The SAGA/TREX-2 subunit Sus1 binds widely to transcribed genes and affects mRNA turnover globally
Eukaryotic transcription is regulated through two complexes, the general transcription factor IID (TFIID) and the coactivator Spt–Ada–Gcn5 acetyltransferase (SAGA). Recent findings confirm that both TFIID and SAGA contribute to the synthesis of nearly all transcripts and are recruited genome-wide in yeast. However, how this broad recruitment confers selectivity under specific conditions remains an open question.
Here we find that the SAGA/TREX-2 subunit Sus1 associates with upstream regulatory regions of many yeast genes and that heat shock drastically changes Sus1 binding. While Sus1 binding to TFIID-dominated genes is not affected by temperature, its recruitment to SAGA-dominated genes and RP genes is significantly disturbed under heat shock, with Sus1 relocated to environmental stress-responsive genes in these conditions. Moreover, in contrast to recent results showing that SAGA deubiquitinating enzyme Ubp8 is dispensable for RNA synthesis, genomic run-on experiments demonstrate that Sus1 contributes to synthesis and stability of a wide range of transcripts.
Our study provides support for a model in which SAGA/TREX-2 factor Sus1 acts as a global transcriptional regulator in yeast but has differential activity at yeast genes as a function of their transcription rate or during stress conditions.
KeywordsSus1 SAGA Transcription GRO ChIP-exo
transcription factor IID
RNA polymerase II
histone acetyltransferase activity
transcription and export complex
environmental stress response
gene set enrichment analysis
relative mRNA abundance
transcription start site
transcription termination site
Expression of protein-coding genes is a coordinated process that consists of a series of linked steps in mRNA fate: transcription, processing, export to the cytoplasm, translation and degradation. Regulation of gene expression is essential to cellular metabolism and to adaptation to stress and developmental processes. A key step in eukaryotic transcription is the binding of a sophisticated repertoire of regulators that collectively function to control the timing of expression of each gene. Eukaryotic RNA polymerase II (RNAPII) genes have been classically divided into two main groups attending to the presence or absence of the TATA sequence in their promoters . While transcription of all genes requires the loading of the TATA-binding protein (TBP) onto their promoter sequences, differences exist in the co-regulators that are used. TATA-like containing genes are mainly dependent on the general transcription factor IID (TFIID), while the SAGA coactivator (Spt–Ada–Gcn5 acetyltransferase) plays a relatively greater role in the expression of TATA-box genes [2, 3]. Accordingly, yeast genes have been called TFIID- and SAGA-dominated, depending on which complex has the strongest influence on their transcription . The SAGA complex is widely conserved from yeast to humans and acts during different stages of gene expression (reviewed in [4, 5, 6]). Interestingly, and in contrast to SAGA, TFIID seems to play a role restricted to preinitiation events. In addition to SAGA, a similar coactivator, referred to as SLIK (SAGA-LIKe), has also been described [2, 3, 7, 8, 9]. The subunit composition is identical to that of SAGA, except that in SLIK the Spt7 subunit is C-terminally truncated by proteolytic cleavage, leading to release of the Spt8 subunit.
SAGA and SLIK are organized in submodules [4, 5, 6, 10] and have two inherent catalytic activities on nucleosomal histones: histone acetyltransferase activity (HAT) [11, 12] and histone H2B deubiquitylation (DUB) [13, 14, 15]. Both activities are essential for the post-translational modification of the histone tails, which provokes a shift from compact/inactive to open/active chromatin states . SAGA is formed by distinct modules that work together to localize the HAT and DUB activities of the complex in the gene promoter, a necessary step for enhancing transcriptional activation, facilitating elongation and promoting nucleosome eviction.
Apart from SAGA-dependent events affecting histone modifications during transcription initiation, several studies have identified SAGA roles other than those of upstream regions. In fact, yeast SAGA subunits are not restricted to promoters, as they also localize to coding sequences of specific genes [17, 18, 19]. In these locations, SAGA subunits co-transcriptionally promote nucleosome eviction through Gcn5’s HAT activity and the deubiquitylation of H2B by Ubp8. Furthermore, SAGA interacts functionally and physically with TREX-2 (transcription and export complex 2), a nuclear pore-associated complex involved in genome stability, mRNA biogenesis and export [20, 21, 22, 23]. Through components of the SAGA deubiquitylation module, SAGA and TREX-2 co-localize at the nuclear periphery . Sus1, a shared SAGA/SLIK/TREX-2 factor, also localizes to promoters and coding regions of some SAGA-dominated genes (ADH1 and GAL1), where it promotes the interaction with export factors, contributing to the coupling of transcription and mRNA metabolism .
