Differential effects of 40S ribosome recycling factors on reinitiation at regulatory uORFs in GCN4 mRNA are not dictated by their roles in bulk 40S recycling

Abstract

Recycling of 40S ribosomal subunits following translation termination, entailing release of deacylated tRNA and dissociation of the empty 40S from mRNA, involves yeast Tma20/Tma22 heterodimer and Tma64, counterparts of mammalian MCTS1/DENR and eIF2D. MCTS1/DENR enhance reinitiation (REI) at short upstream open reading frames (uORFs) harboring penultimate codons that confer heightened dependence on these factors in bulk 40S recycling. Tma factors, by contrast, inhibited REI at particular uORFs in extracts; however, their roles at regulatory uORFs in vivo were unknown. We examined effects of eliminating Tma proteins on REI at regulatory uORFs mediating translational control of GCN4 optimized for either promoting (uORF1) or preventing (uORF4) REI. We found that the Tma proteins generally impede REI at native uORF4 and its variants equipped with various penultimate codons regardless of their Tma-dependence in bulk recycling. The Tma factors have no effect on REI at native uORF1 and equipping it with Tma-hyperdependent penultimate codons generally did not confer Tma-dependent REI; nor did converting the uORFs to AUG-stop elements. Thus, effects of the Tma proteins vary depending on the REI potential of the uORF and penultimate codon, but unlike in mammals, are not principally dictated by the Tma-dependence of the codon in bulk 40S recycling.

Fig. 1. Predicted mechanisms governing MCTS1/DENR mediated recycling and REI at mammalian short uORFs and expected functions of yeast Tma20/Tma22 at typical REI-non-permissive yeast uORFs.

MCTS1/DENR or Tma20/Tma22 are depicted as linked blue ovals whose distinct functions in releasing the penultimate deacylated tRNA (red “L”) and subsequently dissociating the empty 40S from the mRNA are labeled with boxes 1 and 2, respectively, and solid or dashed arrows. A DENR-mediated recycling at short mammalian uORFs depending on the presence of (i) DENR-hyperdependent penultimate codons requiring the recycling factors for efficient tRNA release, or (ii) DENR-hypodependent codons where tRNA can be released at high levels spontaneously. (i) DENR-hyperdependent uORFs: in WT cells, “(+) DENR”, MCTS1/DENR efficiently releases the penultimate-codon tRNA (function 1) but its second function in 40S dissociation is impeded by eIF3 (green oval) retained on post-termination 40S subunits to enable high-level REI at the main CDS. In cells depleted of DENR, “(−) DENR”, diminished release of the penultimate-codon tRNA lowers REI. (ii) DENR-hypodependent uORFs: efficient release of the tRNA and high-level REI occur both in the presence (upper) or absence (lower) of DENR, yielding no decrease in REI on DENR depletion. B Tma-mediated recycling at typical yeast short uORFs depending on the presence of (i) Tma-hyperdependent or (ii) Tma-hypodependent penultimate codons. (i) Tma-hyperdependent uORFs: in WT cells (upper), Tma factors efficiently release the penultimate-codon tRNA (function 1) but also dissociate the empty 40S subunit from mRNA (function 2), conferring low-level REI. In tma∆∆ cells (lower), the absence of Tma-mediated tRNA release helps to ensure low-level REI, yielding no change in REI in tma∆∆ vs. WT cells. (ii) Tma-hypodependent uORFs: Tma-independent tRNA release occurs efficiently in both WT and tma∆∆ cells, but REI is low in both cases because, lacking eIF3, the empty 40S subunits dissociate from the mRNA independently of Tma factors, for no change in REI in tma∆∆ vs. WT cells.

Fig. 2. Summary of all cis-determinants that either promote or inhibit REI on GCN4 mRNA after translation of its four short uORFs.

Schematics of the 5′ enhancers of uORF1 and 2 containing their respective RPEs, some of which functionally interact with eIF3 to promote resumption of scanning. Green color-coding generally indicates stimulatory effects of the corresponding cis-factors on efficiency of REI, whereas red color-coding indicates inhibitory effects (with the exception of RPE ii. of uORF1, which is also stimulatory); the number of asterisks below the inhibitory elements of the uORF2 and uORF3 3′ sequences depicts the degree of their inhibition as determined experimentally. Mutations converting uORFs 1, 3, or 4 to Start-stop elements uSt-st1, uSt-st3, or uSt-st4, described later in RESULTS, are given below the respective WT uORF schematics. Reprinted and modified with permission from ref. .

Fig. 3. Differential effects of various TMA deletions on expression of uORF4-only GCN4-lacZ reporters equipped with Tma-hypodependent or Tma-hyperdependent penultimate codons.

