakly expressed genes, regardless of whether they rely on SAGA or TFIID in non-stressed cells. A similar pattern was found for genes highly induced by oxidative stress, which are not activated by Gcn4. Importantly, depleting Mot1 increased rather than decreased TBP occupancies at both groups of stress-induced genes in non-stressed cells where the genes are transcribed at low basal levels. Interestingly, while Mot1 depletion reduced TBP binding at highly expressed stress-induced and housekeeping genes alike, transcription declined only at TFIID-dependent genes, indicating that stress-activated genes controlled by SAGA have a reservoir of TBP in non-functional PICs whose formation requires Mot1. Our findings reveal that Mot1 redistributes TBP from weakly- to strongly-expressed genes rather than simply shifting TBP from SAGA- to TFIID-dependent genes and indicate further that stress-responsive genes have a unique mechanism allowing them to maintain robust transcription at reduced levels of TBP binding. Differential contributions of TATA sequences to asymmetric activation of divergently-oriented gene pairs by Gcn4 About 100 genes containing a Gcn4 binding site upstream of the transcription start site are not induced by Gcn4 in starved cells, of which 14 share the same binding site in a divergent gene pair (DGP) where the other gene is highly induced. We are probing the mechanisms that block Gcn4 activation of these unresponsive promoters using reporters containing the intergenic regions from various DGPs with either the unresponsive (Pun) or responsive (Pres) promoter driving HIS3 or FLUC expression. Mutational analysis of reporters for the DGP SNO1/SNZ1 revealed that converting two “weak” TATA-like elements in Pun-SNO1 to “strong” consensus TATAs elevates Gcn4-activation to the level of native Pres-SNZ1 containing its single “strong” TATA; and simultaneously weakening the strong TATA at SNZ1 completely reverses differential activation in this DGP to favor Pun-SNO1. The changes in reporter expression were associated with commensurate changes in TBP occupancies at the relevant TATA elements as determined by TBP ChIP-Seq. Thus, except for the requirement for two strong TATAs to fully activate Pun-SNO1, the quality of the TATA elements is a major determinant of differential Gcn4 activation at this DGP. For the YMR1/CPA2 DGP, by contrast, converting any of 5 weak TATAs to strong TATAs only weakly activated the unresponsive Pun-YMR1 promoter; and strikingly, introducing two strong TATAs at Pun-YMR1 and simultaneously eliminating the strong TATA in Pres-CPA2 still conferred greater expression from Pres-CPA2. Thus, preferential Gcn4 activation of Pres-CPA2 vs. Pun-YMR1 is not dictated by exclusive occurrence of a strong TATA in Pres-CPA2, implicating other unknown features in this DGP that strongly favor activation of Pres-CPA2.