Deciphering how SAGA’s different enzymatic activities control the expression of their target genes is a major challenge in fundamental transcription research. Classical results from genome-wide approaches suggested that SAGA binds to a subset of promoters in yeast [26, 27, 28, 29, 30] that are downregulated in SAGA mutants . SAGA predominates at promoters of highly regulated stress-responsive TATA-box-containing genes, which comprise about 10% of the genome. Interestingly, there is not a clear correlation between the binding of SAGA and the misregulation of bound genes [28, 31, 32]. In fact, it has been shown that > 99% of the expressed yeast genome is positively regulated by the overlapping involvement of TFIID and SAGA . Moreover, enzymatic activities of SAGA subunits, i.e. histone acetylation and deubiquitylation, seem to play roles that go beyond the transcriptional initiation of SAGA-dominated genes. For example, a previous study showed that SAGA acetylates histones at promoter regions as well as deubiquitylates histone H2B from transcribed region of all expressed RNAPII genes . Furthermore, the separated roles of SAGA and TFIID in the regulation of TATA-containing and TATA-like genes have recently been reconsidered. New results confirm that transcription of nearly all genes is dependent on TFIID  and that SAGA may act as a general co-factor required for RNAPII transcription in Saccharomyces cerevisiae . In contrast, other studies indicate that SAGA is not a general transcription factor in flies and mammals [36, 37, 38].
In this study, we show for the first time the genome-wide recruitment of the SAGA/TREX-2 factor, Sus1, to RNAPII-transcribed genes. Sus1 accumulates at upstream activator sequences (UASs) at a variety of gene classes: ribosomal protein (RP), SAGA-dominated and TFIID-dominated genes. The amount of Sus1 binding is correlated with the transcription rate (TR) of target genes. Analyses of Sus1 chromatin recruitment under heat shock revealed that while this stress scarcely affected Sus1 binding to TFIID-dominated genes, the binding pattern at SAGA and RP genes was substantially affected. Consistent with the genome-wide association of Sus1 to UAS, we found that transcription rates and stabilities of most yeast transcripts are changed in the absence of Sus1. Lastly, we demonstrate that both SAGA and SLIK participate in Sus1 recruitment to SAGA-dominated genes. In sum, our results support a model in which Sus1 is intimately tied to the expression of yeast RNAPII genes.
Sus1 is recruited genome-wide to promoters of RNAPII-transcribed genes
Sus1 association to UAS depends on growth conditions
Distinct mobilization of SAGA- and TFIID-linked regulators occurs during acute heat shock [28, 30]. The heat-shock response in yeast involves transient changes in transcriptional rate, mRNA levels and mRNA stabilities of approximately 1000 genes within 15 min of the stress [40, 41, 42]. Different gene clusters showed accumulation of different complexes that regulate transcription, and SAGA-dominated genes had the largest increase in occupancy of many regulators . To study the extent to which the genome-wide association of Sus1 to all gene classes is affected by heat shock, we conducted ChIP-exo experiments of Sus1 binding at 37 °C (Fig. 1a, red line) and compared to data at 25 °C (Fig. 1a, blue line). We discovered that while heat shock barely affected Sus1 binding to TFIID genes, it significantly shifted the Sus1-binding pattern at SAGA and RP genes. While Sus1 binding to RP genes was significantly reduced, the presence of Sus1 at SAGA-dominated genes was augmented. Examples of Sus1 association to each class of genes at both temperatures are shown in Fig. 1c. Gene set enrichment analysis (GSEA) of the Gene Ontology (GO) terms of the top in the ordered list of bound genes at each temperature indicated that at 25 °C Sus1 preferentially binds to cytoplasmic translation and ribosomal subunit genes, while at 37 °C protein folding and response to heat genes are most associated with Sus1 (Additional file 2: Fig. S2).
