Grr1 is responsible for the proteasomal degradation of Ubp3.
Ribosome ubiquitination increases significantly upon induction of the UPR in mammals and yeast18,37. In yeast, Ubp3 is the enzyme responsible for deubiquitylation of the ubiquitinated eS7A and is downregulated, which is consistent with the increased level of monoubiquitinated eS7A in response to UPR18. To understand how ribosome ubiquitination is regulated upon UPR in yeast, we performed genetic screening of E3 ubiquitin ligases that are required for Tm-resistance growth and Ubp3 degradation. Among 66 deletion mutants of the E3 ligases, we identified pep3D and grr1D mutants that conferred Tm-sensitive growth (Figs. 1A and S1A). The level of Hac1 protein (Hac1p) normalized to that of tubulin also indicated that Hac1p was significantly downregulated in the grr1D mutant, being approximately 50% of the levels in WT cells 4 h after Tm treatment (Fig. 1B, lanes 1–4). In contrast, the level of Hac1p was not decreased in the pep3D mutant cells (Fig. 1B, lanes 5–6). In the grr1D, pep3D mutant cells, induction of HAC1i mRNA by Tm treatment was intact (Fig. 1C), suggesting that Grr1 contributes to translation of HAC1i mRNA.
Grr1 is the F-box protein component of an SCF ubiquitin-ligase complex38,39, therefore we hypothesized that Grr1 degrades Ubp3 and the in the grr1D mutant cells, the upregulated Ubp3 decreases the monoubiquitinated eS7A thereby the reduces Hac1p level. The level of Hac1p 2 h after Tm treatment was reduced in the eS7 ubiquitination-defective eS7A-4KR cells, confirming that monoubiquitination of ribosomal protein eS7A plays a crucial role in translational control during the ER stress response in yeast (Fig. S1B). Neither uS3 nor uS10 ubiquitination, which is essential for 18S NRD and RQC, affected the Hac1p level upon UPR (Fig. S1C-D), indicating that eS7 ubiquitination plays a crucial role in ER stress response. The downregulation of Ubp3 was diminished in grr1D mutant cells (Fig. 1D, lanes 9–12), indicating that Grr1 is required for Ubp3 downregulation upon UPR. To examine whether Ubp3 is subjected to Grr1-dependent proteasomal degradation, we measured Grr1 stability using cycloheximide (CHX) chase followed by western blotting (Fig. 1E). Ubp3 was degraded under normal (Fig. 1E, lanes 1–5) and more unstable under ER stress conditions (Fig. 1F, lanes 6–10; half-life, t1/2 ~80 min, Fig. 1G) than in the absence of ER stress (Fig. 1F, lanes 1–5; half-life, t1/2 ~120 min, Fig. 1G). Ubp3 was stabilized after treatment with the proteasome inhibitor MG132 (Fig. S2A, lanes 6–10) and in grr1D mutant cells regardless of Tm treatment (Fig. 1F, lanes 1–10; half-life, t1/2 > 120 min, Fig. 1G). These results demonstrated that Grr1 is responsible for the efficient degradation of Ubp3, which may contribute to efficient HAC1i translation probably via an increase in eS7A ubiquitination.
Grr1 interacts with Ubp3.
Grr1 is a modular substrate specificity factor that associates with core SCF (Cdc53p, Skp1p, and Hrt1p/Rbx1p) to form the SCFGrr1 complex33,38–40. The F-box domain was responsible for binding to SCFGrr1, and the LRR domain interacted with the substrate for ubiquitination38,39 (Fig. 2A). The levels of Ubp3 upon UPR were determined in the grr1D cells expressing Grr1 mutants with the deletion of the indicated domains (Fig. 2A). The downregulation of Ubp3 was diminished in the grr1D mutants expressing Grr1 mutant proteins that lack the amino-terminal 1-280 residues (N-ter), the 311–361 residues F-box (Fbox) and the 413–740 residues LRR (LRR) domains (Fig. 2B, lanes 7–10). Consistently, the grr1D mutants expressing Grr1-DF or Grr1-DLRR mutant conferred sensitivity to Tm (Fig. 2C) and the reduced Hac1 protein expression upon UPR (Fig. 2D). These indicated that the F-box and LRR domains of Grr1 are required for the downregulation of Ubp3 upon UPR. The expression level of Grr1-DF or Grr1-DLRR mutant proteins was the same as the wild-type Grr1 (Fig. 2E). The expression level of Grr1-DN was significantly lower than that of wild-type Grr1 (Fig. 2E), but Grr1-DN complemented the grr1D mutant phenotype in Tm sensitivity (Fig. 2C) and the reduced Hac1p level upon UPR (Fig. 2D), indicating that the N-terminal region is dispensable for Grr1 function. To confirm the interaction of Ubp3 with Grr1, we purified the wild-type and mutant FLAG-Grr1 proteins that lack F-box (DFbox) or LRR (DLRR) domains and examined the co-immunoprecipitation of the Ubp3-3HA (Fig. 2F). Ubp3-3HA co-immunoprecipitated with all Grr1 proteins (WT, DFbox, DLRR)(Fig. 2F, lanes 11–16), indicating that Grr1 interacts with Ubp3-HA independent of the F-box and LRR domains. Taken together, we conclude that Grr1 downregulates Ubp3 for efficient HAC1i translation.
