Transcriptional and Translational Regulatory Mechanisms in Nutrient Control of Gene Expression

Poly(A)–binding protein (PABP, Pab1 in yeast) has been implicated in stimulating translation initiation and regulating mRNA decapping and decay by forming a closed-loop mRNP via interaction of PABP bound to the poly(A) tail with eIF4G bound to the m⁷G cap; however, PABP’s importance for promoting the translation or stability of individual transcripts is poorly defined in most eukaryotes, including budding yeast. Using an auxin-inducible pab1-AID degron mutant, we found that Pab1 depletion impairs yeast cell growth and polysome assembly and reduces bulk mRNA levels by about 40%. Strikingly, these phenotypes were suppressed by deleting DCP2 (dcp2∆), which encodes the catalytic subunit of the mRNA–decapping enzyme, implying that diminished mRNA abundance and global translation at limiting Pab1 results primarily from enhanced mRNA decapping and decay. Single-molecule sequencing of poly(A) tails (SM-PAT-Seq) revealed a shift to longer poly(A) tails on Pab1 depletion in DCP2, but not in dcp2∆ cells, which therefore likely results from preferential degradation of mRNAs with shorter poly(A) tails rather than slower deadenylation. Unlike findings in mammalian cells, where PABP depletion reduced mRNA abundance and protein synthesis but had little effect on translational efficiencies of individual mRNAs, we found that Pab1 depletion reprogrammed the translational efficiencies of about 1200 mRNAs and conferred corresponding changes in encoded protein levels. Importantly, these changes in translation were diminished or eliminated by deleting DCP2 and thus arise indirectly from reduced mRNA abundance on Pab1 depletion, possibly from altered ratios of mRNAs to eIF4F, preinitiation complexes, or other RNA–binding proteins. A subset of about 100 mRNAs still exhibit altered translational efficiencies on Pab1 depletion in dcp2∆ cells, suggesting a more direct stimulatory role for Pab1 in their translation. Nevertheless, our results indicate that assembly of the closed-loop mRNP via Pab1–eIF4G interaction is dispensable for wild-type translational efficiencies of the great majority of yeast mRNAs at normal transcript levels in vivo. Interestingly, histone mRNAs and proteins are preferentially diminished on Pab1 depletion, dependent on Dcp2, which is accompanied by activation of internal cryptic promoters in the manner expected for reduced nucleosome occupancies, revealing a new layer of post-transcriptional control over histone gene expression via Pab1.

Degradation of most yeast mRNAs involves decapping by the Dcp1/Dcp2 complex. The Edc3 and Scd6 proteins have been implicated as decapping activators in budding yeast, but few native mRNAs were shown to depend on either protein for stimulating decapping and degradation. Scd6 has also been implicated as a translational repressor, but again, few native mRNAs were shown previously to be translationally derepressed in mutant cells lacking only Scd6. By combining RNA-Seq with ribosome profiling and global proteomics via TMT-mass spectrometry, we demonstrated that Scd6 and Edc3 function redundantly to repress the abundance or translation of a sizeable group of mRNAs, whose abundance is derepressed only in the mutant cells lacking both proteins. By interrogating our previous analyses of mutants lacking the decapping activators Dhh1 and Pat1, we discovered that mRNAs targeted for degradation or translational repression by Scd6/Edc3 are generally also targeted by Dhh1 or Pat1. An especially strong correspondence between mRNAs dysregulated in scd6∆edc3∆ and dhh1∆ mutants supports a recent proposal that Edc3 and Scd6 function interchangeably in recruiting Dhh1 to Dcp2 for activation of decapping, which we corroborated by co-immunoprecipitation analysis of native proteins in cell extracts. We further established that mRNAs preferentially repressed by Scd6/Edc3 encode proteins normally expressed at low levels on rich media and derepressed during growth on alternative carbon sources. These include the mitochondrial enzymes of oxidative phosphorylation, and we demonstrated elevated mitochondrial membrane potential and increased carbon flow from glucose into intermediates of the tricarboxylic acid cycle in mutants lacking Scd6/Edc3, Dhh1, or Pat1. Overall, we found that Scd6 and Edc3 are functionally redundant and cooperate extensively with Dhh1 and Pat1 in post-transcriptional repression of gene expression that promotes fermentation (glycolysis) over respiration in glucose-grown cells.

In budding yeast, TFIID and SAGA are the principal coactivator complexes that recruit TATA–binding protein (TBP) for assembly of preinitiation complexes at gene promoters, and the resident histone acetylation (HAT) activity of SAGA lodged in subunit Gcn5 enhances transcription via histone acetylation. Published evidence suggests that the majority of yeast genes (around 90%) require TFIID for efficient transcription, with only about 10% utilizing SAGA in addition to or instead of TFIID. It was concluded recently that the cohort of genes activated by transcriptional activator Gcn4 in amino acid–starved cells utilize only the HAT activity and not the TBP–recruitment function of SAGA and rely exclusively on TFIID for TBP recruitment. We found that eliminating the two TBP–interacting subunits of SAGA, Spt3 and Spt8, but not Gcn5, reduced TBP binding at most Gcn4 target genes in amino acid–starved cells. In contrast, eliminating Gcn5 but not Spt3 or Spt8, impaired histone acetylation and nucleosome eviction at highly expressed Gcn4 target genes. We further showed that depleting the key Taf1 subunit of TFIID strongly impairs TBP recruitment at most Gcn4 target genes only in cells devoid of Spt3 or Spt8. These results indicate a critical, non-redundant role for SAGA in TBP recruitment, apart from its HAT function, even in cells harboring wild-type TFIID. We obtained similar findings for the most highly expressed subset of genes previously shown to utilize SAGA or TFIID redundantly, dubbed “coactivator-redundant (CR)” genes. In contrast, highly expressed TFIID–dependent genes showed reduced TBP binding on depleting Taf1, even in cells with wild-type SAGA, and eliminating Spt3/Spt8 had little effect on TBP recruitment. These findings deepen our understanding of the relative importance of SAGA vs. TFIID in TBP recruitment, PIC assembly, and transcription of different classes of genes under stress conditions, highlighting the crucial role of SAGA in TBP recruitment via Spt3/Spt8 and the independent contribution of Gcn5 HAT activity for the Gcn4 transcriptome and other highly expressed CR genes in amino acid–starved cells.

Previous studies had suggested that the ATPase Mot1 generally dissociates TBP from promoters that utilize SAGA to augment TBP recruitment in favor of TBP binding at genes reliant on TFIID for PIC assembly. Our findings show instead that promoters are dependent on Mot1 for high-level TBP occupancies when they are highly expressed, regardless of their dependence on SAGA or TFIID for TBP recruitment. Accordingly, TBP binding at condition-dependent SAGA–activated genes is enhanced by Mot1 when they are stress-induced, but antagonized by Mot1 when expressed at basal levels. Among bulk promoters expressed constitutively, Mot1 serves to redistribute TBP from weaker to stronger promoters regardless of their coactivator dependence. Our analysis also uncovered an unexpected response to Mot1 depletion whereby reductions in TBP occupancies at highly expressed/induced SAGA–dependent genes does not reduce their transcription levels. This uncoupling could arise from SAGA–mediated assembly of a reservoir of incomplete PICs containing TBP whose formation is not crucial for promoter activation per se, but may increase the rate or endpoint of induction. These findings substantially alter our understanding of the genome-wide role of Mot1 in regulating PIC assembly and its importance for mounting an effective transcriptional response to stress.