A Schematic of the uORF4-only GCN4-lacZ reporters harboring either WT or mutant penultimate codons. B Yeast strains YSG181 (tma20∆), YSG184 (tma22∆), YSG178 (tma64∆), or YKJ3 (tma20∆tma22∆tma64∆), deleted for one or all three TMA genes and the corresponding WT strain YSG142 (WT BY4741) were transformed with uORF4-only reporters containing WT uORF4 (p226) or uORF4 variants with the penultimate Tma-average codon CCG exchanged for Tma-hyperdependent TTG^Leu (pKJ34) or Tma-hypodependent TGG^Trp triplet (pKJ36). Reporter activity was assayed in whole cell extracts for at least five independent transformants (with the specific n denoted in each graph) and activities in the tma∆ transformants were normalized as described in “Methods”. Data are represented as ratios of normalized mean values ± SD of β-galactosidase activities in tma∆ strains/WT strain. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns not significant; for details of statistical analysis, see “Materials and methods”. The exact p values can be found in Supplementary Data 1. C Same analysis as in (B) except comparing transformants of the double mutants YKJ6 (tma20∆tma22∆) and YSG196 (tma20∆tma64∆) to the triple mutant YKJ3 (tma20∆tma22∆tma64∆). D Serial dilutions of strains of the indicated genotypes from (C) were spotted on minimal SD medium and grown for 48 h at 30 °C.

Fig. 4. Differential effects of various TMA double and triple deletions on expression of uORF4-only GCN4-lacZ reporters equipped with Tma-hypodependent, Tma-average or Tma-hyperdependent penultimate codons.

A Schematic of the uORF4-only GCN4-lacZ reporters harboring either WT or mutant penultimate codons. B–D Yeast strains YKJ6 (tma20∆tma22∆) and YKJ3 (tma20∆tma22∆tma64∆) and the corresponding WT strain were transformed with the WT uORF4-only GCN4-lacZ reporter construct (p226) and variants with the WT uORF4 penultimate codon (CCG) exchanged for the indicated Tma-hypodependent, Tma-average or Tma-hyperdependent codons. Reporter activity was assayed in whole cell extracts for at least four independent transformants (with the specific n denoted in each graph) and activities in the tma∆ transformants were normalized as described in “Methods”. Data are represented as GCN4-LacZ reporter enzyme activities in WT strain and tma∆ strains. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns not significant; for details of statistical analysis, see “Materials and methods”. The exact p values can be found in Supplementary Data 1.

Fig. 5. Changes in reporter expression for groups of Tma-hyperdependent, Tma-hypodependent or Tma-average penultimate codons in tma mutant vs. WT cells.

A Violin plot of the average fold-changes in expression of uORF4-only GCN4-lacZ reporters harboring the four Tma-hyperdependent (Hyperdep), four Tma-hypodependent (Hypodep) or five Tma-average (Average) penultimate codons in the triple deletant yeast strain YKJ3 (tma20∆tma22∆tma64∆) vs. isogenic WT strain. The average fold-change of expression determined from biological replicates only was calculated for each construct from results plotted in Fig. 4. Statistical analysis was performed using two-tailed unpaired t-test *p < 0.05. B Violin plot of fold changes in expression of uORF1-only GCN4-lacZ reporters harboring the six Tma-hyperdependent, nine Tma-hypodependent and 14 Tma-average penultimate codons in tma20∆tma64∆ strain H4520 vs. WT strain BY4741. The average fold-change of expression determined from biological replicates was calculated for each construct from the results plotted in Fig. 6. Statistical analysis was performed using the two-tailed unpaired t-test. *p < 0.05. The exact p values can be found in Supplementary Data 1.

Fig. 6. Differential effects of deleting both TMA20 and TMA64 on expression of uORF1-only GCN4-lacZ reporters equipped with Tma-average, Tma-hypodependent or Tma-hyperdependent penultimate codons.

A Schematic of the uORF1-only GCN4-lacZ reporters harboring either WT or mutant penultimate codons. B Yeast strains H4520 (tma20∆tma64∆) and WT strain BY4741 were transformed with uORF1-only GCN4-lacZ reporter constructs with the WT uORF1 penultimate codon (TGC) exchanged for the indicated Tma-average, Tma-hypodependent or Tma-hyperdependent codons. β-galactosidase activities determined for each of three independent transformants (n = 3) of the tma20∆tma64∆ strain were divided by a factor of 1.316 prior to calculating the mean activities and S.E.M. values plotted here to normalize for differences in reporter expression between the two strains that are independent of the uORF1 variants and observed for an uORF-less reporter (as described in “Methods”). The mean values between mutant and WT were compared for each reporter in a two-tailed, unpaired Student’s t test to assign p values (*p < 0.05). The exact p values can be found in Supplementary Data 1.

Fig. 7. Effects of start-stop elements on downstream REI in the GCN4-lacZ reporter system and differential effects of TMA20 and TMA64 deletion on re-initiation downstream of start-stop elements.