Grr1 facilitates eS7A ubiquitination.
Not4-mediated monoubiquitination of eS7A is required to produce Hac1p during the UPR; therefore, we assessed whether the upregulation of eS7A ubiquitination is dependent on Grr1. We found that the upregulation of eS7A ubiquitination diminished in grr1D mutant cells after the addition of Tm (Fig. 3A), indicating that Grr1 is required for the upregulation of eS7A ubiquitination upon UPR. We also confirmed that defects in Ubp3 downregulation in the grr1D mutants expressing Grr1-DF or Grr1-DLRR proteins results in the reduction of eS7A mono-ubiquitination. The levels of mono- and diubiquitinated eS7 were reduced in grr1D cells (Fig. 3B, lanes 3–4) and grr1D cells expressing Grr1-DF or Grr1-DLRR mutants (Fig. 3B, lanes 9–12). The levels of mono and di-ubiquitinated eS7 did not change in grr1D cells expressing wild-type Grr1 and Grr1-DN mutant protein (Fig. 3B, lanes 5–8), in which Ubp3 downregulation was intact (Fig. 2B). Not4 monoubiquitinates eS7A at the four lysine residues (Fig. 3C), with K83 ubiquitination primarily responsible for mRNA quality control (Ref). We examined cell growth in the presence of Tm, eS7A ubiquitination, and Hac1p production upon UPR in four eS7A mutants containing a single lysine residue, susceptible to Not4-mediated monoubiquitination, eS7A-K72only, eS7A-K76only, eS7A-K83only and eS7A-K84only. K83 but no other lysine residues were responsible for resistance to Tm (Fig. 3D) and Hac1 expression upon UPR (Fig. 3E, middle panels). Consistently, K83 was found to be the major ubiquitination site of eS7A (Fig. 3E, top panels), confirming that K83 ubiquitination is primarily responsible for eS7A ubiquitination thereby Hac1 production upon UPR. The upregulation of eS7A-K83only ubiquitination was diminished in the grr1D mutant cells after the addition of Tm (Fig. 3F), indicating a crucial role of Grr1 in the upregulation of eS7A ubiquitination at K83 upon UPR.
Grr1 is required for HAC1 i translation during the UPR in yeast.
These results suggest that Grr1 plays a role in HAC1 translation by downregulating Ubp3 during the UPR in yeast. To test whether Grr1 is involved in translational control of the UPR, we performed RNA-seq and ribosome profiling. To investigate the regulation of translation in response to ER stress, we estimated translation efficiency (TE) by assessing both mRNA abundance and ribosome occupancy. eS7A ubiquitination-dependent translational regulation was monitored at 4 h. Translational responses were observed in grr1D mutant cells, with statistically significant changes in the translation of 41 mRNAs (Fig. 4A-B; Fig. 4A; q-value < 0.01). To examine the Grr1-mediated ubiquitination dependency of the involved mRNAs, TE was compared between WT and grr1D mutant cells (Fig. 4B), with 29 and 12 mRNAs categorized as up- and downregulated, respectively (Fig. 4C). These subsets were identified using the formula “log2 TE fold change (WT) - log2 TE fold change (grr1)”, with mRNAs scored as > 2 and <-2 defined being up- and down-regulated, respectively.
We identified HAC1 as the most significantly upregulated gene during the UPR (Figs. 4A-D). In comparing TE in cells 4 h after the addition of Tm based on reads of HAC1, we found that the TE of HAC1 mRNAs was approximately 4-fold lower in grr1D than in wild-type (WT) cells (Fig. 4C). Mapping of the footprints throughout HAC1i mRNA in WT and grr1D mutant cells HAC1i mRNA contains potential translation pause sites (Fig. 4D). These results suggest that Grr1 may be involved in translation initiation and/or elongation of HAC1i mRNA.
Translation of HAC1 i mRNA requires eS7A ubiquitination regardless of ER stress.
UPR induces the transcription and splicing of HAC1u mRNA ensuring Hac1p expression upon UPR. We constructed a 5H3 clone containing the HAC1 ORF, 5′ and 3′ untranslated regions (5′ and 3′ UTRs) but lacking an intron (Fig. 5A) and examined Hac1p levels in hac1D mutant cells harboring 5H3 clones under the control of various promoters (Fig. 5B). Hac1p is induced upon Tm treatment in hac1D mutant cells harboring 5H3 under the control of the HAC1 promoter, as reported (Fig. 5B, lanes 5–6). Hac1p was expressed from 5H3 under the control of the GPD1, TEF1 and ADH1 promoters, regardless of Tm treatment (Fig. 5B, lanes 7–12). Unexpectedly, Hac1p was induced upon UPR in hac1D mutant cells expressing 5H3 under the CYC1 promoter (Fig. 5B, lanes 5–6 and 13–14), suggesting that YC1 transcription is induced upon UPR. We confirmed the expression levels of Hac1p in hac1D mutant cells harboring 5H3 clones under the control of various promoters were sufficient for Tm resistance (Fig. 5C).