Sus1 affects the synthesis and stability of mRNAs at a genome-wide level
Sus1 binding shows a positive correlation with transcriptional rate
Our analyses revealed that upon heat shock Sus1 binding to RP genes was significantly reduced, while its presence at SAGA-dominated genes was augmented (Fig. 1b). It has been shown that many SAGA-dominated genes are part of the ESR-up transcripts [2, 40, 45, 46, 47, 48], whereas RP genes are a fundamental part of the ESR-down response. Therefore, we hypothesized that Sus1 would release from RP genes and associate preferentially to ESR-up genes upon stress . To confirm this hypothesis, the binding of Sus1 to RP and ESR-up genes was studied before and after this temperature shift. As shown in Fig. 4b, Sus1 binding to ESR-up genes was lower than the global genome average in a wild-type strain at 25 °C. In contrast, association of Sus1 to ESR-up genes was significantly augmented upon heat shock. In contrast, Sus1 binding to RP genes dropped abruptly when cells were incubated at 37 °C for 15 min (Fig. 4b). These results support an active role for Sus1 in the transcriptional regulation of the environmental stress response in yeast.
SLIK participates in Sus1 recruitment to SAGA-dominated genes
In this study, we demonstrate a general role for Sus1 in transcription and mRNA stability in yeast cells. Sus1 is recruited to the upstream activation sequences of many genes and is necessary for mRNA synthesis of a large number of transcripts independently of the presence of a TATA sequence. The classical model of the role of SAGA/TFIID in regulating transcription states that they possess distinct but overlapping selectivity for promoters, thereby fostering a cascade of events that leads to the correct expression of their targeted genes. In classical studies, TATA-containing genes were relatively more dependent on SAGA, whereas TATA-like genes were relatively more dependent on the general transcription factor TFIID . Numerous ChIP studies of SAGA factors revealed a picture of confined recruitment of SAGA to specific promoters and to some coding sequences [17, 18, 19, 28, 30, 51]. However, recent work that tried to delimit the specificity of SAGA regulation concluded that TFIID and SAGA are more equivalent than initially thought [34, 35]. These studies reported that SAGA-associated enzymatic activities act more widely at all transcribed genes . In particular, it was shown that histone acetylation and deubiquitylation mediated by the SAGA-associated proteins Gcn5 and Ubp8, respectively, occur genome-wide, suggesting that SAGA has a broader role in gene expression. However, we suggest that contributions from subcomplexes that are not part of SAGA could not be excluded. The current view is further supported by two recent publications describing TFIID and SAGA as general cofactors for RNAPII transcription in yeast [34, 35]. SAGA factors Spt7, Spt3, Spt8 and Ubp8 were reported to associate with UASs at a majority of genes, which contrasts with recent results showing much more restricted localization . Sus1 (this study) and SAGA () recruitments are enriched for highly transcribed genes and peak in the -150/-200-bp region coinciding with the UAS and NFR of many yeast genes, with particular enrichment of SAGA subunits at RP and SAGA-dominated genes. Further work is needed to resolve the discrepancy between the discordant studies.
Seminal studies that established the regulatory role of SAGA and the functional clustering of its subunits measured steady-state RNA levels, without differentiating between transcription and degradation rates [2, 31]. In this work, we measured separately transcriptional rate and mRNA abundance, and discovered that i) the absence of SUS1 has a strong negative effect on the transcription rate and ii) mRNAs were generally more stable in the sus1Δ mutant. Similar phenotypes have recently been described for the SAGA/SLIK structural mutants spt7Δ and spt20Δ  and for some other SAGA subunits . The GRO analyses of sus1Δ presented here complement these observations, suggesting that Sus1 is part of the SAGA-related factors that influence the stability of RNAPII-transcribed mRNAs. However, since Sus1 is a shared component of SAGA/SLIK and TREX-2, we cannot attribute unambiguously the observed Sus1-dependent effect on transcription only to SAGA/SLIK. Further work is required to study the contribution of TREX-2 to this phenotype.