A WT strain BY4741 was transformed with GCN4-lacZ reporter plasmids containing the following single WT uORFs: uORF1-only (p209), uORF3-only (pSG61_(2)), or uORF4-only (p226); or containing single Start-stop elements uSt-st1 (pKP78), uSt-st3 (pKP76), or uSt-st4 (pKP77). β-galactosidase activities were assayed in at least four biological replicates (with the specific n denoted in each graph) and data are presented as GCN4-LacZ reporter enzyme activities. B Strains YSG196 (tma20∆ tma64∆) and the corresponding WT were transformed with the same set of GCN4-lacZ reporter plasmids described in (A) and reporter expression was measured as described there. Reporter activity values in the tma∆∆ strain were normalized as described in “Methods”. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns not significant; for details of statistical analysis, see “Materials and methods”. The exact p values can be found in Supplementary Data 1.

Fig. 8. Expected vs. proposed functions for yeast Tma20/Tma22 at REI-permissive GCN4 uORF1 and REI-non-permissive GCN4 uORF4.

Tma20/Tma22 (blue ovals), penultimate deacylated tRNA (red “L”), and the two functions of Tma factors (boxes 1 and 2) in releasing tRNA and dissociating empty 40S subunits, respectively (solid or dashed arrows), are depicted as in Fig. 1. A Tma-mediated recycling at GCN4 uORF4 depending on the presence of (i) Tma-hyperdependent or (ii) Tma-hypodependent penultimate codons. (i) Tma-hyperdependent uORFs. (a) expectation: as depicted exactly as in Fig. 1B(i) (presented again for comparison), in WT cells (upper), Tma factors efficiently release the penultimate-codon tRNA and also dissociate the empty 40S subunits, conferring low-level REI. In tma∆∆ cells (lower), the absence of Tma-mediated tRNA release helps to ensure low-level REI, yielding no change in REI in tma∆∆ vs. WT cells. (b) model: in WT (upper), deacylated tRNA can be released by Tma-hyperdependent or Tma-independent mechanisms, but Tma-mediated 40S dissociation maintains low-level REI. In tma∆∆ cells (lower), Tma-independent tRNA release coupled with absence of Tma-mediated 40S dissociation confers the increased REI we observed in tma∆∆ vs. WT cells. (ii) Tma-hypodependent uORFs. (a) expectation: as depicted exactly in Fig. 1B(ii), Tma-independent tRNA release occurs in both WT and tma∆∆ cells, but REI is low in both cases because the empty 40S subunits dissociate from the mRNA independently of Tma factors, for no change in REI vs. WT. (b) model: in WT (upper), 40S dissociation is catalyzed by Tma factors even at these Tma-hypodependent uORFs, and absence of this function in tma∆∆ cells, coupled with Tma-independent tRNA release, confers the observed increased REI in tma∆∆ vs. WT cells. B Tma-mediated recycling at GCN4 uORF1 depending on the presence of Tma-hyperdependent (i) or Tma-hypodependent (ii) codons. (i) Tma-hyperdependent uORFs. (a) expectation: as depicted for MCTS1/DENR in Fig. 1A(i), In WT cells (upper), Tma factors efficiently release the penultimate-codon tRNA but its second function in 40S dissociation is impeded by eIF3 (green oval) to enable high-level REI. In tma∆∆ cells (lower), diminished release of the penultimate-codon tRNA lowers REI. (b) model: deacylated tRNA can be released from uORF1 variants by both Tma-hyperdependent and Tma-independent mechanisms in WT yeast (upper) and Tma-independent tRNA release maintains high-level REI in tma∆∆ cells (lower), as 40S dissociation is impeded by eIF3 in both WT and tma∆∆ cells, thus accounting for the unchanged REI we observed in tma∆∆ vs. WT cells. (ii) Tma-hypodependent uORFs. (a) expectation: as depicted for MCTS1/DENR in Fig. 1A(ii), Tma-independent release of the tRNA and subsequent high-level REI occurs in the presence (top) or absence (bottom) of Tma factors, yielding no decrease in REI in tma∆∆ vs. WT cells, as we observed.

Fig. 9. Comparison of S. cerevisiae (yeast), D. melanogaster (fly), H. sapiens (human) and C. elegans (worm) Tma20/MCTS1 proteins.

A Multiple sequence alignment was performed using Clustal Omega. This alignment illustrates the conservation of amino acid residues of the ribosome binding interface described previously and marked here by red dots. Similarity score and identity percentage were determined using UniProt BLAST tool. B Comparison of predicted Tma20 structure (upper panel) and structure of MCTS1 in complex with DENR (bottom panel; MCTS1 colored in light gray, DENR colored in dark grey). Conserved residues are depicted in red and regions separating them are colored in blue. Tma20 structure was predicted using the AlphaFold3 algorithm. MCTS1/DENR complex structure was published previously (PDB accession 6MS4). PyMOL 3.0 was used for data visualization.