We then examined whether eS7A ubiquitination is essential for the translation of HAC1i mRNA without ER stress. Hac1 was expressed regardless of UPR in eS7-WThac1D cells expressing 5H3 under the control of HAC1 and TEF1 promoters (Fig. 5D, lanes 5–6 and 9–10) but not in eS7-4KRhac1D mutant cells that lack eS7 ubiquitination (Fig. 5D, lanes 7–8 and 11–12), indicating that eS7A ubiquitination is essential for translation of HAC1i mRNA without ER stress. We then assessed the role of Grr1 in HAC1i translation by using 5H3 constructs. Hac1p was expressed regardless of UPR in grr1Dhac1D mutant cells expressing 5H3 under the control of HAC1 and TEF1 promoters (Fig. 5E, lanes 5–6 and 9–10). Importantly, Hac1p levels in grr1Dhac1D cells (Fig. 5E, lanes 7–8 and 11–12) were significantly lower than those in hac1D cells (Fig. 5E, lanes 5–6 and 9–10), indicating that Grr1 facilitates translation of HAC1i mRNA independent of ER stress.
ORF of HAC1i mRNA requires eS7A ubiquitination-mediated translation stimulation
To understand how eS7A ubiquitination facilitates HAC1i mRNA translation under ER stress, we examined the putative cis elements required for the regulation. The HAC1i mRNA 3′-UTR contains a cis-acting 3′-bipartite element (BE) at nucleotides C1134 to A1192 (adenine of the AUG codon is + 1) that promotes co-localization of the translationally repressed HAC1i mRNA under ER stress41,42. The protein kinases Kin1 and Kin2 contribute to HAC1u mRNA processing by phosphorylation of Pal1 and Pal2 which bind to the 3′-BE of HAC1 mRNA. HAC1u mRNA is stored in the cytoplasm in the absence of ER stress, and its translation is tightly suppressed by a base-pairing interaction between the intron and the 5′-UTR. Excision of the intron by Ire1-dependent splicing in response to ER stress led to robust translation of HAC1i mRNA, with the resulting Hac1p (Hac1p) upregulating UPR target gene expression (Fig. S3A). We constructed HAC1 with the deletion of 3′-BE or 5′-BP, a region responsible for base-pairing with 3′ UTR or both (Fig. S3A, DBE, DBP, DBP + DBE). The Hac1p was induced 2h after Tm addition in the hac1D mutant cells harboring the 5H3 clones under the control of the HAC1 promoter (Fig. S3B, lanes 5–12), indicating that these cis elements are not essential for the induction of Hac1p upon UPR. Consistently, hac1D mutant cells expressing HAC1 conferred resistance to Tm similar to wild-type cells (Fig. S3C).
To identify the cis-elements involved in eS7A monoubiquitination-dependent translation activation of HAC1i mRNA, we constructed 5H3, 5H, H3, and H clones that contain the HAC1i ORF with or without 5′ and 3′ UTRs (Fig. 6A). The hac1D mutant cells expressing the 5H3, 5H, H3, and H mRNAs by the TEF1 promoter (TEF1p-5H3, 5H, H3, H) conferred resistance to Tm similar to wild-type cells (Fig. 6B), indicating that UTRs are dispensable for the expression of Hac1p by the TEF1 promoter at a sufficient level for Tm resistance. The Hac1p was expressed regardless of Tm addition in hac1D mutant cells harboring TEF1p-5H3, 5H, H3, and H clones (Fig. 6C, lanes 7–14). The levels of Hac1p were increased by 5′ and 3′ UTRs (Fig. 6C, lanes 7, 9, 11, 13; Fig. 6D), indicating that UTRs stimulates translation of HAC1i mRNA as previously reported. We examined the role of eS7A ubiquitination in Hac1p derived from these clones. The levels of Hac1p expressed in all HAC1i clones were significantly diminished in hac1DeS7A-4KR cells (Fig. 6C, lanes 7–14; Fig, 6D). Inhibition of proteasome activity by MG132 treatment did not rescue the reduction of Hac1p in eS7A-4KR cells (Figs. 6E-F), suggesting that the downregulation of Hac1p in the eS7A-4KR mutant cells may not be due to the facilitation of proteasomal degradation of the Hac1p. Finally, we propose that eS7A ubiquitination facilitates the translation of HAC1i mRNA depending on putative cis element(s) in the ORF.