We observe that genes more stabilized by absence of SUS1 tend to have higher TR. This could reflect a compensatory crosstalk mechanism in which the general reduction in TR by SUS1 deletion is compensated by increase in HL in order to maintain mRNA levels, or vice versa. Such effect has previously been reported for decay factors [52, 53] and for transcription factors [54, 55]. In this way, Sus1 would participate in the co-transcriptional imprinting on mRNA molecules that regulates transcript fate [52, 56]. The fact that Sus1 is also involved in mRNA export, which is known to be coupled to upstream events , makes it an excellent candidate to participate in the mRNA imprinting of different classes of transcripts. Although the stability of all types of mRNAs (SAGA, TFIID and RP) depends on Sus1, we observed differences among them. On average, Sus1 absence affects more the HL of RPs mRNAs than TFIID-dominated and SAGA-dominated ones. RP and TFIID-dominated genes require a higher increase in HL to buffer for a global mRNA homoeostasis which is expected under standard conditions. Sus1 association to TFIID-dominated genes is barely affected by heat, probably because relative few TFIID genes are heat-shock-induced. In contrast, upon stress Sus1 relocates preferentially to promoters of SAGA-dominated genes. This change is maximal for promoters that are part of the environmental stress response (ESR-up). Since the RP genes are a fundamental part of the ESR-down response, a reallocation of Sus1 from RP genes to ESR-up genes as part of the heat-shock ESR is expected. This behaviour is consistent with the observation that Sus1 associates with elongating RNAPII , which is reallocated from RP to ESR-up genes upon stress. One possibility is that this phenotype is imposed by the chromatin organization of each gene class, being ESR-up genes controlled by SAGA to allow fine-tuned transcriptional changes. We speculate that SAGA-dependent histone modifications might become more crucial in these chromatin contexts than in genes that require less expression plasticity (for instance, constitutive TFIID genes). Additionally, data indicate that SAGA and TFIID have different functional spectra, and while TFIID is exclusively involved in PIC assembly, SAGA subunits also participate in transcript elongation . Consequently, the SAGA pathway provides a greater dynamic range of mRNA output and more opportunities for transcriptional regulation .
SAGA and SLIK share all but two of their components: Spt8 and full Spt7. The distinctive roles of the two complexes are still not clear and are difficult to address because of this high degree of overlap. We examine here an “ONLY-SLIK” phenotype by using mutants of spt8Δ and truncated spt7-1180. We addressed the contribution of SLIK to Sus1 recruitment by demonstrating that SLIK contributes to recruitment of Sus1 genome-wide, along with SAGA. Therefore, these results provide insights into the shared roles of SAGA and SLIK.
We propose here that under standard conditions Sus1 is recruited at a basal level to the UAS regions of expressed RP, SAGA-dominated and TFIID-dominated genes, at levels that reflect their transcriptional activity. Upon heat stress, Sus1 is preferentially associates with promoters of SAGA-dominated genes (especially ESR-up genes) in parallel with gene activation. Since Sus1-dependent deubiquitylation is genome-wide , Sus1 could facilitate the “fast” binding of the SAGA-DUB module in all these genomic locations, supporting a transient interaction between at least some SAGA/SLIK subunits and coding regions. Furthermore, since Sus1 is part of TREX-2 whose presence is enriched towards 3′ ends , Sus1 could contact TREX-2 to ensure a correct coupling between the different steps of the gene expression process. This is consistent with our detection of Sus1 also at the 3′ end of genes. Last, a common feature of SAGA and TFIID is the presence of a shared set of TAFs . A provocative hypothesis is that a preassembled core of these shared TAFs serves as a scaffold for assembly or switching between TFIID, SAGA or SLIK complexes at yeast genes promoters.
Yeast strains, recombinant DNA work and microbiological techniques
The yeast strains used in this study are listed in Additional file 5: Table S1. Chromosomal integration of TAP (URA3 marker) and MYC (HIS3 marker) as C-terminal tags was performed as previously described [60, 61]. For gene disruption, the indicated gene was deleted by high-efficiency transformation using a PCR product amplified from the plasmid pRS400-KanMX4. All deletions and genomically tagged strains were confirmed by PCR and/or western blot analysis. Strains were grown under standard conditions.
Chromatin immunoprecipitation (ChIP)
ChIP was performed as described  with minor modifications. Fifty millilitres of exponential-phase cultures grown in YPD at 25 °C was cross-linked with 1% of formaldehyde solution (Sigma) for 20 min at room temperature. The reaction was quenched with 125 mM glycine. Cells were collected and washed with Tris–saline buffer. Cells were broken in lysis buffer with glass beads, and cell extracts were sonicated in a Bioruptor (Diagenode). Ten microlitres of extract was reserved as the input, and the rest was incubated with Dynabeads® Pan Mouse IgG for 2 h. Immunoprecipitates (IP) were washed, and samples were eluted at 65 °C for 20 min with 100 µL of elution buffer. Inputs and immunoprecipitation samples were incubated at 65 °C overnight to reverse the cross-linking. Samples were treated with proteinase K, and DNA was purified with phenol/chloroform/isoamyl alcohol extraction and ethanol precipitation. The list of primers for ChIP analysis and RTqPCR can be found in the Additional file 6: Table S2).
