We study transcriptional and translational mechanisms in nutrient control of gene expression in the yeast Saccharomyces, focusing on a regulatory system that induces genes encoding amino acid biosynthetic enzymes in response to starvation for amino acids. The transcriptional activator in this pathway, GCN4, is induced at the translational level in starved cells by phosphorylation of initiation factor 2 (eIF2) by the protein kinase (PK) GCN2. Phosphorylation of eIF2 reduces the concentration of the ternary complex (TC) containing eIF2, GTP, and initiator methionyl tRNA, that transfers tRNAiMet to the 40S ribosome. This impedes general protein synthesis but induces GCN4 translation by a reinitiation mechanism involving small upstream open reading frames (uORFs) in the GCN4 mRNA leader. A reduction in TC levels allows 40S ribosomes scanning the GCN4 mRNA leader after translating uORF1 to bypass uORFs 2-4 and reinitiate at the GCN4 start codon instead. GCN2 is activated in starved cells by binding of uncharged tRNA to a histidyl-tRNA synthetase (HisRS)-like region which functions as a sensor of amino acid limitation. Phosphorylation of Ser-577 in GCN2 by another kinase decreases the affinity of GCN2 for tRNA and impedes kinase activation in nonstarved cells. Dephosphorylation of Ser577 and attendant activation of GCN2 occurs in cells treated with rapamycin, resulting from inhibition of TOR1 and TOR2 proteins. Genetic analysis shows that the TOR proteins promote phosphorylation of Ser577 partly by inhibiting SIT4 and other type 2A phosphatases via regulatory protein TAP42. We concluded that TOR impedes eIF2 function and stimulates translation in nutrient-replete cells by preventing activation of GCN2 by basal levels of uncharged tRNA. Previously, we constructed a subunit interaction map for the five essential subunits of yeast eIF3 and found that eIF3c/NIP1 contains a binding site for eIFs 1 and 5. eIF5, the GTPase activating protein for the TC, and eIF1 function in selection of AUG as start codon. The C-terminal segment of eIF5 (CTD), which binds NIP1, interacts with the beta subunit of eIF2, stabilizing a multifactor complex (MFC) containing eIFs 1,2,3, 5 and tRNAiMet. An eIF5 mutation that disrupts these interactions (tif5-7A) impairs translation in vivo in a manner partially suppressed by overexpressing the TC, and it reduces TC binding to 40S subunits in vitro. Thus, the MFC is an important translation intermediate whose integrity is important for 40S-binding of TC. We incorporated an affinity tag into the 3 largest eIF3 subunits and deleted predicted binding domains in each tagged protein. By characterizing the mutant subcomplexes, we confirmed all binding interactions predicted by our model and uncovered direct contact between the eIF3a/TIF32-CTD and eIF2-beta. Overexpressing a CTD-less form of TIF32 produced a slow-growth phenotype that was partially suppressed by overexpressing TC, and exacerbated the growth defect of tif5-7A mutation. Thus, the two independent eIF2-eIF3 contacts make additive contributions to the efficiency of TC recruitment. Using a new technique we developed for cross-linking of initiation complexes in living cells, we showed that overexpressing CTD-less TIF32 in tif5-7A cells reduces TC binding to 40S subunits in vivo. However, it does not have the expected effect of constitutively inducing GCN4 translation (Gcd- phenotype). Thus, disrupting the MFC with these mutations seems to impair a function involved in ribosomal scanning or AUG recognition and obscures the effect of reduced TC recruitment on GCN4 translation. Characterization of mutations in eIF3a/TIF32 and eIF3b/PRT1 provides additional evidence for a critical post-recruitment function of eIF3 in scanning and AUG recognition. The prt1-1 mutation impairs translation without reducing the abundance of 48S preinitiation complexes in vivo and, interestingly, prt1-1 reduces the selection non-AUG triplets as start codons and impairs induction of GCN4 translation in starved cells when eIF2 is phosphorylated (Gcn- phenotype). The latter can be attributed to a delay in scanning or GTP hydrolysis by ribosomes scanning the uORF1-uORF4 interval on GCN4 mRNA. The rpg1-1 allele of TIF32 also produces a Gcn- phenotype but in this case, it seems to arise from a defect in the stability of mRNA-40S complexes. By determining the ability of various MFC subcomplexes to bind to 40S subunits in vivo, we found that the N- and C-terminal domains of NIP1/eIF3c, the N- and C-terminal domains of TIF32/eIF3a, and eIF5 have critical functions in 40S binding, with eIF5 and the TIF32-CTD performing redundant roles. Consistently, purified eIF3 and a trimeric complex of NIP1, C-terminally truncated TIF32 and eIF5 bind to purified 40S ribosomes in vitro. The TIF32-CTD interacts in vitro with helices 16 to 18 of domain I in 18S rRNA, and the TIF32-NTD and NIP1 interacts with recombinant the 40S protein RPS0A. These results suggest that eIF3 binds to the solvent side of the 40S subunit in a way that provides access to the interface side for the two eIF3 segments (NIP1-NTD and TIF32-CTD) that contact eIFs 1, 5, and the TC in the MFC. Our previous genetic analysis showed that optimal transcriptional activation by GCN4 requires the coactivators SAGA, SWI/SNF, SRB/mediator, CCR4/NOT complex, RSC, Paf1 complex, and MBF1 protein. The GCN4 activation domain binds specifically in vitro to all of these coactivators, except Paf1 complex, and chromatin immunoprecipitation (ChIP) experiments show that GCN4 can recruit all 7 coactivators to the same target gene in living cells. Although SWI/SNF subunits SNF2, SNF5, and SWI1 interact directly with GCN4 in vitro, we found that SNF2 is not required for recruitment of SWI/SNF by GCN4 in vivo, nor can SNF2 be recruited independently of other SWI/SNF subunits. SNF5 also was not recruited as an isolated subunit, but was required along with SNF6 and SWI3 for optimal recruitment of other SWI/SNF subunits. Thus, SNF2, SNF5, and SWI1 are recruited strictly as subunits of intact SWI/SNF. Optimal recruitment of SWI/SNF also requires specific subunits of Srb mediator and SAGA, but is independent of the histone acetyltransferase in SAGA, GCN5. It appears that SWI/SNF recruitment is enhanced by cooperative interactions with subunits of Srb mediator and SAGA that were recruited by GCN4 to the same promoters.
Showing the most recent 10 out of 81 publications