ChIP-exo and data analyses
Two biological replicates of each ChIP-exo of Sus1-MYC-tagged protein and a no-tag strain were performed. Cells were grown in YPD at 25 °C. The experiments were performed as in . The heat-shock treatment was performed at 37 °C for 15 min. Prechilled formaldehyde was added to achieve a final concentration of 1%. Reactions were quenched and cells collected by centrifugation. Solubilized sonicated chromatin was prepared from cell extracts to yield a DNA fragment size of < 500 bp. For mapping nucleosomes, data from  were used. DNA was trimmed in the 5–3 direction with lambda exonuclease, until stopped by formaldehyde cross-link. The resulting DNA was subjected to deep sequencing. Sequencing reads were mapped to yeast genome (sacCer3) using BWA (version 0.5.9-r16). For all figures, reads were shifted in the 3′ direction by 6 bp and strand information was removed to better reflect the point of cross-linking. To compare mock and heat-shock data sets, data were normalized to have equal number of reads. The set of SAGA-/TFIID-dominated genes were obtained from  (Additional file 7).
Genomic run-on (GRO) for WT and sus1Δ was done and analysed as described in . This method allows for the determination on nascent transcription rates genome-wide because it labels nascent mRNAs in elongating RNA polymerases. Values were transformed into true mRNA synthesis rates (TR) by dividing by the cell volume (see  for a detailed explanation). Aliquots of the same cell samples were used to determine mRNA levels for all yeast genes as described in . Total mRNA concentration in WT and sus1Δ cells was determined  and used for normalization of mRNA levels (RA). In this way, mRNA half-lives (HL) for all mRNAs were calculated, assuming steady-state conditions, as the ratio between RAs levels and TRs.
SR-N, JEP-O and BFP designed the research. VG-M, JG-M, OA and VV performed the experiments. RR, VG-M, JG-M, BFP, PF-T and AC analysed the genomic data. SR-N and JEP-O wrote the paper. BFP and AC edited the manuscript. All authors read and approved the final manuscript.
We are grateful to present and past members from the Rodríguez-Navarro laboratory for scientific comments and Dr. Antonio Jordán-Pla for his help with graphical representations.
The authors declare that they have no competing interests.
Availability of the data and accessions numbers
Gene Expression Omnibus (GEO) database references are GSE65902 for ChIP-exo data and GSE63769 for GRO data.
Consent for publication
Ethics approval and consent to participate
VG-M was supported by the FPU program from MEC (AP2009-0917). This study was supported by funds from MINECO (BFU2011-23418 and BFU2014-57636 to SR-N, BIO2012-40244 to AC and BFU2016-77728-C3-3-P to JEP-O) and GV (PROMETEO/2012/061, PROMETEO/2016/093 and ACOMP/2014/061 to SR-N, and PROMETEOII/2015/006, ACOMP/2014/253 to JEP-O). The contributions of PF-T and AC were supported by the FP7 Project STATegra 306000. BFP was supported by NIH GM059055. BFP has a financial interest in Peconic, LLC, which utilizes the ChIP-exo technology implemented in this study and could potentially benefit from the outcomes of this research.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 10.Lee KK, Sardiu ME, Swanson SK, Gilmore JM, Torok M, Grant PA, et al. Combinatorial depletion analysis to assemble the network architecture of the SAGA and ADA chromatin remodeling complexes. Mol Syst Biol. 2011;7:1–12.Google Scholar
- 12.Sterner DE, Grant PA, Roberts SM, Duggan LJ, Belotserkovskaya R, Pacella LA, et al. Functional organization of the yeast SAGA complex: distinct components involved in structural integrity, nucleosome acetylation, and TATA-binding protein interaction. Mol Cell Biol. 1999;19:86–98.CrossRefPubMedPubMedCentralGoogle Scholar
- 32.Lenstra TL, Holstege FCP. The discrepancy between chromatin factor location and effect. Nucleus (Austin TX). 2012;3:213–9.Google Scholar
- 62.García-Oliver E, Pascual-García P, García-Molinero V, Lenstra TL, Holstege FCP, Rodríguez-Navarro S. A novel role for Sem1 and TREX-2 in transcription involves their impact on recruitment and H2B deubiquitylation activity of SAGA. Nucleic Acids Res. 2013;41:5655–68.CrossRefPubMedPubMedCentralGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.